DOCUMENT RESUME

ED 244 811 SE 044 536

TITLE Emerging Issues in Science and Technology, 1982. ACompendium of Working Papers for the National ScienceFoundation.

INSTITUTION National Science Foundation, Washington, D.C.REPORT NO NSF -83 -61PUB DATE Oct 83NOTE 142p.; For 1981 edition, see ED 221 364.PUB TYPE Reports - General (140)

EDRS PRICE MF01/PC06 Plus Postage.DESCRIPTORS *Communications; Computer Oriented Programs;

Computers; *Cooperative Planning; Federal Aid;Government Role; Higher Education; *Industry;International Programs; Manufacturing; NationalSecurity; Paraprofessional Personnel; PolicyFormation; *Research and Development; ScienceProjects; *Sciences; Technical Education;*Technology

IDENTIFIERS National Science Foundation; *Scientific andTechnical Information

ABSTRACTPresented are nine working papers prepared for the

National Science Foundation as one means of assisting the Office ofScience and Technology Policy in preparing the administration's"Annual Science and Technology Report to the Congress, 1982." Thepapers explore aspects of three broad themes central to theadministration's science and technology (ST) policies and strategiesfor their implementation: (1) optimizing use of limited resources forresearch and development (R&D) so that ST can be used moreeffectively to achieve national goals; (2) developing a set ofequitable and consistent guidelines, within the administration'soverall ST policy, for dealing with the generation, organization, anddissemination of ST information; and (3) encouraging implementation,by the private sector, of new technologies that can increase theproductivity and international competitiveness of U-.S. industry.Papers (each preceded by an abstract) focus on: internationalcooperation in science--U.S. role in megaprojects; trends incollective industrial research; impact of increased defense R&Dexpenditures on the U.S. research system; training and utilization ofengineering technicians and technologists; national security controlsand scientific information; issues in ST information policy; legaland regulatory implications of the video telecommunicationsrevolution; trends in computers and communications--offices of thefuture; and fostering use of advanced manufacturing technology.(JN)

************************************************************************ Reproductions supplied by EDRS are the best that can be made ** from the original document. ************************************************************************

U.S. DEPARTMENT OF EDUCATION__

ATIONAL INSTITUTE OF EDUCATION_UCATIONAL RESOURCES INFORMATION

CENTER (ERIC)This document has been reproduced asreceived from the person or organizationoriginating itMinor chanties have been made to improve------reproductron quality.

. _ .Points o' view or opinions stated in thisd_o_ctt

mem do not necessarily represent official NIEposition or policy.

SCOPE OF INTEREST NOTICEThe Eric Facility has assignedthis dbbuirreint for proccesingto:

In our judgment, this document_is also_of inierest_to the Clearing-houses noted to the right, Index-ing should telbct their specialpoints of view

co EMERGING ISSUES_--- IN SCIENCE AND(\I2 TECHNOLOGY, 1982

1111

A Compendium of Working Papersfor the National Science Foundation

National Science FoundgtiOn

Contents

PagePrefaceIntroduction vii

International Cooperation in Science: TheU.S. Role in Megaprojects 1

Trends in Collective Industrial Research 15The Impact of Increases in Defense R&D Expenditures

on the U.S. Research System 31Training and Utilization of Engineering

Technicians and Technologists 49National Security Controls and Scientific Information 59Issues in Scientific and Technical Information Policy 73Lzgal and Regulatory Implications of the Video

Telecommunications Revolution 87Trends in Computers and Communication:

The Office of the Future 99Fostering the Use of Advanced Manufacturing Technology . . . 113Acknowledgements 127

iii

3

Preface

In enacting the National Science and Tech-nology Policy; Organization, and PrioritiesAct of 1976 (Public Law 94-282), the Con-gress specified; in a declaration of principle,that the development and implementationof strategies for determining and achievingthe appropriate scope, level, and directionof U.S. scientific and technological effortsshould involve a wide range of participantsfrom both the public and the private sectors.

In keeping with that commitment, and asone means of fulfilling the National ScienceFoundation's responsibility to provide pri-mary assistance to the President's ScienceAdvisor in the preparation of the AnnualScience and Technology Report to theCongress: 1982. NSF convened a series ofad hoc panels of experts from Government,industry, and academia during 1982. Thosepanels explored the policy implications of anumber of current and emerging issues inscience and technology that were selectedby the staff of the National Science Foun-dation in consultation with advisors fromboth inside and outside of Government.

4

The panels' deliberations are summarizedin the nine working papers in this com-pendium. As anticipated, they were usefulto the staff of the NSF Office of SpecialProjects in assisting the Office of Scienceand Technology Policy with the preparationof the President's Annual Science and Tech--nology Report to the Congress: 1982 Sincethe papers also delineate an important setof policy issues on the national agenda, NSFis publishing them separately to stimulatepublic discussion about the roles scienceand technology can play in contemporaryAmerican society. Although all of the paperswere reviewed for technical accuracy, theviews and perspectives they express do notnecessarily reflect official policy positions ofthe U.S. Government or the National Sci-ence Foundation.

Edward A KnappDirectorNational Science FoundationOctober 1983

Introduction

The nine working papers in this compen-dium, like the set published in 1982,' wereprepared for the National Science Founda-tion as one means of assisting_the WhiteHouse Office of Science and TechnologyPolicy with the preparation of the President'sAnnual Science and Technotogy Report tothe Congress. They explore particular aspectsof three broad themes central to theAdministration's science and technologypolicies and the strategies for their imple-mentation, as described in the most recentAnnual Report.2 These themes are:

Optimizing the use of limited resoutesfor research and development (R&D)so that science and technology can beused more effectively to achieve na-tional goals;Developing a set of equitable and con-sistent guidelines; within the Adminis-tration's overall science and technologypolicy, for dealing with the generation;organization, and dissemination ofscientific and technical information; andEncouraging implementation, by theprivate sector, of new technologies thatcan increase the productivity and inter-national competitiveness of U.S.industry.

The compendium papers do not aim toprovide detailed analyses of all relevant issues.Indeed, since the policy context for most ofthese issues is in a state of considerable flux,attempts at such analyses run the risk ofbecoming rapidly dated. Nor do the papersweigh advantages and disadvantages.of allpossible policy options. Rather, each is in-tended to identify and discuss significantnational issues in science and technologythat are either currently on the policy agendaor likely to emerge in the near future.

Optimizing the Use of R&DResources

The first two papers in the compendiumInternational Cooperation in Science: TheU.S. Role in Megaprojects and Trends in Col-lective Industrial Researchexplore the

.1

policy aspects of two types of cooperativeinstitutional arrangements intended to in-crease the effectiveness of the resourcesavailable for the conduct of R&D. The thirdpaper--The Impact of Increases in DefenseR&D Expenditures on the US. ResearchSystemexamines some probable effectson research priorities, particularly in univer-sities, of the Administration's commitmentto strengthen the Nation's defense capabili-ties. The fourth paperTrainirg and Utili-zation of Engineering Technicians andTechnologistsraises questions about theadequacy, in both quantitative and qualitativeterms, of the workforce required to supportthe activities of professional scientists andengineers.

International Cooperation in Science:The EL& Role in Megaprojects

The context within which internationalcooperative scientific activities takes placeis considerably different from what it was20 years ago. During that era, the resourcesfor conducting scientific research in theUnited Statesparticularly basic researchwere relatively unconstrained, and the UnitedStates enjoyed preeminence in virtually allscientific fields. in the present environment,where the scientific capabilities and achieve-ments of several other countries are roughlycomparable to our own, pursuing researchprojects within an international division oflabor framework has an obvious appeal.This is particularly true for so-called "mega-projects;" that is; projects requiring extensivestaffs, complex managerial arrangements,and large budgets.

The Reagan Administration has placed ahigh priority on international cooperativeprojects that can augment limited U.S.resources and yield benefits to U.S. science.So, for similar reasons, have the governmentsof several other industrialized countries. Atthe Versailles Economic Summit in June1982, a Working Group was established toconsider common opportunities, problems,and challenges associated with science andtechnology. A draft report, released in March

VII

and considered at the Williamsburg EconomicSummit in May, includes proposals for 18cooperative science and technology projects,many of which qualify as megaprojects.

The paper on international cooperationin science reviews some of the potentialbenefits, both direct and indirect; that canbe derived from successful internationalcooperative projects. The most obvious ofthese benefits is improved economic effi-ciency. However, financial advantages areoften supplemented by the promise ofenhanced access to intellectual resourcesand unique facilities that are only availableabroad; and possibly by important gainS inthe innovative process itself.

In an international setting, individualgovernments almost always have multiplepolicy goals and objective§ for internationalcooperative project§. These_ can lead todifficultiet that need to be Weighed againstthe potential benefits to be derived fromcollaboration. In addition, countries oftenhave differing rationales for, and rankingsof, goals and objectives for scientific coop-eration. For that reason, a country's overallpolitical culture may be a more importantdeterminant of its science policyand there-fore the ways in which it will enter into acooperative venturethan any objectivescientific benefits that may derive fromcooperation.

The paper argues that the most difficultissue confronting effective U.S. participation inmegaprojects is in identifying specific futureopportunities for effective cooperation. Tomake appropriate decisions about suchopportunities, information must be availableregarding the scientific capabilities of countrieswith which the United States could con-ceivably cooperate, and the priorities thatare likely to be attached to the developmentand deproyment of thoSe capabilities by therespective nations. Unfortunately, such asystematic knowledge base does not exist.

Trends in Collective IndustrialResearchOne of the key strategies adopted by theReagan Administration to implement itsscience and technology policy is to encourageand facilitate cooperative R&D activitiesamong different types of institutions in the

viii

United States. Interest in developing coop-erative modes has been heightened becauseof both the rising costs of R&D and thedesirability of improving the links betweeninstitutions with vested interests in differentportions of the R&D spectrum. ranging hornbasic research to commercialization.

University-industry research collaborationhas been the most highlYpublicized of thesecooperative modes. Additionally, thereappears to be a trend toward new cooperativearrangements among industrial firms, someof which also involve support for universityresearch and education.

There have been numerous examples inthe past of two or more companies in aparticular industry establishing a relationshipto engage in technical activities with no formalcommitment to joint commercial exploitation.Such joint ventures have often been estab-lished through trade associations. Within thepast decade, and particularly the past 4 years;cooperative industrial structures have beenestablished that differ from these olderarrangements in at least three respects:

First, their levels of funding are con-siderably larger:Second; several include cooperationwith and/or assistance to universitiesamong their objectives; andThird, they are based on industrywideconcerns about decreasing marketshares and increasing foreign com=petition.

The paper on industrial research coop-eration examines the operational charac-teristics of several collective industrial researcharrangements, focusing particularly on fourof the newer associationstwo in the energysector, and one each in the chemical andsemiconductor industries. The underlyingobjective of all four groups is to acceleratethe pace of technical change either throughincreasing the number of technically trainedpeople available to the industry; or by con-ducting research in targeted areas, or both.However, the otigins and goals of each groupalso reflect industry-specific characteristicssuch as competitive structure, degree ofregulation, capital requirements, manpowerneeds, and relevant antitrust restrictions.

The paper considers the longer rangeimplications of large-scale industriai coop:erative research for industry, universires, aid

6

Government. It concludes that although thefunds available for collective research amongfirms within an industry are likely to remainsmall relative to the total expended for R&Dby the industry; the focused character ofthe available collective resources could have asignificant impact on industrialand uni-versityresearch directions:

Possible Government actions to encouragethe growth of collective industrial supportinclude: (a) offering indirect financial incen-tives to industrial firms for collective support touniversities, and (b) providing focused supportto universities to strengthen their capabilitiesfor cooperation in areas of interest to col-lective industrial groups. These actions wouldbe broadly consistent with the Reagan Ad-ministration's policy of providing indirectincentives rather than direct subsidies forindustrial R&D, while concentrating Federalresources on strengthening the capability ofuniversities to conduct fundamental researchin areas of importance to the Nation.

The Impact of Increases in DefenseR&D EiMentlittire.S On the U.S.Research System

The third paper in the compendium exploresthe potential impacts on the US. researchsystem, particularly on universities, o f t heaccelerated growth of the defense R&Dsector relative to the civilian sector. Depart-ment of Defense (DOD) support Of universitybasic research during the 1950s and 1960sprovided the foundation for much Of thegrowth and development. in the universityresearch system. This support also laid thegroundwork for such important technologcalinnovations as computers and lasers. Thispaper concludes that overall, the presentdefense buildup is likely to have a far lesssignificant effect than in the past on universityresearch. The Department of Defense'sbudget for research, development, test andevaluation (RDT&E) will increase by anestimated .78 percent between fiscal years1981 and 1984; while all other Federal R&Dexpenditures will decrease by an estimated12 percent: However; whereas the growthin the basic research component of thedefense budget is relatively small; the basicresearch components of the other principalagencies that support universities have been

relatively well insulated against the generaldecrease in civilian R&D budgets. Therecould, however, be substantial impacts onspecific scientific disciplines of particularinterest to DOD as shifts in research priontiesresult in changing support allocations.

The paper argues that DOD's increasedsupport for graduate students in particularfields of science and engineeringthroughboth fellowships and research assistant-shipscould have a substantial effect onthe entire U.S. R&D system; including theuniversity and industrial sectors: This aug-mented support is intended; impart; to helpresolve the problem the armed services havebeen experiencing recently in recruiting andretaining engineers as well as scientists_inseveral critical subspecialties. Despite DOD'sStudent support programs, the pool of highlyqualified scientific and technical personnelin critical fields is unlikely to be sufficient tomeet the demands of both the defense andthe nondefense sectors during the net fewyears. For that reason, there may well be acontinuing and increasing competition forscientific, and particularly, engineering talent.In view of the current defense buildup, thiscompetition could have a deleterious effecton both universities and nondefense indus-tries; and thus on the overall science andtechnology base required to maintain long-term U:S: national security:

Training and Utilization of EngineeringTechnicians and Technologists

Concerns about both the quantitative andthe _qualitative adequacy of scientists andengineers in_the defense and civilian sectorshave been widespread for a number of years.Factors that threaten to undermine the qualityof U.S engineering education havd also beenexamined in detail. Likewise, deficiencies insecondary education have been widely a-clvertised and considered at length in severalrecent reports, most notably those of theNational Commission on Excellence inEducation and the National Science Board'sCommission on Precollege Education inMathematics; Science; and Technology:However, the quantitative and qualitativeadequacy of the technicians and technologistsrequired to support scientific and engineering

ix

activities has received little comparablenational attention.

A primary objective of this paper is toplace questions regarding the training andutilization of these support personnel withinthe broader contextual issue of assuring thatthe overall scientific and technical workforceis adequate to meet both present and long-term national goals. Three potentially seriousproblems associated with technician trainingand utilization are identified. First, there is aconsiderable mismatch between the typesof training and skills that have been acquiredby the existing technical workforce and theskills demanded by industiy. In particular,many technicians, as well as engineers,appear to be underutilized. Second, thereare serious pressures on technician traininginstitutions, including community collegesand proprietary schools, that are broadlysimilar to those currently plaguing engineeringschools. These pressures include chronicfaculty vacancies and lack of state-of-the-artinstructional apparatus. They threaten toerode the quality of the educational offeringsof the institutions. Third, reliable informationabout the supply and demand for techniciansand technologists is inadequate. Theseinadequacies are exacerbated by the factthat the job skills required for technicians inparticular industries are often poorly defined.Thus; it is difficult to place questions abouttechnician training and utilization within thecontext of national scientific and engineeringmanpower goals.

The paper suggests that industry has acentral role to play in clarifying training andutilization opportunities for engineeringtechnicians and technologists. However, newcooperative initiatives involving industry,education, 'government at all levels, andprofsional associations appear to be essen-tial if the national need for an adequatetechnical workforce is to be met.

Scientific and TechnicalInformation Policies

The major problems explored in the nexttwo papers in the compendiumNationalSecurity Controls and Scientific informationand issiies in Scientific and Technical Infor-

mation Policy have emerged as compellingpolicy issues because of their relationshipto two of this Administration's major goalsfor science and technology: (a) relying onadvanced technologies to strengthen theNation's defense capabilities; and (b) delin-eating more clearly the appropriate rolesand responsibilities of the public and privatesectors with respect to science and technologyactivities: Issues associated with scientific andtechnical information are unusually complexfrom both a technical and a policy perspec-tive because they pervade a variety of policyspheres often far removed from science andtechnology.

National Security Controls andScientific Information

One of the byproducts of the Administra-tion's overall policy of enhancing the Nation'sdefense posture has been a well-publicizeddebate focused on the Government's ap-parent willingness to restrict the disseminationof scientific information to foreign nationalsand to limit their participation in R&Dactivities in an attempt to stern the outflowof U.S. military technology and sensitiveinformation. Concerns and uncertaintiesabout policy in this area have become _apotential source of tension that couldundercut the efforts made by the Departrnentof Defense during the past few year toreestablish closer ties with the universityresearch community.

The issues posed by the Administration'sgeneral policy of increasing control over thedissemination of scientific and technicalinformation involve a fundamental valueconflict between basic national securityconcerns and the need to preserve academicfreedom. This paper argues that the resolu-tion of these issues requires a clearer under-standing of whether the various controls nowin use, or under consideration; will be effectivein preventing a loss in this Nation's tech-nological lead over the Soviet Union; andwhether those controls can be implementedwithout imposing unacceptable economic,administrative, scientific, or political costs tothe American scientific enterprise.

The debate over the control of scientificinformation has both substantive and pro-cedural dimensions. The Government

8

acknowledges that whereas the Soviet Unionhas acquired a great deal of advancedAmerican technology by both open andcovert channels, very little has come throughsuch normal modes of scientific communi-cations as university graduate courses, lab-oratory visits, conferences, and publications.However, the Government is concernedthat in the future the Soviets will more con-sciously exploit the openness of the U.S.system of research and graduate educationto gain access to advanced defenserelatedtechnology:

The university community; while con-ceding that the application of existing con-trols has as yet created few real problemsfor research or education; is concerned thatthe ;,aguely defined, open-ended characterof that system could cause-problems in thefuture. The terms of the Export Adminis.tration Ad, for example, could be interpretedas restricting dissemination of scientificinformation directly applicable not only tomilitary technologies, but also to so-calleddual-use technologies. More seriously,there is concern that more stringent restric-tions could be applied to the disseminationof a much broader class of scientific infor-mation in an attempt to limit the outflow oftechnologies to U.S. economic as well asmilitary competitors.

The paper reviews some of the funda-mental issues that need to be addressed toclarify the terms of the current debate andmove toward a resolution: These issuesinclude defining "national security;" thegeographical scope of controls, what tech-nologies are to be controlled; and what typesof scientific information and scientific activityunderlying those technologies are to becontrolled. Procedural questions dealing withwhat forms controls should take and howthey are to be adapted, enforced, andmodified also need to be addressed. Regard-ing this latter set of questions, the paperconcludes that broadly drafted rules ofgeneral applicability are likely to generatemore serious frictions between the Govern-ment and the research community than thosethat can be defined on a case-by-case basisand tailored to a particular setting or aparticular technology. If broader- rules areused the burden for making decisionsabout their applicability falls heavily on the

research community; with narrower restric-tions, those decisions are either in the handsof the Government and/or are negotiablein advance.

The paper argues, finally, that an enhanceddialogue between the Government and thescientific community and an increased reli-ance on contractual restnctioni in federallysponsored research offer the greatest promisefor balancing the conflict between maintainingnational security and preserving maximumscientific openness.

Issues in Scientific and TechnicalInformation Policy

The control of scientific information fornational security purposes is only the mostvisible of a host of complex issues on thenational agenda that are associated withscientific and technical information policy.This paper suggests that policymaking con-cerning scientific and technical informationcannot be considered in isolation. First, sincethe generation and dissemination of scientificand technical information is clearly an integralpart of R&D policy, decisions in this areashould be closely linked with R &D prioritiesand management. Second. scientific andtechnical information policy is but one aspectof the much larger domain of informationpolicy and management. These linkagesresult in complexities in delineating andresolving science and technology informationpolicy issues.

Two sets of issues are reviewed in thepaper: those having to do with access toscientific and technical information; and thoseassociated with the economics of that infor-mation. Both sets of issues have becomeincreasingly urgent as a result of the rapidgrowth and convergence of computer andcommunications technologies.

Access issues include the perennial prob-lem of how to protect the confidentiality ofpe.sonal data gathered as a result of researchactivities such as statistical surveys, psy-chological examinations, and epidemiologicalstudies. A closely related issue has to dowith protecting proprietary rights to com-mercially valuable scientific information.Traditional patent, copyright, and trade secretcopyright laws may, in the case of computersoftware, provide inadequate ownership

xi

9

protection, and thus discourage industry frommaking adequate investments in softwaredevelopment. By comparison, in such rapidlydeveloping fields as biotechnology thatdepend heavily on progress in basic research,the overzealous protection of proprietarytights can limit scientific communication andthereby inhibit the development of the field.

A final access issue is concerned withpossible U.S. responros to restrictions im-posed by several other nations on the flowof information across their borders. The papersuggests that while there are many goodreasons why the United States should notemulate these restrictive policies, it is im-portant that U.S. foreign policies deal withthem more effectively. For that reason, theeconomic and/or cultural rationale under-lying those policies needs to be more clearlyunderStobd.

EcOnornic issues associated with scientificinformation have assumed particular im-portance because of the Reagan Adminis-tration's policy of returning responsibility forthe development of commercializable prod-ucts and services to the private sector. Oneunresolved problem is how to set a price oninformation. collected _ or generated_ by theFederal Government that recovers a reason-able fraction of the cost, while assuringequitable access to that information: A closelyrelated problem has to do with protectinghe private sector from unfair competitionav the Federal Governmerg, while assuring:hat the quality and accessibility of data bases:hat have been the responsibility of theFederal Government are maintained. Centralo the resolution of these issues will be findingmays to place a monetary value on informa-:ion products and ervices, and to determinevhich types of information serve the broadPublic interest and which serve narrowernterests more appropriate for private sector;ommercialilation-

The paper reviews the debate on thesessues, highlighting the major arguments inavor of, and opposed to, further limitinghe Federal role in generating, organizing,rnd disseminating information. It concludeshat there is a pressing need for a systematiceassessment of pricing policies for Federalnformation services, particularly since those,olicies differ considerably among Govern-nent agencies. More generally; the paper

suggests that authority and responsibility fordealing with scientific and technical infor-mation issues within the Federal bureaucracyneed to be more clearly defined. Concurrentwith such a review and possible restructuringof Federal authority and responsibility in thisarea; awareness of scientific and technicalinformation policy issues needs to be en-hanced among Federal R&D policyrnakers:Likewise; rep: esentation of U.S. interests inscientific and technical information oughtto become a more conscious element ofour foreign policy.

Encouraging the Implementa-tion of New Technologies

The three paperS that constitute the finalset in the compendiumLegal and Regu-latory Implications of the Video Telecom-munications Revolution, Trends in Com-puters and Communication: The Office ofthe Future, and Fostering the Use of Ad-vanced Manufacturing Technology areexplicitly concerned with the effects ofemerging technologies on public policy; and,reciprocally;_ with .economici---social;--andpolitical factors that affect the adaptationand implementation of promising new tech-nologies. Understanding these factors isparticularly important in view of the Admin-istration's policy to encourage the acceleratedcommerciali7ation of high-technologyproducts and services and by so doing toincrease the productivity and internationalcompetitiveness of American industry.

Legal and Regulatory Implications ofthe Video TelecommunicationsRevolution

This paper presents a particularly timely casestudy in the way new technological capabilitiescan challenge the rationale underlying along-standing public policy.

The basic premise for the CommunicationsAct of 1934, which still serves as the center-piece for Federal regulation of broadcastingand cable; was the presumed scarcity ofcommunications frequency bands in theelectromagnetic spectrum: This scarcityassumption was entirely appropriate in 1934when foreseeable communications tech-

10

nologies rested on broadcasting within arelatively limited region of the spectrum. Inview of the assumed scarcity of frequencyspace, the framers of the CommunicationsAct agreed that a totally free market wouldresult in a highly chaotic situation that wouldin turn distort the development of the broad-cast industry. The Act therefore sanctionedregulated monopolies by directing thatspecific frequencies should be licensed to alimited number of radio (and later television)stations. In return for the privilege of limitedcompetition, the activities of licenscd stationswere to be regulated in the public interestby a new agencythe Federal Communi-cations Commission (FCC).

During the past decade, new communi-cations technologies have all but made mootthe scarcity assumption on which the 1934Communications Act rests. This is particularlytrue for video telecommunications, wherethe advent of cable; microwave, and satellitecommunications technologies has led to aphenomenal increase in the number ofinterference-free broadcast channels availablein a particular geographical region: The paperargues that, as a result, the former Gov-ernment-sanctioned and regulated regimewithin which the broadcast industry-develz---oped is in the process of becoming a com-petitive, unregulated regime.

The paper considers four broad policyissues raised by this transition from a relativelyuncompetitive video market to one charac-terized by competition and abundance:

How should entry into the video market-place be regulatedfor example, howactive should the Federal Governmentbe in licensing, franchising, and tech-nical standard-setting?To what extent and by what meansshould video content be regulatedforexample, should the Fairness Doctrinerequiring a balance of presentationson controversial issues be maintainedin a nonmonopoly environment?To what extent should the FederalGovernment encourage a truly com-petitive video marketfor example;what should it do about concentration;cross ownership, and joint venturesamong media operations?What dangers to democratic rights maybe posed by the video telecommunica-

11

tions revolutionfor example, invasionof privacy and the widening of infor-mation gaps between rich and poor?

All four issues involve complex economicand social factors. The paper posits that themost likely scenario for their resolution duringthe remainder of the decade Is a shift inboth the substantive focus and the locale ofvideo regulatory action: According to thisscenario, emphasis will move from a concernwith licensing standards toward maintaining acompetitive market. Since policy debates anddirectives would then be carried on withinan antitrust framework, the locale of regu-latory activity would most probably shift fromthe FCC to the Congress and the courts.

The regulatory framework for video tele-communications, involving as it does theinterplay of a number of uncertain economic,social, and technological factors, is changingas the available technologies evolve. Theextent to which any individual new technologyis ultimately adapted and implemented forsuccessful commercial use will depend inlarge measure on the development of thatregulatory framework.

Trends_iniComputers_and Communi-cation: The Office of the Future

According to the Annual Science and Tech-nology Report to the Congress: 1982, theAdministration intends to focus its R&Dsupport more sharply than in the past inthe service sector, in recognition of its growingimportance to the U.S. economy. 3 Recenttechnological advances have considerablepotential to increase productivity in the servicesector. In particular, the use of computerslinked by communications networks couldlead to dramatic changes in the ways thatoffice functions are performed and, therefore,in the ways that officesand the largerinstitutions they serve in both public andprivate sectorsare organized and managed.The relatively recent introduction of free-standing word processing facilities hasalready led to some changes in office prac-tices, although not, as yet, to the dramaticgains in white collar productivity that manyexperts had foreseen: One probable reasonis that word processors have often beenregarded primarily as labor-saving devicesand introduced with little or no advanced

XI

Dlanning about how to reorganize an officeSr redefine tasks to make optimum use of:heir capabilities:

This paper describes a range of currentlywailable technologies. all of which are likely to)e implemented in the office of the future.Mese include word processors with sub-;tantial stand-alone capabilities that can:ommunicate with one another via telephoneinks; facilities to handle electronically all:orrespondence within an organization asvell as a good deal of communication)etween organizations; and teleconferencingacilities to replace face-to-face meetings. The)pportunities offered by these innovationso develop entirely new approaches toJrganization, management, and control3rorruse dramatic gains in service sector)rOductivity. However, the fact that behavioralind organizational factors are crucial to themplementation of computer-based officeechnologies may be a principal reason forheir unexpectedly slow acceptance in thenarketplace.

The paper discusses some of the factorshat have impeded the implementation ofhese technologies. It also explores some ofhe implications for the nature and organi-ation-of-office-work-in-the-futureand-forhe skills that will be demanded at variousevels within organizations if and when theseechnologies become commonplace. The)aper concludes by identifying a number of!conomic and social problems that are likelyo emerge if office automation proceeds asinticipated. while noting that the nature andeverity of these problems will depend onhe pace of implementation. Whatever that)ace may be, the demands on industry andin educational institutions to define and toirovide the technical skills required of futureiffice personnel are likely to be particularlyiressing.

'ostering the Use of Advancedlanufacturing Technologyhe adaptation and implementation ofvailable and emerging technologies forommercial use are particularly important) manufacturing industries. During the pastevade; several new technologies have madeossible fundamental changes in manufac-Iring processes that could lead to significani

IV

enhancements in the quality of manufacturedproducts and to an overall growth of pro-ductivity and international competitivenessin the U.S. manufacturing sector. Substantialcontributions were made by American engi-neers to the development of many of theseinnovative technologies, including robotics,computer-aided manufacturing, group tech-nology, and flexible manufacturing systems.Yet the United States has lagged behindother industrialized countries in the diffu-sion and implementation of advanced man-ufacturing systems. This lag has led, in part,to the erosion of our international competi-tive position.

This paper argues that most analyses ofthe relative failure of U.S. industry to incor-porate advanced process technologies havebeen flawed by an almost exclusive concernwith macroeconomic factors: In contrast,factors that determine whether and how thesetechnologies are implemented at the factorylevel have been largely ignored. Advancedprocess technologies ought to be regardedas complex sociotechnical systems ratherthan as pieces of hardware: It follows thattheir implementation into manufacturingprocesses often involves major changes incorporate-strategies; organizational-design,and human resource requirements.

For example, implementation of a roboticssystem cannot be carried out effectively bysimply replacing human workers withmachines. Rather, making effective use ofrobotics may require that entire productionprocesses be redesigned and new types oftasks assigned to workers on the factoryfloor. Such modifications most often requirethat workers be retrained in new skills. Equallyimportant, successful implementation of anadvanced robotics system may require theactive participation of workers in operationaldecisionmaking, with significant implicationsfor corporate management practices. Thus,implementation of these and other advancedmanufacturing technologies involves inter-actions among different functional units withina firm, with effects that may be radical ratherthan incremental.

The paper argues that current efforts toencourage the implementation of advancedmanufacturing systems in U.S. industry areprobably inadequate, given the inherentdifficu lties in the technology transfer process.

12

Federal efforts have been uneven,particularlysince much of the Federal R&D effort ismission oriented. Universities, while reason:ably active in the R&D underlying advancedmanufactunng technologies, have largelyignored dissemination and implementationproblems. The paper concludes that a sys-tems orientation by both Government andprivate industry will be required if U.S.industry is to take full advantage of thepromise of these new technologies.

While the last set of these papers dealsexplicitly with policies to accelerate theimplementation of new technologies, eachof the other papers in the compendium dealswith parts oi the broader question of howthe United States can make more effectiveuse of its superlative capabilities in scienceand engineering to achieve its nationalgoals. AS all nine papers make clear, attain-

merit of this broad objective will require, inthe words of the Annual Science and Tech-nology Report to the Congress: 1982, "..aStrong !partnership among Government; in-duStry, and academia; with each understand-ing iti responsibilities, assuming them to thefulleSt, and carrying them out completely."4

Refeiefies

1. National Science Foundation. Emerging Issues inScience and Technology. 1981. NSF 82-27. Wash-

. ington; DC: National Science Foundation, June1982.

2. Office of Science and Technology Policy. AnnualScience and Technology Report to the Congress:1982. Washington. DC: U.S. Government PrintingOffite, 1983.

3. Ibid., Chapter I.4: Ibid :, Chapter I.

International Cooperation in Science:The U.S. Role in Megaprojects

Abstract

A promising option for reducing the strain on Amencan scientific resources is cooperationwith other countries in those international activities that demand big staffs, complex man-agerial arrangements; and large budgetsthe so-called "megaprojects." In the future, these"big science" efforts increasingly may be pursued within an international division of laborframework emphasizing the identification of particular areas of expertise or concentrationsof special facilities that provide unique opportunities for the U.S. scientific community.The greatest potential benefit from such cooperation is improved economic efficiency, butthe financial advantages of joint action are supplemented by the promise of importantgains in the innovation process itself and by the possibility that political good will amongthe participants will be enhanced. But scientific cooperation on a large scale is not withoutits obstacles and constraints. Bureaucratic inertia and resistance must be overcome, suf-ficient planning capabilities must be linked to the cooperative venture; and systematicevaluation of the anticipated and realized costs and benefits of the project must take place.Moreover, the fragmented and largely incoherent approach used to date in formulatingand implementing cooperative scientific programs must be modified. At least four sets ofpolicy issues must be faced and overcome: choice of areas of cooperation; choice ofpartners, choice of organizational and managerial mechanisms, and choice of fundingarrangements. Resolving these issues appears to depend upon the ability to develop a database defining the international division of labor in science, to generate greater coordinationamong the various national science organizations, and to be more explicit about the fundinglevels for cooperative projects.

Introduction

The more than 20-year era in which theUnited States was able to dominate theglobal science and technology arena bypursuing an aggressive, broadfront assaulton the entire menu of scientific and tech-nological alternatives appears to be over.Several factors have contributed to the endof American dominance of internationalscience and technology.

First; the long-term condition of reducedrates of economic growth; combined withinflation, has constrained the investment offunds in both the human and the physicalresources available for scientific and tech-nological activities. For example, as earlyas 1981, analysts of the U.S. research anddevelopment (R&D) system were observingthat a "fundamental turning point" had beenreached with regard to Federal funding.Nondefense R&D has been faced with tighterbudgets, and many initiatives have beeneither cancelled or deferred)

14

Second, U.S. leadership in scientific andtechnological fields has given way to shared,or even lost, leadership as the Westerndemocracies and other states have recon-structed and developed in the postwarperiod. This is particularly the case withregard to the OECD (Organization for Eco7nomic Cooperation and Development)economies, but increasing competition alsocomes from a number of Corm-nunist coun-tries and a few industrializing less developednations? The OECD states, especially WestGermany and Japan, have expanded dra-matically their intellectual and productivecapabilities since World War II, and thesestates have gone further than the UnitedStates in exerting a measure of centralizedcontrol over the formulation and imple-mentation of science and technology policies.As a consequence, they have become muchmore competitive in international trade.3

Third. the science and technology policypriorities of most OECD states, includingthe United States, have undergone substantial

1

changes over the past two decadeS. In thiScountry, the pressure to reestablish ecb:nomic groWth rates and to create employEment opportUnitieg has Meant that industrialinnovation increasingly hag come to beseen as the most crucial, and most ignored,nondefense science and technologypriority.`' Compare this orientation with the1970s focus on such priorities as environ-mental protection or social developmentand set-vices.

Recognition of these limitations, implicitin actions of the Federal Government forsome time now has been officially acknowl-edged: For instance; George A-. Keyworth;Director of the Office of Science and Tech-nology Policy (OSTP); has said of the emerg-ing difficulties facing American scientific andtechnological initiatives:

As I have stated on other occasions, thereare a number of good reasons why wecannot expect to be preeminent in allfields;nor is it necessarily desirable:idea that we can't be first across thespectrum of science and technology isnot simply a function of our currenteconomic situation: The fact is that im-mediately after World War II this countrywas alone in developing and pursuingtechnology. Since then the rest Of theworld has been catching upwith muchhelp from us .5

Taken together; these factors have madeinternational cooperation in science andtechnology an especially attractive option:Not only can joint action reduce the strainon American resources; but the capabilitiesof other advanced; industrialized countries;and occasionally those of underdeveloped:ountries, are welcome assets in the pursuitof the benefits of science and technology.Indeed; President Reagan himself has citedhe desirabilitY Of such effOrtS and the need'or this country to identify the "most fruitfulareas of cooperation. "6

International collaboration in Science andechnology encompasses a variety of glib=,tantive activities. These range from support)f military and political alliances throughhe use of more applied R&D, to verynformal linkages among members of the3Iobal scientific community concerned withhe advancement of knowledge and the most

7

basic aspects of research. Such activities arepursued through many different organiza-tional and managenal arrangements, in-cluding bilateral or multilateral governmentalrelationships, or the use of internationalorganizations. Many different participantsperform a variety of roles in these cooperativeventures. The most significant actors arenational governments, private corporations,and universities?

There appears to be some consensus inthe American scientific and engineeringcommunity that one of the most promisingopportunities for U:S: involvement is ininternational "megaprojects," or socalled"big science." These are projects in suchareas as high-energy particle physics, outerspace exploration, or geodynamics re-search that require extremely elaborateequiPrnent and facilities and large teams ofprofessionalsrequisites that typicallydemand complek organizational and man-agerial mechanisms, usually multilateral incharacter, and a variety of participants,including substantial inputs from the privateSector and univerSitieS.9

A good example of a megaproject is theEuropean Center for Nuclear Research(CERN) plan to build a new particle accel-erator as part of its high-energy physicsprogram. This large electron-positron (LEP)storage ring is designed to speed subatomicparticles around a circular tunnel 16 mileslong (by comparison; the Fermi NationalLaboratory's facility in Illinois is 4 miles inlength). The first phase of the new CERNaccelerator is estimated to cost some $610million, and the second phase about $120million more. The final cost might be ashigh as $1 billion. Perhaps 250 physicists,from both Western and Communist scientificestablishments, would be involved in theeffort For the United States, the cost ofparticipation is projected to be some $20million. If it is approved, the LEP wouldrepresent the largest American commitmentof this kind. Given constrained domesticfinances, the LEP proposal has been termedan "acid test" of U.S. involvement in coop-erative international science projects.1°

Uncertainties about the size, scope, andskill mix of megaprojects like the LEP arethe greatest incentives for American involve-ment and; at the same time; the most

Significant sources of opposition to inter-national cooperation. On the one hand,supporters of collaboration argue that thecost of U.S. unilateral action in bi_g Scienceprojects increasingly is prohibitive. Moreover,many proponents of cooperation have beencritical of the traditional "pork barrel" wayin which some funds have been allocatedamong U.S. scientists in big science areas.On the other hand, opponents of cooperationstem the underfunding of American projects,and they ask difficult questions about howthe costs, risks, and benefits of collaborationare to be determined and distributedamong participants)'

Faced with such sensitive and controversialissues, U.S. policymakers have been searchingfor a more systematic way to choose appro-priate targets of opportunity. Increasingly,the Reagan Administration is approachingsuch choices in the context of an "inter-national division of labor" framework. Thatis, there is an attempt to identify specialareas of expertise, concentrations of particularskills or equipment, or other characteristicsof the global scientific infrastructure that poseattractive opportunities for U.S. involvement.As OSTP has argued:

We must now think in terms of an inter-national division of labor; where achieve-ments in one place can complement thosein another. Through cooperation withother developed nations, we can achievea more efficient distribution of the burdenof scientific and technological researchon a world Scale and provide access forU.S. scientists to special or unique facili-tieS abroad that would be prohibitivelycostly to reproduce at home)2

Although hardly a new ideait has beensuggested for years that greater internationaluse should be made of certain nationalfacilitiesthe use of this approach as adecision framework still is new to this country.Nevertheless, enough is known to make itclear that an international division of labororientation has some key advantages. First,it incorporates the requisite of budgetarydiscipline into the choice of appropriatetargets of opportunity. Second, because itfocuses attention on the unique resourcesof other societies, it holds the promise ofsimplifying somewhat the process of making

16

decisions about potential partners, institu-tional mechanisms, and areas of cooperation.Third, it makes explicit a criterion, economicefficiency, by which choices can be made.Clearly, defining operational evaluative criteriaappears to be a very important requirementwhen attempting to assess such complexscientific undertakings as megaprojects.

The purpose of this paper is to ask: Whatissues are likely to confront the division oflabor approach to American participationin scientific megaprojects? And how can theseissues be resolved? Following a backgrounddiscussion of the opportunities, difficulties,and conditions for success in internationalscientific cooperation; the emerging issues .

likely to pose obstacles for U.S. involvement inscience megaprojects are outlined: Finally,.some policy options for contributing to theresolution of these issues are delineated:

Background

The substantial opportunities and benefitsassociated with international cooperation inscience always must be weighed against aSet of very real difficulties and costs, andthe balancing of these factors places nationaldeciSionmakerS in Something of 6 dilemma.This is because in an international Setting,national governments almost- always havemultiple policy goals and objectives regardingthe performance of the Scientific enterprise;and these goalt and objectives often are indirect conflict. To cite only the most obviousexample, a balance must be struck betweenthe use of research to support domestic firmsin the international marketplace and thereliance on cooperative ventures to reducethe high costs of acting In isolation. In short,decisions must be made regarding the inter-national costs and benefits of cooperativeversus competitive science projects:13 Thus,the opportunities and limitations of coop-eration almost always are diverse and tosome degree a function of a particulardefinition of national interest, the hierarchyof public policy priorities, and the perceptionof international comparative advantage. Itis possible, however, to make some gen-eralizations about the relative benefits andcosts of international scientific cooperation.

3

The major opportunities and benefitsgenerated by involvement in cooperativescientific endeavors are as follows:

Making greater resources available, interms of information, knowledge, andknow-how necessary for any scientificactivity;Making possible a wider range of topicsand a broader range of approaches;Reducing the financial burden on allparticipants:Speeding up the entire innovation proc-ess; from basic research to application;Reducing wasteful redundancy; andEnhancing good will and communication among the participants."

These benefits fall into three broad cate-gories: the facilitation of the innovationprocess itself (the first two points), financialadvantages (the next three points), andthe provision of political opportunities (thefinal point).

The development and maintenance ofgood will among partners often is cited asan important political advantage of inter-national cooperation, although this set ofbenefits is of the lowest priority and the mostdifficult to calculate. Highly symbolic aspectsof megaprojects, such as "space handshakes"between U.S. and Soviet astronauts, havebeen justified in terms of political benefitsof this sort. In addition to the assumptionthat cooperation will improve transnationalunderstanding and build an appreciation forand tolerance of various cultural and philo-soohical factors; it is assumed that scientificcooperation can serve as a useful diplomatictool. In other words; collaboration in scientificactivities often is viewed as a useful way tosignal approval of the actions of anothernation, while the withdrawal of cooperativeprograms is presumed to send importantmessages of condemnation)5 Such was thecase with the limited curtailment of Americancooperation with the Soviet Union in thewake of the invasion of Afghanistan. Theare, however, only signals and symbols. Nomatter how valuable they are in the politicalcontext, such advantages of scientific coop.eration are unlikely to provide the basicrationale for engaging in expensive mega-projects, especially in a future likely to bedominated by constrained resources:

Far more significant are the benefits thatcooperation provides to the innovativeprocess itself. In the absence of collaboration,the attainment of a critical mass of expertiseand funding may be impossible: As a con-sequence, some projects will not be under-taken and others will be scaled down. Forexample; the International Geophysical Yearwould have been an unthinkable undertakingwithout global cooperation, Also; a varietyof participants makes available a broaderrange of analytical approaches, methodol-ogies; and research techniques than wouldotherwise be the case. No nation has amonopoly on scientific ingenuity, and theunique resources and talents of many smallercountries are not likely to be utilized in themost effective mariner unless cooperativerelationships are established.

Most attention in the United States hasbeen devoted to the direct financial benefitsof megaprojects. As outlined above, theseinclude the reduction of time eliminationof some duplication of effort, and thespreading of costs over a larger group ofactors. In the past, significant benefits haveaccrued to the United States in each ofthese categories. For example, the StateDepartment cites task sharing with foreignscientists and laboratories in the lunar sam-ple analysis program as having saved theAmerican taxpayer more than $5 million.Similarly; visits to laboratories and ex-changes of technical data in the U.S.-JapanNatural Resources Program were creditedwith saving an estimated $100,000 to$150,000 by helping American scientistsavoid research duplication.16

Several important caveats regarding theseobvious financial advantages must be noted,however. First, empirical evidence on thepayoff from most cooperative activities issketchy, and often the economic benefits ofsuch projects are not subject to quantification.Second, cost overruns in the managementof megaprojects are serious potential liabilitiesand must be taken 'into, account in anyfinancial analysis. And third; despite theproblems encountered with duplication ofeffort; a certain amount of redundancy inthe scientific enterprise clearly is desirableas a crosscheck on the validity of resultsand findings.

17

The most significant difficulties and costsassociated with international cooperation inscience are:

Inherent difficulties in meshing disparatenational bureaucracies;Delays in reaching decisions amongdiffering political and legal systems;Complications of varying decisionprocesses, priorities, and competencies;Costs of international bureaucracy;The danger that political inertia, whichmakes projects hard to start, but evenharder to stop, will dominate;The possibility of drains on nationalresearch budgets because of interna-tional commitments;The tendency to undertake, interna-tionally, only low-priority projects; andThe apparent conflict between coop-eration and improving a nation's com-petitive position.'7

In one way or another, each of these costsof cooperation has to do with the dynamicsof national and international organizations,especially bureaucratic ones; Barriers to theimplementation of megaprojects include ahost of differences among national decision-making procedures, consent mechanisms;and legal frameworks. The degree of cen-tralization of science policymaking, or thecomprehensiveness of national scientificplanning are only two examples of thefactorS that help determine how well par-ticipants in cooperative ventures are able tomesh their bureaucracies.

An even more troublesome obstacle tocooperation is posed by the very differentnational science policy priorities that mustbe integrated into a megaproject. Clearly,nations have very different rationales for andrankings of .basic goals and objectives, suchas the desire to maximize industrial innovationor the need to maintain the national knowl-edge pool, and these divergent priorities posemajor constraints on cooperation. This isespecially the case given the fact that thesegoals and objectives seem increasingly tobe linked to nonscientific foundations. Asthe OECD has observed:

(T)he process of policy formulation forScience and technology increasinglyappears to be based, in some countries

18

at least, upon broader political currents.This is so for a number of reasons but itsimplication is that the policy process isgradually coming to reflect individualnational political values and traditions,rather than something generally charac-teristic of science and technology's

Thus; the general policy style or politicalculture of a country may be a more importantdeterminant of science policy than anyidentifiable importance attached to thescientific enterprise itself or to any specialbenefit assumed to result from the supportof that enterprise. This is not to argue thatthere isn't considerable consistency betweenOECD Sci1/4.1 ice policy priorities, but that thosegoals and objective§ are far from identicaland are shaped by the individual economic,foreign, and social policy concerns of eachcountry. For example,- even the currentemphasis on stimulating induStrial innovationwithin the developed world masks importantnational differences in emphasis, and goy:ernment R&D expenditures in this areaactually have been on the decline in a number,of countries (Belgium, Sweden, and theUnited Kingdom, for example).'s

Given these kinds of complications, it isno surprise to find that bureaucratic inertiais a major concern in megaprojects. Inaddition to the constraints outlined above,four other factors seem to contribute to delaysand higher costs in big science efforts: therequirement to spread financial and man-.power resources over a number of scientificproblems at the same time; the long timeframet within which such projects operate,Which generally means that managementhas trouble Specifying the duration of anyparticular activity; the difficulties involved inmaximizing interaction among a range ofacademic disciplines and other professions;and the rigidity that seems to accompanyprojects overseeing large groups of expertsfrom diverse national backgroundS.2°

The problems encountered in organizingcooperation contribute to the higher totalcost of international projects, although, asnoted above, a significant benefit to eachparticipant is the lower cost. But even thereduced financial burdens of involvementin a collaborative effort may be resisted by

5

domestic research agencies: As in anybureaucratic setting, specialized knowledgeand research funds are major sources ofpolitical powersources that may not easilybe committed to international ventures ifsuch a policy commitment means a transferof budgetary or personnel authority to aninternational organization or a differentagency: This tendency to protect bureaucraticterritory cannot be overemphasized, and, ina time of intense competition among agenciesfor funds, the skepticism aboUt intemationalactivities naturally is heightened. All this helpsto explain why Cooperative efforts tend tobe assigned relatiVely low priorities in mostOECD science polia..s and why internationalcollaboration -generally is placed at a severediSadVaritage in the bureaucratic policy-making process.

Finally, while certainly there is no checklistOf how to succeed in international scientific-cooperation, it is useful to specify some ofthe lessons learned in this arena: A reviewof the literature has revealed the followingimportant conditions:

Intergovernmental cooperation must bebased upon an awareness of the politicalcontext; and the further the programmoves toward applied research, themore precise the political irriPlfEationsmust be.It is important that there should besimilarity between partners, both in sci:entific and technical development, andin economic development.Aims of the joint action must be clearlydefined at the outset.A general preparatory mechanism forcontact and discussion is necessary tolaunch, define, and mount the jointeffort.A detailed cost-benefit analysis of variouspotential institutional frameworksshould be conducted:Direct cooperation between nationalestablishments is preferable to the cre-ation of an international body.A balance between equity f returns inrelation to investment) and efficiency(entrusting work to those most corn-petent to perform it) must be reached.Adequate mechanisms for supervisionand responsibility in monitoring andManagement must be provided.

6

The international program should notcompete with national programsitshould complement thern.Red tape must be minimized and thedelegation of responsibilities maximizedBudgets should extend over a numberof years to ensure financial stability.21

Essentially, these factors point to thenecessity of anticipating the relative costsand benefits of cooperation systematicallyand comprehensively. Planning appears tobe a requisite for success in cooperativeprojects. Also important is the ability tobalance political and scientific variables in apragmatic way. Of course, an underlyingassumption of these conditions is that thereis some consensus about what constitutes"success" in an international scientific en-deavor; In fact; such agreement has beendifficult to achieve. Debate about the mostappropriate indicators of project successappears to be endless, with some analystsarguing that the scientific significance of theresults produced by the prOject is the onlylegitimate guideline, and others pressing forthe application of broader Standards. Thus,it has been suggested that the degree towhich the continuity of scientific activity in acountry has been enhanced by a projectis the best criterion for determiningproject success.22

About the best that can be said for thecurrent state-of-the-art of project evaluationis that something is known regarding thecomponents that many nations appear toconsider as being conducive to internationalpartnership. Thus; scientific projects thatsucceed as international cooperative venturesare typically related to subjects that transcendnational frontiers; are costly, have long-rangeobjectives rather than commercial aims, andcorrespond to the political objectives of thecountries involved.23 Moving beyond thesefactors to more systematic analysiS requireSmore explicit evaluative criteria. It is herethat the Reagan Administration's divisionof labbr approach ultimately may have itsgreatest appeal. Because this orientationmandates an efficiency standard and a hardlook at the scientific capabilities of otherdeveloped societies, success may be easierto defineeconomic payoff in the area ofindustrial innovation; for example: But im-plementing such a approach will not be

easy. A number of emerging issues must befaced and resolved if American involvementin megaprojects is to succeed within a divisionof labor approach.

Emerging Issues

Many U.S. cooperative programs in sciencesuffer from the absence of any coherentstrategy. The approach used to date, in whichdecisions are based as much on the overallforeign policy climate as on any particularscientific considerations, accurately reflectthe primacy of politics. But it is also aconseque ice of the very real problemsinherent in long-range planning for scienceand the dangers of overplanning in anenvironment that changes as much and asrapidly as science:24 Because today's inter-national scientific programs increasingly aredominated by the problem of funding;25 this`%shotgun" approach no longer is viable.Changing this pattern of policymaking,however, will require coming to grips withsome of the most intractable formulationand implementation issues. A big sciencefuture based on an international division oflabor will face at least four Such issues: choiceof substantive areas of cooperation, choiceof partners, choice of organizational andmanagerial mechanisms, and choice offunding arrangements.

Choice of Areas of Cooperation

The most fundamental issue -confitiiiiirig_U.S,participation in megaprojects is the difficulty ofspecifying future opporttinities. Science policyfrequently is shooting at a moving tat-get,and the ability to identify those internationalprojects that will extend and complementdomestic activities is made more complexby the need to take into account not onlyour own dynamic national interests but alsoforeign capabilities (facilities; personnel, andother resources) and the changing sciencepolicy priorities of other countries.

Getting a reasonable picture of scientificcapabilities across national boundaries isrelatively easy; and this is where most of theattention to date has been focused.Deter-mining how these capabilities are likely tobe utilized in a hierarchy of constantly

changing science priorities is not so easy.This is as true of broad shifts in emphasis,Such as the loWer priority attached to socialobjective (health or pollution research) bymost OECD States today as compared tothe mid-19706, as it is of more subtle changesin goals, like the higher ranking of regionaleconomic development in thy science policiesof many industrial societies. Yet the abilityto anticipate and monitor such changes inpriorities, especially in OECD countries, iscrucial for big science efforts. Clearly, shouldthe goals and objectives of these potentialpartners diverge too radically from our own,cooperation would be made more difficult.But too much convergence in priorities alsocould be dysfunctional. If industrial innovationand economic growth have become thecentral concerns in most of the developedworld, as appears to be the case; cooperationmight take a back seat to pressures forcompetitiveness. According to Tisdell:

Several governments and societies see thestrategy of increased international com-petitiveness of domestic industries en-couraged by appropriate governmentSSET (science and technology) policiesas a means to solve unemployment,reduce inflation and increase economicgrowth. Countries such as Japan andGermany appear to have used suchpolicies successfully. They can work butthey are not certain to do so. Further-more, the more countries that indulgein these policies the greater the chanceof the policies not being successful inthe world as a whole. They are not ex-plicitly beggar-my-neighbor policies butthey could become so in an inflexibleeconomic world. Thus new difficulties forthis realpolitik strategy could anse on aglobal scale even ignonng the possibleadverse long-term effects on the environ-ment, the depletion of resources and thesocial fabric of society.26

To avoid this poSSibility, more compre-hensive and systematic priOtity assessmentsin the area of science policy are needed.Not only do we need better information aboutthe social methods by which Science prioritiesare set In various countries, but the mak-imization of cooperative projects requiresbetter understanding of the ways in which

7

iormal goals and objectives are translatedinto funding allocations in science budgetsin other societies?'

- Even assuming comprehension of therange of national Science priorities and thatthese goals and objectives are, on balance,compatible with bur own, there are additionaldangers associated with the current thrustof Arriericari cooperative science strategy.BecauSe a diviSion of labor approach stressesbuilding upon the strengths of the existinginternational science system, there is thepotential that scarce resources will be con-centrated in a few established big sciencefields, like high-energy particle physics. Giventhe past appetite of megaprojects for moneyand expertise, resources may not be availablefor cooperative ventures in less prestigiousfields such as environmental protection orthe social implications of science. It can beanticipated that the great value now placedon cooperation that has the highest potentialfor leading to direct economic benefits orthat facilitates industrial innovation will onlywork to reinforce this tendency.

Choice of PartnersThe OECD states provide the most attractiveopportunities for Amencan involvement inmegaprojects._Since similarity between par-ticipants on scientific, economic, and politicaldimensions/ appears to be a crucial requisitefor project success, the other Western in-dustrialized democracies naturally have beenviewed as the logical U.S. partners in coop-erative science ventures. Because WesternEurope, Canada, and Japan have capabilitiesclosest to our own, there is more under-standing of the potential targets of oppor-tunity, and there is substantial experience inscientific collaboration within the OECDframework and related global institutions.For some time, many of the most productiveU:S. cooperative programs have involvedOECD nations. Examples of such ventureswould include the 1979 U.S.-Japaneseagreement to cooperate on fusion research,and the 1980 French:American oceanog-raphy cooperative prograrn.28 A division oflabor approach to future scientific collabora-tion appears to offer promising payoffs withtheSe countries. There are however, someimportant issues raised by focusing America's

.

8

international science policy on potentialinteractions with other Western nations:

In the first place, enthusiasm about thebenefits to be had from megaprojects maynot be as high in other democracies as it ishere. There is ample evidence, for example;that public spending on big science hasleveled off and even declined in many OECDcounties:29 Even within a cooperative frame-work, there may be the perception thatmegaprojects spread national resources toothin and that too many other opportunitiesare foregone as a consequence of channelingof scarce resources to big science. Moreover,there are doubts throughout the OECDsystem about the level and stability of thevarious national commitments to internationalprojects. This is as true of the United Statesas of any other OECD rnernber State. As arecent study by a committee of the AmericanCongress pointed out, "the United Statesappears to have international cooperationless on its mind than most of the nations;developed or undeveloped, with which itdeals."3° This point has been amplified bySkolnikoff, who has argued that:

Successful cooperation also requiresreliable partners. The record of the UnitedStates in modifying or abrogating agree-ments makes future agreements harderto reach. Most recently, the proposals tocancel the coal liquefaction developmentproject with Japan and Germany and towithdraw from the International InstituteOf Applied Systems Analysis have dam-aged our reputation as reliable partners 31

If scientific cooperation is to succeed, par-ticularly in megaprojects, participants mustsomehow develop more consistent, longerterm commitments; despite the Very realexigencies of domestic and foreign politics.

Finally, it is not clear how an internationaldivision of labor decision framework will affectthe ability of less developed societies toparticipate effectively in megaprojects. It iscertain that a number of the newly indus-trializing states, such as Mexico and Brazil,will play important roles in future globalcooperative ventures: But because the re-sources for research are so concentrated inthe developed world; there is a danger thatsignificant Third World capabilities will beoverlooked in the effort to maximize the

scientific talents of the West. Much betterinformation about the scientific resourcesand policy priorities of less developed coun-tries is needed before major Third Worldinvolvement in a division of labor frameworkcan be anticipated. A comprehensive inven-tory of the availability of scientific facilities,equipment, materials, and expertise in theunderdeveloped world would help to resolvemany of these uncertainties.32 Even with thevery substantial obstacles posed by differencesin levels and rates of economic and scientificdevelopment, with the threat of politicalinstability, and with the pressing needs formore applied research and development inthe Third World, opportunities for collab-oration exist and are ignored in this countryat some cost and risk

Choice of OitaiiiiatiOnal andManagement Mechanisms

The performance to date of megaprojectsand of the U.S. science policy machinerysuggests several important issues regardingcooperative organizational arrangements.First, there are a set of issues having to dowith the appropriateness of existing U.S.planning and policymaking structures forfuture collaboration in big science. If mega-projects are to be tailored for specific eco-nomic and scientific payoffs, then an efficiencycriterion appears to demand a fairly com-prehensive and possibly somewhat centralizedoversight and review capability at the nationallevel. This tendency is made more dramaticby the consensus in this country's sciencepolicy establishment that cooperation isfacilitated by the use of a relatively largenumber of bilateral and multilateral arrange-ments. The conventional wisdom is that thesemodes of organization are easier to designto more precisely match national intereststhan is the use of existing internationalorganizations or the creation of new inter-national bodies.

In such a system, planning and oversightbecome crucial to minimize fragmentation,duplication of effort, and all the otherbureaucratic ills discussed above. Unfor-tunately, the American national science policyapparatus has received mixed reviews of itsperformance in this area While there areclear indicators that the role of science in

22

American foreign policy has grown sub-stantially in the last several decades, there isless evidence that this has resulted in acoherent program in the national interestCritics of the U.S: Government's leadershipand organization in science policy emphasizethat while there are important advantagesassociated with our decentralized system 33there is evidence that our approach oftenleads to poor coordination of initiatives;conflicts in policies and goals, inadequateparticipation from the science community,and a weak sense of international mission.34Each of these shortcomings may prove tobe damaging to a division of labor orientation.Organizing complex big science projectsaccording to anything like the conditjoiii----for success outlined above (the use ofextensive preparatory mechanisms, wide-spread application of cost-benefit analysis,systematic delegation of management re-sponsibilities and provision of stable, long-term budgets) will be difficult in the currentsystem: All this does not imply the need formassive reforms through the extensivereorganization of existing agencies; nor thecreation-of-new-bodies-ea-lolly the lutes, forexample; of the proposed Institute for Sci-entific and Technological Cooperation); but aclearer delineation of the responsibilities forcooperative policy formulation and imple-mentation within the OSTP/Department ofState/National Science Foundation frame-work is needed.

The second set of issues refers to themanagement of big science projects them-selves. Here, the major problem in the pastand the most significant bamer to the successof megaprojects in a resource-constrainedfuture appears to be the inability to establishand maintain satisfactory mechanisms toinvolve nonuniversity elements of the privatesector: One of the most attractive aspects ofan international division of labor in sciencepolicy is the potential for tapping moresystematically the capabilities of industrialfirms at -home and abroad. To date it hasbeen difficult to link corporate resources tocollaborative projects in ways that are in thenational interest.35 Especially troublesomeis the tension between cooperation andcompetition in the intemational marketplace,and the resulting danger of a rise in pro:tectionism among the Western industrial

9

states. Resolving this issue will require comingto grips with the very different scientificpriorities of private firms (the emphasis onmore applied activities, for example), as wellas recognizing the significant disparities indecisionmaking styles and mechanismsamong multinational corporations andsmaller companies.

Choice of Funding Arrangements

Underlying many of the institutional andmanagerial problems encountered in mega-projects is the issue of funding: For sometime, American cooperative programs havebeen dominated by the problem of securingstable financing in an era of budgetary conflict.Each of the most common methods offinancing U.S. collaborative efforts (directpayment by each side of its own costs, eachside paying for all in-country expenses,payment by the country that benefits, etc.)has advantages, but each also poses barriersto cooperation. Most significantly, thesemethods reduce flexibility, induce relativelyshortterm outlooks, and expose intemationalprograms to budget cuts because of difficultiesin specifying benefits. As a consequence,several reforms have been suggested. Theseinclude the creation of an interagency fundfor international cooperation in science andtechnology, the establishment of a specialfund to promote certain cooperative efforts,and the creation of a separately fundedorganization for scientific and technologicalcooperation Each of these proposals triesto reduce the vulnerability of cooperativeprojects by making it easier to demonstratethe relative merits and liabilities of suchactivities and by stressing the strategic sig-nificance of collaborative ventures.

Budgetary reform also would contributeto the resolution of many of the organizationaland managerial issues outlined above. Anyrationalization of the planning process inthe national science policy institutions is tosome extent a function of improvements inthe ability to assess costs and benefits ofcooperation, and, as noted, this is difficultto do in the existing financial managementsystem.

Policy OptionsInternational cooperation in big science hasmoved up the hierarchy of priorities on the

10

national policy agenda, as the benefits andopportunities created by collaboration arecontrasted with the economic realities ofattempting to pursue unilateral projects. Asa representative of the U.S. Department ofState has put it:

Indeed, the advantages appear to be com-pelling enough to suggest that cooperativeapproaches should not be regarded asexceptional; rather, their potential andpromise should be routinely consideredas research plans are formulated:37

Yet, the attractiveness of international col:laboration has not prompted a rush towardthe maximization of the potential benefitsof joint scientific ventures. It appears at ifwe continue to view the promise of coop-erative projects as exceptional rather thanroutine. In this environment, opportunitieswill be missed until and unless policy uncer-tainties, such as those outlined above, areaddressed. At least three policy options mayhold some promise for helping to resolvethese uncertainties for future cooperationin big science.

Data Base Defining the InternatiotillDivision of Labor

Before we are fully able to operationalizean international division of labor decisionframework we must know what it looks like:A major requisite for future advancementsin megaprojects; therefore; appears to bethe generation and compilation of informa-tion regarding the range of scientific capa-bilities in place or likely to be developed inthe world; and more importantly; the variouspriorities likely to be attached to thesecapabilities by the diverse nation st.tes. Wesimply need much more comprehensive dataabout scientific institutions, personnel, re-sources: and needs in the OECD states,Communist countries, and the nations ofthe Third World. And, as discussed earlier,we urgently require systematic informationon the ways vinous countries are likely torank these capabilities in the short-, mid-,and long-term future. Only if the UnitedStates has available this kind of compre-hensive data bank can the international divi-sion of labor in big science move beyond rhet-oric to an operational decision framework.

Of course, much of this information alreadyexists, but it has not been collected andcompiled in such a way as to permit antici-patory choices. Our data regarding capabilitiesand priorities is so limited and fragmentedthat we have been forced to operate in areactive; ad hoc manner: We need a masterroster of al! potential partners; including thosewho for various reasons are today considered"beyond the pale" (for example; Cuba; NorthKorea, or Vietnam) 3s that lists estimates ofcurrent and projected expenditures on arange of scientific priorities; and identifiesoverlaps and gaps with our own situation:

It is essential that this data base go beyonda characterisation of national governmentalcapabilities to include what is known,aboutthe use of laboratories, research institutes,and other mechanisms for, carrying out re-search Overseas. We have learned that theOECD everience in these areas offers somehelpful suggestions for our own case,39 and

uncle statidingofglobal patterns could expand our optionsdramatically. Similarly, no attempt to definethe international division of labor in sciencewill be complete unless better data is madeavailable on the resources and expertisecurrently in existence and likely to be mobi-lized by the private firms operating abroad:

Coordination of LIS; CooperativePrograms

A move toward some greater coordinationof U.S. cooperative programs is implied inthe need for a data base defining the inter-national division of labor in science: Improvedinformation on potential areas of cooperationand possible partners in megaprojects willbe useful only to the degree that there is amore coherent system for building consensusand making choices at the national level. Atthe same_ time, it must be recognized thateach of the major Federal science policyorganizations has a strong sense of its missionand that previous attempts to improvecoordination have been of limited success.The National Science Board has recog-nized the need for far more active Federalcoordinating mechanisms. In its "Statementon Science in the International Setting," atits September 1982 meeting, the Boardargued that:

24

Agencies such as the National ScienceFoundation, as well as universities andnongovernmental professional scientificorganizations, will each have unique andimportant contributions to make towardthe success of cooperative internationalscientific activities. The National ScienceFoundation, by virtue of its fundamentaland broad -based Scientific program,should take the initiative, in cooperationwith the Department of State and otheragencies as appropriate, to bring togetherpotential international partners to accom-plish the necessary planning and imple-mentation for international sharing orcollaboration in fundamental science andengineering research ^0

And the Board had the following to sayabout putting this recommendation intooperation:

The nature of science requires that itsinternational dimension be considered anorganic aspect of the scientific enterprise:This dimension must be actively providedfor in all Foundation programs; fromeducation and fellowships to the variousdisciplinary efforts in the natural sciences;social sciences, and engineering. Planningfor new facilities and the setting of priori-ties for major scientific investigations andprograms should be carried out with thefull recognition of the priorities of othercountries and in an environment whichencourages complementarity or plannedsupplementation, cost sharing, andcoherence of the various efforts of coop-erating countries: National ScienceFoundation organization and manage-ment procedures should reflect theseprinciples 4'

Most of our potential partners in big scienceendeavors already have moved beyond thismodest effort to centralize national policyfor international science; and at least somereforms along this line appear to be a requisitefor this country to be able to take meaningfulaction when the benefits of cooperationmanifest themselves.

Line Item in Each Agency Budget forInternational Science ActivitiesCalls for incorporating international coop-erative programs into the internal bureaucratic

decisionrriakitig process in an "organic" waywill remain just talk until we are more explicitabbut the funding levels for these activitiesamong all Federal agencies. Because politicalp- wer ultimately is linked to the budgetaryprocess. establishing international scienceas a separate line item would have theadvantages of defining the fiscal dimensionsof our international cooperative commit-ments. developing a more refined calculusof the savings and benefits to domesticprograms of our international activities, andremoving international cooperative initiativesfrom their second-class status and therebyreducing their vulnerability to budget cutsin the future.42 Science policy cannot helpbut be improved if we remove the currentuncertainty aboUt the funding levels ofinternational cooperative programs as awhole, and the funding commitments ofspecific agencies.43 We simply must baseour decisibriS on more coherent understand-

5.

6.

7.

8.

in f th t pent-on-vanous_collabg e amoun s-sorative ventures and the rationales underlyingthese expenditures. Otherwise, attempts todevelop systematic preparatory mechanisms;including the more widespread use of cost-benefit analyses. and the expansion of ourcapabilities to evaluate cooperative projectson a comparative basis will continue to bebased as much on faith as on empiricalevidence of the success or failure of jointscientific endeavors.

Notes and References

1. Logsdon, John M. -Introduction: The ReSearchSystem Under Stress:" In John_M,. Logsdonied.).The Research System in the 1980's: Ptiblie Policy

Philadelphia: Franklin Institute Press, 1982.pp. 1.11; and Shapley, Willis H.. Teich, Albert H..and Breslow; Gail J. Research & DevelopmentAAAS Report W. Washington. DC: AmericanAssociation for the Advancement of Science,1981, p. 17.

2. President's Commission for a National Agendafor the Eighties. Science and Technology: Promisesand Dangers in the Eighties. Washington. DC:U.S. Government Printing Office: 1980.pp, 17.27.

3. Tisdell, C. A. Science and Technology Pri-

britieS of Governments. London: Chapman andHall, 1981. pp. 106.148.

4. Organisation for Economic Co-Operation andDevelopment,Science andTechnotogy Poky for_ _

the 1980s. Pads: Organisation for Economic Co.Operation and Development. 1981. pp. 33.37.

12

Keynote Lecture by Dr. George A. Keyworth,OirectOr, Office of Science and TechriblogyExecutive Office of the President, and Science andTechnology Adviior to the President, at the 148thAnnual Meeting of the American Association forthe Advancement of Science, Washington, D.C..January 3. 1982, p. 6 (mMeographed). For aMore general review of the Reagan Administra-tion's approach to science policy, see Barfield,Claude E. Science Policy from Ford _to_Reagan:Change and Continuity. Washington, DC: Amen'can Enterprise Institute for Public Policy Re-search. 1982.Letter of Transmittal for Science. Technology,and Anted-can DiPlontricy. 1982. the Third An-nual Report Submitted_ to the Congress by thePresident Pursuant to Section 503(b) of Title Vof Public Law 95-246. Washington, DC: U.S. Gov.emment Printing Office. 1982, p. ix.National Science Foundation. "Advancing U.S.National Interests Through International Coop-eration in Science and Technology." In EmergingIssues in Science and Technology; 1981. Wash-ington, DC: National Science Foundation, 1982,pp. 65.74.Tisdell, op. cit., p. 67: and Organisation for Eco-nomic Co-Operation and Development. op. cit.,n. 1-6R

9. Adams, John. "Reflections on a Big Science." Sci-ence and Public Policy, vol. 9 (April 1982):81.87.A good illustration of the ways in which the scopeof many megaprojetts demand international col-laboration is provided by geodynamics research.See Deake, Charles G and Muwell. John C. "Geo,dynarnics-Where Are We and What Lies Ahead?"Science. vol. 213 (3 July 1981): 1&22.

10. Broad, William J. "A $20-Million Test of Coop-eration:" Science; vol. 217_(20 August 1982): 710.712: and Organisation for ECOnomic Co-Operationand Development. op. cit., pp. 126.127.

11. Ibid.;and Robinson. Arthur L. "CERN Gives Nodto Four_ LEP Detectors." Science. vol. 217 (20AugiiSt 1982): 722.

12. Office of Science and Technology Policy in coop-eration with the National Science Foundation.Annual Science and Technology Report to theCongress: 1981: Washington; DC: U.S. Govern-ment Printing Office. 1982. P. 137.

13. Bobrow, Davis a_and Kudde, Robert T. "EnergyR&D: In Tepid Pursuit of Coliective Goods."international Organization. vol. 33 (Spring1979): 151.

14. Statement by Nelson F. Sievering. Jr., AssistantAdMinistrator for International Affairs: EnergyResearch and Development Administration, Beforethe Environment. Energy, and Natural ResourcesSubcommittee of the House Committee on Gov-ernment Operations. May 12, 1977. p.. 3: andOrganisation for Economic Co.Operation andDevelopment: Energy R&D. Paris:-Organisationfor Economic Co:Operation and Development.

___ 1975; p: 154.15. Chinn. Herman I. "International Cooperation in

Scientific and Technical Research." Paper pre.

pared for the National Science Foundation.December; 1979. pp. 12-13 (mimeographed).

16. Department of State. International ScientificCooperation-A Summary of Tangible Benefits.Washington. DC: U.S. Department of State, 1974,PP. 3.5.

17: Skolnikoff, Eugene B. "Science. Technology andInternational Security: A Synthesis." In Teich;Albert H: and Thornton, Ray (eds:). Science.Thchnology, and the Issues of the Eighties. Boulder,CO. Westview PreS, 1982, p. 139.

18. Organisation for Economic Co peration andDevelopment; Science and Technotogy Policy forthe 1980s. op. cit.. p. 19.

19. 'bid,. p. 34._20. di Castri. Francisco, Hadley. Malcolm. and

Damlarnia_n, Jeanne, 'The Ecology of an International Scientific Project." Impact of Scienceon Society. vol. 30-no. 4 (1980): 247.260.

21: Orgahis-atitih for Economic CoOperation andDevelopment. Science and Technology Policy for

_ _ the 1980s. op. cit., pp. 153-156.22. di Castri, op. cit pp. 257-258,23. Kerwin; Larkin: "International Science-An Over-

view." Science, vol. 213 (4 September 1981):1069-1072.

24. Press. Frank "Science and Technology in the WhiteHouse 1977 to 1980 science, vol. 211 (16

Outlook: Problems. Opportunities and Constraintsin Science and Technology. Washington; DC:U.S.- Government Printing Office, 1980, pp.292:293.

33. National Science Foundatort: Science USA/USSR. VOluthe I: Science Policy in the United$tates, Washington; DC: U.S: Government Print.ing Office, 1980, pp. 2.4.

34. Denver Research Institute; Center for PublicISsues. Report of the WOrkShop: New U.S. Initia.ayes in international Science and Technology.Denver. CO: University of DenYer. 1977, pp. 11-12.

35. Granger, John V. Technology and InternationalRelations: San Francisco: W. H. Freerrian, 1979,pp. 42.47.

36. Department of State. Bureau of Oceans and In;temational Enyironmental and Scientific Affairs.Funding of US: Government Science TechnologyCooperation With Other countries." May 1931.pp. 10.13 (mimeographed):

37. "United States Statement on International Coop-eration by the Honorable James L Malone. AS.sistant Secretary of State (Designate) for Oceansand International Environmental and ScientificAffairs." Meeting of the CoMmittee for Scientificand Technological Policy at _Ministerial Level;Organisation for Economic CP-Operation andDevelopment, Paris, March 19-20, 1981, p. 5

January 1981): 249-256.25. Holden, Constance. "Science Budget: Coping with 38.

Austerity," Science. vol. 215 (19 February 1982): 39.944.948.

26: TiSdell. op. cit., p. 203.27. Ibid., pp. 1.30; and Organisation for Economic

CoOperatioh and Development. Science andTechnology Policy for the 1980s. op. cit.. pp. 40.26!29:

28. Office of Science and Technology Policy in coop.eration with the National Science Foundation, op.cit.. P. 138. 41.

29. Tisdell, op. cit., p, 192: 42.30. U.S. HOuse of Representatives. Committee on

Science and Technology. Survey of Science andTechnology Issues: Present and Future. Wash- 43.ington, DC: U.S. Govemment Printing Office;1981; p.

31. Skolnikoff, op. cit., pp. 1397140.32. National Science Foundation. "Institutions for

International Cooperation." In The Fiue-Year

(mimeographed).Chinn, op. cit., pp. 27.28.National Science Foundation. "The lnternatiOnalContext of U.S. Science and Technology." In The5. Year OutlookortScience and Teehnorogy. 1981.Washington, DC: U.S. Government Printing Office.1982. pp. 18.25."Statement on Science in the International Setting."Adopted by the National Science Board at h 238thMeeting on September 16;17. 1982. p. 3(mimeographed).Ibid., p. 4.Department of State. Bureau of Oceans and In,ternational Environmental and Scientific Affairtop. cit., pp. 1-11."Comments of the Congressional Res0..atth Serviceon the 1982 Title V Report on Science. Tech.nology, and American Diplomacy." In Science.Technology, and American Diplomacy, 1982, op.cit., pp. 331.332 (see reference 6).

13

Trends in CollectiVe Stria.' Research

Abstract

Collective actions among commercial firms to promote Scientific research are based on anumber of objectives and take a variety of forms. These forms include trade associations thatsponsor research to strengthen a particular industry, groups Of ffirms that work togetherwithin the structure of a Federal agency, and consortia involving several firms within aparticular industry and one or more universities. While older types of collective industrialresearch were motivated almost exclusively by factors internal to an industry, recent collectiveactions have also been stimulated by concerns about international competitiveneSS, productivity,and the need for trained manpower. The structural and financial arrangements of the morerecently established research consortia also suggest that they are becoming an importantcomponent of the national technical effort. Although their total expenditures are smallrelative to total industrial R&D funding, a substantial portion is earmarked for basic re-search in universities.

Several issues associated with collective industrial research are considered in this paper:Will such activity be a significant repla-cement for Federal R&D support?In addition to research, will theSe efforts address the projected shortage of technicalmanpower trained to work in selected fields?Will such activity set the direction for national R&D efforts in the fields affected?What is the relationship of collective efforts to international competitiveness?Can this activity provide a significant increase in university support with regard toresearch, training, or equipment?In cases where universities are directly involved, what are the implications for university/industry institutional ties?What are the main concerns related to patents, licensing, royalty income, and anti-trust regulations?

Introduction

The rapid growth of U.S. research anddevelopment since the 1940S haS occurredprimarily along two separate, though related,streams. One is that of federally Supportedprograms devoted to broad national objec:tives, including support of the underlyingbasic science and engineering structure. Thesecond consists of the sum of those researchand development activities conducted byindividual corporations, constituting thenational industrial research effort. Thus, whileGovernment programs have reflected theR&D needs of society as a whole; allof the industrial effort relates to the resourcesand objectives of individual corporations,_

Almost all, but not entirely. In addition tothe individual efforts each corporation under-takes to pursue specific interests and goals;there are numerous examples where two ormore companies, usually in the same indus-try, have established a working relationship

27

in support of a technical activity withoutany formal link to subsequent commercialexploitation. This type of collaborationinvolving several companies which form anassociation to engage in a technical researchor training effort is commonly referred to ascollective industrial research.

Collective actions are based on a numberof objectives and have taken a variety offonmS. The most obvious category of such arelationship is the trade association, someof which sponsor_ research in various insti-tutions to Strengthen a particular industry.There are examples of groups of companiesworking together within the sanitizing struc-ture of a Government agency. These includethe early stages of the National AdvisoryCommittee on Aeronautics, or NACA, thepredecessor of NASA, and industry sponsor-ship of visiting scientists at the NationalBureau of Standards. Another category ofcooperation is a mechanism for interactionbetween a group of companies in a particular

15

industry and one or more universities. Therecently established Semiconductor ResearchCooperative, a research affiliate of theSemiconductor Industry Association, illus-trates this type of cooperation. And thereare otherS.

What then is new? And what aspects ofcollective industry research are of particularinterest and importance today? Conventionalwisdom suggests that the structural andfinancial arrangements of these associationsmay well imply that they are becoming animportant component of our national tech-nical network:

What appears to be happening from aslirvey of groups is that the focuse,I qualityof these programs; their operational charac-teristics: and the amount of funding tosupport such activities are, in fact, likely tohave a considerable impact on the country'stechnical base. For example, in 1983, fourgroups alone (Electric Power Research In-

stitute. Gas Research Institute, Council forChemical Research, and the SemiconductorResearch Cooperative) will devote approxi-mately $400 million to these programs.Although this will constitute only about 1percent of all industrial R&D funding, asubstantial portion will be earmarked for basicresearch and much will go to universities.This reflects a considerable allocation ofindustry funds to these areas and suggestsimplications for new university/industryinstitutional ties in research and training.

Other considerations that have stimulatedrecent efforts in collective action by industryare concerns about international competi:tiveness, productivity: and trained manpower.The contribution of these cooperative pro-grams to strengthening the baSe of a majorindustry and, relatedly, the base of a technicalarea or discipline is thus a factor to evaluate.

In brief; these activities may affect the directionand; in part. the nature of technical activi-ties conducted within industry and withinthe university.

Given the level and purposes of collectiveindustrial research, there are a number ofissues for consideration:

Will such activity be a significant replace.merit for Federal R&D support?In addition to research. will these effortsaddress the projected shortage of

16

technical manpower trained to work inselected fields?Will such activity set the direction fornational R&D efforts in the fieldsaffected?What is the relationship of collectiveefforts to international competitiveness?Can this activity provide a significantincrease in university support with re-gard to research, training, or equipment?In cases_ where universities are directlyinvolved, what are the implications foruniversity /industry institutional ties?What are the main concerns related topateritS, licensing, royalty income; andantitrust regulations?

These issues are receiving increased atten-tion as a result of broad national concem fornurturing technical leadership and economicvitality. The purpose of this paper is not tooffer a comprehensive treatment of the sub-ject. Rather% it is intended to provide anoverview of the scope of activities and to indi-cate areas requiring more detailed research.

Scope of Collective IndustrialResearch

This section reviews the characteristics andobjectives of several types_ of collectiveindustry associations and their principalactivities. The overview of origins and goalsOf each industry group reflects industryspecific features such as the competitivestructure of the industry, degree of regulation,Capital requirements, manpower needs; andrelevant antitrust restrictions. However; theunderlying objective of all groups is toaccelerate the pace of technical change byeither conducting research in targeted areasand/or increasing the number of trainedpeople required. Clearly; whether the em-phasis rests primarily on research or train-ing. there is a reinforcing effect of one onthe other:

Within the past decade, and noticeablywithin the past few years, there have beennew forms of collective industry R&D activitiesestablished by industry sectors not previouslyinvolved and, in some respects, for newobjectives. One important characteristic ofthese newer structures is the magnitude ofeffort; far abOve the typical earlier develop.

28

ments. As mentioned earlier, four of theorganized activities=the Electric PowerResearch Institute, the Gas Research institute,the Semiconductor Research Cooperative,and the Council for Chemical Research willspend an aggregate total of over $400 millionin 1983.

Thus, in our preliminary survey, we havecategorized the collective industry associationswithin two groups: (a) recent developmentsorganizations established within the lastdecade and (b) older instituhonsorganiza-tions established more than 25 years ago.

Recent DevelopmentsExamples_ of recently established organiza-tions are the Council for Chemical Research(CCR), the Semiconductor Research Coop-erative (SRC), the Electric Power ResearchInstitute (EPRI), and the Gas Re SearchInstitute (GRI) (see Table 1).

Several characteristics of the Semicon-ductor Research Cooperative and theCouncil for Chemical Research are of par-ticular interest:

(1) Both are very recent developments,stimulated in large part by concerns for inade-quate basic research in each field and by aconcerted attempt among major companiesin each industry to respond effectively toboth near-term and projected pressures ofinternational competition in high technology.

(2) Both have related goals addressing theneed for trained manpower in each field.

(3) Both represent industries which arenonregulated and are highly competitive:

(4) The arrangements for collective actionwith the focus on basic research and openmembership are designed to avoid compli-cations with antitrust regulations.

(5) The amount of funding for each coop-erative is significant. Actions of CCR mayresult in a one-third increase in industrialsupport for academic basic research inchemistry and chemical engineering, andSRC may lacome "the largest single conduitfor industrial support.of university research."'

Collective action within the energy in-dustry reflects a different set of characteris-tics, in large part influenced by the role ofenergy in the economy, the adverse impactsof OPEC, and the effects of cutbacks inFederal funding.

The Electric Power Research Institute andthe Gas Research Institute thus representseveral special features of the energy industry.

(1) The requirements for capital and tech-nical resources for energy research, devel-opment, and demonstration are enormousand in many cases present an overwhelmingburden for any one organization. Thesecharacteristics, common to both the electricand gas industries, are reflected in the largemembership of electric and gas utilities inthese organizations to pool needed resources,and in a substantial level of R&D funding.

(2) Due to the regulation of rates for gasand electricity use through public utilitycommissions, the energy industry does nothave the same competitive structure as suchother industries as semiconductors orchemicals: Activity can thus be focused onthe development end of the R&D spectrumwithout extreme concern for infringementsof antitrust regulations:

(3) The pressure to work in developmentand demonstration is reinforced by twofactors: OPEC and Federal funding cutbacks.The pronounced emphasis of both GRI andEPRI on energy generation and efficiencyis an apparent response to the decade-long supply vulnerability and pricing policiespresented by OPEC. The effects of Federalcutbacks are also evident in the_programstructure of both organilationS. EPRI hasincreased its support for near-term devel-opment and demonstration projects from50 percent to 70 percent; GRI is revising itsresearch program to accommodate a 22percent reduction in Federal funds for itscoordinated funding activity.

Older Institutions

Several older institutions and trade associa-tions have also engaged in collective industrialresearch. The magnitude of these efforts,however, in terms of membership andamount of funding are modest in comparisonto the newer organizations. They appear tohave addressed the specific needs of aparticular industry in a specialized area ofresearch and/or training, and the level ofeffort over time is apparently still satisfactoryto the participating members. Table 2 pro-videt an overview of two of these older

17

Table 1

Recently Established Organizations for the Conduct of

Collective Industrial Research

INDUSTRY ASSOCIATION COUNCIL FOR CHEMICAL RESEARCH la nonprofit organization)

GOAL To boost Industrial Ilnancial supeorl f_or_basic reae.Prellen campus al eherhistfland

chemical engineerin_g, and to ensure high quality advanced education in the chentical

sciences and engineering.

YEAR OF ESTABLISHMENT 1979.

MEMBERSHIP

SOURCES OF FUNDING

LEVEL OF FUNDING

FUNDING MECHANISM

MODE OF OPERATING

Over 120 universities and 35 companies representing the major chemical acid Petro

chemical industries 'figures as of 4/1/821.

Membership dues and voluntary participation in CCR's Chemical Science and Engl.

nearing Fund ICSEF1 designed to Increase industry's support of CCRs u_niversity

members, Based on a formula reflecting number of domestically employed chemists

and chemical engineers at each company.

Pledges of additional support to universities as of April 1982 equaled $3.6 million,

Companies participating in CSEF pledge their increase to the CCR but actually dis.

tribute the funds directly to the university of their choice. There is also a central-fund

that receives monies from industry based on a 25 percent commitment of the CSEF

pledge. This Central Fund Is then distributed to CCR university departments based on

a lormula reflecting Phil students graduated.

CCR Serves as a mechanism to promote and facilitate one-to-one interaction between

industry and university members. There is no peer review, no proposal process. and

details of work are handled solely by partners,

RESEARCH EMPHASIS Basic research in Chernistry and chemical engineering,

TRAINING ASPECTS

EQUIPMENT

INVOLVEMENT OF A

FEDERAL AGENCY

OTHER ACTIVITIES

COMMENTS

Interadons. to increase.research are designed to e-rihance ttaining. CCR maintains

current demand and supply data on chenitisfs_and_themical tippers.

Determined by individual partner relationships.

No.

Improved communication links for both information and exchange of scientific Per

sonnel between industrial and university laboratories.

CCR has potential of increasing the percentage of industry support lot university

research from 7 percent to 10 percent.

SEMICONDUCTOR RESEARCH COOPERATIVE la nonprofit research alliliale of the

_S_emico inktorindustry Association j

"Toiniintaln U.S.Jeedershito In serniarifitters. and c.omputers through 0 25.50.per.

cent. lriareaSe In.ePre_ research arid ID add significantly to the supply and qualify Of

degreed_oroltssionals."'

1981.

Me_rehant semiconductor comee_nies and their malor_an_d leading users le.... com

puler companies, instrument companies-, consumer produd companiesi. SRC is an

affiliate of the Semicond uctor Industry Association 150 compan les I,

Membership dues and funds contributed by members Ihrou_gh. a formula based on

total semiconductor sales or value of semiconductors incorporated in products.

Programs to begin in 1982, Funding for 1982.1983 estimated at $10 million and for

1983.1984 at $15 million. Could reach S40.50 million per year by 1986,

.

Funds distributed by SRC. To be concentrated in major generic areas and institutions

rather than spread out over heterogeneous subjects and universities.

Members of SRC will outline program areas and solicit universities lor competence in

each area. Universities will submit proposals to be reviewed by a technology staff

responsible lor technical strategy and planning, Initiation of contracts, and evaluation

of project results. Work will thus be performed by universities.

Areas too basic or too loncj,term for Individual indUstry R&D programs. Possible

specific areas: new techniques for imprinting circuits on silicon wafers. alternative

_semiconductor rnatezialUnd computeraided tircadesign,

Overall 'explicit goal to increase quality and supply CI professional personnel.

Designed to upgradenecessary equipment and to share skills of personnel trained to

operate equipment. The.extent of this commitment is reflected in the proposal for

equipment to receive twice as much binds as research..The.cost of equipment in

creases at an exponential_tate with a.3year.life cycle..whitti.thieatens the ability of

University to remain at the frontier at a field for an extended period of time.

No.

The .open.membersnpi of the SRC implies that_American.subsidiart of. Japanese

firms would be eligible to join, Representaives of the SRC; however; have indicated

that the SRC may establish apolfey for such loreigrnowned subsidiaries of requiring

reciprocal membership in counterpart organizations in other _c_ountries. This could

apply to present members of the SRC such as Fairchild Camera and Instrument owned

by Schlumberger, and Signetics owned by N.V. Philips al the Netherlands as weltas

Potential members representing Japaneseowned companies,

' Botkin, Oimancescu, D. and Slate, Ray. Global Stalk The F1 !vie of nigh Tech.

nofogy in America. Cambridge: Ballinger Publishing Company, 1982, p. 94.

INDUSTRY ASSOCIATION ELECTRIC POWER RESEARCH INSTITUTE la nonprofit national research and devel- GAS RESEARCH INSTITUTE la nonprofit national research program for natural gas

GOAL

YEAR OF ESTABLISHMENT

MEMBERSHIP

SOURCES OF FUNDING

blifileatProgramL supply and utilization)

"To advance capabilities in electric power generation, delivery, and use, with special To increase supply options for natural gas, Improve efficiency for utilization, enhanceregard for salety_efficiency, reliability, economy, and environmental considerations." service, and continue fundamental research.[Annual Report

1972. 1976.

Vetuntary membership_consists of over 600investor-owned, municipal, cooperative, 200 companies including interstate pipelines, distribution companies, and publiclyand _Federal _Utilities. Membership represents 70 percent of the electricity supplied owned municipal utilities,by U.S. utilities,

Membership_ duesbasedon each_membefs_annual sale of electricity. Aggregate pay- Gat rate-payer. Support is calculated on gas volumes sold by interstate pipelinements to EPRI in 1981 were just below $260 million. _ _ members and _oninierstale sales of member companies.

LEVEL OF FUNDING $215 million for contract research in 1981: $40 million for in-house work and pro- $68:5 ritillibit deVoted to R&D.gram management.

FUNDING MECHANISM EPRI's Research Advisory commillee of utility executives and technical stall guides theprogram priorities and funding allocations.

MODE OF OPERATING

RESEARCH EMPHASIS

Detailed strategic planning process is undertaken each year to review_tequirementsand developments of the utility industry. Some work is conducted in-house but most isconducted through contracts let to utilities, manufacturers, national laboratories, anduniversities.

GRI itself conducts no research. It establishes research priorities and program goalsand then contracts work to universities, energy companies, professional service firms,and a variety! of research organizations

GEO_ is regulated by the Federal- Energy - Regulatory Corifriattion IFERCI through aformal application proceeding whereby GRI sobrnitt_a5 -year plan of research andptoposed_sete surchages_each year for_ approval: FERC thus aUtheriieS both ratesan_d_program plans; alter this process. State public utility commissions also aUthorizethe rates.

TRAINING ASPECTS

EQUIPMENT

INVOLVEMENT OF AFEDERAL AGENCY

OTHER ACTIVITIES

COMMENTS

Nearly 70 percent of EPRI's funding is devoted to these near-term program areas(initial payoff is anticipated within 10 years], with the balance allocated between mid-term 110-25 years: 27 percent) and long-term projects lover 25 years:3 percent).

Not an explicit goal of EPRI.

Funds specialized equipment needed to perform work.

Congressional stimulation to establish EPRI in 1972. EPRI receives Department ofEnergy R&D funds.

32

_information dissemination, workshops, seminars.

Federal cutbacks have resulted in shift in EPRI's program structure to reflect greateremphasis on near commercial scale demonstration projects.

Areas and percent of 1981 R&D budget: al fundamental research, 5.6 percent: bj en-hanced service, 8.6 percent: c) efficient utilization. 39.6 percent: di supply options,46.2 percent.

Not an explicit goal of GRI, but funds to universities allow graduate_students_topursueadvanced degrees while performing gas-related research. 1980:9 advanced degrees.

Funded as needed in contract research.

Federal Energy Regulatory Commission grants approval for rates and program plans.

Information dissemination, seminars, workshops.

Primary emphasis on efficient utilization and supply options; least emphasis is onbasic research. Of total R&D budget, approximately $2 million went to universitieswith 53.2 million earmarked for 1982.

Coordinated- funding program_which inctuded Federal funds is being reduced to $75million in 1982 from 596 million in 1981 as a result of cutbacks.

33

INDUSTRY ASSOCIATION

Table 2Older Institutions for the Conduct of Collective Industrial Research

INSTITUTE OF PAPER CHEMISTRY 'independent, privately supported educational TEXTILE RESEARCH INSTITUTE (independent. educational and research organization]and research institution I

GOAL To_address specialized needs of the paper industry for professional talent trained in TO broaden the technology base of Vie textile iridatry with an emphasis on principlee,the field of paper chemistry. MethaniamS, and understanding rather than derrelopriterit of specific preducts and

processes

YEAR OF ESTABLISHMENT

MEMBERSHIP

SOURCES OF FUNDING

LEVEL OF FUNDING

FUNDING MECHANISM

MODE OF OPERATING

RESEARCH EMPHASIS

TRAINING ASPECTS

EQUIPMENT

1Nv3LVEMENTOF AFEDERAL AGENCY

OTHER ACTIVITIES

COMMENTS

31

1929.

Majority of U.S. producers of pulp, paper, and paperboard.

1j annual fees from member companies; 2) contract research performed by-the staffon a nonprofit basis; 31 scholarship and fellowship gifts; and 41 miscellaneous sources.

Annual budget: approximately $7 million per year.

Offers fellowships in scientific and technical areas related to paper processing andproduction.

The IPC is affiliated with Lawrence College in Wisconsin, and the Institute is charteredas a graduate school although Lawrence College grants either an M.S. or Ph.D. degree.

Research is under way in both lundamental and applied areas of interest to the paperindUary, ranging in subjects from forestry to waste treatment.

Major emphasis of IPC. About 90 percent of PC's graduates are employed in the paperindustry with the highett concentration in MD areas.

Funds equipment as needed.

No.

TheIPD _provides_aninfosm_alipn Service-Its_library is_regarded as the wain largestcollection of scientific and technical literature related to the pulp and paper industry_

Stimulus for establishment reflects lack of concentration in university_ curricula onpaper chemistry. Currently, approximately 25 students complete the Master's programand 10 to 12 pursue studies toward the Ph.D.

1930.

60 corporate participants in the textile_and related industries.

Pardcip_anttees-general unresticted_supportand grants; industry-supported research.Government-supported research: and publications.

Budget for 1981: $1.3 million.

Allocates funds internally for different program areas.

Most work conducted in-house with training conducted in collaboration with PrincetonUniversity. Has 5 areas of activities: research, education, services for corporate mem-bers,_e monthly jou, lalThe Textile Research Journal, and a technical informationcenter.

Basic research of industrial relevance In the physical and engineering sciences ofpolymer, fiber, and textile systems. There are live principal areas of current research:fiber structure. physical properties of fibers, dyeing and finishing, fiber assemblybehavior, and textiles In pollution control.

TRI has a long - established link with Princeton's Department of Chemical Engineeringto orient scientists and engineers to fiber and textile science and technology. Theprogram involves both students and faculty at Princeton, and TRI fellowships areawarded for thesis research on textile-related subjects. In 1981, there were 5 Prince-ton TRI Research Fellows and 2 undergraduate students. Fellowships accounted forabout $90,000 of the $1.3 million budget of operating expenses for TRI.

As needed.

TRI has received research grants from the Ehvironmental Protection Agency and theNatitthaLStiente_FOUndation.

A teehnital inferMatien Center, research services for members, and a monthly journalThe Taxtga Research Journal

institutionsthe Institute of Paper Chemistty(IPC); whose main focus is on education,and the Textile Research Institute (TRI),whose main focus is on research.

Trade associations have also been activeto varying degrees in research: As definedby the American Society of AssociationExecutives, a trade association is a "non-profit organization of business competitorsin a single industry, formed to render anumber of mutual aid services in expand-ing that industry's production, sales, andemployment."

Trade associations have primarily pursueda span of activities including disseminationof information to members on such issuesas Government policies and industrywideposition statements, sponsorship of con-ventions and courses, and lobbying of mem-bers of Congress on issues of particularinterest to the industry. However, the technicalunit of a trade association often compilesand distributes statistical data of interest tothe industry and may conduct or sponsorresearch: Most often the research of a tradeassociation relates to testing and standardi-zation of products and processes. Testingfacilities can be shared and some are locatedon university campuses:

Concentrated technical research; as distinctfrom activities related to testing and stand-ardization. has not been pursued extensivelyby trade organizations. The Motor VehicleManufacturers Association, the AmericanPetroleum Institute, and the American GasASSociation are exceptions. There are severalother examples of solid research sponsorshipby trade associations, but not of the mag-nitude evidenced by EPRI or even SRC.Thus, the metals industry supports researchthrough INCRA (International Copper Re-search Association) and ILZRO (InternationalLead and Zinc Research Organization). Otherindustries use collective contributions similarly,largely for support of modest grants toindividual university researchers. There area very few instances where the industryconducts collective research in its ownfacilities: These include the Portland CementAssociation and the Textile Research Institute:Such efforts appear to be far more extensivein Europe:

Alternate Forms of CooperativeResearch

Research Associates. Industries conductresearch through several other cooperativemechanisms. One is the Industrial ResearchAssociate Program established in 1921 atthe National Bureau of Standards (NBS) inwhich scientists and engineers from industrycan come to NBS to work with NBS staff.The Bureau does not develop programs orpursue areas specifically to Meet the needsof industry. However; there are often pro-grams ongoing at NBS that relate directlyto the interests of industry and thus are ofmutual benefit.

Each year approximately 100 researchassociates from industry participate in theprogram for an average stay of 1 to 2 wars.They are selected by the organizations withwhich they are affiliated and the selectionsare reviewed by NBS. The two basic condi-tions of the program are that industry paysall salary, travel, and related expenses of anassociate and that all work done at NBS isin the public domain.

Research associates at the Bureau comefrom both private industrial firms (e.g., BellLabs, Control Data, Exxon Research andEngineering, Lockheed) and trade associ-ations (e.g., The Aluminum Association,American Dental Association, Society forthe Plastics Industry); The overall researchfocus of the Bureau is to generate measure-ment techniques; calibrations; and statisticaldata and to conduct testing:

Currently, there is an effort to double thesize of the programs over the next 5 years;particularly in the areas of materials proc-essing. automation, electronics, and chemicalengineering. This can be significant because,as a result of reduced Federal support forNBS, such an expansion will optimize theuse of the research facilities and augmentNBS staff resources in several key areas.

Mission-Oriented Institutes. As anotherform of cooperative research, several mission-oriented institutes have been established topursue some area of key research to aparticular industry. The Massachusetts In-stitute of Technology (MIT)-Industry

21

Polymer Processing Program is an example:In contrast to the basic polymer industry,which is research intensive; the polymt:rprocessing industry has not engaged inextensive research: This area has also notreceived research concentration at univer-sities: Thus; the establishment of this programwas stimulated by the need to increaseresearch commitments in polymer processing,particularly those areas related to the man-ufacture of plastic and rubber products.

The program began operation in 1973with a 5-year seed money grant from theNation& Science Foundation experimentalresearch and development incentives pro:gram. Three member companies joined thatyear. The program now includes 10 mem-bers' and is entirely industry sponsored bymembership fees based on a formula ofeach firm's plastics output.

The level of effort consists of about 25projects directed by six members of the MITengineering faculty, operating with an annualbudget of approximately $500;000: Sixpatents have been issued to date as a resultof the program. with 12 applications pending.Corporate sponsors participate in the pro-gram by guiding research directions andgaining first access to research results. Allsponsors have royalty-free, irrevocable, non-exclusive license to use any technologydeveloped under their sponsorship.

Policy IssuesThe preceding material serves as an over=view to identify an activity that is increasingin size and variety and that has the potentialfor influencing the rate of technical changein the United States. Within this context,the issues presented in the first section ofthis paper are examined more closely andsome of the related impacts are discussed.

Will_such activity be asignificant replacementfor Federal R&D suppod?There are several instances where collectiveindustry groups can play a significant roleas the pattern of R&D support changes_butthis role is not likely to replace that of Fed-

`For example. GM. KOdak. Xerox, fIT. InstrumentationLaboratory. and Rogers Corporation.

22

eral funding. In 1981, for example. theFederal Government's share of total uni-versity R&D funding was 65 percent, whileindustry in all types of support provided3.8 percent.2 Additionally, the fundamentaltransition from Federal support to privatesector support in several key areas is likelyto involve a period of adjustment in whichthe impacts on R&D and on the pace of tech-nical change remain unclear.

For example, EPRI has revised the struc-ture of its research program to accommodatereductions in Federal energy support fordevelopment- and demonstration-scaleprojects. Whereas in the past few years near-term programs have accounted for 50 per-cent of total expenditures; EPRI is nowallocating 70 percent for near-commercial-scale projects: To accommodate this shift;however: many programs are being elim-inated (for example; support for all work onelectrical systems and energy storage tech-nology R&D) or reduced (for example, fossilenergy development, development of solarand wind energy. and research on healthand environmental effects). In addition, thiscomes at a time when electric utilities arefacing severe financial constraints. This hasresulted in an overall reduction in the scopeof EPRI's projects over the next 5 years;accounting for inflation, the R&D programwill continue at about the 1980 level ofreal expenditure.

In contrast, the actions of the Councilfor Chemical Research seem likely to increasethe percerit of industry support for researchin university chemistry departments from 7percent to-10 percen0 While this representsa one-third increase in the current level ofindustry support, it will not compensatedirectly for possible reductions in Federalsupport: However; what may be particularlyimportant here is the improved relationshipbetween universities and industry; based onone-to-one interaction; which may result ina more productive use of technical resourcesbetween the two partners.

Will these efforts address theprojected shortage of technicalmanpower trained to work inselected fields?It seems apparent that the university/industryrelationships of collective industry associa-

3 7

tions can serve to increase the quality andsupply of trained manpower. The Instituteof Paper Chemistry is a prime example ofa collective effort to meet the needs of aparticular industry. The newly formed Councilfor Chemical Research has, as an explicitgoal, an intention and funding mechanismtv Fn onioteadvancedecicati n in echemical sciences and engineering. Thefunding of research at universities by theSemiconductor Research Cooperative andparticularly the focus on the upgrading ofexpensive, sophisticated equipment sug-gests a positive impact on the training ofindividuals in engineering and computerscience. Less pronounced, but still a positivecontribution to advanced education, is theGas Research Institute's funding to univer-sities, which allows students to pursue agraduate degree while performing gas-related research.

These trends, nevertheless, were not in-tended to represent the major solution tothe need for trained manpower in differentfields. However, they may well provide animportant component to the solution, incombination with such other activities asindustry-directed professional technicaleducation and training provided by pro-fessional societies.

Will such activity set thedirection for national R&D effortsin the fields affected?

Rather than setting the direction for nationalR&D efforts, the collective actions of industrygroups are more_ likely to support or com-plement R&D directions established byseparate industry sectors and the FederalGovernment, which, combined, constitutethe national effort. A consideration of theseactivities thus rests in a context of how re-sources are utilized for the Nation's tech-nical competence and whether or not pol-icies are needed in either the public and/orthe private sectors to modify this allocationrelative to necessary technical requirementsand the overall supply of resources.

The total funding of collective industrygroups for different research programs isonly about 1 percent of all industry fundingfor R&D. However; the allocation of thecollective funding has significance beyond

38

the stnctly financial in at least two importantways: (a) greater concentration of technicalresources on those points of the R&D spec-trum deemed critical by each industry group,and (b) new institutional relationships,particularly between university and industrypartners.

e Council for Chemical-Research andthe Semiconductor Research Cooperativeillustrate both points: The chemical andsemiconductor industry, groups are par-ticularly concerned with ensuring a strongfoundation of basic research as an essentialingredient to keep pace with rapid tech-nological change and the intense pressureof international competition. The focus ofthese groups has, therefore, been on basicresearch, and the mechanisms chosen havebeen new and/or strengthened university/industry interactions.

With the SRC, for example, substantialamounts of money will be selectively chan-neled to universities to upgrade equipmentand to conduct basic research. These twoefforts clearly imply a major beneficialimpact on the training of new professionalsin the field of microelectronics. The projectedspending of the SRC thus represents a majorattempt to strengthen the industry base.Moreover, by funding universities in targetedareas; research strengths and directions atacademic institutions will also be affected;and those effects will require evaluation:

What is the relationshipof collective efforts tointernational competitiveness?

In brief, collective industry actions permit afocusing of financial and technical resourcesin several areas deemed key by consensusof a particular industry group in the hope ofcontributing to the innovation rate and pro-ductivity of the industry, and hence its com-petitive posture. In several industries, thestrength of international competition isthreatening the position of U.S. firms in worldmarkets; thus the capacity to respond to thischallenge depends upon the best utilizationof all resources. Microelectronics is anobvious example.

The record of growth in this industry hasbeen notably high. Over the last severII years,the annual compound growth rate of U.S.

23

industry revenues has been 25 percent.Moreover, this rate is likely to remain high,from 20 to 25 percent, with worldwiderevenues projected to reach $75 billion by1990. The undiminished prospects of suchgrowth are a keen attraction for internationalcompetition, which in turn has placed in-

--creasechderriailds oil U.S. firms fui morecapital expenditures and R&D to remaincompetitive. As an example; capital expendi-tures by the U.S. semiconductor industry arecurrently about 16 percent of revenue: Onthe other hand; capital expenditures by theJapanese semiconductor industry consumeabout 18 percent of revenue; with R&D ex-penditures at 13.2 percent of revenue.4

Other indicators of competition are alsogermane. From 1970 to -1980, there was adecline in the number of U.S. semiconductorpatents issued annually to U.S. companies.During the same period, such patent issuedto Japanese companies doubled. Also inthe early 1970s, 78 percent of the papersin the prestigious International Solid StateCircuits Conference were .by U.S. authors,With 5 percent by Japanese authors. By1980, the percent of U.S. authors haddropped to 60 percent, while that of theJapanese had risen to 30 percent. In addi-tion, the output of electrical and electronicengineering graduates is also significant:Japan is currently graduating about twiceas many engineers as the United States; andthe U.S.S.R. is graduating about three timesas many as the United States: Moreover;trends in both Japan and the U.S.S.R. indi-cate that the numbers of these graduatesare increasing; while in the' nited States thenumbers appear constant.

All of these indicators of internationalcompetition underscore the role of coor:dinated activities to Leverage technic& andfinancial resources. The results of the re-search and training activities of the SRCobviously remain unt,--,ted owing to its new-ness; and thus the actual impact on corn-petitiveness is unclear. Nevertheless, itseems apparent that there is a need forpooling selected resources to address broadcommon problems; this may free othercompany resources for improving productlines and pursuing other interests.

24

Can this activity provide a significantincrease in support for research;training, or equipment?Collective industry programs can indeedprovide a significant increase in universitysupport. The MIT-Industrg Polymer Proc-essing Program represent the es=tablishmentOf majOr new research concentration atMIT, with industry-sponsored fundingdevoted to equipment and research onprocessing and manufacturing of plasticand rubber products. The program involvesabout 25 projects, and 18 patent applica-tions have been developed since the programwas established in 1973: While training isnot an explicit goal of the program; theresearch projects offer an opportunity forundergraduates and graduates interestedin this field to pursue intensive work inpolymer processing.

The CCR and the SRC activities also reflectincreased support for universities in re-search: training; and equipment. Owing tothe operational characteristics and fundinglevels of the SRC, it is possible that univer-sities receiving SRC funding may becomecenters of excellence in the areas of theirspecial expertise. In such a case, one impactof industry support may be the setting ofdirections in university research, thus re-quiring an examination at a university ofappropriability and desirability.

What are the considerations foruniversities and industries forestablishing and/or continuinginstitutional ties?Industry Considerations. The specificnature of factors influencing technical changein each given industry will vary from one toanother. Thus, each industry has a differentset of activities that may be appropriate forcollective actions. Additionally, economicpressures, Government interactions, andinternational conditions play a role. As aresult, some of the factors that need to beevaluated to assess both the desirability for;and the detailed mode of; collective actioninclude:

Type of common research that isneeded and appropriate;

39

Time and cost of program;Common concerns of health, safety, andenvironment;Need for standardization;Extent of Government programs in re-lated fields; andCapa:ity of companies to pursue workinripppnriontly

These factors give rise to questions ofoptimum research management of finiteresources; and of strategic business plan-ning with regard to the areas of oppOrtunityfor industrial growth and leadership: Col-lective action can provide a broader technicalbase for all participants at lower cost percompany. By contrast, individual action canbe more fledble and thus encourage multipleapproaches, more effective integration withother corporate resources, and competitiveadvantage_ for the company supporting itsinternal R&D program.

These factors can aLso he used to delineatea set of criteria so that an industry, as wellas a particular firm within an industry, candetermine:

Whether the industry should initiate orexpand collective research activities;What relative priorities and emphasesshould be given to the possible pro-grams to be undertaken;Whether a particular company shouldjoin in collective research activities;What benefits will accrue to a membercompany;What constraints may exist for a mem-ber company; andWhat research mechanisms and whichuniversity partnersif anyshould besought.

One important pressure for expandedcollective R&D is the combination of thebroadening technical base required for ad-vances in an industry and the finite resourcesof money and people available to individualfirms within that industry. One way to expressthis combination is the increased sensitivityto improving R&D productivity. The needfor progress and cost-effectiveness may thusresult in assigning to collective efforts(a) programs concerned with costly orbroadly based programs of technology de-velopment, as evident in some areas of EPRI

40

and GRI interests, or (b) depending on therequirements of different industries, programsfocused on basic research, such as the CCRand SRC: The impact on each industrysector, as discussed earlier; can be significant:

A second pressure for collective activitiesis an increasing concern with the technical"environment" within which a particularindustry must operate. Specifically, there isa growing uneasiness throughout the in-dustrial research community, of varyingintensity depending on the industry, aboutthe production of adequate numbers of well-trained graduates, about the research facil-ities of universities, and about the abilityof some industries to maintain their inter-national competitiveness. To the extent thatthese "environmental" factors can be im-proved, the technical base of an industrysector can be strengthened. This motivationwas strongly evident in the initiation of theCouncil for Chemical Research, and it under-lies the plans of the Semiconductor ResearchCooperative.

The extent to which similar collectiveactions can be pursued by other industriesis not clear: Nevertheless, the elementspresent in the ones discussed seem validfor (a) industries that face constraints ontheir pursuit of technical improvement suchas energy; metals; and mining; and (b) otherindustries immersed in- rapid technicalchange, such as electronics and chemicals.The actions to date may serve as usefulmodels for the future.

University Considerations. Universitiesare a major participant with industry in theimplementation of several collective industryresearch programs. Since universities are,in a sense, a principal instrument of societyfor providing a common reservoir of scienceand technology, and since they functionwithin a special context of goalS, motivations,and constraints, their participation in collec-tive industry research efforts involves a dif-ferent set of considerations and impactt.

The principal concerns of each universityin this regard are whether and how to en-courage collective industry actions. In thecase of the Council for Chemical Research,university personnel, as active members of

25

the Council, are involved directly in plan-ning and implementing the programs.

The continuing attention of universitiesto the establishment of mission-orientedresearch centers may be stimulated by thegrowth of collective industrial research. Therewould appear to be some attraction in havinga university research facility structured toconsider the same basic science and engi-neering problems related to an industry ora mission that are simultaneously the objec-tives of a collective industry group. In pointof fact, without any formal industry action;each mission-oriented research center at auniversity that sets up close linkages withindustry represents a form of collectiveindustry action.

Similarly, this enlarged industrial activitymay offer the option for universities to ex-amine and possibly to modify the curriculumand degree offenngs to match the changingtechnical base of particular industries. If thiswere pursued cooperatively with a repre-sentative of an industry association in a givenfield, then constructive approaches couldbe considered more easily than would bethe case for a university working alone orwith the advice of a single company:

Funding from collective industry groupsmay serve to stimulate increased supportfor basic research; for training; and possiblyfor instrumentation. Not only are increasedfunds likely to be available, but the collectiveassociations may provide a sound base forlonger term planning and for broader inter-actions with those industries.

However, there is clearly a challenge touniversities in this expansion. When largeramounts of funds flow through new ormodified channels, there can easily emergestrong biases in research within particularfields. This was precisely the situation re-

suiting from the major growth of FederalGovernment support of R&D dunng the1950s and 1960s. For example, dunng thatperiod the fields of metallurgy and mate-rials tilted heavily toward materials science,with relatively less attention to those areasof process metallurgy that appear to beneeded for productivity improvements today.

Despite these influences, the universitysystem tends in general to be reasonablybalanced in research: The newer industrialactions should not have the potential unbal-ancing impact of earlier Government pro-grams for several reasons.

First; industry support of university researchis simply too small, on the order of 4-percentof all university R&D. If collective fundingprograms could double this, it would stillnot be the dominant factor (See Table 3).And, collective industry funding leaves intactan important characteristic of the universityconsideration of R&D as an end in itself,whether in the conduct of basic science orin the solution of particular problems. Thus,there seems to be no basis for concern as toobjectivity or undue influence from indus-trial funding that is still only on the order of4 percent. A university system hardy enoughto absorb and grow with Federal sponsorship,largely from mission agencies, that reached70 percent is surely able to remain equallyindependent with industry support that isonly a small fraction of the total:

Second, and more importantly, the strongobjective of individual companies to developdirect ties with universities will not be sub-merged within the collective associations.The structure of the Council for ChemicalResearch specifically provides for a one-to-one relationship between a company and auniversity. The collective actions in semi-conductors and energy are only modest

Table 3

University R&D Funding (in millions of dollars)1953 1970 1979 1981 test./

Total University R&D $255 $2,335 $5,183 $6,300Funded by Federal Government $138 $1,647 $3,432 $4,100Funded by Industry $ 19 $ 61 $ 194 $ 240Percentage Federal Govemment 54% 71% 66% 65%Percentage Industry 7.5% 2.6% 3.7% 3.8%

Source: National Science Foundation. National Patterns of Science and Technology Resources, 1981. NSF81.311. Washington, DC: U.S. Government Printing Office, 1981:

26

41

additions to the separate industry-universityrelations of IBM, General Electric, and bacon.Thus, the pluralism of research interests willcontinue.

Nevertheless, there will be an impact onuniversity research by the simple leveragethat well-focused collective actions canproduce. For example, support of particularmicroelectronics centers will propel _thosecenters to the forefront of theifspecial areasof excellence, serving to attract Governmentfunds and other private support more easily.Thus, a major impact may be the setting of adirection of university research.

This imposes a continuing obligation onthe university to maintain its own inde-pendence of research choices; building uponthe support available from industry as wellas Government sources. The mechanismsin place for the collective industry groupssurveyed suggest that universities are in factcarefully screening projects to ensure theirappropriateness for academic goals andpurposes, and discussing these issues withtheir industry colleagues.

What are the main concerns relatedto patents, licensing, royalty income,and antitrust regulations?The issues of patents, licensing, royaltyincome, and antitrust regulations are com-mon to all collective research arrangementsand are reflected in the operational charac-teristics and objectives of each organization.When examining these issues, it is useful tokeep in mind that there are two distinctcategories of collective arrangements: groupssuch as EPRI that conduct most researchat private facilities, and groups that Conductresearch in collaboration with universities. Thefollowing is a general discussion of some ofthe major points in each regulatory issue.The specific impacts of these issues on eachgroup and, in particular, the effects of the1980 Department of Justice Guidelines arenot discussed here:

Patents. Patent rights are often a majorissue related to the conduct of research. Theconcern lies between the rights of the inventorand the tights of the host institutionusuallyan employee/employer relationship. Theassignment of patent rights for inventions,innovations, discoveries, and improvements

42

can vary with the circumstances. Within auniversity context, for example, a researcherwho performs patentable work with the useof university facilities or services in the courseof regular duties generally assigns patentrights to the university. A notable exceptionis the policy of the University of Wisconsin,which states that the university "does notclaim any interest in employee inventions."5

Other arrangements for the assignment ofpatent rights are involved in cases whereresearch is sponsored by a third party. Forexample, as a result of the Uniform PatentAct, a university or small business can retainpatent tights to inventions made in the courseof research under Government sponsorship,with three exceptions:

Operation of a Government-owned re-search or production facility;Exceptional circumstances determinedby the Government agency (stringentdocumentation is required from theagency and is submitted to the Comp-troller General to curb abuse by theagency); orWhen necessary to protect the securityof Government intelligence or counter-intelligence activities.

In cases where industry is the third-partysponsor, and particularly where a collectivegroup is the sponsor, arrangements for thetitle to inventions can vary on a case-by-casebasis to accommodate the patent policy ofan industry group and the patent policy ofa university. In the Council for ChemicalResearch, for instance, all arrangements aremade between individual university/induStrypartners, rather than by the group as a whole,owing to the Council's "one-to-one inter-action" modus operandi. For other collectiveassociations, general policies can be madefor the entiregroup of corporate sponsors;these policies are then negotiated with auniversity.

Licensing. One particular point to con-sider is the arrangement for licensing whena university retains patent rights under anagreement with a collective group. Here, aneicclutive or nonexclusive license may benegotiated. For example, the industrysponsors of the MIT-Industry Polymer Proc-essing Program have royalty-free, irrevocable,nonexclusive license to use any technology

27

developed under their sponsorship. Arecently completed study by the New YorkUniversity Center for Science and Technol-ogy Policy indicates that many industrysponsors participating in cooperative researchcenters do not require exclusive licenses:Exclusive licenses may be more important"in areas of research where the outcomemay be a new drug or agricultural product."6

Royalty Income. The division of royaltyincome is also a consideration related topatents and licensing. Again, arrangementsvary with individual circumstances. In caseswhere a third party is involved; such as inresearch sponsored by collective industrygroups at a university; royalty income divi-sions are negotiated as part of an overallagreement. These divisions can include ashare for the university and the sponsoringgroup, and they may or may not include ashare for the inventor.

Antitrust Regulations. Considerationsrelated to possible conflict with antitrust lawsare also present in the research carried outby collective industry groups: "Because jointresearch may involve or create market dom-inating technology; may be conducted bycompetitors or potential competitors; ormay involve restrictive agreements concem-ing the use of the results of the research;antitrust issues can arise."7 Some of themajor antitrust considerations are reflectedin the focus and structure of differentcollective groups.

The major points of reference when re-viewing the legality of joint research venturesare the nature of the proposed research,the joint venturers, the industry, and therestraints on the conduct of research imposedduring the project. With these four pointsin mind, the general case for not offendingantitrust laws involves: (a) research concen-trated at the frontier or basic end of theresearch spectrum, rather than where it mayhave substantial market effects; (b) a largerrather than smaller number of actual orpotential competitors; (c) a narrow field ofjoini activity; and (d) limited restraints:

The assessment of legality rests in examin-ing the effect on the competitive relationshipof individual firms in the collective group. Inthis regard, there are three major effects ofjoint research agreements to consider:

28

Reduced existing or reduced potentialcompetition between firms;Agreement restrictions that restraincompetition; andLimitations on participation that maygive members of the group unfair ad-vantage in the marketplace:

Evaluation of the effects on competitiongenerally involves application of Section 1of the Sherman Act and Section 7 of theClayton Act.

Given these considerations, researchcompetition is a kw: issue. If the researchconducted by a collective group serves todecrease competition, then the innovativeedge and productivity gains spurred bycompetitive advantages may decrease, withadverse effects to the marketplace. On theother hand, if joint efforts make possibleresearch that firms could not conduct indi-vidually, then the technical base of the in-dustry can be strengthened and thus providenew competitive opportunities derived fromthe technical advances. Therefore, main-taining or strengthening competition is aparticularly important consideration.

The specific arrangements of each col-lective group p;re obviously critical and meritdetailed ex7.rnination. An additional pointto consider is the role of a "neutral" party;such as a university, in conducting the re-search. The participation of a university canserve to reduce the anticompetitive potentialof research projects conducted by an industrygroup, particularly in cases where membersof the group belong to the same industry and,as individual firms, are highly competitive.The SRC is an example. Moreover, the traditional interest of the university in dissem-inating results serves to reinforce the anti-competitive potential. University/industryrelationships that are part of or are the basisfor collective industry research may thushave significance beyond the purely technical.

Possible Government ActionsThe preceding comments on the growth ofcollective industry research are based uponthe impact of these activities on the Nation'stechnical base: Since there are a number ofbenefits inherent in these efforts, considera-tion should be given to possible Govemmentactions that might enhance the use of tech-

nical resources in these collective arrange-ments. However, the justification for anyGovernment action to support collectiveindustry research derives from severalobservations about the objectives and effectsof such initiatives.

First, the identification of areas of basicscience or engineering most desirable by anindustry sector sets a higher probability thatadvances in these areas will be converted toeconomic use Thus, the research directionsset by the private sector should be keptclearly in mind when considering the roleof a Government action: Will it provide addi-tional support and therefore strengthen aparticular trend or will it complement adirection to ensure a balanced base?

Second, support by collective action ofcommon, noncompetitive R&D programscan permit advances to be made on costlyand difficult areas that might not otherwisebe attempted by a single company. This isparticularly true in capital-intensive processindustries, such as mining, and can be afactor in raising productivity within anentire industry.

Third, collective industry action to sup-port broad common research interests in acompetitive industry can release individualcorporate R&D resources for competitivebusiness interests, thus advancing the tech-nical level of an industry generally andstrengthening its overall competitive status.

With these observations in mind, there areseveral areas for possible Govemment actionto enhance the use of technical resources inthese arrangements. A key concern relatesto institutional arrangements between in-dustry and universities.

(1) Should the Government providesome form of indirect financial incentives,for example, seed money, matching fundsor tax deductions, to encourage the growthof collective industry actions?

In the case of the MIT-Industry PolymerProcessing Program, the National ScienceFoundation provided seed money to stimu-late research concentration in an area ofspecial interest to a university and a groupof industry sponsors. The program's viabilityand technical contribution have beendemonstrated through the ongoing partici-pation of industrial firms and the program'sself-supporting mechanism. A critical ingre-

4 4

client here appears to have been the interestof the parties concerned for addressing aneeded area of research. Governmentfunding provided the wherewithal to organizethe required resources into a programfocused on polymer processing research.Similar Government incentives in other areasmight merit examination when the interest,need, and conditions of work are appropriate.

(2) Should Government support for basicresearch at universities expand the directionsstrengthened by collective- industry pro-grams, or should Federal funding be usedto emphasize wholly new directions?

This consideration relates to the obser-vation discussed earlier that the identificationof areas of basic science or engineeringmost desirable by an industry sector sets ahigher probability that advances in theseareas will be converted to economic useDecisions for Federal funding of basicresearch at universities should thus accountfor (a) the overall separate efforts of industryand academic institutions in a particular area,(b) the cooperative mechanisms focusedon an area; and (c) the possible gaps in thetechnical base that could adversely affectbalanced economic growth. A considerationof these points may vary from one technicalfield to the next. Government actions musttherefore reflect the sources of technicalchange in a given field or industry and thepoints of leverage that require attention. Insome cases this may involve added supportfor a particular research direction, and inother cases it may involve concentration ina separate area.

(3) Can Government support of selectedresearch facilities at universities serve toencourage similar or related actions by col-lective industry programs in developingcooperative relations with universities?

It seems apparent that excellence in auniversity research facility can attract coop-erative relations with collective industrygroups. Government support that serves tostrengthen a particular expertise at a uni-versity can result in that university's becominga "center of excellence" in a given researcharea. This in turn can make the university amore likely candidate for other sourcesof support.

Of course, the process can also workwhen industry is the initial, principal sponsor.

29

As a research facility builds special expertise,it can in turn become a more likely candidatefor Federal support.

Conclusion

Collective industry groups are an importantcomponent of the Nation's technical base,with implications for overall economic growthand international competitiveness. As such,collective actions and new or strengthenedinstitutional relationships merit (a) detailedexamination of impacts on both relationshipsand technical trends and (b) considerationof Federal policies that can nurture theprocess when necessary;

30

Notes and References1. Norman. Cohn. "Chip Makers Turn to Academe

With Offer of Research Support." Science. vol. 26(6 May 1982): 601.

2. National Science Foundation. National Patternsof Science and Technology Resources. 1981.NSF81311. Washington, DC: U.S. GovernmentPrinting Office. 1981.

3. Norman. op. cit.4. Information in this section was obtained from an

internal proposal for a Semiconductor ResearchCOOPerative prepared by the Semiconductor In-dustry Association. 1281, pp. 5!8. _

5. Peters. Lois S. and Fusfeld. Herbert I. Current(..I.S.Uniuersitygndustry Research Connections.NSF-AR-14.81.18. June 1982, p. 262.

6. Ibid.. p. 268.7. U.S. Department of Justice, Antitrust Division. Ant

trust Guide Concerning Research Joint Ventures,November 1980. p. 1.

45

The Impact of Increases in Defense R&DExpenditures on the U.S. Research System

Abstract

Currently under way is a significant reallocation of Federal budget resources; with theDepartment of Defense (DOD) budget scheduled to increase 54 percent between fiscalyear 1981 and fiscal year 1984, while the overall Federal budget increases only 29 percent:During that same period, the DOD budget for research; development; test; and evaluationwill increase by 78 percent, while all other Federal research and development expendituresdecrease by 12 percent.

This paper explores the potential impacts of these shifts in Federal R&D spending onthose institutions with a primary role in the Nation's basic and applied research activitiesand in the training of scientists and engineers. The data reviewed in the first part of thepaper suggest that the major impacts of the increases in defense R&D budgets will be onindustry, and particularly on those sectors of industry already .engaged in defenserelatedwork. While the overall impact on the Nation's research universities is not likely to be major,there could be substantial impacts on specific institutions and/or among specific disciplinesas shifts in research priorities result in changing allocation of research support.

The paper reviews several more qualitative issues related to DOD research support. Itnotes the willingness of most major universities to take on more defense-related work, butalso notes university concern over recent Government proposals to increase control, in thename of national security, over the dissemination of the results of that research. Thisconcern must be resolved if a satisfactory DODuniversity relationship is to be established. Amajor issue cutting across Government industry, and universities is the impact of thedefense buildup on the Nation's pool of skilled scientific and engineering personnel; thereis a possibility that universities, the armed services, and some parts of the private sector mayexperience difficulty in recruiting and retaining engineers and computer scientists, particularlythose with advanced degrees: There is likely to be increased competition between defenseand nondefense sectors for technical talent;

The paper concludes that Federal investment in all sectors of R&D, not only those clearlyrelated to defense needs, is essential to maintain and improve the technological base ofU.S. national security.

IntroductionBetween fiscal year 1981 and fiscal year1984, Federal funding of Department ofDefense (DOD) research, development, test,and evaluation (RDT&E)* activities willincrease by some 78 percent: over the same

*For the purposes of this paper. the abbreviation R&Dwill be used for both civilian artditepartment of Defenseprograms; when referring to DOD, this abbreviation ismeant to include test and evaluation activities. However itis not really accurate to compare DOD and other agencybudgets at this aggregate level, since other agencieseither do not have significant It and evaluation activitiesand/or do not treat them together with research anddevelopment efforts. This reporting artifact thereforecauses some distortion In understanding the comparativeIncreases In DOD R&D activities.

46

period, Federal funding for nondefenserelated research and development will de-crease by approximately 12 percent. Theincreases in defense funds for R&D are partof an overall acceleration in defense spendingdriven by the desire, in the Congress andthe Reagan Administration, to enhance theU.S. secunty posture now and in the future;overall, the defense budget is targeted for a54 percent increase between fiscal year 1981and fiscal year 1984, while the Federal budgetoverall increases only 29 percent. Table 1presents overall budget patterns for the fiscalyear 1981/fiscal year 1984 period:

Significant reallocations of nationalresources such as these are likely to have .broad societal impacts. This paper explores

31

Table 1

Overall Budget Patterns, Fiscal Year 1981-Fiscal Year 1984FY81Actual

FY84Proposed Change

Federal Bucigeta $6572 billioil $848.5 billion 29.1%Dept. of Defense Budgeta $159.8 $245.3 53.5%Dept. of Defense RDT&Eb $ 16.6 $ 29.6 78.3%Other R &D' $ 19 0 $ 17 0 11 8%Dept. of Defense Basic Researchb . . $0.610 $0.869 42.4%Other Basic Researchb $4.497 $5570 27.9%

Notes:a Outlays:

bNew Authority. not including construLtion of facilities.

Source: Shapley, Willis, Teich, Albert. and Weinberg, Jill. Research and Development: AAAS Report VII.Washington, DC: American Association for the Advancement of Science; 1982, _pp. 6, 27: and_IntersocietyWorking Group. R&D in the Fiscal Year 19$4 Budget A Preliminary Analysis. Washington, DC: AmericanAssociation for the Advancement of Science. 1983. pp. 13. 15. 22. 33.

one such set of potential impacts: It discussesthe possible effects on the research systemof recent increases in defense R&D spending:The research system, as conceptualized here;is defined as "the set of institutions; facilities;and most importantly; people; whose activitiesboth increase society's storehouse of knowl-edge a'iout physical, biological, and socialreality and investigate ways in which thatknoWledge can be used for human pur-poses."' This definition, it should be noted,emphasizes basic and applied researchactivities, rather than development efforts.Such an emphasis is appropriate for thispaper, since substantially more attention isgiven herein to issr and impacts relatedto ` research carried out within the Nation'suniversities than it is to development activitiescarried out within industry.

A number of recent reports on the stateof the U.S. research system have noted "signsof stress, including resource constraints,demographic trends affecting higher edu-cation; escalating instrumentation costs; andpressures for short-term returns on researchinvestments::::"2 This paper attempts to castsome light on the interaction between shiftsin Federal R&D funding patterns and thepolicies that underpin those shifts, on onehand; and emerging problems in the researchsystem, on the other. Will increases in defenseresearch spending ameliorate, or possiblyexacerbate, some of the emerging stresses?Will the increased role of the Departmentof Defense in Federal R&D support create

32

new stresses. new issues? Or are the sys-temwide impacts of accelerated DOD R&Dfunding likely to be minimal? The followingpaper provides some preliminary answersto these questions.

Trends and Developments

Historical Perspective

Before examining current and future impactsin detail, it is worth reflecting briefly on theimpacts of Department of Defense supporton the creation and evolution of the U.S.research system in its current form and, inparticular, on the development of the FederalGovernment's relationship to that system.The point of such a historical review is toprovide a basis for examining whether whatonce was, will be again: Although VannevarBush's 1945 vision; in Science: the EndlessFrontier, centered around the creation of acivilian National Research Foundation toserve as the keystone of a post-World WarII Govemment-science partnership, in reality itwas the _military services that took the19454950 initiatives to create that partner-ship. In particular, the Office of Naval Re-search (ONR), established in 1946, was by1950 supporting over 40 percent of U.S.basic science and had developed a varietyof means of providing this support. Most ofthose means are still in use today. Theyinclude:

Funds for construction of large facilitiesoperated by a consortium of universities;Funds for large single-university labbra-tories, with the research agenda set bya single laboratory director;Acquisition of expensive, specializedequipment;Funding for unique institutions such asWoods Hole Oceanographic Institute;andProject grants to individual investigatorsto pursue a particular line of research.3

The Office of Naval Research supportedresearch, not development; and its majorpartner was the U.S. university system. Thecharacter of that partnership and othersbetween Government and universitiesdeveloped during the 1940s has beendescribed by the current president of a majorresearch university: "In an overall sense theAmerican university was mobilized for warby the Federal Government in 1941, anddemobilization did not occur until twenty-five years later." As a result, he notes, "theAmerican research university yearns for the1950s and early 1960s....The fact is that wewere spoiled. We took a great deal forgrantedaffluence, growth, the respect ofsociety; a clear sense of purpose."

Although other channels of Governmentsupport for research and development weredeveloped or grew in significance duringthe 1940s and 1950sthe Atomic EnergyCommission (AEC), an enlarged NationalInstitutes of Health (NIH); the NationalScience Foundation (NSF), and the National

Aeronautics and Space Adr.i.,istration(NASA)the Department of Defense con-tinued as a dominant Federal R&D supporterinto the mid-sixties. The national shocksfollowing the Korean invasion, the Sovietintercontinental ballistic missile (ICBM)buildup, and the launch of Sputnik reinforcedthe national security rationale for Federalfunding of R&D and, in the case of Sputnik,triggered an across-the-board concern forU.S. standing in science and technology:In addition to support of university research;during the fifties and sixties Federal fundswent to a new type of high-technologyfirm organized to develop and produce thetechnology-intensive systems required byDOD and NASA, and the "aerospace"industry became a major performer of fed-erally funded R&D. Edsting, more traditionalfirms created new divisions to performcontract work for the Government. Also, anew kind of institution, called a federallyfunded research and development center(FFRDC), was developed to carry out spe-cialized research tasks, usually for a singleGovernment sponsor. Table 2 contrasts therole of defense-related R&D outlays overthe past decades and in more recent times.

In terms of scientific advances and tech-nological progress, this Government-university-industry partnership in R&D proveda powerful success: For example; a recentanalysis of the results of three basic researchprojects at the Massachusetts Institute ofTechnology sponsored by ONR in the late1940s and early 1950s identified a "flood

Tablt 2Role of Defense Funding in U.S Research and

Development SupportFederal Sharp

YearFederalTotal

DefenseRelated

SpaceRelated

CivilianRelated

Non.FederalShare

1953 54% 48% 1% 5% 46%1960 65 52 3 9 351965 65 33 21 11 351970 57 33 10 14 431975 51 27 7 17 491980 (est) 49 24 8 17 511982 (est) 47 27 6 14 53

Sources: For 1953.1980; National Science Foundation. Notional Patterns of R&D Resources. 19& NSF 80.308Washington. DC: U.S. Government Printing Office, 1981: for 1982, author's estimate.

4833

of knowledge and practical accomplish-ments," and concluded that, "without suchsupport, these benefits would very likely havebeen postponed for many years or perhapsnot have oeen attained at all."5 It is probablethat similar results would follow from carefulstudy of other DOD-supported basic researchefforts of the 1950s. Since 1960, some 20Nobel Prize winners have drawn directsupport from the Department of Defense .°At the level of technological achievement, itwas the high-technology industries nurturedby DOD (and then NASA) funds that bothU.S. and European observers in the 1960sidentified as the secret of U.S. economicpower; the U.S. leadership position in suchareas as aviation, space, microelectronics,computers, and advanced materials was seenas a counter to the erosion of the country'sposition in other technologically intensivesectors that did not receive direct Govern-ment financial support.

Another product of close DOD-sciencerelationships during the 1950s and 1960swas the involvement o_ of the Nation'sleading scientists with national securityprograms, either as DOD-funded investigatorsor as advisers to the Defense Departmentor the White House on scientific and tech-nological issues related to defense. Thislinkage meant that some of the country'sbest minds were familiar enough with ideasfor new weapons systems to provide bothsupport and constructive criticism of suchproposa Is.

Even as the momentum of the partnershipof the fifties appeared to be increasing, signsof tension appeared. In his farewell addressas President, Dwight Eisenhower warned thecountry of the potential of undue influenceon the part of both the "military-industrialcomplex" and a "scientific-technical elite."During the mid-1960s, as national prioritiesshifted toward domestic concerns and asU.S. involvement in Southeast Asia becameincreasingly unpopular on universitycampuses and in society at large, the part-nership between DOD and the Nation's majoruniversities largely came apart. There was adownturn in DOD R&D investments overall,as the costs of the Vietnam War dominatedthe defense budget. Between the mid-1960sand the mid-1970s, DOD support of basicresearch, in constant dollars, was cut in half,

34

and DOD funding of R&D overall, again inconstant dollars, declined by one-third (SeeTable 3). Students and faculty on thecampuses of many of the country's leadinguniversities questioned the appropriatenessof close DOD-university ties. Finally, theMansfield Amendment of 1969 prohibitedDOD research support unless there was a"direct and apparent relationship" to someestablished DOD function or mission. Thoughit is difficult to trace the specific impacts ofthe Mansfield Amendment on particularDOD research investments, the spirit of theAmendment,* coupled with university dis-affection and budget constraints, served aseffective limits on DOD involvement in overallresearch policy during most of the 1970s.

The point of this compressed discussionis to point out the central historical role thatthe Department of Defense has played inthe post-World War II U.S. research system;particularly in the Government-science part-nership. Will the current step up in defense-related R&D investments have similar broadimpacts? Because of the availability of addi-tional DOD funds, will there be:

Major shifts in the character of, policiesfor, or mechanisms for Govemment in-vestment in research and development?:More activity in the frontier areas offundamental science and engineeringinquiry, which will result in major sci-entific advances in the 1980s and 1990s,from DOD-supported research ?; and/orA lessening or removal of current prob-lems such as obsolete instruments, toofew graduate and undergraduatestudents in science and engineering,and shortages of funding for worthyresearch?

Will an increased DOD presence in theresearch system be a source of new con-troversy, a cause of the diversion of high-quality scientific and engineering efforts awayfrom promising lines of civilian-orientedresearch, and/or a major contributor toemerging shortages of or bottlenecks in thesupply of scientists and engineers? It shouldbe noted that the recent increases in DODresearch support are planned to continue

*Although the amendment was legally binding for onlyone year, its impact persisted.

Table 3

Trends in Department of Defense R&D, Fiscal Year 1965-Fiscal Year 1984(in millions)

Current Doll6is Cori Stant FY 1972 Da Hari

Fiscal YearBasic Total

Research R&D&Sit

ResearchTotalR&D

1965 $17 $6865 $506 $100161966 341 7099 476 99031967 362 8136 490 110081968 318 7908 411 102091969 353 7890 436 97521970 323 7491 371 86071971 318 7654 82281972 328 8482 328 84821973 304 8541 285 80111974 303 $578 265 75051975 305 9167 244 73241976 328 9770 245 73081977 373 11385 257 78561978 412 11760 262 74681979 474 12751 283 76121980 552 14150 305 78251981 617 17050 312 86101982 695 20(144 322 94611983 788 24300 323 98001984 869 29500 339 11500

Sources: Shapley. Willis. Teich. Albert. and BresloW,Gail. Research and Development: HAAS Report VI.Washington. DC: American Association Lor the Advancement of Science. 1982. p. 99 and author's calculations.Report hereafter cited as AAAS. Report VI. The fiscal year 1984 figures are from PreSident Reagan's budget submission.

in subsequent years, and thus their impactmay be growing.

Or, will the current and potentially con-tinuing upswing in the availability of DODfunds for R&D have only a marginal, thoughnot insignificant, impact on the researchsystem overall, although impacts on specificinstitutions or on specific disciplines maywell be substantial? The increases in DODfunds may well have the intended effectson the Nation's security posture, and it isprimarily on this basis that those increasesshould be evaluated (though not In thispaper): But their systemwide effects maynot be as great as it might be expected,given the large increases in R&D spendingIn particular; at the level of basic researchand of effects on the Nation's research (asopposed to development) institutions; in-creases in- OD funding are in fact not allthat large. The following sections of this papercontain a preliminary analysis of th6Sesystemwide effects.

Patterns of Defense RDT&EExpenditures'

Tracing the patterns of DOD expendituresfor research and development is at best animprecise art, particularly if there is an attemptto compare them to overall national patternsof R&D expenditures and performance.*Definitions of various categories of R&Dactivities are different within DOD than theyare for civilian agencies, and DOD statisticsinclude test and evaluation efforts in thesame accounting system as research anddevelopment activities. Reporting systems

It should be noted that this analysis examines ontythe Department of Defense research budget. Nationalsecurityrelated expenditures by the Department ofEnergy. NASA. NSF, and other Federal agencies arenot Included in the analysis. A full examination of thetotal national security research budget is needed, butis beyond the scope of this paper: the assumption hereis that such analysis would not markedly change theconclusions of this paper.

5035

and time lags before reports are availablealso differ. What follows; therefore, shouldbe interpreted as an impressionistic sketchof the current patterns of DOD spendingand of performers of DOE) research activities,placed in the overall context of nationalR&D patterns.

Perhaps the central insight from this sketchshould be identified at the outset. Duringthe 1945-1965 period, DOD sup_port was adominant feature of the Federal R&D budget;DOD funding of universities and industrywas essential to the overall national R&Denterprise. The picture is somewhat differenttoday. As Table 2 suggests, defense relatedR&D funding in recent years is just overone-quarter of national R&D expenditures.compared to 52 percent in 1960. Moreover,this quarter of national R&D expendituresis relatively narrowly concentrated. Over 70percent goes to industry (including FFRDCs),and the industries that receive major defensecontracts tend to specialize in defense andspacerelated work; they are; with a fewexceptions; not the industrial giants of thecountry or the major U.S. actors in inter-national trade. Of research universities re-ceiving Federal R&D support, only one ofthe top ten DOD recipients gets more than50 percent of its funds from DOD, and theaverage "DOD share" of research supportamong the top ten is 16 percent (fiscal year1980 figures). (The reality is that the Depart-ment of Defense, even during the period itwas the major source of Federal R&Dfunding overall, never was the primary sourceeither of Government investment in basicresearch or of Federal funds going to uni-versities for R&D. For example, Table 4provides a historical perspective on the DODshare of funding for university R&D.)

These data suggest the major impacts ofthe increases in defense R&D budgets willbe on industry; and particularly on thosesectors of industry already engaged indefense-related work. While the overallimpact on the Nation's research universitiesis not likely to be major, there could besubstantial impacts on specific institutionsand/or among specific disciplines as shiftsin research priorities result in changingallocation of research support. The followingparagraphs provide evidence for this general

36

Table 4

Sources of University R&D Funding

YearNon-Federal

SourcesTotal

Federal DODa

1955 59% 41% 19%1960 37 63 211965 27 73 181970 29 71 101975 33 67 51980 35 65 6

Note:a If DOD funding to universities were to increase bythe mid-1980s to more than 10 percent of total sup-port land this is the implication of recent increases).its Impacts could be substantial.

Source: Author's calculations based on National ScienceFoundation. National Patterns of Science and Tech-nology Resources. 1980. NSF80-308. Washington.DC: U.S. Government Printing Office. 1981.

conclusion and suggest where it must bequalified or refined.

From fiscal year 1981 to fiscal year 1984,the DOD R&D budget increased by some78 percent while all other Federal R&Dexpenditures decreased about 12 percent.When one examines where these increaseshave gone, the relative narrowness ofpotential impact becomes even more evident.Table 5 provides the relevant data. Basicand applied research activities for their DODequivalents, roughly DOD budget categories6.1 and 6.2) increased much less than theDOD average, and systems developmentefforts in such areas as strategic programsincreased at well above the average rate;further, these increases in hardware devel-opment programs began with a much largerfunding base. For example, increases instrategic programs alone required 44 percentof the total DOD R&D budget increase inthe past 2 years. While the dollar amountsof budget increases do not have a one-to-one correlation with the potential impactsof those increases, the fact that most in-creases are going to development, test, andevaluation activities puts some limits on theinfluence of DOD increases on basic andapplied research activities overall.

There have been governmentwide at-tempts during the last three administrationsto provide real growth each year in Federalinvestment in basic research. The increase

51

Table 5

DiOtibution of DOD R&D Budget

CategoryFY81Actual

FY84Proposed

% ofTotal

Inc_rea_se -

1981.1984

Basic Researcha _$615 million $850 million 2.9% 382%Exploratory Developments $1985 $2963 9.1% 49.3%Advanced Technology Development $593 $1233 4.2% 107.9%Strategic Programs $3440 $9160 30.9% 166.3%Tactical Programs $6130 $8850 29:9% 44.4%Intelligence and Communications $1632 $3576 12.1% 119.1%Mission Support $2238 $3260 11.0% 45.7%

Total $16,634 _$2%622 78.1%

Note:

aThese two categories are what DOD defines as the "technology base:" Over the past 3 fiscal years, the techriOlogYbase budget has increased by 46.6 percent.

Source: American Association for the Advancement of Science. Fiscal Year .1984 Budget Report, p. 33.

in overall Federal support for basic researchover the past 7 years is 159 percent; thecorresponding increase in DOD basic re-search support is 133 percent: In other words;DOD support of basic research has increasedat about the same rate as Federal supportoverall. The pattern is a bit different in recentyears, with DOD support of bask researchincreasing 42 percent between fiscal year1981=fiScal year 1984, while bask researchsupport from other Federal agencies in-creased 28 percent. Still, the dollar amountsinvolved are not overwhelming; DOD supportfor basic research over that period increasedby only $191 million. Indeed, recent increasesin DOD basic research funding mayJ be begunderstood not as part of the national securitybuildup under the current Administration,but rather as the continuation of a trendthat began in the mid-1970s and was givenparticular attention under the Carter Ad-ministration: For example; in 1976 theDefense Science Board conducted a reviewof fundamental research in DOD; and in1978 the Office of Science and TechnologyPolicy issued a report (the "Galt Report")that focused on bask research within DODand called for a reinvigorated DOD. basicresearch effort. An office to oversee DODresearch efforts was established under theUndersecretary of Defense for Research andEngineering in 1978, and there was sub-

stantial real growth in DOD research budgets

in the late 1970s, growth that has continuedin recent years.

The Department of Defense spends itsR&D money very differently than otherFederal agencies do. A recent AmericanAssociation for the Advancement of Science(AMS) report noted that DOD "dependson and supports a major segment of theU.S. scientific and technological communitythrough a fairly comfortable pattern ofworking arrangements, mostly with industry,that has evolved over the years: Other partsof the scientific and technical community;especially in universities; have remainedlargely outside the military orbit."8 Forexample, currently 24 percent of the R&Dbudgets of nondefense agencies go to uni-versities, while only 3 percent of the DODR&D budget is spent in academic institutions.The Defense Department puts 74 percentof its R&D budget into industry; nondefenseagencies, only 46 percent. The allocationsto in-house laboratories are similar: 21 per=cent for DOD, 30 percent for other agencies.

American businesses receive by far themajority of DOD's extramural researchawards. Table 6 lists the top 10 recipients ofDOD research contracts. As mentionedearlier, most of these recipients are highlyspecialized; advanced technology firms thathave been created or have been adapted toperform DOD (and NASA) sponsored work;most of them do not have a diversified

37

Table 6

Major Department of DefenseContractors, Fiscal Year 1981

Compari or InstitutionDODRalik

Fortune 500Rank

Martin Marietta Corp. .... 1 130Boeing Co. 2 31Rockwell International 3 48Hughes Aircraft 4 213General Electric CO. 5 11General Dynamics Corp. 6 76TRW. Inc. 7 71United Technologies Corp. . 8 20Boeing Aerospace Co. 9 a

Aerospace Corp 10 b

Notes:a Not separately ranked.bNonprofit corporation IFFRDC).

Source: Department of Defense. 500 Contractors Re-ceiving the Largest Dollar Volume of Prime ComractAwards for Research. Development. Test. and Eval-uation: Fiscal Year 1981. p. 2.

product line, nor do they sell their productsin commercial markets: DOD is included inthe provisions of a new law requiring allFederal agencies to set aside an increasingportion of their R&D budgets exclusivelyfor small businesses: Other agencies have 4years to increase the small business shareof the R&D spending to 1.25 percent; DODhas 5 years to meet that target figure. Whatelements of the DOD budget (for example,the test and evaluation categories) will besubject to this set-aside requirement ha% enot yet been determined.

Of the top 500 defense R&D contractors,only 12 are universities; Table 7 lists the 10universities receiving the most DOD R&Dsupport. Four of these institutions are amongthe top 10 university recipients of FederalR&D funds overall, but with the exceptionof Johns Hopkins University (because theoffcampus Applied Physics Laboratory isincluded in the Johns Hopkins total); infiscal year 1980 none of these or the othertop 6 research universities received morethan a fifth of their Federal research sup-port from DOD:

In summary; then, the Department ofDefense provides over half of all FederalR&D support proposed for fiscal year 1984;and the DOD R&D budget is increasing

Table 7

Universities Receiving DODSupport for Academic Science;

Fiscal Year 1980

Institution

Rank asRecipientof DODFunds

Rank as DODRecipient Funds as_of all Percent ofFederal TcitalFunds Support

Johns Hopkinsa 1 1 65%MIT 2 2 17%Georgia Institute of

Technology 3 47 68%Stanford University 4 3 17%Pennsylvania Stateb . 5 17 28%University of Texas 6 32 36%University of Dayton . 7 86 90%University of

Washington 8 4 12%University of Southem

California 21 23%University of Califor-

nia, San Diego 10 5 12%

Notes:a Of Federal support: 71 percent goes to AppliedPhysics Laborator,.... fn-merly a FFRDC.bOf Federal support, 25 percent goes to AppliedResearch Laboratory. formerly a FFRDC.

Source: National Science Foundation. Federal Support to Universities: Colleges: and Selected NonProfitInstitutions: Fiscal Year 1980. Washington, DC: U.S.Government Printing Office: 1982:

much faster than the rest of Federal R&Dexpenditures. Although the short-term in-fluence of this growth on the Nation's re-search system may be limited, if the defensebuildup continues, and if DOD research anddevelopment expenditures continue to grow,the longer term impacts, particularly onhighpriority areas of the physical sciences,mathematics, and engineering, could be-come significant.

Policy IssuesThe industrial base for national securityprograms is not in robust condition; andconcentrated attention is being given toimproving that situation. 9 But, (a) this is notprimarily an issue of science and technolocgypolicy, and (b) as already noted, defenseindustries are highly specialized and existsomewhat in isolation from the mainstreamof American industry. Thus, increased

53

budgets for weapons system developmentare not likely to have widespread impactson the research system overall. Some prob-lems are likely to emerge from the industrialmobilization to support an enhanced nationalsecurity posture; particularly with respect towhether sufficient industrial capacity andadequate supplies of skilled scientific andtechnical personnel will be available fordefense requirements without creating sig-nificant shortfalls or bottlenecks in the non-defense sector. There are also over 70 DODresearch laboratories under service man-agement. Recent studies have identifiedsignificant problems within these laboratories,and DOD managers are taking steps toimprove their peiformance. These steps arealso unlikely to have significant systemwideimpacts, again with the exception that theymay make DOD laboratories more suc-_cessful competitors for the limited supply oftechnical talent.

With respact to the academic sector, arecent Defense Science Board Study notedthat

the universities and DOD need each other.DOD needs the scientists and engineerstrained by universities; it needs the facultypool of scientists and engineers workingin the DOD area as originators of newideaS and as expert consultants and ad-viSerS. The university research base fordefense preparedness is in considerabledisrepair and therefore in need of up-grading in faculty, equipment, facilities,and support. The problem is muchbroader than DOD, but DOD has a spe-cific interest and responsibility and acritical need to see that a solution is foundand that the solution is enduring.1°

In its efforts to develop an "enduringsolution" to creating an effective DOD-university partnerShip, the Department ofDefense has taken a number of program-matic initiatives in addition to increasing itsinvestment in academic research. It is in thechanging character of the DOD-universityrelationship that many of the new or emerg-ing issues discussed in this paper can befound. Five such issues are discussed below;following these discussions; issues associatedwith DOD support of industrial research and

54

issues associated with science and engi-neering personnel are highlighted.

University-Related Issues

Are Universities Willing to UndertakeMore DOD-Funded Research? In thecurrent fiscal year, Federal R&D support touniversities will total approximately $5.2billion; of this amount, about $0.9 billion(17 percent) comes from the Departmentof Defense: Major research universities seethemselves in a funding crisis; and thus thepossibility of substantially enlarged DOD R&Dsupport is quite attractive to the leadershipof the academic community: In recent Con-gressional testimony, a panel of universitypresidents suggested that over the next 15years the Department of Defense "shouldbegin making up" a $4 billion underinvest-ment in basic research over the past 15years.n One member of the panel told theCongress that "universities today should beand...are willing to do all within their capa!bilitieS and limited resources to be involvedin meeting national security needs. The greatcrisis of ideology during Viet Nam has allbut evaporated among faculty, Students, andstaff."12 Other observers are not convincedthat the kind of differences in values andperspectives that drove the academic andnational security communities apart in thelate 1960s have totally disappeared; recentdemonstrations related to a campaign fornuclear freeze may suggest lingering universityhostility to DOD programs: A blue-ribbonpanel that recently examined national securityR&D programs noted "the emotional can-over of the Vietnam era: students whophilosophically reject the concept of strongdefense as deterrence combined with facultywho have put aside certain technical fieldsto pursue investigations less likely to haveimplications for armament. The institutionalsuspicion of the military in some schoolsseverely limits their role in the great adventureof keeping the peace."13 At a minimum,suggests the president of the university thattops the list of recipients of DOD funds, "anonmobilized community of research uni-versities can and should be more cautiousand selective with respect to initiatives fromgovernment for new research activities...."14

39

University Concerns over Export Con-trol Requirements. One issue of centralimportance to restoring a mutually satisfactoryDOD-university relationship stems fromrecent Government proposals to increasecontrol, in the name of national securi ,

over the dissemination of technical infor-mation. Since a full discussion of this issueappears in another paper in_ this com-pendium; its major elements will simply behighlighted here.t5

The recent controversy is a reflection of alongstanding tension betWeen the notionsof free communication of scientificmation and of the need toprotect informationrelated to national security_ interests._ Thereis general agreement -on the need_ for theUnited States to protect engineering_ andtechnological information related to nationalsecurity; the controversy arises over sug-gestions that since relatively fundamentalscientific discoveries can quickly be incor-porated into technology for military systemsand since the United States depends for itssecurity on technological superiority; thereis a need for controlling access to scientific;as well as technological, information. This isespecially the case; argue those concernedwith national security; when the Soviet Unionand its Warsaw Pact allies are undertakingsystematic efforts to acquire U.S. scientificand technological information.

The scientific community has been resistantto attempts_ to control the flow of basicscientific information;_ arguing that opendissemination of results and subsequentreview and criticism of new findings areessential elemerils of the scientific-- processand that any barriers to such flow will imposemuch greater costs,_ in terms of slowingscientific diScoVery and perVerting the practiceOf science; than the potential national securitybenefits. As Government concerns over theleakage of technical information have in-creased, the scientific and research universitycommunities have been outspoken in theirresistance to additional controls over theactivities'of the basic research community.

During 1981 and 1982, Administrationofficials expressed concern that bilateralU.S.-U.S.S.R. agreements for scientificexchanges were one-sided, with the Soviets

40

gaining access in areas where they are weak;that scholarly mhanges were being misusedby the Soviets; who were sending senior tech-nical people; some from closed militaryinstitutes; to the United States; that muchdefense-related information was being inad-vertently disseminated at professional con-ferences and scientific symposia attendedby scientists from the Communist countries;and that open publication of scientific findingsin certain fields was transferring sensitiveinformation to U.S. enemies. The concerncame to public attention in January 1982,when the Deputy Director of the CIA ad.dressed a scientific meeting and warned ofa "hemorrhage" of U.S. technology to theSoviet Union and of a "tidal wave" of publicreaction if the scientific and technologicalcommunities did not develop voluntarymeans for ensuring that sensitive informa-tion was not accidentally made availableto U.S. enemies.

These recent pressures for increasedcotrol over technical information have beenmet by strong reactions from spokesmenfor the scientific community, who argue thatany moves toward restricting the activitiesof basic researchers are not justified andmay be both unwise and unconstitutional.

There has been continuing discussion ofthe "science vs. secrecy" issue in the pastyear, and a number of groups, such as theCommittee on Science, Engineering, andPublic Policy of the National Academy ofSciences and the Committee on ScientificFreedom and Responsibility of the AmericanAssociation for the Advancement of Science,have launched studies aimed at finding acommon ground for agreement betweenthe national security and scientific com-munities. This is a major agenda item forthe DOD-University Forum, discussed below.Although there is little doubt that the De-partment of Defense can impose whateverconditions it chooses over research it fundsdirectly (just as researchers can reserve theright not to accept DOD funds with unac-ceptable conditions attached), the broaderconcerns of the research community over ageneral restriction of scientific communica-tion on national security grounds must beallayed if the major research institutions of

the United States are to be willing to engagethemselves actively in an accelerated DODresearch effort.

Character of DOD Support to Univer-sities. As discussed previously, the increasesin DOD funds going to universities, whenadjusted for inflation and for the funds setaside for DOD's "instrumentation initiative"(see below), are less than 20 percent ofoverall Federal funding for university research.Still, DOD university funding is almost equiva:lent in dollar amount to that of the NationalScience Foundation and thus far frominconsequential. In addition, over the nextfew years the "DOD share" of universityresearch _support may continue to growfaster than other elements of the basicresearch budget.

If a particular institution were to receivesignificant new amounts of DOD funds forresearch, instrumentation; and student sup-port, the concentration of such funding couldsubstantially alter the character of thatinstitution's research and graduate teachingefforts. Furthermore, DOD support is likelyto be concentrated in a few disciplines andfields of particular relevance to nationalsecurity applications,' and this could havemajor impacts within the basic researchenterprise in terms of status, ability to attractthe best students, and indeed the pace ofscientific progress overall.

In the immediate poSt-World War II period,DOD funds, particularly those channeledthrough the Office of Naval Research, werein effect general Government support of,especially, basic research. Such has not beenthe case for the past two decades, althoughONR is still seen as the least restrictive ofany Government agency supporting research.Even though the requirement of the Mans-field Amendment that all DOD-funded re-search have a "direct and apparent rela-tionship" to a military function has beensoftened to require only a "potential rela-tionship," DOD R&D investments; even thosein the 6.1 "research" category, are bestunderstood as targeted long-range researchsupport in areas of perceive d national securityimportance with priority research areasSelected by DOD. This has been the case

for some time; for example, a 1974 study ofDOD-supported projects at Stanford Uni:versity concluded that "the military haddeveloped a rational, well-administeredprogram to define research priorities in termsof current and projected military needs andto purchase (emphasis dded) R&D fromuniversities based on those needs. Thus, whilethe scientific purpose as reflected in eachindividual project proceeded objectively,funding availability biased scientists' choiceson which projects to pursue." To the authorsof this study, such external criteria for projectSelection were problematic; they raised"serious questions about the university'seffort§ to fulfill its role of protecting theprocesses by which people search for scientifictruth. For nonscientific standards set outsidethe scientific community to have a heavyinfluence on the choice of which projectsare undertaken may be proper and desirablefor industry or Government; but... it is notcompatible with the universities' role asagency to protect the scientific process."16

This is a rather idealistic view of thescientific enterprise. By accepting externalfunds for research support, universities atleast tacitly also accept some set of externallyderived research priorities. Perhaps morerealistic is the question of whether an in-creasing trend toward harnessing researchpriorities to national security requirementsis in the national interest: Certainly DODresearch needs are likely to differ somewhatfrom those defined in the civil sector, andthus there could be shifts in the existingpattern of basic research activities as DODfund§ are injected into the system and othersources of funding have deereasf?d budgets.This kind of reallocation of basic researchpriorities appears to be well under way. Notonly research priority decisions but alsodecisions on which specific projects to fundare, in general, made by DOD technical staff;peer review is rarely used by the Departmentof Defense.

As part of its accelerated R&D effort,DOD has identified its highest prioritytechnologies as:

Very high speed integrated circuits,High-energy lasers,Manufacturing technology,

41

Precision-guided munitions, andRapid solidification alloys.

DOD is pursuing a research investmentstrategy that will support rapid progress inthese and other critical technological areaswith "order of magnitude" impacts on futuremilitary systems.17

Historically, research done under nationalsecurity auspices and in response to nationalsecurity requirements has been a majorsource of scientific breakthroughs that arethe bases for significant technological inno-vations of general economic significance.Most of the high-priority research areastargeted for DOD funding also may havepotentially broad civilian applications. How-ever, military systems are becoming increas-ingly sophisticated and specialized, andthere may not be as much transfer frommilitary to- applications as has oc-curred in the past. There will be a needfor continued attention on the part of DODand other research managers to making theresults of DOD research accessible to theprivate commercial sector and other Gov-ernment users. consistent with securityrequirements.

Also; as particular research prioritiesdominate DOD investments at even the basicresearch level, areas of research currentlybeing supported by civilian agencies couldbecome candidates for DOD funding, andlines of research not falling into areas ofDOD interest may require_particular attentionfrom such agencies as NSF if the best science,regardless of external relevance, is to besupported. There will be an increased needfor govemmentwide coordination of researchsupport at the disciplinary or program officelevel, as well as more generally, and NSFmay well be required to return to somewhatof a "balance wheel" role as a Federalresearch support agency. Already in placeis a mechanism by which DOD. NSF, andthe Office of Science and Technology Policy(OSTP) can coordinate research activitiesin areas of mutual interest, and this type ofcoordination is necessary to achieve somesemblance of coherence in Federalresearch policy.

As this discussion suggests; the rolc ofDOD in the support of university-basedresearch will increase in importance, butDOD, as an agency with a nonresearch

42

mission, is a poor candidate to reassumeleadership as the support agency for aca-demic science. The WS notes that somepeople are asking "since national prioritiesare being tilted toward defense, and sincebasic science is so important to nationaldefense in the long run, why shouldn't theDOD budget take on a really major shareof support of research at universities?" Thisview is described as a "pipe dream," onethat does not recognize "that it is highlyunlikely that DOD could get a nationaldefense priority within DOD itself or inCongress for the general support ofbasic research." 18

DOD Instrumentation Support. Inrecent Congressional testimony, the thenDeputy Director of NSF reported "an emerg-ing consensus in universities. the FederalGovemment, and private industry that there isa critical and growing need to replace ob-solete and worn-out research apparatus andlaboratory facilities in the Nation's researchuniversities: Although its precise dimensionsarc not known; there is strong, qualitativeevidence that the problem is pervasive andlarge in scope. A rough, but reasonable.estimate of the lower level of the deficit is$1.0 billion. Upper boundaries of the problemhave been placed in the $3.0-$4.0 billionrange."'" Among a number of Fc'deralagencies addressing this problem, the De-partment of Defense stands out by proposinga 5-year. $150 million initiative to funduniversity instrument purchases in areasrelated to DOD scientific programs.

The first grants under DOD's new instru-mentation program have recently beenannounced. The concept is that each servicewill have $10 million per year. in addition toits research support budget, to allow univer-sities to purchase scientific instrumentsneeded to conduct new, or to improveexisting, research efforts in areas of DODinterest:20

Both the Defense Science Board anduniversity administrators had called for aDOD instrumentation program equal to 25percent of DOD's basic research budget, orat least that portion of the budget going touniversities. By generous calculation, theproposed initiative is some 4 percent -of theDOD basic research budget, and over 5 years

57

would add up to only 15 percent of theminimum estimate of the current need. Still, itis the most identifiable Federal response toa widely perceived problem in the researchsystem; and it could make a meaningfulcontribution to ameliorating that problem.The DOD instrumentation initiative may pro-vide an example for other Federal agenciesto emulate; recent Congressional testimonysuggested that DOD should "take the leader-ship role in establishing programs for the sup-port of facilities and instrumentation."2'

University-Industry Relationships forDefense R &D: The Department of De:fense is attempting to encourage closer tiesbetween defense contractors and researchuniversities. Each defense contractor conductsan Independent Research and Development(IR&D) effort; with about one -third of thecosts of this effort being provided by DOD.The purpose of IR&D is to allow defensecontractors to develop their own ideas inareas of potential value to national securityobjectives, as distinct from the R&D theycarry out to meet DOD contract require-ments. Currently, essentially all of this !R&Deffort is conducted in-house by defensecontractors, but the Defense Science Boardrecently recommended that "industry beencouraged to support work at universitiesthrough the IR&D route."22 Such a proposalis currently under discussion within DOD,and various ways of providing incentives toconvince industry to channel some portionof its current IR&D funds (DOD provideSdose to $1 billion/year in IR&D reimburte:ment to its contractors) or more likely, totransfer newly provided funds to universityresearchers The financial and policy implica-tions of such an initiative could be significantin coming years, as part of the emergingpattern of industry-university connectionscenters around defense-related R&D.

Issues Related to DOD Support ofIndustrial R&DEarlier, it was suggested that defense indus7tries were highly specialized institutions andthat their R&D activities may have littleconnection with nondefense progress inscience and technology. In one sense, thiswas not an accurate characterization; in thepast 10-15 years, according to one recentDOD report

58

the character of the defense industry haschanged significantly. The large primecontractors and major subcontractors areno longer stand-alone organizationsdevoted primarily to defense business.The companies have become elementsof large multi-market organizations andmust compete internally for the limitedcapital that is available....There are strongindications that the return on investmentin the defense sector has deteriorated...and that investment is going to the non-defense sector because of higher yieldsand lower risk The situation is exacer-bated by the instability in the defensemarket; as evidenced by changing pro-gram requirements. As a result, thedefense industry is under-capitalized.23

Whether or not this is an accurate pictureof the present situation, the current defensebuildup, if sustained, is likely to make thedefense business an attractive propositionto corporate managers. The results of areemphasis on defense industries, from anoverall national perspective, require moreextensive examination than is possible withinthe scope of this paper. A prominenteconomist has suggested that one pressingissue is

how the U.S. can maintain the industrialstrength to compete with other countriesin civilian production and sales. The basicproblem here is not so much one of ob-taining critical raw materials and equip-ment, although there may be shortagesof both, but is one of skilled workerscraftsmen; engineers, and scientists.Such people will tend to be attracted tomilitary production: Defense contractorswill entice workers away from civilian firmsby paying higher salaries as they buildup their work forces on a crash basis.But even if the salaries were identicalthere would be a tendency for the mosthighly qualified people to move intodefense. For most engineers, such workis Simply more exciting 24

Close observers of the defense industry areslightly more optimistic about the capacity ofU.S. industry to perform additional defense-related work without dislocations in thecivilian economy. For example, the Defense

43

Economics Research Report concludes that"the large groWth rates demanded of thedefense-supplying sectors relative to theirrecent levels of growth, the fact that theselarge grotAAh rates must be achieved acrossnumerous industrial sectors, and the factthat they must be sustained for a lengthyperiod all combine to provide a quantitativebasis for the concerns regarding potentialbottlenecks, lengthening leadtimes, and pricepressures." The report recognizes that "theproblem may be a real one of non-trivialsignificance." but suggests that "effectivemanagement; investment, and worker train-ing could lead to the avoidance of many ofthese problems."25

The major issue relevant to this paperemerging from thiS analysis is the likelydemands from the defense industry on theNation's pool of skilled scientific and engi7neering personnel. This problem is discussedin more detail below.

Another issue related to increased DODsupport of industrial research is whether theresults of that research will find their wayinto the nondefense sector, so that they canbe considered in terms of their potentialeconomic or social benefits. Histoncally, thistransfer process has been a major source ofcivilian technological innovation, at leastaccording to most analysts. As defensesystems become more specialized and distinctfrom nondefense analogs, DOD researchinvestments may be more difficult to turninto results of broader general benefit tothe economy. On the other hand; as hasbeen the case in the past; new and currentlyunanticipated opportunities for civilian appli-cations with major economic payoffs couldresult from lines of research that DOD intendsto support. Ensuring that the country getsmaximum payoff from investments of publicfunds in R&D support, by DOD as well bynondefense agencies, will require continuedGovernment sensitivity to opportunities forthe floW of defense technology into thecommercial sector. Heightened concernabout the need to control access to 8en8i:tive technical information is another elementthat could impede the process of technol-ogy transfer from defense to nondefenseapplications.

Issues Related to Scientific andEngineering Personnel

An issue on which the interests of universities,Government laboratories, and industry con-verge is the demand for, and supply of, skilledscientific and technical personnel. Currentlythere are some 1,600 vacant faculty positionsin U.S. engineering schoolt, and the Depart-ment of Defente currently estimates it hat5.000 unfilled civilian and military openingsin science and engineering. The Governmentand the universities, each with their ownalary congtraintt, are competing with healthy

and growing defense and civilian industriesfor a limited supply of scientific an,' par-ticularly. engineering professionals. TheDefense Science Board reported recentlythat "DOD and the country face a crisis inthe availability of technical personnel." TheBoard also noted that "over the long runthe universities and DOD will have to respondto market pressures in upgrading their scienceand engineering staff:" 2b However, as onerecent analysis reported;

If present undergraduate enrollmenttrends persist; there should continue tobe enough new graduates in most broadfields of science and technology to satisfyanticipated demands through the decade.However, spot shortages do exist in certainsubspecialties, and others may develop.The greatest problems at present appearto relate to engineers and computerscientists. University faculties, the armedservices, and, in some critical fields, privateindustry are likely to continue to experi-ence difficulties in recruiting and retainingqualified engineers and computer scien-tists, particularly persons with advanced&Greet."

The DOD research buildup includes in:creased support for undergraduate andgraduate education in science and engi:neering. In 1981, while DOD employedalmost 230,000 scientists and engineers, onlysome 21,050 students were receiving DODsupport of some type, and only 50 graduatestudents were recipient of DOD fellowships.28In the main, the services are increasing theirsupport of undergraduate education through

59

their ROTC programs. The Navy requiresthat 80 percent of the recipients of ROTCscholarships major in science and engineer-ing: the Air Force; 70 percent. The Armydoes not have a similar policy, although it isunder some pressure to do so. The servicesare also attempting to increase the size oftheir ROTC programs.

Beginning in fiscal year 1983, DOD sup-port of graduate education is being sub-stantially increased. All three services planfellowship programs with awards in mostcases substantially larger than comparablefellowships from nondefense agencies. Thefellowship stipend will be $12,000, with thehost institution receiving an additional $8,000;the total number of ttudentt to be supportedis approximately 100. DOD fellowthip pro-grams are planned to increase in subsequentfiscal years, but dearly they will support only asmall fraction of the graduate studentsneeded to meet DOD i equirementt foradvanced training in science and engineering.University officials have suggested that DODsupport up to 1,000 new graduate studentseach year-. but this seems unlikely in thecurrent economic climate. 29 Of course, othergraduate students are supported as researchassistants on DOD-funded projects: thecurrent estimate is that some 4,000 studentsreceive such support. One issue here iswhether the comparative financial attrac-tiveness of DOD fellowships will attract adisproportionate share of the best under-graduate science and engineering studentsinto DOD:related work. Prbposals have beenmade that DOD fellowship programs includethe provision that recipients work one yearin DOD laboratories for every year of grad,uate support they receive, though this is notcurrently a requirement.

One attempt to steer promising youngerstudents toward defense-related researchdeserves mention. In July 1981, the De:partment of Defense established a scienceand engineering apprenticeship program forhigh school students to stimulate broaderinterest among students in science andengineering careers and to establish individualworking relationships among students andactive researchers. The program is executedby individual DOD laboratories and by the

tcientific officers responsible for the Army,Navy, and Air Force research programs. Theminimum age limit for the apprenticeshipprogram was relaxed by the Office of Per-sonnel Management to allow employmentof high school freshman and sophomoresaged 14 and 15 years. DOD sponsors ofapprentices are particularly encouraged torefer promising graduates of the apprentice-ship program to other DOD laboratoriet inthe communities where the student intendsto attend college. If successful, these programsand referrals will direct student scientists/engineers toward defense-related researchand issues and, perhaps, toward ultimateemployment by the Department of Defenseor its contractors:

The role of DOD in support of graduateeducation in the United States is likely toremain relatively limited; however; a recentreview of defense R&D concluded that "aninvestment in 20 thousand more Ph.D.s inscience and engineering todaycostingsociety perhaps $2 billionwill be worth, interms of military deterrence and nationalsecurity, many times the $2 billion cost of afuture division or air Wing."3° It seem unlikelythat this kind of argument will carry muchweight as the country considers how best toenhance its defense posture in coming years.

Conclusion

The major findings of this analytit have benidentified earlier, but are worth restating.The tentative nature of these finci.ngs shouldbe emphasized: as the WS R&D analysisremarks. "in many respects it is still too earlyto see the real impacts of changing fundingpatterns...the real impacts...will not be felt atcolleges and universities until some time in1983. and perhaps later. "31

All of those who have considered emergingissues in the U.S. research system of the1980s have identified the supply of well-trained engineers. and particularly of engi-neert with graduate degrees and/or whoare U.S. citizens. as a major concern.'2 Thispaper reinforces that concern, and suggeststhat a major impact of increases in DODR&D budgt overall is related to availability

0045

of and competition for engineering talent:DOD funding will go primarily to the aero-space industry for development of newmilitary systems, and this may well create; in acompetitive personnel market with limitedsupply-, a bias among individuals toward DODwork; since it will offer both exciting workand high salaries. Universities and perhapseven civilian industry may find it increasinglydifficult to attract the best people, given thisdefense-industry competition.

To meet the growing demand for scientistsand engineers, U.S. colleges and universitieswill have to attract more well-qualifiedstudents to technical concentrations andincrease the flow into the economy ofgraduates at the bachelor's and, particularly,the postbachelor's level. There are likely tobe. insufficient numbers of qualified facultyavailable, especially in engineering schools,unless the current state of affairs with respectto faculty salaries and working conditionsis significantly improved. The Departmentof Defense has recognized this situationand is attempting to find ways to helpameliorate it.

Attempts to reestablish a mutually satis-factory DOD-university relationship will mostlikely continue to create policy controversy.More generally; university leaders are ex-hibiting some ambivalence and skepticismabout the changing character of Government-science relationships in general; someobservers think that "recent cutbacks innondefense R&D are unprecedented andhave significantly eroded the faith of manymembers of the scientific community in theunderlying stability" of Government policyfor academic science."" The prospect ofincreases in research, instrumentation,and fellowship funds from DOD is veryattractive to universities that perceive them-selves in difficult financial condition. Yet,there are concerns about the conditions ofDOD support, such as export control require-ments. and about excessive dependence on asource of funds that has had a "stop andgo" record of university support over thepast two decades. A DOD-University Forumhas been created as a vehicle for discussionof such issues as science and engineeringeducation; export control requirements; andforeign language and area studies efforts;DOD is attempting to involve itself more

46

actively in science policy discussions at anumber of levels;

This paper has concentrated on the, effectsof increased DOD support on the U.S.research system overall. But the relationship isa two-way path; there are important questionsrelated to the contribution of the Nation'sresearch system to its overall national security.A recent review of how to assure this con-tribution concluded that "it is not possibleto separate purely military activities in scienceand engineering development from thosethat have broader economic, exploratory,academic, or social rationales for their pursuit"and stressed the important relationshipsthat tie our military R&D programs closelyinto the larger technical problems of oursociety....Healthy military R&D can flourishonly in a healthy overall R&D environment. "'

Increases in DOD R&D funding; by them-selves, will not achieve the desired objectiveof improving the technological basis of US:national security: The effects of those in-creases must be evaluated in a systemwidecontext; and national science and technologypolicy should be adjusted to facilitate thethe adaptation of the U.S. research systemto changing priorities, requirements, andfunding pattems. Only by coordinated policydevelopment can the United States receivethe full benefits. in both defense and non-defense sectors, of Federal investment inresearch and development.

Notes and References

1. Logsdon, John M. (ed.). The Research System inthe 1980s: Public Policy tssues. Philadelphia:Franklin Institute Press, 1982 p. 2,

2. National Science Foundation: The 5Year Out-look on Science and Technology, 1981. Wash-ington, DC: US: Government Printing Office;1982. p. x.

3. Piore. Emmanuel. "Science: Government; andPolicy: A Four.Decade Perspective." In Logsdon,op. cit.. p. 34

4. Muller, Stephen. "A New American University." InHouse of Representatives. Committee on Scienceand Technology. National Science and TechncitogyPolicy Issues 19_79, April 1979, p. 93.

5. Olds. Bruce. "Return on Investment in BasicResearchExploring a Methodology." The Bridge(Spring 1982): 27.

6. Bloustein. Richard, President. Rutonrc the StateUniversity of New Jerscy. in House of Representa-tives. Committee on Armed Services. Hearings on

61

Military Posture.Pati. 4; Research and Developmei% 1979. p. 891. These hearings are hereafterreferenced as House Hearings,

7. Unless specifically in-die-Med otherwise, numbersin this section are the author's calculations basedon data in Shapley,WilliS, Teich, andBreslow, Gaii. Research and Development: AAASReport W. Washington. DC: Aitieritan ASS5ciaticirifor the Advancement of Science, 1981, and Shaplesr,Willis Teich; Albert. and Weinberg, Jill. ReSearchand DeVerciPinent: AAAS Report VII. Washington;DC :-American Association for the Advancement

Science, 1982. The volumes are hereaftercited at AAAS, Report VI and AAAS. Report VII.respectively.

8. /1AAS. Report %/Lop. cit4 p. 108.9. Defense Science Board. Summer Study Panel o-t

Industrial Responsiveness. January 1981. This report discusses the probleMS of mobilizing U.S. industry to support the defense buildup; _

10. Grove, H. Mark, Assistant Deputy Under Secret,*of Defense for Research and Advanced Technology,testimony before the Defense Subcommittee ofthe House Committee on Appropriateness.dune 16. 1982, p. 11.4:

11. Bronstein, in House Hearings, op. cit., p. 948.12. Sproull, Robert; President; University of RcicheSter,

in House Hearings, op. cit., p. 908.13. Williamson, David (ed:): R&D for Mineola! Sfrangth.

Washington, DC: Georgetown University, Centerfor Strategic and International Studies; 1982, p. 10:

14. Muller, op. cit.. P. 95.15. These issues also are fully discussed in National

Academy of Sciences. National Security and Sci-entific Communication, Washington, DC: NationalAcademy of Sciences. 1982.

16. Glanz. Stanton A. and Albers, Norm V. "Depart-ment of Defense R&D in the University." Science(22 November 1974): 706.

17. Defense Science Board. 1981 Summer StudyPane! on Technology Base. November 1981.ChaPter II diScussei technologies that could make"an order of magnitude difference in terms ofdeployable operational capability." It is in theseareas that DOD intends to make its technologybase investments.

18. AAAS. Report VI, op. cit., p. 1_09,19. Quoted in Natibrial ReSearch Council.Revitalizing

Laboratory Instrumentation._ Washington. DC: Na.tional Academy Press, 1982. p. 89.

20. ibid., pp. 20-21.21. Wright: John: President, University of Alabama.

Huntsville, in House Hearings. op. cit., p. 923.22. Defense Science Board: 1981 Suinmer Study Panel

ott TeehnOlogy Base. op. cit., p. V-5.23. Defense Science Board: Summer Study Panel on

Industrial ReSPOnSiverieSS, op. cit.. p. 7.24. Tliurow. Lester. "How to Wreck the Economy."

New York Revieth of BoOks (14 May 1981): 3.25. Data Resources, Inc. Defense Economics Research

Report. April 1982.2b. Defense Science Board. 1981 Summer Study Panel

on Technologi, BaSe. op. cit., p.27. National Science Foundation. op cit., p. t.28. Defense Science Board. Report of the Task Force

on University Responsiveness to National SecurityRequirements, January 1982, pp. 3-12.

29. Bloustein, in House Hearings, op. cit., p. 948.30. Wiromson; op: cit; p. 12:31. AAAS, Report VII, op. cit., p. 103.32. Karl Willenbrock, his essay "Engineering-The

Neglected Ingredient," in Logsdon, op. cit., discusses the problems of engineering faculties andof supply of engineers.

33. AAAS. Report VII. op. cit., p. 101.34. WilliamSon, op. cit., pp. V, 3,

47

Training and Utilization of EngineeringTechnicians and Technologists

AbstractThe training and utilization of the "technical workforce"engineering technicians andtechnologists ..vho support engineering activitieshas emerged as a serious national policyissue. Three dimensions of the issue are explored in this paper: its relationship to theenhancement of productivity, the mismatch of technical training with jobs. and the crowingfinancial constraints on the institutions that train the technical workforce. Trends anddevelopments of the last decade underscore the need for better supply-demand data onthis workforce and on the job skills defined by geographic and industrial service areas.The complementary roles of four institutional actorseducational institutions, :ndustry,Government, and professional associationsare described and linked to the individualprocess of career choice. Industry's role appears to be pivotal in clarifying training andutilization opportunities for engineering technicians and technologists in the United States.However, new patterns of cooperation and initiatives on the part of all four institutionalactors are likely to be essential if the projected national need for associate and baccalaureatedegree technologists is to be satisfied.

IntroductionAt least since Sputnik I, science and engi-neering manpower has been a national policyissue. One component of this manpower isthe "technical workforce"engineeringtechnicians and technologists who supportengineering activities. Although the impor-tancg of this non-Ph.D. and mostly indus-trially employed workforce has been recog-nized, a gap in our manpower knowledgeexists. This paper seeks to analyze trendsand developments in the supply and demandof this workforce during the last decade,discussing the unique but complementaryroles of vailitis institutional actors and pro-posing specific policy options for clarifying,and perhaps creating; training and utilizationopportunities for engineering techniciansand technologists in the United States:

There are at least three dimensions totechnical training as a national policy issue:The first relates to the mismatch of trainingwith jobs (the so-called underemploymentquestion). the second to the growing financialconstraints on the irstitutional producersof the technical workforce. and the thirdto the enhancement of productivity. Thissection provides a brief overview of theseissues. The next section discusses evidencefor the tentative conclusions reached in thisoverview.

Technicians and technologists enhancenational productivity by augmenting the

engineering workforce. Technicians can beeducated in 2 years rather than 4 years.While a technologist holds a 4-year bac-calaureate degree, a technician holds a 2-year associate degree in an engineering oran industrially related (nonhealth, non-business, or nonagricultu re) technology.'But the definitional distinction associatedwith technician-technologist training isroutinely ignored by industrial employers.For example, technologists are often classi-fied as engineers. To blur the distinctionfurther, associate degree technician graduatescan easily continue their education inbachelor degree technology programs.2However. it is very difficult for techniciansto transfer to engineering curricula andreceive credit for their past academic work.3Consequently, there are discontinuities cre-ated by the educators (or producers) of thetechnical workforce that are not observedby the consumers of that workforce.

A related problem concerns a possiblemismatch between student enrollments intechnical courses and personnel demand.Technical educators can often identify areasof manpower shortage and are often in aposition to sense the aggregate effects andtrends of employers' demands over a ratherlong term. However, student enrollmentsrarely reflect these assessments. Most tech-nical educators agree, for example, that thereare critical shortages of technicians andtechnologists in computer software and

49

hardware, digital instrumentation and con-trol, computer-aided drafting, computer-aidedmanufacturing, energy, and environmentalsystems. Yet, other educational areas thathave a persistently low demand often enrolllarge numbers of students.

The mismatch of training with jobs meansthat American industry uses many engineersin positions more appropriate for techniciansand technologists. As a result, many engi-neers are underutilized by being assignedduties for which they have had little trainingor education. This, in turn, affects morale,productivity, and turnover: A short-termchallenge, in a nutshell; is to set engineersfree from such tasks: The persistence oftnese practices on the part of American in-dustry seems to be due to three factors:insufficient knowledge of the capabilities oftechnical personnel. the inertia of pastpersonnel practices; and the growth (plusinevitable variability) in technology educa-tion itself.

The costs for providing such educationcontinue to escalate. The 2- and 4-yearinstitutions that train technicians todaymust cope with severe constraints on theirability (a) to attract and retain quality faculty,and (b) to purchase new equipment. Tech-nology faculty need to be familiar with thelatest industrial techniques; therefore, theyneed ongoing industrial experience. Indi-viduals with such skills are reluctant to leaveindustry to go into education full-time wheresalary levels are low. And many educationalinstitutions do not differentiate in teachersalaries between high-demand and low-demand technical fields:

Faculty for technician education programshave master's degrees, typically in educationor engineering; or an equivalent bachelor'sdegree and licensure: But neither engineer-ing nor education master's degree workadvances the knowledge of the teacher inthe technology in which she/he must teach.Likewise: the few master's degree programsin technology are teaching programs ratherthan technology programs!'

Because technology education is at leastas, if not more, equipment-intensive than isscience or engineering education, the wide-spread introduction of computers has mademuch existing equipment obsolete and re-placement equipment very expensive. Few,

50

if any, educational institutions seem to haveenough money to devote to the training offaculty, the development of instructionalmaterials, or the purchase of ,new equip-ment. Whereas engineering education leadsin the faculty shortage crisis, technologyeducation is most affected by the equip-ment cnsis. If technicians are being trainedto become productive practitioners imme-diately upon graduation, they need to leamhow to operate state-of-the-art industrialequipment.

If technology educational institutions arenot able to afford modern equipment andtraining of faculty; an educational systemthat should lead the Nation's technologyis apt to fall behind prevailing practice inindustry: Indeed; industry is compelled toinstitute extensive on-the-job (re)trainingprograms; diverting talented individuals(both the industrial trainer and the companyneophyte) from the productive work theycould otherwise do. Thus, the financialconstraints felt in the training capacity ofeducational institutions could have far-reaching effects, contributing both to theunderemployment and "worker-readiness"problems and, ultimately, to lags in company,industry, and national productivity.

The present tendency to misuse both engi-neers and technicians could also affectproductivity. Inasmuch as engineers in theUnited States are no longer educated forthe shop (now the province of the "crafts-man"), the industrial plant, or the I-boratory,it is technicians and technologists whomustdo this work Yet, at present. far fewer tech-nicians and technologists are graduated thanengineers: With the advent of the "hightechnologies"which typically involveelectronics and computers in their implementationindustry must recognize who istrained to do this work, and then title andutilize these personnei accordingly. Theeffectiveness with which this is done hasimplications for the larger issues of produc-tivity and technological competitiveness.

Trends and Developments

Technical Manpower Statistics

Planning educational programs that pre-pare individuals to enter specific niches in

64

the workforce requireS reliable statisticalinformation and projections. To the extentthat statistical information and correspondingoccupational definitions are lacking, humanand economic resources are inefficientlyexpended and the needs of individuals andemployers alike are not met.

Depending upon the sources of the sta-tistics; and there are many, information onspecific occupations varies greatly. Tradi-tional professional personnel, such as sci-entists, engineers, accountants; and architects,and easily identified craft occupations, suchas barbers, machinists, and carpenters; tendto be accurately counted. The less under-stood and inconsistently defined occupations,such as technicians, receive short shrift,usually in the form of gross statistical aggre-gation or partial coverage. A review of Fed-eral statistical sources (Department of Labor,Department of Commerce, and the NationalScience Foundation) and commercial andprofessiona! research publications (Con-ference Board's Help Wanted index, Scientific Manpower Commission, AmericanAssociation of Engineering Societies [AAES]and National Society of Professional Engi-neers [NSPE]) yields the conclusion thatthere is a national dearth of informationon engineering technology and that what isavailable generally lacks the detail neces-s for eauca Iona an -reer planning.

Nevertheless; some general trends, bydegree level and curriculum category, canbe discerned. For example, the Departmentof Education reports that for the 10.yearperiod between 1970-71 and 1979.80; thenumber of associate degrees (at least 2 butless than 4 years of postsecondary work)conferred Showed the largest numericalincrease among all levels of degrees. Thelargest increases aLso occurred in science- andengineering-related occupational curriculumcategories, with women experiencing greatercategorical gains in the percentage earningassociate degrees in technological areas.Overall, 2-year institutions conferred anaverage of six times more associate degreesthan 4-year institutions over the decade, the1979-80 ratio being 350,000 to 60,000.Public institutions conferred 85 percentof these associate degrees .6

The most recent, and distaggregated, profileof associate degree conferrals is presented

in Table 1. It summarizes occupational cur-riculum totals divided first by "science- andengineering-related" vs. "nonscience- andnonengineering-related" categories, andthen by technological area within each. Theseareas are ghoWn by sex and curriculum typeto illustrate the current Supply-Side configura-tion as we entered the 19806. Note thatthe "mechanical and engineering tear-nologies" area is second only to "healthservices and paramedical technologies" intotal associate degrees awarded and first inawardAs based on 1-2 year curricula.

Demand statistics are more difficult tofind. Operational definitions of the manytypes of engineering technologists are stillnot widely available and hence precludesufficiently large numbers of employers fromincorporating the title "engineering technol-ogist" into their manpower plans: EstimatesOf the technician workforceengineeringand science technicians are treated as cate-gorically the samehover around 1 million.A 1980 Bdreau of Labor Statistics (BLS)bulletin offers a 1978 employment figureof 600,000,6 while a more recent Occupa-tional Employment Statistics projectionmatrix lists the same category of the work-force as totaling 1.25 million in 1980. About80 percent worked in private industry. Andwithin the manufacturing sector, the principalemp oyers w al equipment,chemical, machinery, and aerospace indus-tries: The Federal Government employedapproximately 100,000 technicians in 1981,with the largest number located in theDepartment of Defense.

According to a joint 1980 National Sci:ence Foundation/Department of Educationreport,7 the dcritand for technicians, tech-nologists; engineers, and scientists willremain strong through the 1980s. Therewill be over 30 percent more new jobs inthose fields during the decade. A 1981 Sci-entific Manpower Commission reports pegsthe demand at 375,000 new jobs for engi-neering and science technicians and placesthe overall growth rate at-38 percent from-1978 to 1990. The employment outlook isuniformly ekellent, as summarized in Table 2.

The lack of a specific category for tech-nologists that they are included inunknoWn proportions in the engineer andtechnician data. HoWever, by deriving the

5165

Table 1

Associate Degrees and Other Awards Based on Occupational Curriculums, byLength and Type of Curriculum and by Sex of Recipient:

United States and Outlying Areas, 1979-80Awards based on organized occupational cuniculums

Curriculum

All awardsAt least 2 years but less

than 4 yearsAt least 1 year but less

than 2 years

TOW Men Women Total Men Women Total Men Women

Occupational cur:riculums: total 353.333 158.647 194.686 278.555 127,191 151.364 74.778 31.456 43.322

I. Science and engineering:relatedcurriculums 193.532 96.510 97.022 144.703 70.948 73.755 48.829 25.562 23 2-67

A. Dataprocessingtechnologies 15.147 7.525 7.622 12.560 6.616 5.944 2.587 909 1.678

B. Health servicesand paramedicaltechnologies 86.647 10.322 76.325 66.452 8.222 58.230 20.195 2.100 18.095

C. Natural science'echnologies 19.214 11.536 7.678 14.431 8.743 5.688 4.783 2,793 1.990

D. Mechanical andengineeringtechnologies 72.524 67.127 5.397 51.260 47.367 3.893 21.264 19.760 1.504

II. Nonscience andnonengineeringrelatedcurriculums 159.801 62.137 97.664 133:852 56.243 77.609 25.949 5.894 20.055

A. Business and comcoerce technologies 124.485 44.512 79.973 102.557 40.591 61.966 21.928 3.921 18.007

B. Public servicerelated technologies 35.316 17.625 17.691 31.295 15.652 15.643 4.021 1.973 2.048

Source: U.S. Department of Education. National Center for Education Statistics. Digest of Education Statistics.1'182. Table 123 (adapted)

projected average annual groWth in thetechnician labor force from Table 2, underLow Trend and High Trend II (conservativeeconomic optimism) assumptions. and com-paring the supply of graduates from Accredi-tation Board for Engineering and Tech-nology (ABET)- accredited programs, anindication of how demand is being met canbe inferred. There appears to be a consid-erable undersupply of technology graduates.but the data include neither Bachelor ofScience in Engineering Technology (BSET)nor any industrial technology degree statis-

--- OM a total of more than 19,000 additionaltechnology degrees. When these are con-sidered, the AAES- estimated shortfall of16.000-20.000 disappears.9 This suggeststhat there is an ample supply of technologymanpower to cope with demand. It alsoleaves the supply from all nonaccredited

52

schools not included in the AAES estimatesto compensate for elasticity, deaths, andreirements of the technicians and technol-ogists who move into jobs that are mistakenlycounted in the "engineer" category.

The statistics indicate that any expansionof technology education programs shouldbe very carefully evaluated, not only on thebasis of national projections, but on theassessed needs of the geographic servicearea of the program.'" Although nationaldata may suggest that a near balance oftechnician supply-demand exists, needs formore individuals with certain technicalspecialties in specific geographic areas canalso exist to justify new or expanded tech-nical programs. Clearly, a major obstacle toeducational planning is the very limitedbreakout of technical specialties reportedin national statistics. This underscores the

66

Table 2

Labor Force Projections for Engineers and Technicians (1978-1990)

Employment (in thousands) 1978Technician/

EngineerRatiosOccupation

1978 1990Low Trend

1990High Trend

1990High Trend II

AU Engineers 1:071 1,504 1.624 1:531

AeroAstronautical 57 98 104 100Chemical 53 68 73 70Civil 149 208 218 2nElectrical 291 441 479 448Industrial 109 146 159 148Mechanical 199 274 300 279

All Engineering & ScienceTechnicians 1.160 1,577 1.700 1.609 1.08

Drafters ......... ......... 293 412 446 419Electrical & Electronics 319 464 512 478 1.10Industrial 31 40 41 0.28Mechanical 45 61 67 62 023Surveyors 51! 73 75 76 0.36

Source: Scientific Manpower Commission. Odober 1981.

need to consider an educational institution's-geographic service area in the planning oftechnician training programs..

Institutional Actors and Roles

Several institutional actors have already

become formidable: One language that isunderstood by both industry representativesand technical educators is a job task state-ment inventory for each technical field;Despite the investment of considerableFederal funds in computerized job taskbanks intended to serve occupational edu-caturs--w-erigineer-in-9

beinridentified as_participants in The processof educating and employing the technicalworkforce. Foremost, these include educa-tional institutions and industrial employers.Underlying their effectiveness, however, arethe linkages and barriers that affect careerchoice, attracting and repelling individualsfrom occupational niches they are more orless trained to fill. The matching of trainedtalents to jobs goes well beyond nationalsupply-demand statistics to changing well-entrenched perceptions and practices.

The Industrial Coanection. In his bookon occupational education and industry,"the late Samuel M. Burt cited "confusionon the part of industry concerning how towork effectively with the schools" and "dis-illusionment on the part of industry...toestablish effective relationships with edu-cators" as major obstacles to responsiveoccupational education programs. In tech-nology education, those obstacles have

Statements existed for any field as late as1976. Since such statements are critical toimplementing a joblelated, individualized,computer-generated and computer-scoredcertification examination system, the Insti-tute for the Certification of EngineeringTechnicians (ICET) undertook in 1976 tocreate job task inventories for the many

of engineering technology. TheSe jobtask inventories proved to be an effectivelanguage for communicating with many dif-ferent employers, examination committees,Government officials, educators, and engi-neers. They also appear effective as careerguidance tools.

Similarly, job descriptions reflect em-ployers decisions in structuring specificpositions: These positions often differ fromemployer to employer because of firm size,type of business; capabilities of employees,and other factors. Therefore; managernei it

67 53

decisions are made to accomplish a missionutilizing particular manpower configurations.As a result, employers tend to assign dif-ferent duties to persons with the same tech-nical background.

These differences are especially notice-able in descriptions used by Governmentagencies and private firms as opposed tomilitary organizations. The mismatch of edu-cational requirements for job assignmentsis due largely to time constraints in themilitary training program and the narrowjob tasks associated with most military tech-nician assignments. But the mismatch canbe identified and managed by utilizing amatrix of courses by job tasks. Some of thejob tasks and courses; of necessity; need tobe broken down to fit the military system;but a considerable portion of the require-ments tams out to be common to the militaryand civilian sector; thereby creating newopportunities for military-civilian educa-tional cooperation.

Industry also responds to its training needsby developing materials on specialized topics.ASide from those- of a proprietary nature,these materials can be made available totechnolocgy schoolS. Contemporary technicaleducators, pressed by the rapid expansionof technological knowledge, must realizethat an entire field of technology cannot beimparted to students in 2 years. In any tech-

tasks overshadow the meager samplingfound in 2-year technician curricula. Civilengineering technology, for example, in-cludes highway design, construction, mate-rials testing, surveying, traffic engineering,buildings, structures, water systems, rail sys-tems. wastewater systems, environmentalprotection, safety, dams, waterways, andports. Beyond an underlying core, theeducator is confronted with an array ofalternatives from which she/he must select amajor part of the instructional program: Atthe point of specialization; employer inputof specif c job tasks can assist the educatorin making critical choices of curriculum andcourse content for the school's geographicservice are& Once the choices are madefor required and optional content; industry-produced training materials can contributeto the quality and relevance of the program.

54

Even using a job task inventory, however,the technical educator should obtain em-ployer input throughout the service area.With ever-shrinking school resources, edu- .

cators have neither the time nor funds todesign and implement cost-effective pro-grams by themselves. The altemative is mis-aligned technical education programs,unrecognized employer potential, and thelost opportunity to bring employers andpracticing engineers, technologists, andtechnicians into curriculum design and theteaching process per se.

Career Choice; Where does the technicalworkforce originate? Early educational ex-periences and aspirations initiate a processthat results in "career choice:" This decisionis a subjective response to the cumulativesocial images and pressures that family;friends; teachers; and others bring to bear:These "others" are often professional organi-zations. For example; guidance materials forengineering and technology_are producedby the National Executive Committee onGuidance INECG) on be_half of the engi-neering profession. NECG includes repre-sentatives of NSPE, ABET, and the AinericanSociety for Engineering Education (ASEE).Millions of booklets and brochures havebeen distributed to secondary schoolsthroughout the United States. Exhibits are

displayed-at-the-an nual-conferences-of_theNational Science Teachers Association andthe American Personnel and GuidanceAssociation. Films and slide shows aremade available, and kits for presentationsby engineers are sold for a nominal sum.In addition, NECG provides input to literallyhundreds of commercial guidance publica-tions that include engineering and tech-nology. But many guidance counselors knowlittle about career opportunities in engi-neering technology: And much of what theydo know seems to be outdated or derivedmainly from media comments and portrayals:

Perhaps the chief problem confrontingtechnical education as a career choice isfhe general-lac k-of understanding of theassociate degree: Parents; educators; andstudents identify the baccalaureate degreeas the key to success in life. Compoundingthis are salary surveys that consistently con-

firm the financial advantages of higherdegrees. Little it publiciied concerning jobsatisfaction, working conditions, and outletsfor creativityall of which do not auto-matically follow receipt of a 4:year degree.

Students must be reminded, too, that atechnical education is but a start in one'scareer course. Cooperative programs provideexperiences that familiarize students withactual work conditions, practices, equipment,and problems. This facilitates the transitionfrom school to work and gives the student achance to rectify shortcomings in his/herpreparation. Despite the operating expense,cooperative education experiences receivealmost universal praise from students andemploye s.

When one considers that technology andits career opportunities are constantly ex-panding, increased support; especially frombuSirieSS and industry, should be forthcomingto the institutions entrusted with trainingthe technical workforce. The forms of thissupport and cooperation are explored nextas policy 9:_ltions.

Policy Options

In this concluding section, several specificpolicy options are proposed. Each is dis-cussed with reference to one of the fourinstitutional actors introduced above=ptofessional associations, industry, educa-tional institutions, and Governmentwhoare seen as the loci of initiatives for fostering,coordinating; and implementing actionamong the producers and employers of thetechnical workforce:

Professional Associations

Because engineering-based associations aremost sensitive to trends and needs occur& Alin technician/technology training, they arein a key position to inform, debate. andexert policy pre!,sures on both the educatorsand the employers of the technical work-force. But the empirical basis for such actionlies in the collection, analytis, and dissem-ination of manpower statistics. Building onthe efforts of the National Sotiety of Pro:fessional Engineers and the Engineering

Manpower Commission to monitor theproduction of technology associate; baccalaureate, and master's degrees; profes-sional associations should act as informationclearinghouses. They can focus attentionon industry needs, especially sectoral andgeographic differences in opportunities fortechnician/technology specialties. Devel-oping data through employment surveyson supply and demand is an immediate need;augmented job task inventories would clarify,over the long term, job titles and descriptions,prerequisite SkillS, and career paths.

A continuing function, too, is the provi7sion of guidance information about technicalcareers. Professional associations recognizethat the maintenance of technical skill is asmuch a problem as maintenance of equip-ment. As products become obsolete, sodoes the workforce that designs, makes,sells; installs; tests, and services those prod-ucts. Thus; as production changes, so muststate-of-the-art technical training. Accredi-tation exerts influence here, as do profes-sional associations. But they can do more,namely, by acting as a broker between theprincipal institutional actors to bridge gapsand identify strategic points of articulation:

Industry

Industry_should invest boldly in underwritingtechnical educationespecially equipment,time-shared personnel, and cognitive inputsto -ctiritc;i1a. They would_ be doing this fortheir own good. For only by blurring furtherthe e'iStitclibri 1;etWeeri education as pursuit-of-degree and training as onthe -job workexp;ttietice. will industry spare itself the laterexpense of extensive retraining. Nevertheless,bringing the job into the classroom is nosmall feat. It requires an outreach effort thatidentifies a service area in which educationalinstitutions are Willing to engage in tbbp-erative internship programs. Inasmuch asfaculty shortages; equipment deterioration..and crowded facilities are common todayin technology schools; such industry initiativesshould indeed be welcomed:

Industrial subsidy of technician/technologytraining- .could emulate the "academic-industrial.' model occurring in bioengineeringand microelectronics.12 High-technology

69 55

corporate sponsors are donating equipment,providing matching funds, sharing labora-tory facilities, offering personnel as part-timefaculty. and contributing student fellowshipsand loans. These are innovations that arebecoming necessities. They carry induce-ments of cost-effectiveness and rejuvenationof personnel, while redefining the boundariesseparating institutional actors. Given eco-nomic realities and the rockets of successfulcooperation between education and industrythroughout the United States, the earlierand more salient introduction of industry tothe training process looms as a basic remedy.

Educational InstitutionsThe reciprocal role that educational insti-tutions should play if industry initiatives areto succeed should be obvious: If such rela-tionships are seen as partnerships; withindustry supplying most of the financialcapital and education most of the humancapital. everybody profits:

Technology education is still largelydependent upon engineering school grad-uates for its teachers. The limited numberof baccalaureate (8.469) and master's (30)degrees in technology in 1981'3 severelyrestricts the pool of potential instructors fortechnology schools. Industry representsa source of skilled part-time faculty. In addi-tion, active industrial advisory committeesshould become an essential element of everytechnician curriculum. Institutions offeringtechnology curricula need to have thecapability of revising, updating, initiating, orphasing out programs. The AccreditationBoard for Engineering and Technologycontributes to this review process. Yet, atpresent, obsolete curricula continue foryears while new ones are deferred becauseinstitutional procedures promote conserva-tism. Institutional response time can oftenbe cut by using computers and instructionalmodules in the educational approach andby creating departments where emergingtechnologies can be tried with experimentalcurricula; cooperative education; and eventemporary faculties.

Finally; good students are attracted tocareers that afford opportunities for advance-ment. To ensure a supply of quality associate

56

degree technicians; it is therefore advisablethat 2-year programs articulate with 4-yearbaccalaureate technology programs. About20 percent of the associate degree graduatesgo on to advanced technology degrees.About 50 percent will seek to advance theireducation at some time later in their careers.This "career ladder" could be reinforced byformal agreements between associate degreeand bachelor degree institutions. (In stateuniversity systems, unfortunately, this hasbeen difficult to achieve.) The career ladderbegins in high school with vocational edu-cation tracks. Here, the math and sciencecontent must increase so that graduates ofvocational programs will be able to raisetheir aspirations upon graduation from highschool and enter a technician program in ahigh-technology field. In sum, the evidencethat students enter vocational programsprimarily because of perceived job statusand intrinsic interest, and not because ofacademic ability." should be heeded bytechnology educators:

GovernmentThe Federal Government has an interest inthe development of the technical workforcefor the defense and productivity of theNation. Indeed. Washington should maintainthose programs that contain incentives forcooperative funding involving nongovern-ment actors. One way is through smallmatching fund grants that could be madeavailable for instructional laboratory equip-ment. The other half of the match could beprovided by either industry, a private foun-dation, or the educational institution, giventhat the technology program is being con-sistently supported by institutional capitalequipment funds. That way, the new fundscan make a significant difference.

Perhaps the greatest short-term assistancethat the Federal Government can offer isthe dissemination of guidance informationon technician/technologist careers withinthe Federal Government itself. Civil serviceclassifications for technicians and tech-nologists should be refined in such sourcesas The Dictionary of Occupational Titles,Occupational Outlook Handbook. andOccupational Census Data, and then"advertised."

The brunt of Government action; however,should occur at State and local levels. Forexample. recent efforts in ten States tomake legislators aware of the dimensionsof the engineering education crisis resultedin special appropriations_ in excess of $40rnillibn of aid. In four States. differentialpay increases were approved and imple-mented for engineering faculty membersan achievement long believed to be i.inat7tainable. Unfortunately, technicions and

. technologists have not achieved a similarlevel of organization engtio.e.:ing aid totheir type of edricati In. Stiqe universitysystems, it is difficu'it to convey the equip-ment, staffing, and per-5tticient instructioncosts to legislatures and boards of regentsdominated by a "liberal arts" or. conversely,a "Proposition 13" mentality.

There are several specific actions that Stateand local governments should consider. First.they could earmark more educational fundsfor instructional equipment in technology.These funds could then be awarded to

-technology departments as line items in thebudget for equipment or made available forState equipment grants in response tosolicited !proposals. Second. State boardsshould monitor institutional responsivenessto the rapid changes inherent in technicaleducation. Curriculum innovation shouldbe encouraged, especially in thoSe institu-tions that regularly undergo periodic external(ABET) accreditation review by specialistS.The awarding of Associate in Arts (AA)degrees in technology by vo -tech SchobISwill intensify competition among accreditedand State-funded 4-year technology institu-tions already laboring under severely strainedbudgets. Third, State boards should alsoencourage articulation between high schoolvocational programs and higher educationtechnology programs. A career ladderapproach to developing a highly qualifiedtechnical workforce should be exploited.High school vocational programs shouldprovide the option of entry into collegetechnical programs. Associate degree pro-grams should not prevent graduates fromtransferring. with advanced standing; tobaccalaureate degree programs in tech-nology. Graduate programs in technology,as opposed to engineering or education,

should be carefully nurtured to provide astream of faculty for technology programsat levels commensurate with local andregional industrial needs:

Policy Conclusions

State agencies, in concert with nationalprofessional associations and their localchapters, can act as catalysts to facilitateinteraction among all institutional actors ontopics of mutual interest: educational pro-,grams, employe-1 needs; occupational titles,faculty qualifications; and equipment needs.Through such interaction; the training andutilization of engineering technicians andtechnologists will be fully appreciated as apolicy issue distinct from, but integrallyrelated to. the productivity of science andengineering inanpower in the United States.

Notes and References

1. Accreditation Board for Engineering and Tech.nology. 49th Annual Report. New York: September 1981.

2. Wolf. L.J. "Articulation Between Associate andBaccalaureate Programs in Engineering Tech-nology." Engineering Education (December 1977).According to the "two.plus.two" concept, afterearning an assOda,0 degree. the student can getfull-time work while completing the other 2 yearsthrough part -time study. The trend, however, seemsto be toward 4 years of full-time study.

3. Greenwald. S.H. and Wecker, I.R. "Transfers toEngineering, 2 + 2 = 6." Engineering Education(May 1975)

4. Wolf. L.J. "M.E./M.E.T.Crossroads of Coales-cence?" Mechanical Engineers Division, Amer-innSOciety for Engineering Education, 88th AnnualConference. Universityof Massachusetts; June 1980.

5. National Center for EchiCatiOri Statiitici. "Asso.date Degrees: A Look the 70's." Statistics Bul.letiri (1981): 1, 8.

6. Occupational Projections and Training Data,1980 Edition. Bulletin 205:... Bureau of LaborStatistics. p. 60.

7. National Science Foundation. Science and Engl.neering Education for the 1980`s and Beyord.Washington: DC: October 1980.

8. Scientific Manpower Commission. Scientific,Engineering; Technical Manpower Commen.Washington, DC:October 1981.

9. American Association .of Engineering Societies,Engineering Manpower Commission. Engineeringend Technology Degrees. November 1981.

57

10. A good example is the growth of microelectronicsfirms in the Silicon Valley. See Useem. E. "Educa-tion and HighTechno(ogy Industry: The Case ofSilicon Valley." Unpublislied paper. August 1981.A student from a oneindustry irea may seektechnical education as an escape from, not a visainto, a local plant. The school itself accepts one-industry domination at its peril, tying its fortunesto the success of that industry.

1 Burt. S.M. industry and VocationalTechnrcalEducation: A Study of Industry Education-AdvisoryCommittees. New York: McGraw Hill; 1967:

12. Norman, C. "Electronics Firms Plug into the Uni-versities." Science. vol. 217 (6 August 1982):511.514.

58

13, Scientific Manpower Commission. Scientific.Engineering. Thchnical Manpower Comments.Washingtom DC:June 1982:

14. Alexander, R.L., Cook. M., and McDill, E.L. "Cur.riculum Tracking and Educational Stratification,"American Sociological Review. vol. 43 (February1978): 4766.

15. For contrasting views, see Rath, G.A. "The Crisisin Engineering Technology Education." Engineer-ing Education (May 1982): 799-800. and "NoFederal Programs are Designed Primarily to Sup-port Engineenng Education, But Many Do." Reportby the U.S. General Accounting Office, May 1982.GAO/PAD-82.20:

National Security Controls andScientific Information

Abstract

As part of a broad effort to stem the outflow of U.S. technology to the Nation's militaryadversaries and to maintain our technological lead; the Government appears increasinglyWilling to restrict the dissemination of scientific research information to foreign nationalsat scientific conferences, in American classrooms and laboratories; and through publication.The research community, bcith acaderrild and industrial; has expressed serious concernabout the substance and procedures of those restrictions; about recent enforcementpractices, and about the process by which the rules of the game are established.'

The issues posed are difficult and cornplex, and their resolution requires a clearerdefinition of our national security objectives. It also requires a clearer understanding ofwhether the vanous controls now in use or under consideration will be effective in prevent-1.1u P r", in Amenca's technological lead over the Soviet Union and can be administered:,r.rhout imposing unacceptable economic, administrative, scientific, or political costs onAmerican society. It is likely that broadly drafted rules of general applicability will generatemore serious problems than those that can be refined on a case -by -case basis, tailored tothe particular setting and the particular technology. More specifically, it appears that exportcontrols on scientific information are not likely to be coSt:effectiVe. Although somewhatgreater use of the Government's authority to classify GoVernment;oWned or Government-controlled information or to deny foreigners access to the United StateS might be appro-priate; it appears that enhanced dialogue between the Government and the Scientificcommunity and increased reliance on contractual restrictions in federally sponsoredresearch offer the greatest promise of success.

Introduction

Advanced technology underpins the Nation'smilitary strategy and its economic strength.Ca military strategy depends on maintainingtechnological superiority to counter thequantitative superiority of the Warsaw Pactnations; our trade position, heavily relianton the export or domestic purchase of goodsand services involving sophisticated tech-nology, depends on maintaining a tech-nological advantage over current andpotential commercial competitors. Bothdepend primarily on creative scientificresearch and continuing technologicalinnovation and leadership. But both alSodepend on denying, or delaying the tranSferof, some of that new science and technologyto certain users and for certain purposes.Because rapid scientific and :echnologicalprogress flourishes best in an open environ-ment; there is a constant in-built tensionbetween creation and suppression.

Laws and regulations restricting the dis-semination of scientific and technological

7

information for national security reasons haveexisted for many years. Classification andrestrictions on the export of technical datahave been well-established features ofAmerican society since World War 11. Suchcontrols, however, were traditionally appliedto a very narrow range of scientific informa-tion: Four interrelated developments haveboth spurred efforts to apply controls morebroadly and made it more likely that suchcontrols will seriously affect American scienceand technology.

First, America's relative strength comparedto that of its allies and of its adversaries haschanged substantially. In the early postwarperiod, the United States, undamaged bythe war, had both an overwhelming com-mercial advantage and a parallel techno-lOgical advantage resulting from the wartimeeffort Within the United States and the inflowof talented scientists who had fled Hitler'sEurope. That economic and military positionhas eroded. Where once a unilateral U.S.decision to withhold goods or technologiesfrom the Soviet Union might have been

59

effective, today the successful imposition ofnational security controls requires the coop-eration of our allies, principally through theCoordinating Committee (COCOM). Sim-ilarly. where three decades ago Amencanscience and technology were supreme andsubstantially self-sufficient, today Amencanscience, technology, and industry frequentlyrely on the knowledgeand the man-powerof other nations. This relative ero-sion of American power and an accompany-ing perception of vulnerability seems to haveincreased the Nation's urge to control evenas its ability to do so unilaterally declined.

Second. technology export is more frequently perceived as an important contributorto America's military and commercial vul-nerability: The Government is increasinglyconcerned that purchased or stolen Americantechnology is an important component ofSoviet military strength: Large segments ofAmerican society believe that unwise techno logy exports have accelerated commrrcialcompetition both from industrialized nationssuch as Gerrnany and Japan and fromadvanced developing nations such as Taiwanand South Korea. Those who advocatestricter national security contrc:Is may makecommon cause with those whose principalconcerns are commercial,_ thereby blurringthe important disrinc:ions between the twosets of concerns.

Third, as pressures to control informationincrease, some have begun to question thelong-accepted premise that unclassified,nonproprietary research in general anduniversity-based research in particular wouldnot be controlled except in the most unusualcircumstances. Although all agree that mostuniversity research raises no national securityconcerns; many universities are now per-ceived as doing the kind of applied researchonce found only in commercial Qr govern-mental laboratories; and even basic researchis more likely to be seen as having nearterm applicebility. Recent acceleration ofindustry/university cooperation in a numberof fields, important to continued technologicaladvance, contributes to this perception. Ifuniversity research resembles other research,there is a strong argument that it should besimilarly controlled.

The last major development, closcly relatedto the weakening of Arnerican technological

60

supremacy, is the increasingly internationalface of science and of the American researchcommunity and the heightened role of themultinational corporation. Over 40 percentof all citations found in U.S. journals are toforeign publications, and large numbers ofcollaborative research projects cut acrossnational boundanes.2 The Amencan campusis similarly international. In some fields,almost half of all doctorates are awarded toforeigners"many of whom then obtainimmigrant status and work in either academiaor industry. The R&D efforts of many cor-porations reflect both the increasing multi:nationalization of corporate activities andthe international complexion of recentdegree recipients:

The Nation's military and commercialposition will continue to be challenged inthe years ahead: Concern about the loss ofscientific and technical information is there-fore also likely to persist. Such concern;I:owever, does not lead to a single, obviouspolicy solution. The dissemination of scientificacid technical information occurs in a multi-tude of ways ranging from espionage topublicaiion in the open literature, intrafirrndiscussions, patent applications, or scientificconferences: a single form of control isunlikely to be equally effective for all typesof dissemination. Although building forseveral years, the current debate about thecontrol of scientific information was throwninto sharper perspective by application ofexport controls to scientific conferences, andmuch of the debate since has focused onexport controls. There are. however, anumber of other mechanisms that have beenor can be used with greater or lesser effec-tiveness to stem technology outflow. Thispaper reviews both export controls and thoseother mechanisms: First: however, it examinessome of the specific reasons; beyond thedevelopments outlined above: wIly Goceni-ment and the scientific community areconcerned and asks some fundamentalquestions about what technology needs tobe controlled.

Trends and Developments

The question of whether increased Govern.mental restrictions are needed or acceptable

must be judged in light of the threat toAmerican security and to American society.In judging the threat; it should be remem-bered that while the Soviet Union hasundoubtedly acquired much advancedAmerican technology: almost all has comethrough normal commercial channels. diver-sion from legal sales; or espionage: Littlehas come from normal scientific communi-cat:on.4 The Government. however; worriesthat the improper dissemination of certainstate-of-the-art scientific research; even ifoccurring rarely. could seriously damage theNation's military position. Beyond that, manyin the Government fear that in the futurethe Soviets will more consciously exploit theopenness of the research environment toacquire advanced technology.5

Just as the Government worries aboutthe future as well as the present. so toodoes the scientific community. So far. theexport control laws have had only limitedimpact on the research community: thepresent regulations provide considerablelatitude for most exchanges of informationwith countries other than the Soviet Unionand its allies. and they have only occasionallybeen applied to scientific activity. A greaterimpact. however; appears to be in the offingfor at least two reasons.

First. researchers are becoming increasinglyaware thai a potentially controlled "export"takes place when they discuss their researchwith foreign colleagues here or abroad: mailan unpublished paper to a foreign scientist;present a paper at a symposium with inter-national participation, or hire foreign graduatestudents to work on an advanced researchproject. Such awarene.,s enhances the like-lihood of compliance and therefore of affect-ing research. Additionally, recent increasesin the number of foreign graduate students,faculty, and researchers on American cam-puses of course mean that the u.port rulesapply more frequently.

Second, controls may in fact be tightened.Some see this as having hapoened already,pointing to enhanced enforcement of existingexport control laws and last winter's well-publicized but perhaps misinterpreted revi,_sion of the Executive Order on nationalsecunty classification. Others simply forecastgreater restrictiveness; citing the suggestionof the then Deputy Director of the Central

Intelligence Agency that broad prepublicationclearance might be required!' The Govern-ment's visible difficulty in narrowing andrefining the scope of the Defense Depart-ment's Militarily Critical Technologies List(which will be the core of the Nation's controlsystem in the years ahead) is alSo a sign offuture restraints.

Uncertainty about the present and thefuture begets worry. The absence of specificsalso hampers rational debate. Although therecent report of the National Academy ofSciences (see reference 1) greatly enhancesthe likelihood of useful arid collaborativediscussion among the Government, industry,arid academia; it is not yet dear how theGovernment will answer several questionsand how it will reconcile competing interests.

Fundamental Questions to beAnswered

Export controls on scientific information arecurrently the main focus of attention. Toclarify their applicability or to assess alternativemechanisms, the Government must addressfour fundamental substantive questions: Howshould "national security" be defined? Whatdestinations call for controls? What tech-nologies will be controlled? And, what fornisof exchange require control? It Should alsoaddress the procedural question of how therules of the game are adopted or modified.

Defining "National Security`-':

If national security controls are to bebounded; "national security" must be defined.It is. unfortunately; an elusive term, used inmany senses. it can refer to short-term militarystrength: it can also refer to the Nation'slong-term strategic; political; and economicposition in world affairs. Current law isrelatively clear that "national security con-trols," in contrast to foreign policy controls;are to focus on those exports thai contributesignificantly to the military potential of ouradversaries. That, however, does not entirelydelimit the reach of nafional security controls.Soviet military strength, like that Of the West,depends on both a military establfshirientand an underlying industrial economy. Ex-ports that bolster t; se Soviet economy there-

75 61

fore enhance Soviet military capability inthe long term. It is widely believed. however.that controls focused on long-term Sovieteconomic strength are likely to be neithersuccessful nor cost-effective7 and that nationalsecurity controls should therefore concentrateon short-term military consequences. How-ever. even if one accepts this short-termmilitary focus. there is, as will be discussedbelow. the further practical difficulty ofdelineating those civilian goods and tech-nologies that are militarily relevant in theshort term from those that are not.

Geographic Scope of Controls

Scientific exchange and technology transferwithin the non-Communist world are largelyuncontrolled unless they involve military ornuclear goods or data. or classified infor-mation. Under the International Traffic inArms Regulations (ITAR), military goods anddirectly associated technical data as well asall classified information to all foreign desti-nations are strictly controlled; exports to mostCommunist-controlled nations are forbiddenentirelsi. However; only a relatively smallpercentage of U.S. exports, and an evensmaller percentage of exported scientificinformation, falls within the scope of thesecontrols. Export of most U.S. goods andtechnical data is controlled instead underthe Export Administration Regulations (EAR).Under these regulations, national securitycontrols seriously restrict technology exportSonly to the Soviet Union. Eastern Europe(e:cluding Yugoslavia), the People's Republicof China. and a group of smaller nations(including Laos, North Korea, Vietnam,Kampuchea, and Cuba). Although goodsand technology that are controlled for exportto Communist nations are also controlledto non-Communist destinations, controls onsuch non-Communist destinations are gen-erally designed only to prevent transshipmentor re-export to Communist nations; not toprevent the original export:

The volume of exports to Europe; Japan,and Third World nations makes any require-ment of specific Government approval ofeach transaction wholly impractical. Imposingsuch a requirement would in practice banmost exports by making them prohibitivelyexpensive in time and money. Moreover,

62

limitations on what can be freely shared withWestern Europeans and others in the non-Communist world without Govemment per-mission would require far-reaching changes inthe nature of American society (includingits campuses).

No se .h requirement for specific Gov-ernment approval of all technology exportsexists. The Export Administration Regulationsprovide several mechanisms that effectivelyexempt many transactions from any require-ment of specific Government approval andmany other transactions from the require-ment of case-by-case Govemment approval:All published technical data and most un-published scientific and educational data arecovered by one of the General Licensesand therefore require no specific Govemmentapproval: Technical data that do requirespecific Government approval (in the formof a Validated License) may be eiigible for a"bulk" license (for example; a Project License)which can cover multiple transactions. Gen-eral Licenses effectively exempt most funda-mental scientific researchand thereforemost academic researchfrom the needfor specific Federal authbrization. GeneralLicenses together with bulk licenses providesimilar freedom for most corporate exchangesof scientific research and technologicalapplications in the non-Communist world.

The system nominally controls almosteverything, but in practiCe requires specificlicenses of much less, and actually prohibitsthe export of very little. The critical questionfor the future is whether technology transferwithin the non-Communist world will remainso unrestricted. It is an important questionfor at least two reasons. First, the scientificand technological links among the non-Communist nationsand particularly amongthe industrial nationsare far stronger andfar more central to day-to-day scientific andcommercial activities than technologytransfers with most Communistcontrollednations: For example; while almost half ofall U.S. engineering doctorates in 1980 wereawarded to nonimmigrant aliens, only .3percent of those aliens were citizens of theSoviet Union; Eastern Europe, or the People'sRepublic of China. Industrial employmentexperience is presumably similar. Givencurrent strains within COCOM, U.S. attemptsto control more technology may simply lead

76

to greater U.S. isolation from friends andfoes alike.

Second, while many Americans haveaccepted the need for stringent controlson technology transfer to the Communistworld, the imposition of meaningful controlson transfers among non-Communist na:tionsor West-West tansfers--could under:mine the existing consensus that nationalsecurity controls are legitimate and sensiblepublic policy. Within the United States, suchcontrols on West West transfers would likelybe met with considerable skepticism; withinother industrialized nations, such controlsmight be perceived as yet anoti ier nontariffbarrier, as commercial policy draped innational security bunting.

Substantive Scope of Conti-61S

Under the current rules; unpublished tech-nical information may require Governmentexport approval either if it is directly relatedto an item on the Commodity Control LiFtor the Munitions List or if, under Part 3 /9of the Export Administration Regulations; itrelates directly and significantly to anyindustrial process. All published and otherunpublished technical data are either legallyor practically uncontrolled.

The current scope of controls on technicaldata has been widely criticized, principallyon two counts. First, it is argued, much ofonly marginal national security importanceis controlled and thus neediessIy subjettedto cumbersome and expensive bureaucratiCprocedures: Although this is also a problemfor hardware, such overbreadth is particularlyacute for technical data; because Part 379of the EAR excludes from the most liberalGeneral License much unpubliShed technicaldata relating to indusbial processes regardlessof the national security importance of theprocesses. And despite continuing effortsto rid the control lists of goods or technologythat are widely available abroad; the listsremain long and Complicated.

Second, as all agree, the current scope ofcontrols is simply not well understood. Thisis in past beca...-ie the technical data regu-lationsand particularly Part 379 of theEARare almost incomprehensible to theaverage reader. More importantly, both theperception of overcontrol and the fad Of

incomprehensibility stem from a lack ofconsensus about what should be controlled.The congressionally mandated MilitarilyCritical Technologies List was supposed todetermine the scope of controls: Followingthe recommendation of the 1976 BucyReport8 that controls should focus on tech-nology and manufacturing know-how ratherthan on hardware; the 1979 Export Admin-istration Act instructed the Defense Depart-mei !7 to develop a list of critical technologies"which, if exported, would permit a significantadvance in a military system" of countriesSubject to national Security controls.Sub-equelit effort to construct such a list reflect a

continuing and unreSolyed tension betweenadvocates- of a very Short liSt of patentlycritical technologies and advocates of a muchlonger list including most modern tech-nologies that undergird any advanced in-dustrialized economy. The obstacles toconsensus include not only differing conceptof national security, but also the nature ofmany advanced dual-use (i.e., civilian andmilitary) technologies and the difficulties ofreducing general conceptual agreement toregulatory language.

Computers are an example of a dual-usetechnology. In the United States; most majorindustriesand probably all militarily sig-nificant industriesuse computer technologyin all aspects of the life cycle of a product:definition of product requirements; devel-opment and design, production and opera-tional Support, and utilization. Computer-Aided InduSteal Process Control (CAIPC)technoldgy, even When developed for purelycommercial uses, provides a strong mobili-zation base by permitting the rapid conversionof industrial capacity from civilian to militaryuses. The Same is true of Computer-AidedManufacture and Test (CAM,'CAT) tech-niques. Both CAIPC and CAM/CAT_illustratethe difficulty of drawing the line betweencontrolled and uncontrolled technology. First,both techniques are strategically important,but both also have broad commercial appli-cation; much, if not most, of the researchand development related to these technol-ogies is being done by the private sector forits own use. Particularly because Americanmanufacturing leadership may depend onsophisticated factory automation; efforts tocontrol the dissemination either of the

7763

technology or of the products embodying itwould have immediate and irtiportarit tradeconsequences.

Beyond that, the Government's own useof these technologies depends in goodmeasure on their development by the com-mercial sector. And that development de-pends importantly on university-basedfundamental research in a wide range ofscientific fields (including dynamics; stressanalysis, computer architecture; computa-tional techniques; and microstructures); wherethe lag between basic research and com-mercial application is likely to be very short,Restrictions that discourage academia fromworking with industry in these area willtherefore have important national securityas well as trade consequences,

Translating into regulatory language thelimited consensus that does exist about_whatshould be controlled has also been a problem.For example; although there is wide agree:merit that most_fundamental research shouldnot be controlled, defining "fundamentalresearch" is difficult. The Export Adminis-tration Regul ,Ions .speak of scientific andeducational inkomation not "related directlyand significa_ntly to design. production, orutilization in industrial processes." ExecutiveOrder 12356_bri national security informationspeaks merely of "basic scientific research."This implied distinction between basic and

research is not helpful. becausewhether a given research result is basic orapplied depends both on the purpose ofthe research and on the judgment of theobserver: If publishing generic definitions ofwhat is controlled runs into insurmountabledefinitional problems; publishing specificguidance runs head on into the "blueprintproblem:" If Government defines specificallythe line between fundamental research thatneed not be controlled and other_ researchthat may require .controls,_ that definitionprovides a great deal of inkrmation aboutAmerican technological capabilities and theGovem.ment's strategic concerns. Publishing itmight therefore give our adversaries a "blueprint" of those technologies of greatestimportance_to the United States and allowthem to reallocate their own R&D resourcesinto more promising ar3s. Deciding whattechnologies need be controlled in thefuture will take time itanslating tht:;t decisioninto i-goiatibn will be yet more diacult,64.

When does technology transfertake place?

The fourth question asks, What exchangesof scientific information effectively transfertechnology? Those that do not; need notbe controlled. (Of course; not all effectivetransfers should be prohibited even to ourmost serious adversary: Most scientific ex-changes work in two directions: Therefore;once it is determined that a technology lossis likely; it is also necessary to judge whetherthere will be an offsetting technology gain.If there will be; the exchange should usuallybe permitted;) Ordinary, though perhapsdifficult; observation can establish whethera given piece of hardware has been trans-ferred to an adversary. It is much harder totell when technology has been transferred.Whether technology transfer takes place ina given situation depends on the nature ofthe information, the skill and training of thegiver and the receiver, and the nature andduration of their interaction.

For some technical information, simplepossession is enough. Steal the recipe andyou should be able to produce a reasonableimitation of Coca-Cola. Although there areimportant exceptions, the theft of blueprintsfor hardware to be produced abroad isgenerally not the pnncipal national securityConcern today. That kind of technologicalpiracy will usually assure preservation ofAmerican leadtime; precisely the objectiveof controls: Similarly, while the simple knowl-edge that something can be done mayoccasionally he the problem, more usuallythe concern today is whether an adversarywill be able to apply sophisticated scientificand principles and information toits own needs and then build further onthem. Occasionally, a casual exchange willtransfer a critical concept or important piereof information. More frequently it will not.As the Bucy Report concluded ii 1976 andas American foreign aid programs learnedthrough hard e>perience, effective technologytransfer does not occur casually or quickly.Rather, it requires that the 'giver and thereceiver actively interact with each other overa sustained penod of time.

Despite wide agreement on this point;current export control regulations on technicaldata generally do not distinguish "exports"that will transfer technology effectively from

78

those that will not. An hour of formal pres-entation of unpublished research findingsto a small group that includes foreign sci-entists or a quick walk-through of a laboratorycontaining advanced computer equipmentare as likely to fall within the scope of controkas an inte: training program. As a result,specific approval is required for many "ex-ports" to Communist countries that mostagree to be harmless. And once within thescope of controls; there are no clear andpublicly known criteria to guide the Govern-ment's decision when to approve or disap-prove a license.

Designing the Rules of the Game

Much of the scientific community's concernarises from its conviction that the Governmentcannot properly assess the scientific andeducational costs of various kinds of restric-

- dons because of inadequate input from theScientific community itself. While it is generallyundeSirable for large segments of theAmerican populace to feel excluded fromthe Government'S decisionmaking processon issues of importance, it is particularlyunfortunate when the Government must relyin great measure on voluntary complianceand cooperation rather than on legal com-pulsion. It is therefore important for theGovernment to address the issue of howthe scientific community can most produc-tiely participate in establishing the systethunder which all must live.

In short; there is still no natie---,a1 consensusabout how broadly to define national security;what destinations require control: how manytechnologies should be controlled; what kindsof research; if any; warrant controls and howthe line between controlled and uncontrolledis to be drawn; and what kinds of exposureto American technology are sufficientlydetrimental that restrictions are likely to becost-effective. Neither is there a clear mech-anism fur attempting to build that consensus.

Assessment of ControlMechanisms

These questions would be difficult to answereven if export controls were the only way torestrict scientific information. They are not.At least four other control took have beenused in the past and could be used more in

79

the future: national security classification;stricter use of the Government's visa authorities to deny or set conditions on the admis-sion of foreigners; contractual restrictionsfor research performed with Federal funding;and various forms of voluntary self-control:

Without agreement on what a mechanismshould control, it is hard to decide whatcontrol mechanism to use. For example;mechaniSmS designed to deny the Sovietsa mere handful of technologies may begrossly inappropriate if the policy is to denya larger number of nations a broad range oftechnologies. On the other hand, thesemechanisms have different characteristicsand impose different costs on Americansociety. It is difficult to know what to controlwithout knowing the costs of controlling it,and that depends on the mechanism used.It may be useful, therefore, to analyze themerits and demerits of the mechanisms.

Many criteria are relevant. Four cluSterS,however; appear to be most Important:

(1) Effectirmness. Will the system ac-curately identify what needs to be controlled.and will there be sufficient American andCOCOM cooperation that it can be enforced?Alternatively; if COCOM cooperation is notlikely, will unilateral controls merely isolatethe United States?

(2) Administrative and Economic Burden.Can the system be administered withoutimposing unacceptable administrative andeconomic costs either on the U.S. Govern-ment or on the American population? Canthe prbcedures be simplified and the un-certainty reduced?

(3) Scientific/Technological Costs. Cana system be designed that avoids substan!tial slowing of scientific or technologicalprogresswhether by compartmentalizingresearch into controlled and uncontrolledareas, by further hindering industry/universitycooperation, or by damaging the intellectualclimate?

(4) Political Values and Consensus. Cana system be designed and enforced that isgenerally accepted by the American popu-lation as a reasonable and legitimate responseto a shared threat? Will industry and aca-demia perceive the Governmentparticularlythe nation- security r vioe:,vzies and thosecharged with enforcernul h adversaryor as a partner in a Cdas.Jorative effort?Muting the current adversarial climate will

65

require not only public understanding ofthe nature of the threat, but alSo publicacceptance that the means adopted arelawful. predictable in their application andenforcement, and appropriate to the mag-nitude of the problem.

Clearly. throe clutters are related. A systembuilt on domestic political consensus is morelikely to be effective an stemming technologyoutflow. A consensus is easier to build if thesystem is seen as effective, as administrativelyefficient, and as imposing the minimumpossible cost on other soda! values. However,while all four clusters are important and needto be considered. this discussion emphasizesthe effects on science and technology.

Export Controls

Export controls that reach scientific infor-mation unrelated to a commercial transactionor only remotely related to a controlledcommodity fare badly when judged in termsof either effectiveness or political acceptability.They are too broadly and imprecisely definedto give Americans a dear understanding ofproscribed conduct; in addition, a numberof COCOM members are unwilling to controlsuch "disembodied" technology. Further,because of the complexity and breadth ofthe rules. administration is cumbersome,compliance costly in dollars and time, andenforcement difficult.

While ineffectiveness alone would beenough to undermine a political consensusthat controls are valid and leg:tirnate, thepzobk.ni is egacerbated by the increasinglywidespread belief that controls on such"disembodied" technology are unjustifiedinfringements of constitutionally protectedFirst Amendment rights. Both the JusticeDepartment and a US. Court of Appealshave raised serious questions on that issue.The Justice Department addressed theseissues twice. first in a 1978 Memorandumto the Science Advisor9 and again in a 1981Memorandum to the Department of State.ft)On both occasions, Justice stressed that priorrestraints such as those represented by theITAR and EAR licensing systems will beupheld only upon a governmental showingof grave danger to the Nation and concludedthat the regulatory scheme under review"cast such a broad regulatory net" through

66

a system of prior restraint that it was "pre-sumptively unconstitutional." The NinthCircuit Court in United States u. Ed /erIndustries. Inc.," a cnminal case, also stronglysuggested that restnctions on technical datawould be constitutionally acceptable onlyas long as the data were significantly anddirectly related to specific articles on a controllist.

On the other hand, because so muchinformation is practically uncontrolled formost destinations and bemuse enforcementhas been limited, export controls to datehave had little direct impact on progress inscience and technology. Tighter restrictionswould presumably affect science and tech-nology much as classification has done (seebelow). The more these controls are usedthe less benign their effect.

Classification

Since World War II, the Government hasused its classification power to control certainkinds of scientific information. The advan-tages of classifying information are clear.Although claS.sification requires defining whatneedS to be controlled on a case-by-casebasis. the question of to whom is clear. Onlythose with a security clearance and a "needto know may receive classified information.Denying information to almost everyone isan effective means of denying it to one'sadversaries.

Perhaps because the limitations are sosevere, relatively little scientific informationhas been classified. Po a result; althoughthere are serious administrative costs toworking on classified research fas well assome disputes about what is classified),those costs are imposed only on a smallsegment of the scientific community ratherthan on the population at large. The Gov-ernment's relative self-control in classifyingscientific information has also created anaura of legitimacy, which itself encouragescompliance.

Classification's effectiveness in denyinginformation to adversaries is bought at theprice of denying it to nonadversaries--scientific colleagues, for example. Its effectson scientific and technological progress aretherefore quite severe.

ClaSSification dividesor compart-mentalizes the scientific community into

those with clearance and a need to know(which frequently includes nationals of theNorth Atlantic Treaty Organization and otherfriendly countries) and those without; Asthe United States learned from the experi-ence of atomic energy, intellectual exchangebetween the two carrrp:s is limited. The normalprocesses of collegial criticism, of learningFrom the successes and failures of one's peers,and of using research to train the nextgeneration of scientists are all impeded.Restrictions on publication may make work-ing in a classified field less attractive, so itmay be difficult to recruit and retain peoplein the areas of greatest national security need.

The divisive effects of more extensiveclassification are likely to be even greaternow than during the early Cold War period,due to the lingering effects of Vietnam. Manymajor American research universities prohibitclassified research on campus. Off-campusclassified research is detached from main-stream university activities; student partici-pation is minimal. These policies are notlikely to change soon. Therefore, thoseattracted to academic life will likely avoidwork in classified areas.

With classified research thus largely con-fined to industry and Government labora-tories, close collaboration between scientists inindustry (who will & classified research) andthose in academia (who will not) may beimpeded. if so, industry will be decreasinglyable to rely on the academy to generatewail-trained talent' in relevant areas; toperform nonproprietary research of interestto industry, and to provide greater accessfor its researchers to current internationalscientific thinking. A parallel result could befurther separation of academic science fromnational security issues. From mission-relatedresearch, and from research of direct rele-vance to emerging industrial needs, becausemost of the technologies of national securityconcern have commercial applications.

The simplicity, clarity, and precision ofclassification are its great strengths. Atomicenergy information may be "born classified,"but in other areas the Government mustaffirmatively and unambiguously decide toclassify information: Even in atomic energy,disputes over what is classified have beenrelatively manageable; These strengths alsoengender problems. Unless the Government

classifies an entire broad field, like atomicenergy, classification decisions are unending.Further, assuming the Government wantsto classify information before it is generallydisseminated, it must somehow shift someof the burden of identifying potentiallyclassifiable information to the scientists whogenerate it. That may not be too difficult forresearch performed in Government labora-tories: Scientists there are likely to be awareof possible national security ramifications oftheir research and; like some industrialresearchers; are more willing to acceptpublication and other restrictions as condi-tions of employment.

However; as the controversy generatedby Executive Order 12356 on nationalsecurity information attests, identifying poten-tially classifiable information will be muchharder for private sector research. First; andprobably most serious, there is the problemof defining what may require classification.The term information "relating to the nationalsecurity" has no self-evident meaning, par-ticularly for people not in daily contact withnational security issues; all the uncertaintiesof the current export control laws wouldre-emerge here. Second, people outsidethe Government and not working on De-partment of Defense contracts do not expecttheir work to be classified and do not con-sider classification an occupational hazard.Prepublication Government review of asyet unclassified research, especially whennot federally funded, suggests censorshipto many:

Visa Controls

The Federal Govemment has broad powersto bar aliens from entering the United Statesor to set conditions on their stay. While theGovernment may legally bar an alien to avoidan undesirable technology loss, visas arerarely denied on that ground. Intelligent useof the visa authority requires generalizedanswers to the questions of what technologyneeds to be controlled, to whom, and inwhat form. Without them, case-by-case reviewof all requests for nonimmigrant visas wouldbring the entire process to a halt. Intelligentuse also requires more information than isnow typically available for nationals of somecountries. While costly to collect, that same

81 67

information is required to selectively imple-ment any form of control.

Denying a visa usually keeps an alien outof the United States. But it prevents tech-nology transfer only if the visit was essentialto an effective transfer. For someone whointended to come as a degree candidate orpostdegree researcher, the denial probablydoes prevent transfer. For someone whowould have come for a very short stay, for asymposium, or to hear a paper deliveredthat will soon be published in the professionalliterature, the denial probably does not bartransfer (particularly if Americans may freelyconvey the same information at conferencesand .symposia abroad and there are norestrictions on domestic publication).

The private sector generally finds visadenials or restrictions attractive; particularlywhen contrasted with a governmental sug-gestion that the export control laws makeuniversities in some way responsible forensuring that legally admitted aliens aredenied access to controlled but unclassifiedinformation. Such an interpretation of theexport control laws would require universitiesto assume a role for which they are tempera-mentally and physically ill-equipped, namely,that of monitoring the activities of theirstudents. researchers, or visiting faculty andrestricting the access of some merely on thebasis of nationality. In contrast. governmentaluse of visa denials requires the Governmentto shoulder most of the administrative burdenand avoids conflict with strongly held val. !es.

Visas as a control mechanism have certainattractions for the Govemment as well. Whilebroad criteria can (and should) be publiclystated, decisions on individual applicationscan be made by the State and JusticeDepartments behind closed doors in con-sultation with the national security agencies.The Government therefore does not haveto fully explain why a particular visa wasdenied and, when it does explain its reasons;can do so outside the glare of publicity.Definitional and "blueprint" problems,linked to the publication of criteria stiff'ciently detailed to guide public behe

to have major consequences for science andtechnology (although there may be foreignpolicy concerns).

Special problems, however, are likely tooccur when the foreigner is requesting avisa to attend an international scientificconference sponsored by the InternationalCouncil of Scientific Unions (ICSU). BecauseICSU policy prohibits adhering membersfrom holding conferences in countries thatdeny entrance to bona fide scientists; a U.S:visa denial may only serve to force suchconferences abroad and isolate Americansfrom their scientific peers. Such a possibilityshould clearly be considered when theGovernment weighs whether; on balance:the attendance of the scientist or scientistsin question will result in a net technologyloss to the Nation.

Granting or denying aliens admission tothis country is clearly an appropriate gov-ernmental function. If visa controls areexercised with restraint, they are unlikely tobecome a major source of contention. How.ever. closed-door decisions influenced bythe rational security agencies inay have anunhealthy bias toward overcont la tly

if information about the benei . of the visitis not readily available within the Govemment.it any event. visa authority cannot be theprimary means of control: it is too easy forour technology to be transferred outside ourborders by other means. Consequently. visadenials are perhaps best viewed as a way ofreducing the objectionable domestic effectsof export controls.

Contractual Restrictions on FederallyFunded Research.

In early 1982. the Report of the DefenseScience Board Task Force on UniversityResponsiveness to National Security Re-quirements suggested greater use of con-tractually imposed restrictions to avoid someof the uncertainty and contentiousness ofother means of control.* For example, aresearch agreement might limit or requireapproval of foreign participation or require

In general; relatively small numbersdenials or restrictions, particularly if hi A contract. for this ptupose, includes any contractualto nationals of countries that are proscribed instrument whether labeled a contract; grant; or coopunder the export control laws, appear unlikely erative agreement

68

prepublication review of research results.Although recommending this now only forthe Defense Department, the report sug-geSted that such restrictions be used by otherFederal agencies if the need is clearly estab-lished. Presumably. contractual restrictionsapplicable to industry/university collaborativeresearch would bind both industry anduniversity.

With the Federal Government fundingalmost half of the R&D performed in theUnited States,12 contractual restrictions area powerful tool: Contract clauses are likelyto effectively restrict dissemination of theparticular information developed undercontract. Researchers usually heed contractterms. at least when it is made clear that thefunding agency will enforce them: However.if similar or related research is either notfederally funded or funded by an agency thatdoes not impose these restrictions: similarinformation will be disse_rninated unless it isotherwise controlled. The Government'sability to rely on contractual restrictions thusdepends in great measure on whether similarresearch is likely tote otherwise controlled.The answer appears to be that it is. Industry-funded research (including that on campus)promising clear, short-term commercialapplications is usually subject to proprietaryrest& lons. If the firm seeks a patent, theGovernment can impose a patent secrecyorder: Most other research, while not -con-trolled; is unlikely to be sensitive on nationalsecurity grounds;

Contractual restrictions impose only limitedadministrative burdens. Instead of controllingentire fields, as the Atomic Energy Act does.or entire technologies_ as export controlsmay with the Militarily Critical TechnologiesList, contractual restrictions can be tailored'o the particular research project. Areas ofconcern can be identified quietly and ex-plained to the researcher without risking apublic "blueprint problem." Because con7trartual restrictions are likely to be draftedand monitoreel by research sponsors, thereis reason to hope the restnctions would bereasonablealthough guidelines will still beneeded to rha;;nel the discretion of ins_iividualpiogram managers wl ' twise beoverzealous;

The effect these restrictions would haveon science and technology depenc:s on the

way they are used, If they are appliedfrequently or with a heavy hand, their effectscan approach those of classification. Indi-vidual scientists will have to decide whetherto work in tightly controlled areas: Individualinstitutions will have to decide whether theconditions are compatible with their philos-ophy and objectives. The more infrequentlysuch conditions are imposed, the easier it islikely to be for an institution to accept them.If the universities but not industry decline towork on such terms. industry/universitycollaboration once again becomes moredifficult.

The affected public is more likely to acceptbroader contractual restrictions than mostother forms of restrictions, particularly if therestrictions are accompanied by better com-munication and their permissible scope isclearly laid out in agency policy. As withuniversity decisions to accept delays inpublication to preserve patentability inindustry-funded research, there is an auraof voluntarism. Because the conditions canbe tailor-made: they are more likely to appearappropriate to the situation: Because theno.gotiations can be conducted privately. theGovernment is better positioned to explainwhy the information is sensitive. Becausethe contract must be signed before research isbegun. restrictions are less likely to seemarbitrary and unpredictable than those im-posed in midstream.

Voluntary Restraint

The term "voluntary restraint" is not, con-sistently defined and two uses are mostcommon. Sometimes, it refers to arrange-ments, such as those in cryptography, bywhich researchers volur submit theirwork to the Government for prepublicationreview. These arrangements as yet have nolegal basis; the National Security Agency(NSA) concluded that it lacked any groundfor legal compulsion: At other times, the termrefers to the motivation for compliance withlegally imposed restrictions for reasons otherthan fear of penalties. This usage suggeststhat if the Government did a better job ofexplaining to American scientists what itwas worried about and why, it could applyvery tight controls (for example, classification)to only the most critical technologies; alert re-

69

searchers to other possible problem areas:and rely on their general patriotism to restrictdissemination or seek consultation whenquestions or problems arise:

Although the Government should do abetter job of communicating legitimate na-tional security concerns to the Americanpublic (including the scientific community).it would be difficult to rely solely on the firstform of voluntary restraint for several reasons.First, there are the now familiar difficultiesof the "blueprint problem" and administrativeburdens. Willing to submit to prepublicationreview. public-spirited cryptographers askedNSA to define its areas of direct concern:NSA finally concluded that publication ofsuch a list would be more damag: .o thenational security than publication of theresearch results. The Defense Department'sdifficulty in publishing an unclassified versionof the Militarily Critical Technologies Listsuggests that the problem is a general one:if the Government cannot alert researchersto its specific concerns: it could alert themto broad concerns and then scan a hugequantity of voluntarily submitted material.This may be possible in a small, highlysensitive discipline like cryptography. Butwhere the number of researchers is largeand the security implications less compelling(or at least less obvious), this broad netapproach would be slow. costly, and unlikelyto lead to an adequately high level ofvoluntary compliance.

While perhaps also inadequate as the soteGovernment strategy. the second form ofvoluntary restraint has considerable meritas an adjunct strategy. Most Americans donot want to aid America's military adversaries.If they believe the Government :vhen it saysthat the release of certain '.!(,-..:.rnation willbe damaging to the Nation, mostdespitethe importance of publication in the researchenvironment and reward systemwill try tofind a way to accommodate the Govern-ment's concerns; This kind of mutual trustdoes seem to be emerging in the cryptog-raphy fielel. The Government would be well-advised to trt.;., to create it elsewhere.

Better communication. mutual trust, andvoluntary restraint r_ .ay be the only practicalor effective way of dealing with emergingtechnologies. that are not Governmentfunded. Clearly the broad approach of export

70

controls is not working: the classification routeis also limited to information the Governmentowns or controls: and contractual limitationstoo presuppose a Federal financial role.

Policy Considerations

Two general conclusions dominate. First.effective ,nstruments for controlling thedissemination of scientific information wouldimpose great costs on progress in Americanscience and technology (and thus on long-term national scurity). And even ineffectivetools. such as export controls, may discouragescientists from working in areas subject torestriction, impose financial and administrativecosts on science, affect the research en-vironment. and drive a wedge between theGovernment and important segments of thepopulation. Indeed, mechanisms that areineffective because of their broad but uncer-tain sweep may impose higher costs thaneffective mechanisms: Second, clear andnarrow restrictions that put the definitionalburden squarely on the Government'sshoulders are more likely to be accepted aslegitimate and appropriate Governmentpolicy than are broad restrictions that attemptto shift the task of identification to thescientific community.

These general conclusions have specificpolicy implications for each of the severalcontrol mechanisms discussed above. First,it would be both very difficult and sociallydisruptive to apply export controls morebroadly to scientific research. Second.although classification is obviously an effectivecontrol mechanism, its costs to science andtechnology suggest that it cannot be usedsubstantially more frequently without endangering the scientific endeavor thatunderpins our economic and military health:Third, selective use of visa denials whenpotential net technology loss is clearlythreatened would appear to impose littlescientific cost and v.,ould pry ?bably rneet withrelativeil, little scientific: hostility. And fourth,contractual restrictions, although not withouttheir dangers, are a reasonable approach.Because specificailti; net lotiated L.oniract termsare more likely to be appropriate thangenerally applicable export control regula-tions. the regulations snout i be revised to

84

make it clear that compliance v. 2derallyimposed contractual restrictions at leastwhen imposed or reviewed by a nationalsecurity agencyrelieves the research per-former of further obligation under the broaderexport control laws and regulations.

Nit and References

1. The National Academy of Sciences recently pro-vided a thorough and thoughtful analysis of manyof the issues involved in its report. "ScientificCommunication and National Security.** Septem-ber 1982_

2. National Science Foundation. Science Indicators1976. Washington; DC, 1977: pp. 13-16.

3. National Science Foundation. "Foreign Participa-tion in U.S. Science and Engineering Higher Edu-cation and Labor Markets." 1981. and unpublished1980 data.

4. Statement of Mm. B. R. Inman to the U.S. Senate,Committee on Governmental_ Affairs,_ Subcom-mittee on Investigations. May 11, 1982.

5. (CIA) Soviet Acquisition of Western Technology,April 1982; p. 12: NAS Report, supra. pp. 23-25.

6. Statement of Adm. B. R. Inman at the AnnualMeeting of the Arnerican Association for the Ad-vancement of Science, January 7, 1982.

7. Gustafson. T. "Selling the Russians the Rope?Soviet Technology Policy and U.S. Export Con.trots." Rand Report Nt:5 R-2649-ARPA, Apnl1981. p. 9.

8. Department !_a_Defense. "An Analysis of ExportControl of US riogyA DOD Perspective."February 4, 1..1k

9, Dopartment Memorandum to the Sci-ence Advisor. "Constitutionality Under the FirstArnendment of ITAR Restrictions on Public Cryptog-raphy." May 11. 1978. Reprinted in-The Govern-ment's Classification of Private Ideas: Hearingsbelow a sai 'committee of the House Committeeon trove: lent Operations." 96th Congress. 2dSession. 1980. pp. 268.84.

10. Department of Justice. Memorandum for William BRobinson, Office of Munitions Control. Departmesltof State, "Constitutionalry _of the Proposed Revi-sion of the Technical Data Provisions of the Inter-national Traffic in Arms Regulations," July 1: 1981;

11. 579 F.2d 516 (1979).12. National Scienc.e Foundation. National Patterns

of Science and Technology Resources 1982.Washington, DC: March 1982, p. 2.

8571

Issues in Scientific and TechnicalInformation Policy

Abstract

The generation, collection, organization, processing, and distribution of scientific and technicalinformation is a significant but neglected area of R&D policy. Rapid technological changesas well as the growing economic and social importance of information in society are makingthe formulation ofpolicy in this area both necessary and difficult.

Issues related to:control of and access to scientific and technical informationincludingprotection of personal privacy and proprietary information, restriction of information flowfor national security reasons, and the emergence of international barriers to .nformationfloware high on the agenda. Heightened awareness of national security concen:s hasstimulated intense policy debate over the merits of controlling the flow of potentiallysensitive scientific and technical information relative to the value of open communicationamong scientists: Other debates have arisen over the proper U.S. response to economicallymotivated moves by other countries to limit transborder data flows and otherwise impedethe international flow of scientific and technical information:

Another set of issues centers on the economics of scientific and technical information-Id conflicts between the roles of the public and private sectors in providing informationservices. Pricing policies of Government information centers have stimulated considerabledebate, reflecting policy inconsistencies as well as deep seated philosophical differences:Some approaches to resolving the issues, based on differentiating types of uses and users;are suggested.

The various roles of the Federal Government in scientific and tee-- ical information are---- highlighted, and a number of options for further examination and action are proposed.

Introduction

Scientific and technical information is a majorpreduct of research a. id development (R&D).The results of R&E' projects sponsored bythe Federal Government---sorne .140 billionin fiscal year 1983are of value only itcan be communicated and put to use. Theymust be communicated to other researcherswhose work is based on the existing knowl-edge base and who must have rapid ci_cceisto up-to-date, accu:?re scientific and technicalinformation to avo.,:; unnecessary duplicationof effort, to speed up the p:ocess or limo:vation, arid to make best use ofresources. They must also be communicatedto users who can put the findings to work inpractical applicationsnew products and

for the cor .nercial marketplace,and new programs and policies in the publicsector: Yet; for all thy; attention devoted toR&D budgeting; policymaking, and man-agement; ielar,/ely little attention has been

paid to information policy in the context ofFederal R&D.

Scientific and technical information policyshould not be considered in isolation. First;it is clearly an integral part of R&D policyand should be closely linked to decisionson R&D priorities and management. Theproduction, disseminatiGn, and use of sci-entific and technical information are criticale'er:lents in the efficient and effective useof Federal R&D funds. Federal R&D policymust thus be concerned not just with theconduct of the research itself, but with thedissemination and utilization - -in both thepublic and the private sectorsof the fruitsof that research. Second, although discus-sions of it have sometimes failed to aLknowl-edge the fact, scientific and technical infor-mation iv, :icy is also one aspect of thebroa ler domain of information policy ?- .dman :gem' it. The latter is a domain ofglowing scope and importance which en-compasses such diverse concerns as the

86 73

operations of computerized data bases,copyright and patent policy. telecommuni-cations regulation, privacy protection, andtransborder data flows. The implications forscientific and technical information of deci-sions made in these realms must be recog-nized and taken into account in the formu-lation of R&D policy.

Scientific and technical information is asignificant area of economic activity in theLiPited States: King (1981) estimates thatnearly $15 billion is spent each year onauthoring, publishing; storing; distributing;and using scientific and technical information:Of this amount: the Federal Government isreported to support about 45 percent throughdirect publication of books, journals andtechnical reports: funding of over 200 sci-entific and technical information clearing:houses, three national libraries and 3,000other libraries: operation of several nationalstatistical renters and other informationactivities: and support of scientific. com-munication through payment of page chargesand attendance at professional meetings forresearchers under grants and contracts.According to King. the Federal expendituresinclude $1.2 billion a year for library. ab-stracting. and indexing activities, $1 billionfor numeric databases.

Making informa'ion policy is particularlydifficult at present. in ;.art. this is a result ofthe fact that we are becoming increasinglyan information-based society and that tech-nological advances are leading to the mergingof information and communications tech-nologies: Consequently; society is now facedwith a "variety of new technological capabilities,some with important social consequences,for which it is relatively unprepared. Anotherfactor's that most segments of the informa-tion industry (e.g., data processing) havebeen essentially unregulated. while com-mur'cations have been subject to hr.: iyCove nment regulation. The problemshow to establish appropriate leg?' end regulatory frameworks for the new, mergedte,;:inologies are just beginning to be sorter,oat. In,:reasingly, also, information is beingrecognized as a resourcea resource of sig-nificant econoni..:-. value, particularly in aneconomy shifting from manufacturing toservices. The economic importance of the

74

information sector is only beginning to berecognized.

The remainder of this paper reviews brieflya select set of policy issues relating to scientificand technical information. It covets the US:policy context; economic issues and thepubk/private sector debate; access to sci-entific and technical information; and therole of the Federal Government: The issuesdiscussed here are complex and overlap inmany ways: The structure within which theyare treated here is but one of many possibleapproaches to their analysis and discussion.

Policy Context for Scientificand Technical Information

Federal responsibility for scientific and tech-nical information has evolved gradually, oftenassociated with legislation that had neitherinformation nor science and technology asa principal focus.' Each of the agenciesthat produces scientific and technical infor-mation as part of its operationsincludingthe Department of Defense (DOD), the De-partment ,oi Energy (DU3), the Environ-mental Pr-utection Agency (EPA); the Na-tional Instiaiies of Flea..h (NIH); the NationalLibrary of Medicine (NLM); and the NationalAeronautics and Space Administration(N-SA)traditionally has had its ownmethods eid policies for managing distri-bution and access to that information. Theevolution of the National TechniciA Infor-mation Service (NTIS) in the Departmentof Commerce in the 1960s (formerly the

-e of Technical Services, then the Clear-inghouse for Scientific and Technical Infor-mation), wa a major milestone in thedevelopment of Federal scientific an,: tech-nical info:mation activities. NTIS collects,organizes, and disseminates echn:cal reportSresulti, ig from Federal R &L contracts and_grants. The National Science Foundation(NSF) also has developed many scientificand technical information activities associatedwith its responsibilities in funding basic

For a detailed chronology of events and reports relatedto scientfic_and technical ink-- ..ion. see CongressionllResearch. Service. 19/8.

research. rhey have included supportingresearch in information science and policyas well as promoting the dissemination ofresearch results by providing seed moneyfor the development of modern, computerized information systems and databasesoutside of Government, such as ChemicalAbstracts Service (1965 75).

The National Science end TechnologyPolicy, Organization, and Priorities Act (P.L.94:282), under which the Office of Scienceand Technology Policy (OSTP) was estab-lished, includes provisions establishing severaladvisory groups to provide a more coherentmeans of managing and disseminating sci-entific and technical information. Further-more, the Act establishes the basis for Federalscientific and technical informatiorties as a part of R&D policy. It spci..:icallystates that

. . it is recognized as a responsibility of theFederal Government not only to coor-dinate and unify its own science andtechnology information systems, b:.,1facilitate the close coupling of instittscientific research with commercial appli-cation of the useful findings of science.(90 Stat. 461)

Despite the explicit recognition, here andelsewhere, of Federal responsibility for thecoordination of scientific and technicalinformation policy, mo.t Federal agencieshave systems that reflect and support theirown needs and missions and are not wellintegrated into the lager context of scientificand technical information, R&D policy, orgeneral information policies.

It is becoming apparent that existingframeworlzs do not WO accomm,date thedeveloping and expamling information andcommunications techno;ogies. especially withrespect to the problems pt re iulation andpciblic/private sector interacti ms. Scientificand tecIinical information relic& s only as:.bset of the larger universe if informationpolicy: and policy choices made in relationto emerging electronic or banking .5,1.ems,for :ixample; may in turn affect policy delibera-tions more directly related to scientific andtechnic,. information. The convergence ofcomputer and communications technologies;developments in storage, processing. and

88

distribution techniques: and the growth ofinformation services in the economy haveimplications for a wide range of enterprises.

Scientific and technical information policyis often viewed and developed within thecontext of specific programs or projects: Thisis in part a reflection of the diversity ofinterests involved and that no real organiza-tional focus for policy formulation exits. Inaddition, decisions regarding scientific andtechnical information generally seem to bemade by mid-level program managers whoadminister these information systems, acidseldom involve senior level policymakers whohave hrivider concerns and perspectives.

As tht Nation moves increasingly into aninformation era, as society becomes moretechnolociically oriented, and as changesoccur in the direction of R&D efforts, theneed to link scientific and technical infor-mation policies to R&D policies, as well asto broader information and communicationspolicies; becomes more critical. For example,as civilian R&D activities in the FederalGovernment focus increasingly on basicresearch rather than applied research ordevelopment, and as defense-or,Inted R&Dreceives Increasing funding; demands onscientific and technical information systemsare bound to shift and Federal programswill need to adapt. At the same time; as theGovernment reduces its information coilec-*tion and dissemination effortseither as aresult of budget cuts or such initiatives andlaws as the Paperwork Reduction Act--_--bothusers and producers of scier-tfic and technicalinformation will need to make some difficultchoices. The ability of di ase involved :nentific and technical into:Irv-am to Ileightenthe awareness of other polic'imakers to theconsequences of their decit:i,ins will be akey determinant in maintains g the vitalityof scientific and technical inf o-nation vograms as broader policy chatis occur.

The United States differ: isignificantlp fromother nations in its approach to policiesaffecting information; comir!-nications, andR&D. In the United States, the private sectoris the basic provider of 6 to processing andtelecoinmunicatIons products and services,the supporter o' the major share of theNation's R&D effort, and the performer ofan even larger share. The significance of

75

this arrangement has been highlighted inrecent years as policymakers have soughtto enhance marketplace competition andincrease the "privatization" of Governmentactivities: Even where the Government re-mains a major player: however, the approachhas been decentralccl and often fragmented.This is due in part to the traditional U.S.antipathy to centralized planning schemesand the belief that diversity of ideas andopen competition ultimately produce thebest results.

The U.S. approach contrasts sharply withthe approaches of many other nations ofthe world where governments develop na-tional strategies, target specific industries forsupport. and control critical elements of theeconomy. For example, most telecommuni-cations facilities in Europe and Japan areeither government-owned or government-operated. Likewise, the French governmenthas recently announced that it will be sub-stantially increasing its R&D spending :opromote high technology industries, whilethe key position of the Japanese in the

m i con d u ct or industry is attributed to therolo 0: that nation's z in fosteringincrust 'al &ye' -

rhe fact r. a icse challenges are emergingfrom abroad does not mean; as some havesuggested; that the United States shouldforsake its ways of doing things and .z.: :oemulate such foreign practices as il-re Jap-anese sty, of management. The strengthof the United States lies in great measure inthe diversity, independence, and competitionof ideas inherent in our way of governingand in our economic structure. The areasof R&D, communications, and informationtechnology are _particularly noteworthy ex-amples of this. The importance of the cur-rent situation is to alert the United State:,to the need to acknowledge the diffenngapproaches of our major trading DE-r.:inersand our ideological oppor:z2rIts .1broa.: sothat effective policies can be' formulated torespond to thorn.

Econom -; Issues and thePublitiPrivete Sector Deb

1.4ith the increa4ir..3ii..nreance the Infor-mation sector to the U.S. ai., econo-mies, and the growing recognition by those

76

involved in the oeneration, collection, trans-mission, dissemination, and use of informa-tion of its actual and potential value, comea host of theoretical and practical problems. Anumber of the underlie current policydebates in the area of scientific and technicalinformation:

Pricing Policies

The issue of pricing policy and the attendantissue of unfair competition are raised fre-quently and applied to several aspects ofGovernment involvement in scientific andtechnical inforthation. Govemment scientificand technical information centers vary sub-stantially in the prices they charge for thc:rservices. Certain users are not charged atall, while charges to others differ accord-ing to each center's policies and the typeof user.

General policy guidz.rice on ser chargesfor Government information se,vices is con-tained in Office of Management and Budget(OMB) Circular A-25, "User Charges," whichrequires that user charges be imposul when aGovernment service

...enables the beneficiary to obtain moreimmediate or substantial f,....ins or values(which may or may measurable inmoneta.-y terms) tlitm those which accrueto the general public...

This circular establishes the general policyof the Executive Branch that a reason,:blecharge "should be made to each identifiablerecipient for a measurable unit or ainountof Government service or property fromwhich he derives a special oenefit." Thecircular specifies that the charge shouldenable the _agency "to recover the full costto the Federal Government of rer.deringthat service" and must include "the directand indirect costs to the aivemment ofcarrying out ;he activity,' as well as a fairalk-_ution of such :'-ems as salaries, research,and supervisory cos' ')MB Cir-cedar A-25,Paragraphs 3, 5). L a 1974 SupremeCourt decision, however, the reps very ofindirect costs is limited to those co Ets Cziatare actually associated with the specificservices provided, .T;ricl excludes those thatbenefit the public at large or are incurred inrtablishing the whole program (U.S. GeneralAccounting Office; 1979: 2%

89

In a survey of 38 information centers infive departments and agencies, th .2. U.S.General Accounting Office .,G,-.;0) foundthat the centers actually recovered only asmall fraction of the costs of providing theirservices and that, when charges were made,cost recovery policies were not appliedconsistently. The same GAO report con-cluded that private firms that purchaseGovernment databases and market theservice should be assessed "fair and equitablecharges": failing to do so would result in ageneral taxpayer subsidy to non-Govemmentusers of the service (US: General AccountingOffice; 1979: 27-29):

OMB Circular A-76. "Policies for AcquiringCommercial or Industrial Products andServices Needed by the Government," re-vised March 29; 1979, is based on the notionthat the Government's business is not to bein business. The circular limits and definesGovernment competition with the rsector. stressing reliance on the piiva.for the provision of goods and se

In a democratic free enterpri::,;i SySt':Government should not compete wit:, ascitizens.. in recognition of this principle,it has been and continues to be the generalpolicy of the Government to rely on cornpetitive private enterprise to supply theproducts and services it needs. (OMBCircular A-76, Paragraph 2)

Several points are made by critics of currentGovemment cc st recovery practices:

(1) The system is inequitable: By notrecovering the full costs of services fromusers; the Government is setting artificiallylow prices and using general tax revenuesto benefit the few (i.e., special interests) whouse the information services.

.2) Government scientific and technicalinformation services misallocate resources.Since the services are priced without ref-er..:nce to either their 'true" cost or theirvalue to users, the Government has no wayof telling whether it is funding tie servicesat an optimal level.

(3) The presence of Government distortsthe information marketplace. Governmentpractices (e.g., pricing) affect the supply anddemand for information services, make itmore difficult for a free private sector marketto operate, and threaten the viability ofprivate firms.

(4) Government involvement discouragesinnovation by creating uncertainty and re-ducing the rewards for introducing newservices and technologies. The same factorshamper the entry of new firms into the fieldand discourage pnvate sector investment.

Several rejoinders are made to thesecharges:

(1) Production and distribution of scientificand technical information are part of thestatutory missions of many Federal agencies,especially those that operate in R&Dintensiveareas: For example; the statute under whichthe ':ntional Institute of Education operatesauthorizes its director "to conduct educationalresearch...[and] assist and foster such re-search; collection; dissemination..." Similarly;NASA is required by the Space A+ct to providethe widest practicable and appropriate dis-semination of information concerning itsactivities and their results. Under broadmandates such as these, agencies have

'wally favored -1 distribution of scientificvd tecinicel information over full cost

recovery. since wide distribution contributesto the purposes the agencies are tryingto accomplish.

(2) Government scientific and technicalinformation activities serve the :public intere.stin important ways. For example, in discussionsof Medlinc, NLM's online bibliographicservice, advocates of the systerh stress thatlowcos access to the medical literature forresearchers; students, and practicing phys-icians yields dividends in improved healthcare and accelerated biomedical research:These dividends; it is claimed; far outweighthe costs to the Government of providingthe services; and are more important elanthe issue of unfair cr.,:npetition with morecostly private sector bibliographic servircs.

(3) The private se :tor contains both pro-ducKs and users of scientific and technicalinf rmation. It is not coincidental that it isgr:nerally the prc.viu cers who see themselvesin competition with a e Government and whoargue most strongly a3ainst the Government

.

role. The private sector users, on the otherhand, who benefit from rile Gove:nmentservices, favor their ..ontinuation and do notusually regard unfair competition as animportant issue.

One major point too seldom acknowl-edged by either side in this de'v.ite is thefundamental difficulty with tiie concept of

77

fullcost recovery. According to economistYale Braunstein. the "apparent simrlicitvand concreteness of the fullcost calculation isentirely misleading' (National Tel.corn.munications and Information Administration.1981: 67). Braunstein points out that theexistence of joint costs (costs that are sharedamong several products) in any multiproductorganization and the problem of demandelasticities make full-cost recovery an elusivenotion. He quotes a pricing manual preparedfor NSF to demonstrate the complexity ofcreating full-cost recovery policies:

...if there are economies of kale or otherpatterns of responsiveness of costs tovolume of sales. demand data will alsobe needed if the prices selected areactually to end up covering costs: De-mand infc oration cannot be dispensedwith, for in cal:::o! 'ing the pertinent costthe management must be able to ascertainwhat volume of sales can be expectedat full cost. If a calculation of full costsis based on cost data for the past and.for example, it seems to require a sharpincrease in price. the resulting fall inquantity sold may lead to a loss of scaleeconomies, and the alleged hill-cost pricewill in fact fail to produce revenues equalto costs as it is intended tc. (NriA.1981: 67).

Braunstein also suggests that there maybe an inconsistency between the objectivesof OMi3 Circular A-76, which is intended toprevent Government entry into enterprisesthat can be better conduct9r4 b;; -tire privatesector, and OMB Circular A-25, which re-quires full-cost recovery and thus encouragesthe Government to provide self-supportingserices. AS he obServes with respect co A-25:

This guidance conicts with the basicpremises that the government should ranprimarily appropriated funds; and thatif an activity can be self-sustaining, itshould be con lucted in the private; ivherthan the public sector (NTIA, 1981: 68).

Public/Private Sector Conflictsin Perspectiv

Discussions of such issues as full-cost re-covery. which concern the relationships

78

h,etween the roles of the public and privatecertors. are among the thorniest problemsin scientific and technical information policy.Underlying these discussions are basic phil:osophical differences concerning the properrole of the Fedemi Guvemment in providingknformation services products, and resources.

Those who favor restricting the role ofGovernment and relying more on marketforces place a high value on the Americantradition of competitive private enterprise.and fed that the private sector can distributeinformation (originating from both Govern-ment and nonGovernment sources) moste.:onomically and most widely. They believethat the presence of Government "can have achilling effect on private sector investment,"and can reduce the efficiency of the market-place in allocating resources (National Com-mission on Libraries anal Information Science.1982: viii).

Those who would prefer not to restrictthe role of Government emphasize differentsets of values incIuding the need for "equita-ble. open access...to information which hasbeen generated, collected, processed, and/ordistributed with taxpayer funds," and theimportance of that information in ensunngbroad public participation in the affairs ofsociety. regardless of individual ability to pay.They stress that is a proper Governmentrole to meet those information needs notserved by the me-.etplace and to stimulate"the development of information as a re-source for dealing with societal problems"(National Commission on Libraries andInformation Science. 1982: ix).

The most productive course of action liesnot in attempting to decide which of theopposing views is correct in any absolutesense. but rather in seekine a middle groundthat makes the best use of both Govemmentand private sector capabilities to serve theNation's scientific >nd technical informationneeds. This means determining the condi-tions and circumstances' under which eachsector is most capable, and establishingpolicies based on these determinations.

In part, these dilen-ima.s in public-privaterelationships seem to have beer, brought

by the vet-; success of Governmentscientific and tec1;.if.;:i! inforniation programs.Few concerns about unfair competition fromGovernment were raised in the early days of

91

programs like NTIS and MEDLARS, chieflybecause start-up costs for such programswere high and the markets for their serviceswere not large enough to interest manyprivate firms. The Federal programs, takingadvantage of advancing technology, helpedexpand and cultivate the markets, makingthem much more attractive today (Williams,quoted in Schuman, 1982: 1064). Now themarket is more mature and the technologyis less costly. thus the field is much easierto enter: The current need is to adjust tothe new situation in a manner that bestserves the public interest:

While there are no simple ways to ac-complish this; some potentially productiveapproaches have been suggested in recentyears. The common theme of these approaches is differentiating among types ofinformation arid their uses_ and allocatingpublic_and private roles differently in eachcase. One approachthat of Giuliano. etal. described three basic modes of scientificand technical informatioi: transfer (termed"eras"), each corresponding to a differentvalue vstem. Giuliano showed how differntuses ci expectatons in each era affect theways in which scientific and technical infor-mation is managed and disbibuted. Traditionsand in:3+ft: ittonr:1 mechanisms in the scientificand technical information field developed.in connection with D. traril3ter system in whichthe worth of infoirrztion fs seldom measuredin economic terms. A variety of problemsincluding conflicts between public sectorand private sector rolesresult from themisapplication of systems developed underone set of values to situations in whichdifferent values are prevalent (Giuliano;1978: 2-11).

A related approach was taken by a Na-tional Commission on Libraries and Infor-mation Science Task Force on Public Sector/Private Sector Interaction in ProvidingInformation Services, chargeci with recom-mending means tc. resc:ive conflicts bvatweenthe sector5. The tisk force developed a"schematic of contaits," which includes withfactor; as the social value of the infor n;ation,;tf., ,econkJnic utility, the immediacy or itsvale' thc! ability of users to pay, cic.,

to help understand how conili ;:tsconcerning the appropriate role ol t!2Federal Govern' vent in providing infor

92

mation services occur and how they mightbe resolved.

F. W. Horton. in an unpublished discussionpaper circulated in conjunction with theInformation Industry Association's new hand-book, Understanding U.S. Information Policy,proposes a "hierarchy of needs" that infor-mation is expected to serve and suggeststhat pricing policies (and presumably publicand private sector roles) be developed onthe basis of it. The hierarchy begins with"coping information" and goes through"helping:. "enlightening." and "enriching ;"to "edifying" information. In this scheme;"coping information" would be disseminatedfree of charge (presumably by the Govern-ment) while "edifying information" wouldbe priced at fully competitive marketplacelevels. Other categories would fall in between.

None of the approaches described hereprovides a definitive answer to this complexset of problems. but together they highlightthe need to develop new ways to resolvethe public/private sector controversies ininformation policy generally and in scientificand technical information in particular.

Access to Scientific andTechnical Information

Thy, growing commerce in information,prompted by the recognition of its value(social. political, and economic), has raiseda number of concerns related to control ofand access :o information. The matter ofaccess ii s special implications for scientificand technical information, since much of itis the product of a substantial Federal invest-ment in research and development. Problerroanse because of the often-conflicting Federalgoals of p,oviding information resulting fromGovernment-sponsored research to thepublic, rwritecting t'ne rights of individuals toprivacy and of firms to confidentiality; assuringthe Omen ir 'nt access to it-Ain-nation itrequires for mal-tirig regulatory and othcidecisk,hi, safeguarding ratirtnal sec; ;ridpromoting donre.,.:;, i:ri'rrioti,,:r1,.:!

DuringWor;d1:ar II,were placer. ;)!

information for reasons of national security.These controls conflicted with the traditionalscier tific emphasis on open communicationand the free flow of information:* Thetraditional position was stated in 1945 inScience The Endless Frontier. a reportprepared 'and submitted to President Trumanby a task force headed by Vannevr.,

Basically there is no reason to bef.,1.-inwscientists of other countries will t,ire discover everything we now .

sounder foundation for our nutiunalsecurity rests in a broad dissemination ofscientific knowledge upon which furtheradvances mt be more readily made thanin a policy of restrictions which wouldimpede our further advances in the hopethat our potential enemies will not catchup with us. (Bush. 1945: 29)

Balanced against the desire for scientificGfreedom and the need for Government

accountability: which generally favors openaccess to information; has been the need toprotect national security. Controls have takensuc!. forms as increased classification ofscientific and technical information andapplication of _export rrzgulations; and haveplaced particular emphasis on res!iicting the.acquisition by the Soviet Union of Americantechnology and scientific and technicalinformation with potential military value.

The security classification system hasevolved through_a series of Executive ordets.Since 1953. each successive EXecufive orderhas narrowed the definition 4iai. theGovemment could classify. Howev. Y. themost recent one, Executive Order 12356,issued bt.PresidentReagan in April 1982,reversed this trend by expanding the cate-gories of classifiable infotination and revisingvarious classification procedures. The newExecutive order excludes classification of basicscientific research information not clearlytied to national security. However, Jstinctionsbetween basic and applied research areblurring in many fields, and the implications

_____of_thc.,_or_det-in-certain-sensitive areas:- suchas ctyptootaphy. are not yet clear.

*Free Flo..; is used here to mean we "unrestricted"flow of information, rather than "no-cost" transfer.

81)

The concern about foreign access to U.S.scientific and technical infonon has alsobeen addressed through the t!.echi...lism ofexport reguladonF, incto&-..t the ExportAdministration Act (50 (..; App. 2401- -2413) and the Arms Export Control Act(22 U.S.C. 2751-2794): These statutesimpose licensing requirements on the exportof certain goods; technology: and defensearticles from the United States; and alsorestrict the access of foreign nationals tosuch material in the Unii?cl States: Bonistatute: are defined broadly enough toinclude technical__ information and docu-mentation as well as goods and servicestcongressionai Research Service. 198ZRelyea: 7). The International Traffic in ArMsRegulations (ITAR) (22 C.F.R. 121.01 etseq.. 1981). authorized by the Arms ExportControl Act. require Government approvalfor publication of technical data with potentialmilitary significance. These regulations havebeen applied to embargo the export ofadvanced technologies, including computers,t- the Soviet Union to protest its involvementin Afghanistan. and to restrict the discern=ination of information at scientific andtechnical meetings:

Enforcement of these regulations (primarilyby the Deparh%lents of Commerce andDefense) has been increasing under theReagan Administration, causing some con-cern among_ many scientists unused toconsidering the implications of their workin terms of national security issues. Que_stionsabout the appropriate use of Governmentcontrols will continue to gain currency: theExport Administration Act expires at the endof September 1983.

One recent action reflecting in part the --------Administration's concern about the use ofU.S. information processing and telecom:munications systemsand scientific andtechnical information in generalby foreignnations was the withdrawal of U.S. Govern-ment supnntt for he International Institutefor ;ystern..; Analysis MASA): Amongthe-reasons for the decision was the Admin-istration's desire to prevent Communist blocresearchers from obtaining unauthorizedaccess to Western databases through IIASAcomputers (Walsh; 1982: 35).

Other laws intended to promote nationalsecurity through control of information

include the Atomic Energy Act (68 St-at.919: 42 U.S.C. 2011-2296), which containsthe "born classified" concept with respectto atomic energy restricted data; and theInvention Secrecy Act (35 U.5 C. 181-188).which provides authority for Wthholdinr, pat-ents containing sensitive information in orderto keep the inventions in question secret.

A sorriewhat different means of controllingdissemination of information for reasons ofnational security is evolving with the devel-opment of a system of voluntary prepubli-cation review of manuscripts on cryptographicresearch. The system is an attempt to balanceScientific and commercial interests in thecryptographic and cryptoanalytic fields withconcerns about the vulner..:Aity of U.S.cryptographic systems. The extension of thispolicy to other areas of scientific researchwas proposed by former deputy CentralIntelligence Agency director Bobby R. Inmanat the 1982 annual meeting of the AtneticbhAstmciation for the Advancement of Science(AAAS). Inman suggested that "a potentialbaiance between national security and sciencemay be in an arrangement to include in thepeer review process (prior to the start ofresearch and prior to publication) the ques-tion of potential harm to the nation." If such;Okintary ;security safeguards are not adopted;

Inman warned, public outrage over theresulting "hemorrhage of tti, nation's tech-nologies" would result in laws to restrictPUblication of scientific work consideredsensitive on national security grounds.*

Dis;.ussion of the conflicts bet; een secrecybaSed on national security and the unim-peded flow of scientific and technical infor-inc.:ion is under way in many forums and isthe topic ior another pa; eY in this ;com-pendium. Communication andNational Security, a recent rencitt by a panelof the National Acathei::±,- of 5,:iences, onthe whole echoes the sentiments expressedin the Bush report, while facing realisticallythe need to restrict illegal acqui ;Mon oftechnology by other nations ( Nationa!Academy of Science. 1982). Thc. p. elconcluded that in by far the largest shareu;university research; "the benefits of total

For Inman remark ancLthe response of PO,ASexecutive officer William D. Carey. see Aviation Weekand Space Technology. February 8. 1982, pp. 10.11; ff.

94

-7-zynn,,p; overshadow [the] possible neart nn rnilitary benefits to the Soviet Union."At the same time, it noted that there arecertain areas of research that should clearlybe classified. It devoted special attention tothe small "gray area" between the two, forwhich it felt "limited restrictions short ofclassification are appropriate." In this grayarea. the panel suggested several specificmeans of limiting unauthorized foreign accessto potentially sensitive researchmeans thatit regards as consistent with the values ofopen science. It emphasized the need forthe Government to define in concrete termsthe areas in which the International Trafficin Arms Reclulations and the Export AdminTiStratiori Regulations are applicable andindicated that these regulations should notbe invoked to deal with gray areas in Gov-ernmentfunded university research.

Other International IssuesThe international context for informationpolicy includes many issues beyond nationalsec;....ity concerns. The global flow of infor-mation has become increasingly possible asa result of rapid advances in data processingand computer systems and in telecom-munications networks. The merging of thesetechnologies has facilitated internationalcommerce in infor:nation, including scientific and technical information; and hasraised a range of associatc.d economic;political, technological concerns.

Responsil::Pty for inter, itional informationpolicy in the United States, like domesticintermation policy, has traditionally beendispersed among many parts of the Gc,i.,,:frn-ment, including the Departments of ,

Corns-nerce, and Defense, the FederalCommunica;.ions Commission, and theOffice of the U.S. Trade Representative. ThiSdiviFio.n of authority contrasts sharply withthe increasing efforts of other countries todevelop and implement unifom nationalpolicies t., regulate information and com-lunication. ,-echnoloc;es and the form and

content of transborder data -lows. This :snot to intoly that the United States .3hcsuljnecessarily ?rntrit.i...: the approaches vi other

ns in it.iortnation policy. Indeed, theremany good reasons nrst to do so. V /hat

is important is gaining an appreciation of

81

the strengths and motivations behind theseapproaches to informat:jn policy so that U.S.policy can deal more effectively with them.

National information policies are usuallydesigned to build a nation's informationindependence and increase economic gainsby developing indigenous information indus-tries. While the desire to protect the privacyof individuals and to promote data securityled to some of the early efforts by Europeansto control the flow of data, today a widearray of cultural. political. and economicmotivations are behind such moves. Swedenenacted the first national data protectionlegiSlation in 1973; other countries, includingCanada. France. and Germany, have sinceenacted information privacy or data protec-tion legislation as part of their nationalinformation policies. The laws and enforce-ment provisions vary, but in most casesregistration, public disclosure, and licensingrestrictions are detailed. These regulatoryefforts frequently form impediments to thefree flow of information.

In C tober 1980, the Organization forEconowic Cooperation and Dev dopmentOFe3; adopted voluntary "Guirielines on

Privaz's Protection and Transborder DataFlows." About the same time; the 21-memberCouncil of Europe issued its "Conventionfor the Protection of Individuals with Regardto Automatic Processing of Personal Data."Both codes govern. transborder flows ofpersonal data; the extent tc which thetransmission of corporate c.ia is affected isoften unclear.

Sortie restrictions, including the impositionof taxes. tariffs, and user fees, constituteeconomic barriers that may effectively pricesome providers of scientific and technicalinformation services and equipment out ofinternational markets. Such economic -sare of particular concern to the United Sta.which has achieved a competitive edge ininternational markets for communicationsand information services (CongressionalResearch Seivice/Bortnick: 3: National Com-mission on Libraries and Information Science1982: 27). Other regulations that restrictthe flow of information between countriesinclude requiring "domestic" it tobe processed within the originating country(e.g.; West Germany); requiring the purchaseof host countries' equipment (e,g,; Brazil);

82

or. in some cases, complete denial of marketentry (e.g., Canada. Great Britain) (Spero.1982: 143: U.S. House of Representatives,Government Operations Committee.1980: 18).

Potential requirements that data trans-missions be monitored to ensure compliancewith privacy regulations are also of concern,Business and governments are wary of thesecurity of confidential and proprietary datawhen it is open to examination by othergovernrn f-its or when it is transmitted overpublic da't..1 networks (U.S. House of Repre-sentatives Government Operations Com-mittee, 1 L)80: 16). Closely tied to this arethe qt2,:,,qions of national sovereignty and

security. Many nations feel threat-ened the loss of control over databasesthat stored and processed in foreigncorn 'ter systemsespecially in U.S.systems -- subject to foreign laws (Congres-siorlai Research Service, 1982, Bortnick: 3).They are concerned about the ability to accesstheir own data and the ability of other nationsto access critical information when it isremoved from local facilities.

Apart from the question of legal juris-dictioit; many nations are concerned aboutsuch things as sabotage; equipment failures,and political decisions that might inhibit theirability to access their own data. They contendthat once critical information is removed fromlocal facilities, their vulnerability is increasedas foreign nations and multinational corpora-tions have greater potential to access it.

There ,.; an additional concern that trans-border data flows can lead to "culturalerosion." Although the United States main-_tains fundamental beliefs in the value ofcultural diverky, other nations fear that theinfluence of the United States and otherWestern nations may disrupt their indigenouscultural heritage. These nations contend thatforeign databases, as well as mass media,contain cultural biases that are potentiallyharmful to their societies. Although thisperspecive has been voiced by some Westernnations. specifically Canada and France, iti of pani:ular concern to developing coun-tries; many of which lack a highly educatedpopulwiorl and are consciously seeking toencourage adherence to taditional mores.

Indeed; the importance of access to foreignscientific and technical information for pur-

95

poses of development and modernizationversus the perceived economic and politicalvulnerability caused by reliance on foreignfacilities and information is a particularproblem for developing nations (U.S. Houseof Representatives; Government OperationsCommittee, 1980: 21):DatabaseS and ad-vanced information systems are often viewedwith ambivalence ''av developing countriesfaced with th.c.. assuring nationalsove wi7+, 'requiring information

dnd P di lo!ogies for development(b. t981: 32).

One illustration of this is the conflict overcollection of agricultural, environmental, andgeological data by remote sensing satellitesystems such as Landsat. Nations collectingthis data benefit from research applicationsand trade and investment opportunities;developing countries benefit from applica-tions of the data to problems in such areasas agricultural productivity and naturalresources management. However, somedeveloping nations have tried to claim thisinformation as a national resource, andsought to tax and limit ifs collection so as toprevent what they feel is exploitation by moreadvanced nations (Congressional ResearchService; 191: 61):

Protection of Domestic ProprietaryInformation

Among access issues on the domestic frontin scientific and technical information is theproblem of protecting proprietary information.Such protection is a centralcommercial enterprises andwith clairin of the "right" of .information, Patents and tr,ade st?.-the two most ,rnmon means of protectingthe ownership of commercial informationwhile promoting its creation and application:Patent laws attempt to encourage com-munication by requiring disclosure of infor-mation about inventions in return for a17 year legal monopoly. Trade secrecy law;on the other hand; allows employers toenforce nondisclosure clauses in contractswith employees or license&:s to protectspedalized knowledge that gives a firm somecompetitive advantage.

Questions of ownership and control ofinformation are increasingly complex, par-

ticularly in high-technology fields such asmicroelectronics and biotechnology, wherethe lines between basic and applied researchare narrowing. For example, patent prospectsmay affect release of data and inhibit corn-munication among scientists who have avested interest in research results (Nelkin,1982: 706). Findings that may be importantto the advance of basic research may bewithheld from publication because of com-mercial interests in potential applications.

The notion of scientific ideas and data asintellectual property; and the related question

..of ownership of intellectual property; areboth critical to scientific and technical infor-mation policy. While precise definitions ofintellectual property vary, it is generallyassumed that investigators have the right"to enjoy the fruit of their intellectual labor"(National Telecommunications and Infor-mation Administration, 1981: 76). This rightis p :otected in varying degrees by patent,trade secrecy, and copyright law. The copy-light laws protect individual or group rightsto unpublished and published works for aspecific period of time and permit compen-sation to 7 copyright holder for use ofthe materials.

Government support for R&D leads toquestions about ownership and disclosureof data produced in non-Government insti-tutions under Federal funding: Grant pro-visions favor public disclosure of researchresults; but those guidelines are consideredsubject to the researcher's right to decidev.:hen results are ready to be published(Nelkin, 1981: 704). Funding agencies may-equest access to data to verify the progresse'f research: this infon nation may, in turn;

subject to disclosure upon request underthe Freedom 01 Information Act or otherreguiz-Jons (Gordon, 1982: 10). These arebut a few of the problems cif safeguardingthe disclosure of :nional and proprietaryinformation that has `-ieen provided to theGovernment.

Scientific and Technical11,.formation and Federal',

R&D Policy

The preceding iections of this paper haveoutlined a number of issues involving scien-

83

tific and technical information currentlyfacing policymakers and the scientific andengineering communities. Several factorsare combining to give these issues a spe-cial urgency.

The division of labor between the publicand private sectors in the provision ofscientific and 'e.chnical information serviceshaf.,` been unco.i discussion for several years.It --',..ence has increased considerably,j';,; .never. during the recent past, as the; ieagan Administration has sought to reduce

role of the Federal Govemment in societyand to transfer to the private sector thosefunctions that it believes can be performedmore efficiently and effectively outside ofGovernment. At the same time as thisredefinition of Federal roles and responsi7bilities has been taking place, awareness ofthe importance of R&D and technologicallybased industry in economic recovery and inthe solution of social and national securityprobluns has been growing.

In this context; there has been somediscussion of shifting responsibility forfunctions served by Federal scientific andtechnical information systems such as NTISand NLM to the private sector. In general;these discussions have come up on a case-hy-case basis and have not been related toa systematic examination of Federal scientificand technical information policy. Further;they have given little consideration to possiblelong.term consequences to the U.S, R&Dsystem of limiting access of U.S. users toscientific and technical information andpossibly allowing foreign control of U.S.scientific and technical information resources.

Issues of access to scientific and-technicalinformation have come to the fore mainlyin terms of growing emphasis on nationalsecurity concerns and increased recognitionof the importance of science-based tech-nology to U.S. securityboth military andeconomic. Although studies such as thatconducted by the National Academy ofSciences have helped to focus the argumentsin this area; the need to address nationalsecurity concerns may drive policy in amanner not well linked to broader questionsof scientific and technical information policy(National Academy of Sciences; 1982).

Finally; the technology of information itselfis advancing rapidly and offering both prob-

84

lems and opportunities for the U.S. researchsystem. The most recent Five-Year Outlookon Science and Technology suggests thepotential importance to scientific and tech-

communication of such developmentsas electronic "mail" (for rapid exchange ofscientific data); the growing capabilities ofelectronic data bases; and new; electronicforms of rapid publication (National ScienceFoundation; 1983). The continued develop-ment and implementation of such technol-ogies. although likely to be based in theprivate sector, depends in part on the Federalroleboth directly through Federal supportfor the development of associated tech-nologies, and indirectly through incentivesand disincentives to the use of such systemsbuilt into Federal R&D policies.

In these and other policy issues affectingscientific and technical information, theFederal Government plays many differentroles, including:

Generator of scientific and technicalinformation. through federally fundedR&D efforts;Provider /disseminator of scientific andtechnical information. through Federalinformation centers and data bases;User of scientific and technical infor-mation: in its performance of R&D andin other technologically based activities,such as regulation:

e International negotiator i.e.; repre-sentative of U.S. interests in variousinternational forums;Mediator of competing interests:Policymaker, a locus of decision forcontroversial issues and authoritativeallocator of public resources; andSource of technical advance, throughsupport of R&D in communicationsand information technology.

The Federal role is not exclusive in anyof these areas. Commercial enterprisesasR&D performers, as sources of informationtechnology, and as providers and users ofscientific and technical information servicesare involved throughout, as are universities,nonprofit organizations, and professionalscientific and engineering associations.Nonetheless. Government involvement ispervasive and inescapable as part of therecognized Federal responsibility in thesupport of R&D and application of its

97

products to the Government's and othernatio_nal needs. This pervasiveness meansthat Federal policy for scientific and technicalinformation will be made and will have animportant influence on the course of eventsin this area. The choice is between allowingthat policy to be shaped by unintendedconsequences of choices made in otherdomains, making it the net result of numerousuncoordinated efforts, or, alternatively, havinga policy developed on the basis of conscious,purposeful decisions, based on data andunderstanding of the issues and structuredto lead to a desired outcome. The importanceof scientific and technical information to theR&D enterprise dictates that it be the latter.

Policy Options

To address the issues outlined above willrequire a concerted effort on the part ofpolicymakers in the Federal Governmentand among the numerous other stakeholderswhose interests are associated with the areaof scientific and technical information policyMaking the importance of the issues in thisarea better known and gaining the attentionof individuals in organizations whose interestsare involved is an important first step towhich this and similar papers may contribute.Beyond this, it may be useful to suggest anumber of specific options for furtherexamination and action:

(1) Reexamine pricing policies for Federalinformation services. Many of the conflictsbetween the public sector and the privatesector in scientific and technical informationpolicy center on the prices charged users byFederal information services. Although thereare a variety of high-level policy statementson the subject, agency policies differ con-siderably and both the information servicesand their users could benefit from an overallsystematic examination of pricing. The valuesof implicit and explicit subsidies to usersshould be included in the examination, asshould alternative means of providing thosesubsidies. The concepts of differential pricingdiscussed above under the heading "public/private sector conflicts in perspective" meritconsideration in this context.

(2) Make representation of U.S. interests inscientific and technical information a con-

98

scious element of foreign policy. Actions offoreign governments in the information fieldmay in some cases have significant negativeimpacts on the U.S. R&D community andon U.S. scientific and technical informationservices; in both the public and the privatesectors. Traditions and circumstances in theinformation field in the United States aredifferent from those in other countries, andthis country should not necessarily emulatethe actions of others. It is essential, however,that U.S. foreign policymakers recognize theimportance of U.S. scientific and technicalinformation interests and actively representthem. The flow of scientific and technicalinformation is included among these interests,both as an element of foreign trade and asan intellectual resource for the U.S. researchand industrial communities.

(3) Examine the policy structure for sci-entific and technical information in order todefine more clearly authority and responsi-bility. Too often problems in scientific andtechnical information seem to "fall throughthe cracks" because organizations capableof dealing with them do not exist at theproper hierarchical level or because highlevel policymaking bodies have too manycompeting interests to devote adequateattention to them. Creation of a new "Officeof Information Policy." as has been pro-posed from time to time is not likely to bethe answer. But a systematic review of theadequacy of existing organizational ar-rangements and a possible reassignment ofresponsibilities seem both appropriateand timely.

(4) Enhance awareness of scientific andtechnical information issues among R&Dpolicymakers. Of the several contexts in whichdecisions on scientific and technical infor-mation issues should be considered, that ofR&D poli -cy is particularly important. Yetdecisions on scientific and technical infor-mation issues seem all too often to be madeindependently of R&D policy decisions. R&Dpolicymakers should be encouraged torecognize and pursue their interests inscientific and technical information issues.

(5) Provide a forum for continuing dis-cussions among parties with different view-points and interests in scientific and technical .

information policy. Apart from the othersuggestions contained here; it would appear

85

that if real progress is to made in resolvingthe issues of scientific and technical infor-mation policy; many organizations andindividuals with widely differing viewpointswill need to subordinate conflicting interestsand act on the basis of larger commoninterest. An important step in this directionwould be the establishment of a setting forinitiating and maintaining discussions amongthese parties and a mechanism for assuringthat the policy process is informed by theresults of these discussions.

References1. 13rown, George E.. .Jr. "Information Issues Con-

fronting the Congress." ASIS Bulletin (December1981): 31-35.

2. Bush. Vannevar. Science-The Endless Frontier.A Report to the President on a Program for Post-war Scientific Research. Washington. DC: NationalScience Foundation. 1980 (orig. 1945).

3. Congressional Research Service. Library of Con.gross: information and Telecommunications: AnOverview of Issues. Technologies and Applica.tions: Prepared for the Subcommittee on Science.Research and Technology of the Committee onScience and_Technology: U.S: _House of Representatives. Washington. DC. July 1981.

4. Congressional Research Service: Library of Com_gress. loternational Data Flow Issues. Issue BriefNumber 1B81040. by Jane Bortnick. revised Sep-tember 1. 1982.

5. Congressional Research Service. Library of Con-gress. National Security: Con trots and ScientificInformation. Issue Brief Number 1B82083. byHarold C: Relyea. revised September 7. 1982.

b. Congressional Research Service. Library of Con-gress. Scientific and Technical Information (STI)Activities: Issues and Opportunities. Prepared forthe Subcommittee on Science: Research andTechnology of the Committee on Science andTechnology: _LIS: House of Representatives:December 1978.

86

7. Giuliano, Vincent: et al: Into theinformation Age:A report prepared by Arthur D. Little. Inc., for theNational Science Foundation,- Chicago: IL: Amen:can Libraryi Association. 1978.

8. Gordon, Robert S. "Data _Control ConsiderationsUnique to Clinical Trials." Paper presented at theAmerican Association for the Advancement ofScience Annual Meeting. January 5. 1982.King. Donald W. "What the Indicators are TellingUs." Presented to a Special Meeting of Informa-tion Managers. New Executive Office Building.February 4. 1981:

10. National. Academy of Sciences. Scientific Corn:munication and National Security: A report prepared by the Panel on Scientific Communicationand National Secutity, Committee on Science.Engineering, and Public Policy. Washington. DC:National Academy Press, 1982.

11. National Commission on Libraries and Informa-tion Science, Public/Private Sector Interaction inProviding Information Services. Washington. DC:February 1. 1982.

12. National Science Foundation: Five-Year Outtookon Science and Technology. 1982. Washington.DC: U.S: Government Printing Office. 1983:

13. National Telecommunications and InformationAdministration. U.S. Department of Commerce.Issues in Information Policy. Washington, DC:February 1981:

14. Nelkin. Dorothy. "Intellectual Property The Controlof Scientific Information." Science. vol. 216 (14May 1982): 704 -718.

15. Schuman: Patricia Glass. "Information Justice."LibrarY Journa/ (1 June 1982): 1060.1066.

16. Spero. Joan Edelman. "Information: The PolicyVoid," Foreign Policy (Fall 1982): 139 156.

17. U.S. General Accounting Office. Better InformationPolicies Needed: A Study of Scientific and Technical Bibliographic Services. PSAD79:62.August 6. 1979:

18. U.S. House of Representatives. Committee onGovernment Operations: InternationalinformationFloW: Forging a New Framework. H.R. 961535.December 11:_1980.

19. Walsh. John. "Lack of Reciprocity Prompts HASACutoff." Science. vol. 216 (2 April 1982): 35.

that if real progress is to made in resolvingthe issues of scientific and technical infor-mation policy; many organizations andindividuals with widely differing viewpointswill need to subordinate conflicting interestsand act on the basis of larger commoninterest. An important step in this directionwould be the establishment of a setting forinitiating and maintaining discussions amongthese parties and a mechanism for assuringthat the policy process is informed by theresults of these discussions.

References1. Brown. George E.. .Jr. "Information Issues Con-

fronting The Congress." ASIS Bulletin (December1981): 31-35.

2. Bush, Vannevar. Science-The Endless Frontier.A Report to the President on a Program for Post-war Scientific Research. Washington. DC: NationalScience Foundation. 1980 (orig. 1945).

3. Congressional Research Service. Library of Con.gress. information and Telecommunications: AnOverview of Issues. Technologies and Applica.firms: Prepared for the Subcommittee on Science.Research and TechnOtogy of the Committee onScience and_Technology: U.S: _House of Representatives. Washington. DC, July 1981.

4. Congressional Research Service: Library of Con:gress. ternational Data Flow Issues. Issue BriefNumber 1B81040. by Jane Bortnick. revised Sep-tember 1, 1982.

5: Congressional Research Service. Library of Con-gress. National Security. Controls and ScientificInformation. Issue Brief Number 1B82083. byHarold C: Relyea. revised September 7. 1982.

6. Congressional Research Service. Library of Con-gress. Scientific and Technical Information (STI)Activities: Issues and Opportunities. Prepared forthe Subcommittee on Science: Research andTechnology of the Committee on Science andTechnology: _LIS: House of Representatives:December 1978.

;6

7: Giuliano: Vincent: et al: Into theinformation Age:A report prepared by Arthur D. Little. Inc., for theNational Science Foundation,- Chicago: IL: Amen:can LibrarY ASsociation. 1978.

8. Gordon, Robert S. "Data _Control ConsiderationsUnique to Clinical Trials." Paper presented at theAmerican Association for the Advancement ofScience Annual Meeting. January 5. 1982.

9. Kinq.Dunalc1 W. "What the Indicators are TellingUs." Presented to a Special Meeting of Informa-tion Managers. New Executive Office Building.February 4. 1981:

10. National Academy of Sciences. Scientific Corn:munication and National Security. A report pre.pared by the Panel on Scientific Communicationand National Security, Committee on Science:Engineering. and Public Policy. Washington. DC:National Academy Press: 1982.

11. National Commission on Libraries and Informa-tion Science, Public/Private Sector Interaction inProviding Information Services. Washington. DC:February 1, 1982.

12. National Science Foundation: Five-Year Outtookon Science and Technology. 1982. Washington.DC: U.S: Government Printing Office. 1983:

13. National Telecommunications and InformationAdministration. U.S. Department of Commerce.Issues in Information Policy. Washington, DC:February 1981:

14. Nelkin. Dorothy. "Intellectual Property The Controlof Scientific Information." Science. vol. 216 (14M4i 1982): 704 -718.

15. Schuman: Patricia Glass. "Information Justice."Library Jouma/ (1 June 1982): 1060.1066.

16. Spero. Joan Edelman. "Information: The PolicyVoid," Foreign Policy (Fall 1982): 139 156.

17. U.S. General Accounting Office. Better informationPolicies Needed: A Study of Scientific and Tethnical Bibliographic Services. PSAD79:62.August 6. 1979:

18. U.S. House of Representatives. Committee onGovernment Operations: International_informationFlow: Forging a New Framework. H.R. 961535.December 11:_1980.

19. Walsh. John. "Lack of Reciprocity Prompts HASACutoff." Science. vol. 216 (2 April 1982): 35.

FCC felt the Act mandated it to protectexisting service, which by definition (sincethe FCC was on the job) served the publicinterest. This often meant a cautious andeven negative attitude toward any technicalinnovations that might threaten the financialhealth of current broadcasters. Services thatcould have provided additional outlets fortelevision, causing no interference andameliorating scarcity to some degree, werehampered by regulatory restrictions. Theregulatory environment in turn dampenedinvestment and innovation. The FCC's planfor allocation of Very High Frequency (VHF)and Ultrahigh Frequency (UHF) channelsto television stations wound up discourag-ing the formation of more than three nationalnetworks; Severe restrictions on the broad-cast signals that cable systems could carryand on the markets they could serve; strictand expensive specifications for satellitereceiving stations, and limitations on paytelevision held back the growth of cable (17:see also 4; and 18: 11-16).

These regulations may have protectedthe profits and services of existing broad-casters. But, by the 1970s, it had becomeincreasingly clear that technology was leapingbeyond the old assumptions. Scarcity didnot seem inevitable. New outlets for videotelecommunications were making it possibleto envision a genuine competitive videomarketplace to replace Government-regu-lated scarcity. Influential scholarly andpolitical analyses of the banes of overregu-lation began appearing (29; 31), reinforcingthe perception that a market approach wasboth feasible and preferable. Deregulationin such areas as trucking and airline travelenhanced its acceptability among com-munications policymakers.

Prompted by these developments; theFCC began to question the assumptionsand loosen the bonds. The process ofderegulation continues. Precisely where itshould apply; and where regulation shouldbe enforced or modified; is the major ques-tion this paper addresses.

The Geostationary Satellite and theNew Abundance. Although it is impossibleto identify all the technical developmentsthat undermined the persuasiveness of thescarcity assumption; probably the most sig-

88

nificant was the proliferation of geostationaryearth satellites. Perched 22,300 miles abovethe Equator, these small machines receivesignals transmitted from earth stations andretransmit them to parabolic dish antennaslocated across wide regions (a time zone ortwo) (see 33 for a description of satellitetechnology). They are now relatively inex-pensive to build; launch; and operate (26:219-23). As more were pat into orbit; morechannels became available at progresshielylower cost; This allowed television network-ing at a price significantly lower than thatof the previously used land relays. HomeBox Office (HBO), now the largest paycable service with 9 million subscribers (7),was the first to utilize the potential. It beganfeeding movies to cable systems via satellite.At first, HBO was hampered by FCC regu:lations. When the Commission droppedsome restrictions, HBO took off, to be joinedby dozens more satellite networks.

It is difficult to overstate the magnitudeof change: the United States went from fourtelevision networks in 1977 to some 50 injust 5 years. Perhaps 15 of these reach sig-nificant numbers of subscribers. With fewexceptions, the cable networks currentlyreach much smaller audiences than the TVnetworks; many of the cable networks willnot survive the decade. Indeed, the premiercultural network; CBS Cable, shut down inDecember 1982: Still; the new networks areprofoundly changing video telecommuni-cations; Simultaneously; cable and othernew delivery systems are creating capabili-ties besides mere relaying of entertainment.

The Current Market and theNew CompetitorsBroadcast Television. The core of videomedia remains the three commercial broad-cast networks and their affiliates. There areabout 610 stations associated with the threecommercial networks, some 260 publictelevision licensees, and 155 independentcommercial stations. The industry had morethan $8 billion in revenue in 1979; overhalf of Which went to the three networksand the five VHF stations that each ownsand .operates (40: 283), Over-the-air tele-vision continues to garner by far the most

101

viewers and revenues. But its audience shareis declining as the newer services expand.

Cable Television; As of 1981; therewere some 4,350 cable systems with totalrevenues exceeding $1.7 billion (only aboutone-fifth of broadcast television's) (40: 291).Cable began as a rural service providingclear television signals to townsfolk beyondthe reach of big-city broadcasters, As lateas 1969, there were only 3.1 million sub-scribers. By October 1982, basic cable sub-scription had risen to 27 millionaboutone in three of households with television.Pay cable services (premium programmingsuch Z.7 that offered by HBO for an extramonthly fee) are taken by nearly one infive TV households (7). These figures shouldcontinue to grow rapidly, especially sincemost of the Nation's largest dries are stillin the process of being wired and several(Detroit; Baltimore, Washington, Chicago,and New York's four outer boroughs) havebarely begun:

Cable service is organized into "basic"and "pay" services: Basic includes broad-cast stations; automated services providingreadouts of time, stock quotes; headlines;and weather; channels for locally originatedshows; and a chOice among some threedozen satellite networks. The networks earnrevenues from advertisers and some chargecable operators small fees (1-2(X/subscriber/month). Among these are one "cultural"service, two aimed at women, one specializ-ing in black affairs and culture, two headlinenews services and two more in-depth publicaffairs channel:, two emphasizing sports,one for children, and another for teenagemusic afficionados. Basic service averagedabout $7.69 per month in 1980 (14: 1).About a dozen pay services, which are pro-vided for additional monthly fees averaging$8.73 per month (14: 1), emphasize fairlyrecent movies; Broadway shows, and LasVegas. type entertainment shown withoutcommercials.

The older cable systemsstill one-half ormore of the totalonly offer 12 or fewerchannels. Advances in cable technologynow make it possible to provide 54 chan-nels on a single cable, and systems usingdual cable offer up to 108.

102

Besides entertainment and news, thesenewest systems promise to offer an arrayof services that bring the widely heraldedinformation age to fruition. The basic tech-nical innovation is the microprocessor. Thesilicon chip er..Ables the cable to go two-wayessentially installing small computersat the subscriber's end and a larger computerat the system end. Central computers mightbe linked in turn to computers run by banks,newspapers and other data bases, mer-chandisers, and security alarm services.Shopping and banking at home, electronicnewspapers, interactive information retrieval,and home and business security are pro-vided on some cable systems. In addition,cable can be used to hook up institutionsand businesses for high-speed, high-volumedata transmission as well as teleconferenc-ing. Such a network has existed in Manhattanfor 8 years. All of this can be done muchless expensively than by telephone becauseof the larger frequency capacity of coaxialcable compared to traditional uses of thetwisted pair of copper telephone wires. Half-inch coaxial cable can handle as muchinformation as 30,000 twisted pairs (26:327-37).

All the channels cable offers may be sup-plemented by several other infant or about-to-be-launched services.

Subscription Television; In recent years,the FCC also relaxed its restrictions on sub-scription television (STV), which broadcastsa scrambled signal over one channel andrents the descrambler for $20-25 per month.The fare is similar to that of pay cable. Asof September 1981, there were estimatedto be 1.4 million subscribers (41). The futureof STV is problematic. Where cable is avail-able, consumers can purchase basic cableservice and one or two pay channels for theprice of one STV channel.

Multipoint Distribution Service Mul-tipoint Distribution Service (MDS) uses anomnidirectional microwave transmission forone channel of commercial-free premiumentertainment: Most MDS systems offerHBO or other cablelike services for about$20/month. There are 400,000 MDS sub-scribers currently, with a projection of 1.4million by 1985 (40: 303). Althoi,,D,h one-

89

channel service would appear little compe-tition to cable, the FCC has authorized five-channel MDS. If five channels with high-gradeofferings were priced below or with cable,MDS could prove a vigorous competitor.

Direct Broadcast Satellite. The FCChas authorized a new service that will em-ploy satellites broadcasting directly to homesequipped with small (under 1 meter) re-ceiving dishes. The equipment will cost$200-500, with a monthly charge of $20for three or four channels. The nascentindustry is explicitly targeting those homeslikely to be beyond the reach of cable (19).The Direct Broadcast Satellite (DBS) serviceshould begin by 1984.

Low-Power Television. The FCC has alsoauthorized a low-power television stationservice (LPTV) and is currently processingseveral thousand applications. Tnough usinga fraction of the power of full service sta-tions (thus avoiding interference) the typicalLPTV operation could cover a radius of 12-15 milesenough to serve Washington,D.C., and many of its closer suburbs. Equip-ment start-up costs are as low as $100,000(16: 80; 25).

The authorization of LPTV illustrateschanges in Commission thinking that arehelping to stimulate abundance. The FCCstaff report recommends that "consumersshould be able to take full advantage of thetechnologies available in the marketplaceunder the presumption that competition willbest serve the interests of the public" (16:182): Far from regarding competition as athreat to the public interest, the FCC staffreport holds it to be the best way to achievethe public interest.

In accepting these recommendations, theCommission was influenced by technologicalchanges. The cost of production equipmenthas decreased significantly, due in part tothe drop in the price of microprocessors.Those components have also increasedequipment reliability and diminished thedanger of interference. Simultaneously, thenew availability of inexpensive satellitechannels and earth stations enhances thepossibility of forming national networks ofthe low-power stations (25: 39-58).

A Note on Telephone and Other Media.Space limitations preclude consideration of

90

equally epochal alterations of the telephone,computer, and radio branches of telecom-munications; also omitted are videodiscsand videocassettes. Lines between the tech-nologies grow less and less distinct, however,and developments in these areas will ulti-mately affect television and cable.

The telephone industry will alter in espe-cially dramatic fashion as a result of recentantitrust rulings. AT&T will no longer providelocal phone service. It will instead expandfrom long distance service and equipmentprovision into information and computingservices: In all these markets; AT&T willface increasing competition: The companyis not likely to become active in the videofield in the near future; except perhaps indelivering data to be displayed on TV screens(teletext or videotext). In the longer term.AT&T and its telephone competitors maywell nurture new technologies that competemore. directly with cable and broadcast video,but these are not predictable now.

Emerging Policy Issuesand Options

The rise of the competitive video systemsis the major fact of life for the three com-mercial networks and for public television.Growing competition is the driving forcebehind virtually all of the legal and regu-latory issues and initiatives that pervadebroadcasting (see 28). Only cable and broad-casting now pose significant competitivethreats to each other, but all the alterna-tives figure in current policy debates andindustry planning. Underlying most policyissues in both cable and broadcasting, infact, is one central question: Just how vibrantis the competition between them? Forrepresentatives of each industry, the strengthof the other provides the sustaining themeto calls for deregulation. With cable systemsoffering dozens of outlets, broadcasters saychannels are no longer scarce; rer rovingthe need for regulations based on scarcity;For cable operators; the continuing popu-larity (and wealth) of broadcast networksprovides a competitive check preventingabuses of the cable system's local monopoly.

The transition from a regulated and rela-tively uncompetitive video market with ahandful of channels to competition and

1 o 3

abundance is raising four interrelated setsof issues:

How should participation in the videomarketplace be regulated? This issueincludes Government licensing andfranchising and technical stan&rd-setting;How should the video contenttheentertainment and information mes-sages themselvesbe regulated? Thisissue includes questions about ensuringbalance and diversity in broadcastingand cablecasting.What policies will best ensure that atruly competitive video market emergesand persistS? This problem encompassesregulating cross-ownership and jointventures among media corporationsand ensuring programmers' access tonew outlets.What dangers to democratic rights andassumptions are posed by the tele-communications revolution? The twomatters of greatest significance hereare the protection of privacy and thepossibility of an information gap openingup between rich and poor:

These will be the subjects of the remainderof this paper:

Participation in the VideoMarketplace

The arguments for deregulation of entryand exitfor extending license or franchiseterms indefinitelyare based largely on theincreasing competition between broadcastand cable television: Limited license andfranchise terms were designed to subjectperformance to evaluation every few years.Some argue the effect has instead been todiscourage efficient investment.

Television License Terms. Although, inpractice, television licenses are virtuallynever revoked, the need to face the Commission triennially was felt by some to haveinduced more cooperation with viewergroups and thus better service. Now, thederegulation advocates argue, competitionfrom other video providers creates anincentive for high-quality performance bylicensees. Besides, they say, since the Com-mission hardly ever revokes, renewal paper-work reduces to little more than an expen-

1 04

sive ritual. Apparently buying that argument,Congress extended television license termsto 5 years in 1981, and the FCC Chairmanappears amenable to extensions that wouldessentially create a private property rightin licenses. Government renewal would notbe necessary and current restrictions onbuying and selling station licenses would belifted. Market entry and exit would be leftto the market:

Cable Licenses Are Local FranchiseAgreements. Regulation of participationin the cable market has been quite different.With few exceptions, most of the regulationis performed by State and local govemments.The FCC has discontinued most of its cableregulations. Municipalities negotiate 15-20year franchise agreements with cable com-panies, which they (sometimes with a Stateagency) oversee and enforce. Not surpris-ingly, the industry confronts a set of demandsand strictures that varies widely from localeto locale. Citing the competition they facefrom broadcasters and other technologies,deregulation advocates have requestednational legislation to codify and limit thejurisdiction of local governments.

Such a bill, S.2172, passed the SenateCommerce Committee in July 1982. Asimilar bill, S.66, was passed by the Senatein 1983. Perhaps the most controversial pro-vision is the requirement that cities renewthe franchise agreement upon showing ofreasonable compliance with its provisions.Fractious renewal negotiations have alarmedthe industry. Industry members feel smalltowns and cities with older systems aremaking unreasonable demands at renewaltime for totally new 108-channel state-of-the-art systems: They feel these are notgood investments outside densely populatedmetropolises (and perhaps not even there).Broadcasters; MDS, STV; DBS, and the restface no such close controls on investment:The Senate bill; they claim; will allow allthe video providers to "compete on a levelplaying field." The legislation would, inessence, give cable operators indefinite locallicenses of the sort television stations arealso seeking.

City governments, with experience innegotiating and administering franchiseagreements, point to studies showing virtuallyno franchise renewals denied. Like the

91

advocates of continued FCC broadcastlicensing, they argue nonetheless that therenewal processes keep the companieshonest in a way the marketplace simplycannot (see 39). The National Cable Tele-vision Association (NCTA) will certainlycontinue to seek deregulation.

Government Standards Versus MarketReliance. On FCC standards, ChristopherSterling has written, "For decades, theCommission's role was clearcut: industrydeveloped potentially competitive standardsfor a given spectrum-using service and then,under FCC guidance, comparatively testedthem. Based upon the results...the FCCwould then decide which transmissionstandard best served the public interest...."(37: 138). Now; however; the introductionof competing new technologies has beenso rapid "as to defy careful policy consid-

,

eration." (37: 139). In addition, budget cutsat the FCC have severely limited its abilityto engage in the complicated process ofstandard development. And, with proliferatingmarket alternatives, many question the needfor Govemment to make detailed proclama-tions as to what is in the public interest.

The FCC seems to be moving away fromsetting standards on the assumption thatstandards decrease the speed and raise thecost of establishing an innovation in themarketplace. The FCC's recent decision notto set a standard for stereo AM radio re-flected, in part, this belief. Critics have arguedthat the effect could be just the opposite,however. Because of the inevitable periodof competition among alternatives, all butone presumably becomes obsolete: Thatdiscourages investment: Opponents alsoassert that antitrust rulings limit the abilityof private manufacturers to set their ownstandards; that large radio manufacturersand not individual consumers will make theultimate selection, and that the period ofcompetition will allow foreign manufacturersto gain a foothold (37).

Deregulation advocates counter that thereis no clear evidence for any of these claims.In addition, in some uses setting standardswould have negative effects on specificpopulation groups. Existing low-power tele-vision standards, for example, could priceLPTV stations beyond the reach of rural

92

and urban minority audiences who are sup-posed to be prime beneficiaries of the service(37: 144). Finally, the inevitable problemwith standards is that they may rigidly shutJut future technical innovations. One sug-gestion has been to enforce a standard fora few years to allow the technology to takehold, and then "sunset" the rules. If anysuperior techniques arise, they are thenallowed to participate and succeed or failin marketplace competition against the oldstandard. The debate is likely to continue,with stereo television transmission standardslooming as a major test of whether theCommission will bow to the pressures ofthose who want it to set standards to speedinnovative market entry; or will stay thederegulatory course (36; see also 38 and42: 83-86 on videotext standards). In eithercase, the relatively slow pace of FCC de-liberation contrasts sharply with the hyper-activity of communications technology. TheFCC will likely lag behind technologicaldevelopments for some time to comeyet another argument for thoroughgoingderegulation.

Content Regulation

There can be no doubt that cable has greatlyreduced scarcity of channels for its 27 millionsubscribers. But that fact in and of itselfdoes not answer the concerns of those whofavor continued "public interest" regulationof video content. Rising competition betweencable and broadcast television does notguarantee accomplishment of the originalgoals of content regulation:

"Fairness" of Telecasting. Several rulesdesigned to promote public enlightenmentregulate the content of telecasts. Amongthese the most contested is probably theFairness Doctrine (13). The doctrine stipulatesthat stations must devote time to controversialissues and must offer reasonable opportunityfor the discussion of contrasting views onthem. Many observers believe the effect ofthis rule is to discourage issue programming.Stations fear any controversial view they showwill subject them to demands for time fromopponents, or complaints to the FCC if theydeny requests for access (30; 31). Publicinterest groups counter If the FCC vigorously

X05

advocates of continued FCC broadcastlicensing, they argue nonetheless that therenewal processes keep the companieshonest in a way the marketplace simplycannot (see 39). The National Cable Tele-vision Association (NCTA) will certainlycontinue to seek deregulation.

Government Standards Versus MarketReliance. On FCC standards, ChristopherSterling has written, "For decades, theCommission's role was clearcut: industrydeveloped potentially competitive standardsfor a given spectrum-using service and then,under FCC guidance, comparatively testedthem. Based upon the results...the FCCwould then decide which transmissionstandard best served the public interest...."(37: 138). Now; however; the introductionof competing new technologies has beenso rapid "as to defy careful policy consid-

,

eration." (37: 139). In addition, budget cutsat the FCC have severely limited its abilityto engage in the complicated process ofstandard development. And, with proliferatingmarket alternatives, many question the needfor Govemment to make detailed proclama-tions as to what is in the public interest.

The FCC seems to be moving away fromsetting standards on the assumption thatstandards decrease the speed and raise thecost of establishing an innovation in themarketplace. The FCC's recent decision notto set a standard for stereo AM radio re-flected, in part, this belief. Critics have arguedthat the effect could be just the opposite,however. Because of the inevitable periodof competition among alternatives, all butone presumably becomes obsolete: Thatdiscourages investment: Opponents alsoassert that antitrust rulings limit the abilityof private manufacturers to set their ownstandards; that large radio manufacturersand not individual consumers will make theultimate selection, and that the period ofcompetition will allow foreign manufacturersto gain a foothold (37).

Deregulation advocates counter that thereis no clear evidence for any of these claims.In addition, in some uses setting standardswould have negative effects on specificpopulation groups. Existing low-power tele-vision standards, for example, could priceLPTV stations beyond the reach of rural

92

and urban minority audiences who are sup-posed to be prime beneficiaries of the service(37: 144). Finally, the inevitable problemwith standards is that they may rigidly shutJut future technical innovations. One sug-gestion has been to enforce a standard fora few years to allow the technology to takehold, and then "sunset" the rules. If anysuperior techniques arise, they are thenallowed to participate and succeed or failin marketplace competition against the oldstandard. The debate is likely to continue,with stereo television transmission standardslooming as a major test of whether theCommission will bow to the pressures ofthose who want it to set standards to speedinnovative market entry; or will stay thederegulatory course (36; see also 38 and42: 83-86 on videotext standards). In eithercase, the relatively slow pace of FCC de-liberation contrasts sharply with the hyper-activity of communications technology. TheFCC will likely lag behind technologicaldevelopments for some time to comeyet another argument for thoroughgoingderegulation.

Content Regulation

There can be no doubt that cable has greatlyreduced scarcity of channels for its 27 millionsubscribers. But that fact in and of itselfdoes not answer the concerns of those whofavor continued "public interest" regulationof video content. Rising competition betweencable and broadcast television does notguarantee accomplishment of the originalgoals of content regulation:

"Fairness" of Telecasting. Several rulesdesigned to promote public enlightenmentregulate the content of telecasts. Amongthese the most contested is probably theFairness Doctrine (13). The doctrine stipulatesthat stations must devote time to controversialissues and must offer reasonable opportunityfor the discussion of contrasting views onthem. Many observers believe the effect ofthis rule is to discourage issue programming.Stations fear any controversial view they showwill subject them to demands for time fromopponents, or complaints to the FCC if theydeny requests for access (30; 31). Publicinterest groups counter If the FCC vigorously

X05

and diversity. More important, the eco-nomic future of cable networkS is far fromguaranteed.

If, after a shakedown period, only a handkilof basic, mass-targeted, ad-supPOrted cablenetworks survive along with a few payservices, the market will be less abundantthan promised. Serving narrower interestaudiences, as cable was supposed to do,may simply turn out to be a bad investment.Kay_ KoplciVitZ. president of the ad- supportedUSA cable network, believes that "the eco-nomics are very tough for narrowcasting,and the narrower the programming thetougher it gets" (11: 60). A major economicstudy released in September 1982 predictedonly 8-10 viable ad-supported cable networksby 1985 (22).

Beyond this question is the potentialproblem of channel scarcity. Several payand basic networks started up in 1983,including a Disney channel and a countrymusic service. A substantial proportion(upwards of 50 percent) of systemsthosewith 12 or fewer channelswill not havethe capacity to carry them. Even some rela-tively new 36-channel systems are full.

Of course, more optimistic scenarios canbe constructed. And the benefits of a halfdozen or dozen new networks, even if mass-targeted, should not be gainsaid. The ques-tion remains whether the market as it is likelyto emerge actually will serve the same ends asFCC regulatibh was designed to do. Willdiverse views flower? Local programming?Political debate? What should be done ifthe market fails and if we remain skepticalabout moving back toward a more heavilyintrusive FCC (see 21)?

One option is to abandon the goals them-selves as unrealistic. Another option is torely upon the new videotext services moreattuned to information than entertainment,to multiply the access of citizens to wide-ranging information; The competitive re-sponse of the newspaper industry to videotextmay enliven them both. Indeed a reasonablescenario may be substantial competition invideo mass entertainment via brOadCaSt,cable, and the technOlOgies abOrning: andcompetition in information provision viavideotext. newspaper, magazines. and books.While less than the nirvana of competitive

abundance some envision, this system wouldoffer significantly more than the presentregulated market.

Ownership Restrictions. In the late1970s, the FCC appeared to be movingaway from content regulation but towardstricter oversight of market structure, with agreat concern to prevent anticompetitivecombinations of media entities. By 1982that tendency had all but disappeared.Deregulation in this area too is a clear goalof the Commission majority:

Current restrictions on ownership takeseveral forms: Single entities are not allowedto own more than seven television stations(a maximum of five can be VHF) (47 C.F.R.73.636(b)). Television stations cannot owncable systems or newspapers located in thesame area (with some "grandfathered"-inexceptions): nor can the TV networks owncable companies; although this restriction isone of those in process of abOlition (1:35 =36).The FCC places no limit on the number ofcable systems a single entity can own anddoes not prohibit a cable company fromcommon ownership with a newspaper. Pro-ponents of further deregulation ask, preciselywhat harm would come of a free market intelevision stations and other media? Com-pared to many industries, television stationand cable ownership are far from concen-trated. Economies of scale might well improvethe service offered _by those independentTV stations and/or cable systems absorbedinto large conglomerates (see 17).

On the other side stand those fearful ofthe concentration of political power that mightaccompany large-scale acquisition andmerging. Those who favor continued FCCregulation also point to numerous potentialabuses by cross -owned newspapers, televisionstations, and cable systems, whether in settingadvertising rates or covering local issues andcampaigns. These possibilities, they_feel, arenot necessarily susceptible to cure by com-petition from other media outlets. Regulationproponents further base their legal case onthe "Associated Press principle" (326 U.S.1 (1945))that the underlying assumptionof the First Amendment is that the Americanpeople will receive information from diverseand antagonistic sources.

107

Some complex interactions between ef-ficient investment and innovation makechoosing policy quite difficult: The establishedTV networks have all been moving vigorouslyinto the new technologies. If, as some suggest;this entrance should be foreclosed, CBS, etal., might be doomed to stagnant profits atbest, bankruptcy at worst. Keeping thenetworks out of cable, MDS, DBS, and therest would prevent them from respondingfully to the evolving market. On the otherhand, by letting them into joint ventures incable program production or direct ownershipof cable systems, the Government risksanticompetitive behavior as well as violation ofthe "Associated Press principle."

Deregulation advocates appear confidentthat the burgeoning marketplace wouldindeed prevei it transgressions, that the Anti-trust Division of the Justice Department willbe on gbard; and that private antitrust suitswill be effective enforcers of competition. Inapparent agreement; the FCC has beenmoving quite forcefully toward deregulationof ownership restrictions. Congress has notyet spoken on the issue.

Access to Cable Systems. A repre-sentative fear of those who endorse con-tinued ownership oversight is that local cablesystems may become bottlenecks that blockprogram and service providers who desireaccess to subscribers. The barriers may resultfrom vertical integration, where one companyhas an interest both in the system and inprogram services. Such integration is com-mon. American Television and Communica-tions (ATC), the largest cable system operator,for example, is owned by the same corporateparent as Home Box Office: Time Inc. WhileHBO is offered on essentially every ATCsystem, its competitors are offered on onlya few.

There are numerous tie-ins of this sortbetween program suppliers and cable sys-tems. In addition; many cable programnetworks are joint ventures between two ormore cable system operators: Such networksmay gain easy access to their parent corpora-tions' cable systems. With a continued short-

h rsage of open carmeon most systems; cableprogrammers who are not related to cableoperating corporations may be shut out.

One solution proposed for bottlenecks ismandatory leased access: setting aside aportion of each system's channels as opento the highest bidder. The cable industrystrenuously opposes such ideas: With somereason, it fears enforcement of leasingrequirements would lead to treatment ofcable as a public utility, with all of its well.known inefficiencies. Other alternatives havebeen proposed (for example; separatedsubsidiaries or compulsory arbitration). Mostbelieve that required leasing or other formsof access could be mandated only with closeovens ht of rates and leasing practices. Thecosts OT such an apparatus may outweighany benefits of unclogging the bottleneck(see 5). A preferable alternative may bereliance upon private antitrust suits, wherecourts can impose treble-damage settlementson anticompetitive firms.

Protecting the Rights of IndividualsIn the glow of enthusiasm over the new ageof multifaceted information, two central issuesof individual rights will not be overlooked.The first is privacy; the second is a potentialinformation gap.

Privacy and the Nature of TwoWayCable. The privacy problem exists in directproportion to the enormous potential oftwo-way interactive cable. Essentially unlim-ited amounts of information can be storedin computers and transmitted via cable:Videotext subscribers can send orders to thecomputer for specific itemsfrom newsreports to recipes to bcok chapters and muchmore (see 23). The same marriage of com-puter and two -way capability enables cablesystems to conduct instant polls includingthe widely publicized "Qube" polls on someWarner-Amex systems; banking and shop-ping at home; and "pay-per-view"showingsof first run movies and prize fights.

The computers for these interactive sys-tems, then, will collect detailed informationon program preferences, finances, shoppinghabits; and political opinions of subscribers.These data would be enormously valuableto merchandisers; political candidates, em-ployers; credit raters; and; conceivably, publicofficials ferreting out dissidents: Cableoperators will be unier considerable eco-

108

nomic or political pressure to sell or sharethese valuahle data. Individual cable systememployeeS may be subject to blandishmentsfrom those who want embarrassing infor-mation about a specific individual (perhapsa candidate). Legal action may be broughtto bear by police forces or parties to lawsuitsto force the release of data. As computernetworks become increasingly intertwinedand interdep, odent, the damage that mis-takes or incomplete entries might do to anindividual's credit rating or job prospectsincreasesas does the difficulty of discov-ering or erasing an error (see 42).

Government regulations already covercredit rating services, although they mightnot cover all the contingencies raised by thenew technologies. Local cable franchiSeagreements often include proviSionS forprivacy protection. Wamer-Amek, the ownerof Qube, has its own voluntary code of goOdpractices. Local oversight and voluntary selfregulation may prove sufficient.

Yet one aSpect of the dilemma is thatviolations of policy are difficult to detect.Computers will be exchanging informationWith each other and may not leave an easytrail for privacy guardians to follow. Anotherdifficulty is that the conceptual line betweenlegitimate uses of the data, say in assessingcredit reliability, and forbidden uses may notbe easy to draw.

Possible Ameliorating Factors. Video-text is in the earliest stages of development.It may never become widespread. To thedegree it is not, privacy concerns recede.Currently, the cost is quite high-410, forexample, for a half-hour'S use of theCompuServe data baSe (which includes theWashington Post and other newspaperS) (8).In England, home use haS fallen drasticallyshort of projectionS (38).

A second force that may protect privacyhe same technology that threatens it.

criber signals could contain codes that.e their identity or cause automatic

*taw, from the central computer's memory.rrw ,4 rotections could be mandated by

, 3nforced if built into the computer40.400., and software. Then the Issue will

441 au political one: which data shouldkit koto ected and which should be open

ercial exploitation? This is likely

to be an issue for some time. In some waysthe battle lines will resemble those betweenecological conservationists and developers:

An Information Gap? A second policyconcern is the possible development orwidening of an . iformation gap among thecitizenry. If cable and network television offermore entertainment choices than ever before,and videotext and other forms of two-waycable remain expensive, two clasSeS couldarise. One group would be even moreattracted to and dependent upon televisionentertainment than is already common.Another, much smaller segment would havean expanding world of information at itscall. Only for a small group who could afford itwould the competitive video marketplacebe fully implemented. For the restthoseliving in rural areas beyond the reach ofcable, or families unable to commit $50 ormore per month to videothere wouldperhaps be some more entertainment chan-nels, but visions of an efflorescing world ofcompeting ideas and alternatives wouldsimply not pertain.

This seems an undesirable outcome, butpolicy solutions are elusive. "Electronicinformation stamps" or tax credits could beoffered to the poorer classes in recognitionof the inherent right (or democratic desira-bility) of all citizens to have equitable accessto information. A universal service doctrinecould be implemented in cable as it fortelephone service, where business and urbanlong distance (in other words, middle andupper class) users paid higher rates to keeprural and urban local calling inexpensive.

But even if such policies were imple-mentedand they would be costlycitizenscould not be required to seek or use infor-mation (8). Nor could they be preventedfrom seeking gratification in entertainment.The information gap could arise whateverpolicies are attempted. Debate over thisproblem may come to replace the wellwomarguments over television?, encouragement ofviolence and debasement of reading skills(see 12; 32).

Conclusions

The video telecommunications industry isin great flux. Technologies that seemed

109

unmoving are now progressing rapidly.And assumptions and policies that onceseemed sacrosanct are now vigorouslydebatedand altered. Indeed the regula-tion of telecommunications is now changingalmost as rapidly as the technology oftelecommunications:

Though by no means certain; the mostlikely legal and regulatory scenario for therest of this decade is a shift in both thesubstantive focus and the locale of the videoregulatory action. From a concern withlicensing standards and regulation of content,the emphasis will move toward maintaininga competitive market structureand decidingjust how freely competitive the market shouldbe. In particular, the policy debates anddirectives will be in the realm of antitrust.Telecommunications players will attempt toprotect competitive positions, sometimes bylimiting competition. And they will opposeas anticompetitive those actions and actorsthey perceive as threatening their growth.As a concomitant, the scenes of greatestregulatory activity will switch from the FCCto the Congress and courtrooms. The Com-mission will be relegated to increasinglytechnical (though hardly apolitical) mattersof standard-setting;

On the question of regulation of entryand participation in the marketplace; itappears Government will remain involvedbut to a lesser degree. A cautious attitudetoward new entrants will be replaced by awarmer welcome. Setting technical standardsfor entry may remain a major FCC function.But even here the Commission has recentlyevinced a distaste for intervention.

Debate about content regulation will centerin the near term on the advisability ofoverseeing broadcast television. In the slightlylonger run, it seems quite possible that mostFCC content regulation of TV will be elim-inated de facto if not de jure, and that fewrules will be applied to cable and the emerg-ing video media.

The matter of maintaining a competitivemarketplace will probably become increas-ingly complicated; the hopes of free mar-keteers for wholly hands-off governmentnotwithstanding. For many years, the estab-lished members of the telecommunicationsindustry enjoyed a quiet, quasimonopolisticexistence, where each respected the other's

turf. Now, conflicts between telecommuni-cations companies in the legislature andcourtroom may grow more intense as poten-tial competition and stakes in the marketgrow. For example, the NCTA recentlyannounced that it may seek legislative pro-tection from the entrance of AT&T into theVideo and videotext markets (27).

individuals' rights to privacy and to infor-mation access will emerge as major concemsfor some groups. This area may generatethe most publicized controversy, the mostphilosophically puzzling implications, and theleast definitive solutions: The FCC will havelittle role here; the Congress and courts willbe the arbiters of any policy that emerges.

Government officials who make the finalpolicy decisions will be weighing complexuncertainties and tradeoffs. If they choosewisely, the outcome should be a reshapedindustry contributing substantially to arevivified American economy.

References and Bibliography

1. 'An Index in the Act on Multiple Ownership."Broadcasting (19 July 1982): 3536.

2. And Now; a Word From_Our Sponsor." Broadcast.ing (19 July 1982): 68 -70.

3. Arlen. G. "Computers_and Cable: An Ovetview"In Cable/Broadband Communications Book. vol.2. Edited by M. Hollowell. Washington; DC: Communications Press, 1980.

4. Besen, S. and Crandall. R. ''The Deregulation ofCable Television." Law and Contemporary Prob.lents. vol. 44 (Winter 1981): 77.124.

5. Besen. S. and Johnson, L. "An Economic Analysisof Leased Channel Access for Cable Television."Santa Monica; CA: Rand. forthcoming.

6. Breskin. I. "In Canada, Content Country." Mul.tichannel News/ Programming (19 July 1982)11:47.52.

7. "Cable Stats." Cablevision (27 December 1982):103.104.

8. Carey. J. "Videotex The Past as Prologue." Journalof Communication. vol. 32 (Spring 1982):80.87.

9. Cole. B. and Oettinger, M. Reluctant Regulators.Reading, MA: AddisonWesley, 1978.

10. Compaine, B., et al. Who Owns the Media?Secondedition. White Plains, NY: Knowledge IndustryPublications. 1982.

11. Disante. E. Who Will Survive ? " MultichannelNews/Programming (19 July 1982) II: 8.15,59.61.

12. Entman, R. "To Vote or Not to Vote: Media Im.pacts on Political Participation and Rep! ,enta,Hon. A paper presented at the Confer ,1Le ofthe International Association for Mass Communication Research; Paris, September 6.10. 1982.

1no 97

unmoving are now progressing rapidly.And assumptions and policies that onceseemed sacrosanct are now vigorouslydebatedand altered. Indeed the regula-tion of telecommunications is now changingalmost as rapidly as the technology oftelecommunications:

Though by no means certain; the mostlikely legal and regulatory scenario for therest of this decade is a shift in both thesubstantive tocus and the locale of the videoregulatory action. From a concern withlicensing standards and regulation of content,the emphasis will move toward maintaininga competitive market structureand decidingjust how freely competitive the market shouldbe. In particular, the policy debates anddirectives will be in the realm of antitrust.Telecommunications players will attempt toprotect competitive positions, sometimes bylimiting competition. And they will opposeas anticompetitive those actions and actorsthey perceive as threatening their growth.As a concomitant, the scenes of greatestregulatory activity will switch from the FCCto the Congress and courtrooms. The Com-mission will be relegated to increasinglytechnical (though hardly apolitical) mattersof standard-setting.

On the question of regulation of entryand participation in the marketplace; itappears Government will remain involvedbut to a lesser degree. A cautious attitudetoward new entrants will be replaced by awarmer welcome. Setting technical standardsfor entry may remain a major FCC function.But even here the Commission has recentlyevinced a distaste for intervention.

Debate about content regulation will centerin the near term on the advisability ofoverseeing broadcast television. In the slightlylonger run, it seems quite possible that mostFCC content regulation of TV will be elim-inated de facto if not de jure, and that fewrules will be applied to cable and the emerg-ing video media,

The matter of maintaining a competitivemarketplace will probably become increas-ingly complicated; the hopes of free mar-keteers for wholly hands-off governmentnotwithstanding. For many years, the estab-lished members of the telecommunicationsindustry enjoyed a quiet, quasimonopolisticexistence, where each respected the other's

turf. Now, conflicts between telecommuni-cations companies in the legislature andcourtroom may grow more intense as poten-tial competition and stakes in the marketgrow. For example, the NCTA recentlyannounced that it may seek legislative pro-tection from the entrance of AT&T into theVideo and videotext markets (27).

individuals' rights to privacy and to infor-mation access will emerge as major concemsfor some groups. This area may generatethe most publicized controversy, the mostphilosophically puzzling implications, and theleast definitive solutions: The FCC will havelittle role here; the Congress and courts willbe the aibiters of any policy that emerges.

Government officials who make the finalpolicy decisions will be weighing complexuncertainties and tradeoffs. If they choosewisely, the outcome should be a reshapedindustry contributing substantially to arevivified American economy.

References and Bibliography

1. 'An Index in the Act on Multiple Ownership."Broadcasting (19 July 1982): 3536.

2. -And_Now; a Word FrontOur Sponsor." Broadcast.ing 119 July 1982): 68=70.

3. Arlen, O. "Computers_and Cable: An Ovetview:"In Cable/Broadband Communications Book. vol.2. Edited by M. Hollowell. Washington; DC: Com-munications Press, 1980.

4. Besen, S. and Crandall, R. ''The Deregulation ofCable Television." Law and Contemporary Prob.terns. vol. 44 (Winter 1981): 77.124.

5. Besen. 5. and Johnson, L. "An Economic Analysisof Leased Channel Access for Cable Television."Santa Monica; CA: Rand. forthcoming.

6. Bre Skin. I. "ln Canada, Content Country." Mul.tichannel News,' Programming (19 July 1982)11:47.52.

7. -Cable Stats." Cableuision (27 December 1982):103.104.

8. Carey. J. "Videotex: The Past as Prologue." Journalof Communication. vol. 32 (Spring 1982):80.87.Cole. B. and Oettinger, M. Reluctant Regulators.Reading; MA: Addison.Wesley, 1978.Compaine, B., et al. Who Owns the Media?Secondedition. White Plains, NY: Knowledge IndustryPublications. 1982.

11. Disante. E. Who Will Survive ? " MultichannelNews/Programming (19 July 1982) II: 8.15,59.61.

12. Entman, R. "To Vote or Not to Vote: Media Im=pacts on Political Participation and Rep!,tion" A paper presented at the Confer ,1L e ofthe International Association for Mass Communication Research; Paris, September 6:10. 1982.

9.

10.

1097

Trends in Computers And CommunicationtThe Office of the Future

Abstract

Office automation and communication technologies have just begun to 'penetrate the U.S.market. They promise dramatic gains in productivity in the service sector of the economyand equally dramatic gains in the productivity of individual firms and agencies. The promiseof productivity gains from office automation comes at a time when the U.S. economy isshifting from a manufacturing to a service base and is experiencing an associated decline inproductivity growth rate. Because the expected promise of the new office automationtechnology has not yet been realized, several issues confront policyrnakers in both publicand private institutions. This paper describes the new technology and some of the reasonsthat have been offered for its unexpectedly slow acceptance in the marketplace, anticipatesSome of the implications the new technology has for the nature and organization of officeWork in the future, and offers some thoughts on the broader societal consequences of officeautomation. Different problems and policy issue arise, depending upon whether the newtechnology will be accepted rapidly or slowly. The paper identifies these problems anddiscusses actions that could be taken by industry, Government, educational institutions; andothers to address them.

Introduction

Context of the Problem

In the office ofthe_present, routine fUnctionssuch as payroll, billing, inventory, and 6-6counting are carried out by computer.the office of the future, an enormouslyexpanded range of functions will very oteba=bly be canied out by computer and linkedvia communications networks. in somescenarios; no "office" per se exists, officefunctions are performed by machines andhuman operators without close proximityto one another: One familiar technology inthe office of the future will be word proc-essing. In the future; documents will fre-quently be typed in one location and pro-duced automatically at another: Processorsthemselves will have substantial stand-aloneComputing capabilities and will communicateWith one another. Many printing shops willbe replaced by "reprographics" installationsusing computerized typesetters, video_ dataterminals for photo composition; and high-speed, ribii=iMpact printers. Computer outputwill be stored directly on microfilm or otherrecording media. Face-to-face meetings willgive way to teleconferencing. All internalcorrespondence will be handled electronically,

112

as will much communication with otherorganizations, and all these technologies willbe interconnected via telecommunications.

These new capabilities mean that the officeof the future will have to be organized andmanaged very differently from the present:

Hierarchical levels can be eliminated,spans of control extended, middle man-agement personnel utilized more effec-tively, better coordination introduced inresponding to changing business condi-tions, etc. Thus, the Office of the Futureconcept is not just the automated officeor the electronic office: rather; it is onein which new technologies give seniormanagement the opportunity to considerentirely new approaches as to how bestto organize, manage and control theenterprise)

Together; these technological, organiza-tional; and managerial changes promisedramatic increases in productivity in non -

manufacturing sectors of the U.S. economy.They Will occur as the economy undergoesPrOfoUnd changes that began early in the20th century.

During this century, the United Statesmarked its passage into the postindustrialera. In 1900, only about 25 percent of the

99

U.S. labor force was composed of white-collarand service workers, with the bulk of workersengaged in blue- collar (35 percent) and farm(40 percent) employment. By 1980, morethan 60 percent of the labor force wasworking in white-collar and service jobs, withonly a decreasing minority of workers em-ployed as blue-collar and farm workers. Asthe white-collar workforce has grown duringthe past 30 years, particularly rapid expansionhas been occurring in three job categories:professional and technical workers, clericalworkers, and managers and administrators :2These trends are expected to continuethrough the 1980s and beyond:

A central feature of the contemporaryemployment setting is the emergence ofinformation handling as the major taskamong workers:

Gradually and almost imperceptibly, theU.S. economy since 1940 has been en-gaged in a transformation that is uniquein the history of mankind: By the mid-1950s our working population was pre-dominantly engaged in informationhandling: more people were involved inthe manipulation of information thanwere employed in mining, growing crops,raising cattle, manufacturing goods, orproviding personal services. The "infor-mation society" became in fact a properexpression of predominant societalcharacteristics.3

Those whose work contributes to the infor-mation economy include professional andtechnical workers, managers and admin-istrators (the two classifications together aresometimes labelled "knowledge workers:"despite the fact that managers "handle" mostinformation orally); and clerical workers whosupport them: As the passage into thepostindustrial era becomes more complete,the number of knowledge workers is expectedto grow much more rapidly than the work-force average. Despite the massive intro-duction of electronic data processing equip:ment, some observers argue that clericalworkers will not be displaced by computersand that many clerical jobs will evolve intodifferent kinds of support positions in closersymbiosis with knowledge workers. Othersargue that the new computer technologies

100

will eliminate many clerical jobs (held pri-marily by women) and create positions payingonly the minimum wage!'

Policy A0ects of the ProblemA key problem in the information economyis how to identify, obtain; and manipulateneeded information effectively and efficiently.The explosive growth of the informationprocessing industry since 1950, which ac-companied the labor force shifts just de-scribed, created expectations of swift andeasy solutions to this problem and of con-comitant increases in productivity in thesegrowing sectors of the economy. However,the expectations have been largely unfulfilled,with U.S. productivity growth actually de-clining. Indeed. U.S. labor productivity forall industries is increasing more slowly thanin most other industrial nations, and growthin labor productivity in the private sectorpeaked in the mid-1960s.5 Ever more power-ful information processing technology stillappears to offer promise of greatly increasedproductivity growth in the postindustrialUnited States, but current trends suggestthat there is, and will continue to be forsome time, a gap between expectations andreality. As Strassmann puts it:

Clearly there is something amiss if divert-ing workers from industries with highproductivity and effective use of capitalsuch as agriculture, mining, manufac-turing, and utilitiesinto overhead jobsin business and government fails to in-crease aggregate economic performanceas measured by accepted economicindicators!'

This paper describes the new computerand communication technologies thatpromise greatly enhanced productivity inwhite-collar and service work, examines someof the factors that facilitate and inhibit theintroduction of these new technologies intothe workplace, and suggests some actionsthat might be taken by industry, users, andGovernment to overcome barriers to realizingmore completely the potential of the newtechnologies. In addressing these topics; abnef overview of the technological potentialof devices and systems (for example; micro-

1 1 3

computers, teleconferencing, electronicmail) is provided; the question of whetherintroduction of the new technologies istechnology-driven or user-driven is examined;some of the major implications of the newtechnologies for human-machine interaction,the organization of work, and human inter-action are identified; and some of the broadersocietal implications of the new technologiesare suggested. Although the scope of thepaper includes communications technology,the focus is on office automation, particularlyon computers.

Trends and Developments

Approaches to Analyzing White-Collar Productivity

In the United States in 1980, more thanhalf of those employed worked in an office.There are many organizations, such as banksand insurance companies, where almosteverybody is an "office" worker. In manu-facturing, retailing, mining, and other basicindustries, office workers constitute less than50 percent cf the workforce. Nevertheless,even in these industries the percentage ofoffice workers is increasing, and the intro-duction of robotics may cause the percentageto soar as production workers are displacedby automation. As the proportion of staffdevoted to information exchange grows, pro-ductivity gains will have to come from thebulk of the people who work, the office staff.

According to one estimate, 60 percent ofthe $1.3 trillion paid for wages, salaries, andbenefits in the United States in 1980 wentto office workers.' The enormous proportionand amount of money paid to office workerswill be a substantial, increasing incentive forindustry to substitute capital for labor byautomating the office. Furthermore, thecapital to labor cost ratio in offices is currentlyestimated to be between onetenth and one-twentieth of that in highly productive man-ufacturing industries, with the cost of officeautomation equipment actually falling. Giventhis it seems that introducing informationprocessing equipment into the office wouldproduce dramatic productivity gains. Man-

114

ufacturers argue that business should find itvirtually irresistible to substitute machinepower for human labor: Nevertheless, thelink between increased use of computersand communication equipment in the officeand enhanced office productivity is not asdirect as these arguments suggest:

One reason the linkage is; at best. indirect isthat the appropriate level of aggregationthat is, the unit of outputby which whitecollar productivity should be measured isnot clear. Productivity changes measuredat the national level tell us little about whichsectors of the economy are experiencingchanging productivity growth. Differencesin the nature of sectoral outputs makecomparisons of productivity between evenlarge sectors of the economy such as man-ufacturing and Services difficult to interpretand possibly highly misleading. At the otherextreme, productivity change measured atthe individual level can also mislead, espe-cially when new technology alters the struc-ture of the work, thereby influencing grouprather than individual productivity. Inter-mediate levels of aggregation such as industryor "office" (workgroup with specific mission)probably are the least misleading for purposesof estimating the potential and actual effectsof new technology on productivity growth:Unfortunately; efforts to develop these kindsof data for the service industries are at arelatively primitive stage. Clearer linkagesbetween productivity gains and office auto-mation must await improvement in data;especially in service sector output meas-ures at intermediate (workgroup) levelsof aggregation!'

Attainment of Gains in White-CollarProductivity

Demand for computer and telecommun-ications equipment for office applicationshas developed more slowly than expected.9The office automation "revolution" is lookingmore like a gradual transition that willprobably not reach fruition until the end ofthis century. Vendors of office automationtechnology currently are seeking explanationsfor the slow pace of acceptance.

Several reasons have been suggested foroffice automation's gradual penetration of

101

computers, teleconferencing, electronicmail) is provided; the question of whetherintroduction of the new technologies istechnology-driven or user-driven is examined;some of the major implications of the newtechnologies for human-machine interaction,the organization of work, and human inter-action are identified; and some of the broadersocietal implications of the new technologiesare suggested. Although the scope of thepaper includes communications technology,the focus is on office automation, particularlyon computers.

Trends and Developments

Approaches to Analyzing White-Collar Productivity

In the United States in 1980, more thanhalf of those employed worked in an office.There are many organizations, such as banksand insurance companies, where almosteverybody is an "office" worker. In manu-facturing, retailing, mining, and other basicindustries, office workers constitute less than50 percent of the workforce. Nevertheless,even in these industries the percentage ofoffice workers is increasing, and the intro-duction of robotics may cause the percentageto soar as production workers are displacedby automation. As the proportion of staffdevoted to information exchange grows, pro-ductivity gains will have to come from thebulk of the people who work, the office staff.

According to one estimate, 60 percent ofthe $1.3 trillionpaid for wages, salaries, andbenefits in the United States in 1980 wentto office workers.' The enormous proportionand amount of money paid to office workerswill be a substantial, increasing incentive forindustry to substitute capital for labor byautomating the office. Furthermore, thecapital to labor cost ratio in offices is currentlyestimated to be between one-tenth and one-twentieth of that in highly productive man-ufacturing industries, with the cost of officeautomation equipment actually falling. Giventhis it seems that introducing informationprocessing equipment into the office wouldproduce dramatic productivity gains. Man-

114

ufacturers argue that business should find itvirtually irresistible to substitute machinepower for human labor: Nevertheless; thelink between increased use of computersand communication equipment in the officeand enhanced office productivity is not asdirect as these arguments suggest:

One reason the linkage is; at best; indirect isthat the appropriate level of aggregationthat is, the unit of outputby which whitecollar productivity should be measured isnot clear. Productivity changes measuredat the national level tell us little about whichsectors of the economy are experiencingchanging productivity growth. Differencesin the nature of sectoral outputs makecomparisons of productivity between evenlarge sectors of the economy such as man-ufacturing and services difficult to interpretand possibly highly misleading. At the otherextreme, productivity change measured atthe individual level can also mislead, espe-cially when new technology alters the struc-ture of the work, thereby influencing grouprather than individual productivity. Inter-mediate levels of aggregation such as industryor "office" (workgroup with specific mission)probably are the least misleading for purposesof estimating the potential and actual effectsof new technology on productivity growth:Unfortunately; efforts to develop these kindsof data for the service industries are at arelatively primitive stage. Clearer linkagesbetween productivity gains and office auto-mation must await improvement in data;especially in service sector output meas-ures at intermediate (workgroup) levelsof aggregation!'

Attainment of Gains in White-CollarProductivity

Demand for computer and telecommun-ications equipment for office applicationshas developed more slowly than expected.9The office automation "revolution" is lookingmore like a gradual transition that willprobably not reach fruition until the end ofthis century. Vendors of office automationtechnology currently are seeking explanationsfor the slow pace of acceptance.

Several reasons have been suggested foroffice automation's gradual penetration of

101

in electronically supported work stationsare not practical in the 1980s and perhapsnot by the year 2000....As our knowledgeworkers devote more and more time tounstructured communications associatedwith the management of change insteadof to well-defined computational or pro-cedural tasks; the difficulty of changingtasks from manual to electronic processingwill escalate....All this will probably happengradually in the next 5 to 15 years; whenthe current concentration on word proc-essing. text processing, and distributedcomputing will run its course. At thatpoint, the era of the personal work stationas the principal means for interorgani-zational communications for the majoLtyof white=collar WorkerS Will be poSSible.'3

The Technology

The progress of computer technology is im-pressive. It is awesome to !take that somehandheld calculators in use today have moreprocessing power than many computers inuse in the 1960s. There are extremely'ver-satile and powerful central processing units(CPUs) in widespread use today that occupya space about the size of a stick of chewinggum. This same CPU power would haveoccupied a closetful of space no more than15 years ago: The price-performance ratioof memory; including permanent memory;volatile memory; and direct access storagedevices such as disks; has fallen at an evenfaster pace than the CPU, with a much moredramatic decrease in space requirements.Although this technology is the key buildingblock for office automation, it is not essentialto understand the hardware in bider toexamine trends in office automation. Weneed only assume that computer hardwarewill get ever faster, cheaper, and moreversatilea fairly safe assumption for theforeseeable future.

It is important, however, to understandthe function of this technology in the officeenvironment. Fortunately, several investi-gators have attempted to classify the use ofcomputer systems in businesses, principallyin the context of management informationand decision support systems." By consid-ering these earlier classifications and giventhe purpose and context of office automa-

11 6

tion, we offer a typology of computerizedoffice system functions as a means to thisunderstanding:

(1) Transactions. At this level, fixed inputsare processed according to a determinatescheme to yield a highly structured set ofoutputs. This may be an order entry system,a payroll system, a bank card system, oreven a word processing system. Transactions,in general; is a data-oriented function.

(2) Analysis. An analysis system usescomputational; statistical, or more generally,logical and mathematical techniques tomanipulate an existing data base; A typicalanalysis system might use multiple regressionto explore trends in data or analys;s ofvariance to examine the validity of hypoth-eses. Other analysis systems in daily useallow searching of accounts for overduepayments_ or examination of perpetual in-ventories for economic order points. As withtransactions, this is a data-oriented function.

(3) Projection. Projection systems maypermit the user to evaluate the consequencesof planned actions, add data into the systembased on personal judgment, or search forsolutions with certain constraints. That is,the user can explore the variety of outcomesthrough "what if?" types of entries, whichare not necessarily tied to any data orinformation preexisting in the system. Manyof the popular profit-analysis models areprojection systems; most material require-ment planning programs are of this typealso Interestingly, microcomputer programssuch as the popular VISICALe and itsfunctional equivalents are often employedin this mode.

(4) Communication. The final type offunction of a computer system is that ofcommunication. These types of systems actas smart conduits among individuals withaccess to the same compute:, timesharesystem, network, or storage media. Functionsof communication systems include mail,messages, calendars, and data storage andretrieval in general.

Given this typology, the automation ofthe office can be seen not as the imple-mentation of a monolithic technology, butas the introduction of several technologicallybased functions in succession. Although farfrom exhaustive, the following table providessome illustrative applications:

103

Table 1

A Typology of Computerized OfficeSystem Functions

Type Function User

TRANS- PayrollACTIONS Invoicing and

Accounts ReceivableInventory andAccounts Payable

Order Entry (Key-punched)

Word ProcessingPerpetual InventoryPoint of Order Entrywith Confirmation ofStock Available

Staff

ClencalStaffSales

ANALYSIS Goods Ordered StaffMost Frequently

Best WarehouseLocation

Warehouse Routing

PROJEC- ProfitSimplified ManagersT1ON Model

ProfitInteractiveMade'

Extension of Credit Staffto Customers Staff

Orders to Stock

COMMUNI-CATION

Interoffice Messages AllCalendarManagement

During the early years of computers, fromthe first designs up to about 1974, themainframe was all there was in the world ofbusiness. This was a big computer, whichtook up most of a floor in a typical officebuilding and cost millions of dollars. Net toemerge was the minicomputer. At first thismachine had a CPU with only a fraction ofthe power of the mainframe. However,integrated circuits advanced so rapidly thatmany minicomputers soon emerged withfar more power than some of the mainframes.Then, in the late seventies, very large scaleintegrated circuits became economicallyfeasible. This spurred the popular acceptanceOf the microcomputer. The microcomputeroriginally had a very limited memory; perhapsa 16;384 "word" capacity; and minimalprocessing power: However; in the last fewyears; even the microcomputer has become abehemoth in terms of memory and power.Today; some have millions of "words" ofstorage, hard disks, and processing power

104

that equals or even surpasses some of themodem mainframes. Along some significantdimensions, the distinction between themainframe, the minicomputer, and themicrocomputer is blurred. Thus, the personalcomputer user, despite using a "simple"computer, can be very sophisticated: Awholesale infusion of today's personal com-puter users into the ranks of managementmay alter the office of the future more thanany new hardware or software by itself:

Human-Machine Interaction

Symbiosis between the knowledge workerand his/her electronic work stationhuman-machine interactionis one of the essentialingredients of the office of the future. Thedevelopment of better hardware at ever-decreasing costs for automated office systemsis expected to continue at its inexorable pacefor the foreseeable future, but the paralleldevelopment of adequate software is pro-ceeding somewhat less rapidly:

The fruits of the electronics revolution aresuch that we can afford to do anythingwe wantif we can only program it.Software is the dominant challenge inoffice automation....Software will be themajor development cost element for theforeseeable future.'5

In particular, the development of techniquesto interface computers and their peripheraldevices into a human environment lags farbehind other aspects of automation. Thispresents significant problems in effective officeautomation and may well develop into themajor bottleneck in the technology's imple-mentation. Four underlying trends are de-scribed in the following paragraphs.

(1) Hardware ma nu fact uring costs aredeclining, but software development costsare rising. Increased manpower costs andincreased complexity of software systemshave on balance, more than offset advancesin effective programming tools. Recentadvances in programming environments,structure editors, software engineering, andpersonal computing work stations may.however, mitigate the trend toward increasingsoftware cost.

(2) The ever-expanding number of corn-puter users is increasing at a much faster

11 7

Table 1

A Typology of Computerized OfficeSystem Functions

Type Function User

TRANS- PayrollACTIONS Invoicing and

Accounts ReceivableInventory andAccounts Payable

Order Entry (Key-punched)

Word ProcessingPerpetual InventoryPoint of Order Entrywith Confirmation ofStack Available

Staff

ClericalStaffSales

ANALYSIS Goods Ordered StaffMost Frequently

Best WarehouseLocation

Warehouse Routing

PROJEC- ProfitSimplified ManagersTION Model

ProfitInteractiveMOdel

Extension of Credit Staffto Customers Staff

Orders to Stock

COMMUNI- Interoffice Messages AllCATION Calendar

Management

During the early years of computers, fromthe first designs up to about 1974, themainframe was all there was in the world ofbusiness. This was a big computer, whichtook up most of a floor in a typical officebbilding and cost millions of dollars. Next toemerge was the minicomputer. At first thismachine had a CPU with only a fraction ofthe power of the mainframe. However,integrated circuits advanced so rapidly thatmany minicomputers soon emerged withfar more power than some of the mainframes:Then, in the late seventies, very large scaleintegrated circuits became economicallyfeasible. This spurred the popular acceptanceOf the microcomputer. The microcomputeroriginally had a very limited memory; perhapsa 16,384 "word" capacity; and minimalprocessing power: However; in the last fewyears, even the microcomputer has become abehemoth in terms of memory and power.Today; some have millions of "words" ofstorage; hard disks, and processing power

104

that equals or even surpasses some of themodem mainframes. Along some significantdimensions, the distinction between themainframe, the minicomputer, and themicrocomputer is blurred. Thus, the personalcomputer user, despite using a "simple"computer, can be very sophisticated: Awholesale infusion of today's personal com-puter users into the ranks of managementmay alter the office of the future more thanany new hardware or software by itself.

Human-Machine Interaction

Symbiosis between the knowledge workerand his/her electronic work stationhuman-machine interactionis one of the essentialingredients of the office of the future. Thedevelopment of better hardware at ever-decreasing costs for automated office systemsis expected to continue at its inexorable pacefor the foreseeable future, but the paralleldevelopment of adequate software is pro-ceeding somewhat less rapidly:

The fruits of the electronics revolution aresuch that we can afford to do anythingwe wantif we can only program itSoftware is the dominant challenge inoffice automation....Software will be themajor development cost element for theforeseeable future.'5

In particular, the development of techniquesto interface computers and their peripheraldevices into a human environment lags farbehind other aspects of automation. Thispresents significant problems in effective officeautomation and may well develop into themajor bottleneck in the technology's imple-mentation. Four underlying trends are de-scribed in the following paragraphs.

(1) Hardware manufacturing costs aredeclining; but software development costsare rising. Increased manpower costs andincreased complexity of software systemshave on balance, more than offset advancesin effective programming tools. Recentadvances in programming environments,structure editors, software engineering, andpersonal computing work stations may,however, mitigate the trend toward increasingsoftware cost.

(2) The ever-expanding number of com-puter users is increasing at a much faster

11 7

Furthermore; mature users can be quicklyfrustrated by too much "help."

(4) Well-designed; stylized command inter-faces require user training but could bedesigned to achieve high functionality withlow complexity. These seem to be best formature or frequent users.

Interfaces are also required for tech-nologically complex devices that are notcomputerbased work stations. However, thetype of interface required depends stronglyon the particular properties of each device.Uniformity of operating conventions acrossall softWare and hardware systems is clearlya highly desirable objective.

User Training Techniques. Thoughtneeds to be given to methods for trainingusers of new systems. Some casual users ofcomputer systems whose time is a scarceresource (such as managers and scientists)will require interfaces that make no priortraining demands. This clearly limits the scopeof their ability to exploit new computertechnology directly: For other users; a com-bination of forgiving; friendly interfaces andshort; handson training periods is likely toprove much more cost-effective than eithertraining all users to become computer expertsor inventing and providing the ultimate.foolproof, natural (and as yet mythical)interface. However, little effort has gone intoexamining the best methods for trainingcomputer-naive people to become effectivecasual users of automated office systems.Although the tradeoff betWeen laboriouslydeveloped. friendly interfaces and usertraining has not been quantitatively analyzed,the expected size of the user population, itsfrequency of interaction with the system,and the expected system lifetime appear tobe determining factors of cost-effectiveness:How this tradeoff might be evaluated andwho will pay the costs under each alternativeare unresolved questions that merit investi-gation. Another dimension of the trainingissue involves consideration of a widelydistributed training capability; one that in-coiporates the enormous variety of officesituations and; thus; of training needs.

Human Interaction in theOrganizational ContextWe do not yet know how to restructure officework to take advantage of office automation

106

technology. Indeed, variations in the natureof offices and of office work across and withinindustries virtually guarantee that the desiredknowledge will appear piecemeal, made upof a series of contingent statements ratherthan a single. grand "theory." But the outlineSof some of the consequences of the newtechnology for human interaction in theworkplace are beginning to emerge. Astechnology provides more and more func-tions, knowledge workers (particularly man-agers) will be using terminals. This, forexample, will alter the role of the secretaryas the manager substitutes the computerfor tasks previously reserved for clericalworkers. The professional may. in fact, c!ohis/her own "typing" at the keyboarddrafting and editing a document; final hardcopy production becomes trivial. Alternatively:as more office automation is implemented;systems frequently are put in place undersecretarial control. This; in itself; is an im-mediate impetus for changing the secretarialrole and career path to management orcreating; in the office context; a new inter-mediate category of paraprofessionals. Thebest analogy is in the legal field, where manyparaprofessionals were secretaries. In theircurrent jobs they use the same equipmentand technology that they used previouslyas secretaries for routine legal work. Peoplewho arc willing to become paraprofessionalSare themselves in a more professionallyoriented role.

As a larger proportion of our populationuses a microcomputer at home or plays videogames at the tavern, more will be willing touse a terminal at the office. These peoplemay be very impatient if they are not providedwith the capabilities c), an office system tocreate text, do electronic filing, handlecalendars, and receive electronic mail. Wehave become accustomed to an officesystem staffed by workers who are slow tochange, but in the near future this may notbe the case:

The office of the future may significantlyalter the role of time in work. For instance.given the present physical document distri-bution system, nobody really knows when aparticular individual gets a document unless itis hand carried with return receipt requested.This is relatively rare. Even within the internalmail system it takes a day or two to receive

119

a document. In a multilocation organization, itoften takes a week to be sure that everyplant; warehouse, and sales office has re-ceived a particular document. Today peoplefeel they need time to think about theirresponse; some "buy time" by saying a letter isin the mail even before it has been prepared.Just as there is a certain amount of "float"in the present banking system; a "float" existsin the physical mail system. Of course,reducing transmission or transaction timealone may have little effect if dec-isionmakingcapacity remains untouched.

As electronic filing and electronic distri-bution are implemented, management willknow that information has been receivedby all of their staff. The electronic mail systemshave the capability of notifying the senderthat the document has been received. This,there will be a much higher level of expecta-tion that the response will be returned within aspecified time, such as 24 or 48 hours. Thiswill be a much smaller "window" than iscurrently possible with physical mail. Havingto respond more quickly can increase thepressure on individuals and might also resultin a less carefully considered response. Doingthings more quickly is not necessarily doingthem better; yet speed rather than effective-ness of information may be how managersevaluate their staff.

On the plus side, time no longer affectscommunication when electronic mail andmeSSage systems are in place. A telephoneconversation requires the simultaneousavailability of at least two parties. Electronicmessages can be sent and received withOutregard to such constraints. In addition, thenew technology allows periodic reports andrecords to be updated until the laSt minuteand produced at the punch of a button.This can reduce time pressure on analystsand writers, who can place more emphasison the quality of their products. Conversely, itcould allow them to turn out a sloppy initialproduct for repeated revision.

Another consequence may be a reductionin day-to-day human interaction. individualswho can retrieve information from the systemelectronically do not have to ask their secre-tary or their staff to get the information forthem. It is often in the very process of askingfor information that management spends afew minutes discussing personal concerns

120

and family activities. This is an importantpart of the human interaction that makeswork a pleasant experience. If office workersinteract with machines more exclusively, theoffice may assume a more sterile or factory-like atmosphere.

The' advent of office automation, however,will not necessarily lead to hermit officeworkers, communicating only via their com-puter terminals. Such has not been the case atuniversities and industrial research environ-ments where electronic mail systems havebeen functional for many years. In fact,electronic mail displaces interoffice mem-oranda more than it displaces face-to-facediscussion or telephone conversation.Whether this will remain true as officeautomation spreads to other settings andas the restructuring of office work to en-hance productivity growth proceeds remainsto be seen.

The flekibility and freedom of the officeof the future lead to another significantproblem managers will face. How do youmonitor what people are doing, especially ifthey are not doing it in a specified place ata specific time? In the past, managers werecognizant of almost all of their white-collaremployees' operations. In an automatedenvironment, lack of proximity and simul-taneity of work may create new managementtensions and dilemmas.

Anticipated SecondaryConsequences

The so-called office in the home can offerstriking opportunities or create a centralproblem in the technological age. Decreesingcosts of terminals and communication arethe key factors. As costs drop, the reach ofthe office will indeed extend into an indi-vidual's home. This will alter the workstylefor executives, as they would not be con-strained by time when dealing with worldwideoperations but would have to rethink howand when they we personal time.

An office in the home can represent almostthe ultimate in energy savings as far ascommunication and transportation are con-cerned: Individuals will certainly get used tohaving a greater fund of information availableto them through such information servicesas Teletext and Videotex.16 Additionally, all

107

of the office files can be accessed from ahome terminal. Today people are able touse their telephone at home just as if theywere in the office, and extensive telecon-ferencing is not far off. Increasingly, peoplemay question the need to go to the office.

There are some people, of course, whowant to get away from their homes. Thetradition of mobility among nuclear familycommuters may be resisted, even if homesare comfortable to work in and the samelevel of information and ability to com-municate with people exists. Clearly; therewill be intermediate steps between the officeas we know it today and the office of thehome: Such organizations as Satellite Busi-ness Systems and Bell Telephone are pro-posing to set up video conferencing centerswhere people would travel to some centralsite to take advantage of the facilities. Theywould not have to travel to another city toconduct a conference. Value added networksare being set up as well by a number oforganizations to provide the facilities ofelectronic mail. These include GT&F's rele-main System, IT&T's Faxpack System,Tyrnshare, and Tymnet. Remote work centersalso may emerge as intermediate arrange-ments between the office-at-home and asingle, centralized location. The alternativesare clear. but the actual geographic distri-bution of office work in the future is not;

Energy costs will also affect the use ofpaper within the office; Paper costs aredirectly related to the energy involved inchanging wood to pulp and then drying thatpulp to make paper. So the cost of paperescalates almost as fast as the cost of energy.The office of the future can minimize theuse of paper with displays and magneticstorage. Unnecessary printing can be avoidedby revising and distributing informationelectronically and printing it only whenit is needed.

In the past, management felt that officestaff members, particularly principals, werenot hourly paid workers. If a specific problemor project were being handled, people wereexpected to come in early and stay late untilthey could solve the problem. When workingat home becomes a reality, there will be awhole new level of flexibility introduced intothe office system. People could work at anytime they wanted to, not necessarily in eight

108

or nine hour shifts. Under these conditions,only the product will matter, not how itwas produced.

Policy OptionsThe potential that technology holds forincreased efficiency in information handling inthe office setting has been documented, andthe technology that promises to realize thispotential is being developed. In a few in-stances it has been introduced on an expen-mental basis. Yet market demand for thenew information processing and communi-cation technology is developing slowly, moreslowly than might be expected given theintensity with which the need is expressed,even accounting for the effects of a sluggisheconomy and surplus of labor. Realizationof the benefits of a new technology at firmand industry levels (in terms of increasedrates of productivity growth) requires trans-lation of need into demand. In this paper;different perspectives that implicitly placethe burden for action at different doorstepshave been described: In this concludingsection, different perspectives are explicitlylinked with different institutional actors; andsome of the actions that might be taken tofacilitate the introduction and use of newinformation handling technologies into officesettings; assuming that such facilitationis a policy goal of the actor involved, arediscussed.

One perspective currently focuses attentionon problems associated with implementingthe new technology. This is the process duringwhich it is introduced into an organizationand, over time, becomes part of organiza-tional routine. To some, the pace with whichthe new office technology will penetrate themarket is a function of how rapidly users,particularly managers, become acclimatedto it, begin to recognize its benefits first-hand;overcome their technophobia, and adjusttheir work habits. This perspective suggestsaction by the marketing divisions of industryto develop improved indicators of the tech-nology's benefits to the firm; to offer trainingprograms and seminars to potential cus-tomers (particularly management), and toprovide analytical support to firms and otherorganizations as they restructure work flowsand responsibilities to accommodate the new

121

technology: These strategies largely acceptthe technology A S given and focus on useradaptation at individual and organizationallevels.

Another perspective views the problemin terms of human-machine interaction. Morerapid acceptance of the technology dependsupon how completely and rapidly hardwareand software can be developed to supportmore "user-friendly" information-handlingprotocols. Applied research activities nowbeing conducted in industry and universitieson such topics as faulttolerant softwaresystems, natural language codes, and voicerecognition input devices are directed towardreducing this barrier to more rapid intro-duction of the technology. In the past, thefocus on user friendliness largely acceptedthe individual user as relatively fixed andsought adaptations in the technology, gen-erally through the software interface, to theuser's skills, knowludge, and habits. We nowrecognize that as the user becomes moresophisticated; the user-friendly aspects ofthe system must adapt accordingly:

A third perspective emphasizes the organi-zational conditions that must exist beforethe new technology will be implementedextensively. This perspective arises less fromthe need for the organization to adapt tothe technology than from the view that large,complex organizations must make funda-mental changes in the way they manageinformation before the benefits of automatedinformation handling can be realized. Verylittle is known about how this restructuringshould occur, especially across the varietyof different industries (including Government)in which information handling has becomea major expenditure. A significant role forresearch therefore emerges, perhaps underindustrial or joint Government-industrysponsorship.

The policy implications of office automa-tion depend upon the pace with which thenew technology is accepted in the market-place: If the pace is rapid; then productivitygains (with salutary consequences for thenational economy) will be evident: Thesewill mitigate or perhaps reverse the declinein the Nation's rate of productivity growth,despite the fact that manufacturing con-tributes less and less to the gross nationalproduct.

122

Should this occur, however, a number ofproblems may emerge that may warrantattention by Government, industry, andeducational institutions. As noted earlier,agreement does not exist on the conse-quences of office automation for officeworkers. If change is rapid, substantial unem-ployment of workers with one set of skills(for example, clerical) could exist simultane-ously with unmet demand for workers withdifferent skills (for example, analytical).Government may be called upon to easethe personal costs of worker displacement,perhaps by extending unemployment com-pensation to certain classes of workers. Newjob training programs that address the needsof both the unemployed and the agenciesand firms seeking persons with new skillsmay be needed. To meet the demand forpersons to work in the office of the future,equipment vendors may have to supplytraining programs as an integral part of thetechnology they sell. Alternatively; users ofoffice automation equipment may bear theburden of training costs. In any event, inthis fast-paced scenario, industry (bothvendors and users) would have an incentiveto supply training opportunities so that bothsales potential and productivity growthpotential can be realized. Government wouldwork to retrain workers who could not easilymeet new job requirements.

If the pace of acceptance is slow (at leastinitially), another set of policy issues arises.We may assume that relevant institutionssuch as individual firms, industry and tradeassociations, vo-tech schools, and publicschools will be able to adjust quickly enoughto keep labor supply and demand balanced.But costs will accrue to the Nation as awhole if productivity gains do not occur asrapidly as expected or desired. The publicpolicy question then becomes: What steps,if any; should Government take to facilitatethe spread of office automation? Also; whatactions by other institutions could speed thepace of acceptance, thereby realizing ex-pected benefits to both vendors and usersof the new technology?

Again, different diagnoses of the problemof slow market penetration have differentimplications for action. Currently, imple-mentation costs (both monetary and non-monetary) are seen by many as a major

109

!iese costs could be reduced throughAN014W :dence of how the new technologyAmmo,' !fit the firm and by _better under:

j of the several behavioral and,,,,;4!!!! tional factors that impede accept:

ey could also be reduced if vendors" .g to bear_the costs of training and

'Aug support to firms desinng to restaff".40.44' ucture themselves to take advantage

!chnology. Vendor firms currently!acting research that seeks to improve

" ierstanding of the implementation4114, '" Government could provide addi-

44,'""*. sport for this activity by encouraging9-industry cooperative research cm>ct; Industrial and trade associations,:,elop and offer continuing educations for clerical workers interested in>ffice automation to their skills; and

`1****"4-1 ,gers interested in planning for office'''"*" ion or actually using the new tech-

lemselves; Virtually anything industryJ reduce "computer naivete" amongcommunity would facilitate imple-

.4.,, on of the technology. Seminars;programs; handboOkS, and hands-onrations would be useful contributions.

allot I

^

are other research programs that-lent could support to facilitateduction of office automation. Private

unlikely to undertake such workthey could not capture sufficient

to make it worthwhile. Government)onsor research that would identifyye structural forms that information4 could take in different industries,es of Government agencies andheir consequences for enhancedvity. Such research programs shouldwork on definitions and measuresI activity for white-collar workers,y knowledge workers and managers.on of the Government's own ex-?S with office automation might alsoelpful, as would evaluation of state-in installations in industry and uni-omputer science departments. Gov-

, t could study needs for information,,,,04, Ts and support personnel and pro-44.,. 2 data to Mdustry and educational

ins. Finally, the Federal Government,040. , apport the development of software

-is for office automation that would460.0, provisions for collecting data relevant

to information system management research.Regardless of the pace with which office

automation proceeds, educational institutionsmust move to ensure computer literacy, justas they now address reading, writing, andarithmetic. But there is little agreement onhow to address this problem, and there arefew resources for experimentation. Schoolsfrom the elementary level through the uni-versity are not succeeding uniformly orsatisfactorily, and successes achieved in oneschool are not communicated effectively toothers or easily adapted by them. Thus, thefirst priority for all schools should be todetermine reasonable objectives for computerliteracy and support their attainment withsubstantial resources: In addition to offeringnew and modified curricula; universities andtechnical schools should develop continuingeducation courses, perhaps in conjunctionwith industry, directed toward informationmanagers and support personnel. Whetherthe office of the future becomes pervasiveM 15 years or in 30, the next generationshould be computer literate to take fulladvantage of the benefits of the new officetechnology and to mitigate some of itspossible costs.

Notes and References

1. Connell; J. J._"The Office of the Future. Journalof Systems Management. vol. 30 (February 1979):6.10.

2. U.S. Bureau of LabOr Statistics. Employment andEarnings. 1980.

3. Strassmann, Paul. "The Office of the Future: In-formation Management for the New Age." Tech.no/ogy Review. vol. 82. no. 3 (December 1979/January 1980).55.56.

4; "Worry Grows Over Upheaval as TechnologyReshapes Jobs." New York Times (4 July 1982).

5: Abernathy; William and Rosenbloom; Richard S."The Institutional Climate for Innovation in In-dustry: The Role of Management Attitudes andPractices." In National Science Foundation. The 5-Year Outlook on Science and Technology. 1981.NSF 81.42. Washington, DC: U.S. GovemmentPrinting Office, 1982: pp._407-420.

6. Strassmann. op. cit. p. 56. "Accepted economicindicators" may not be valid measures of white.collar productivity. See following text.

7. Uttal. Bro. "What's Detaining the Office of theFuture." Fortune (3 May 19821: 176.

8. Ginsberg. E.. "The Mechanizatioi. of Work." Scientific American (September 19827-75.

9. Uttal. op. cit.

123

lese costs could be reduced through4410114* :dence of how the new technology*bobs !fit the firm and by better under:'!,-* j of the several behavioral and

tional factors that impede accept-ey could also be reduced if vendors. " .g to bear the costs of, training and

414 4441' support to firms desinng to restaff"404" ucture themselves to take advantage4444' 4" ?chnology. Vendor firms currently4"1414 i !acting research that seeks to improve

-4!'"" ierstanding of the implementation""' 4'" Government could provide addi-

41. Sport for this activity by encouragingy-industry cooperative research on>ct. Industrial and trade associations,:,elop and offer continuing education

11* ,s for clerical workers interested in4"-414" >>ffice automation to their skills; and""4"4-1 ,gers interested in planning for office

'''"*" ion or actually using the new tech-lemselves. Virtually anything industryJ reduce "computer naivete" among

.44 44 community would facilitate imple-.4 on of the technology. Seminars.

Drograms, handbooks, and hands-onrations would be useful contributions.are other research programs that

.

.tent could support to facilitatesuction of office automation. Private

tow"?. unlikely to undertake such work

44.1.4 41

they could not capture sufficientto make it worthwhile. Government

*44.,444,44i)onsor research that would identify

,bibirt tt

4044,..:4ye structural forms that informationg could take in different industries,es of Government agencies andheir consequences for enhanced

wonmvity. Such research programs shouldwork on definitions and measures!activity for white-collar workers;y knowledge workers and managers.on of the Government's own ex-?S with office automation might alsoelpful; as would evaluation of state-irt installations in industry and um-omputer science departments. Gov-

, t could study needs for information 6.

Ts and support personnel and pro-I 2 data to industry and educational

ins. Finally, the Federal Government,4410. , apport the development of softWare

is for office automation that would4.0,404, provisions for collecting data relevant

to information system management research.Regardless of the pace with which office

automation proceeds, educational institutionsmust move to ensure computer literacy, justas they now address reading, writing, andarithmetic. But there is little agreement onhow to address this problem, and there arefew resources for experimentation. Schoolsfrom the elementary level through the uni-versity are not succeeding uniformly orsatisfactorily, and successes achieved in oneschool are not communicated effectively toothers or easily adapted by them. Thus, thefirst priority for all schools should be todetermine reasonable objectives for computerliteracy and support their attainment withsubstantial resources: In addition to offeringnew and modified curricula; universities andtechnical schools should develop continuingeducation courses, perhaps in conjunctionwith industry, directed toward informationmanagers and support personnel. Whetherthe office of the future becomes pervasivein 15 years or in 30, the next generationshould be computer literate to take fulladvantage of the benefits of the new officetechnology and to mitigate some of itspossible costs.

1.

2.

3.

4,

5.

7.

8.

9.

Notes and References

Connell. J. J._"The Office of the Future." Journalof Systems Management. vol. 30 (February 1979):6.10.U.S. Bureau of Labor Statistics. Employment andEarnings. 1980.Strassmann. Paul. "The Office of the Future: In-formation Management for the New Age." Tech.no/ogy Review. vol. 82. no. 3 (December 1979/January 1980).55.56."Worry Grows Over Upheaval as TechnologyReshapes Jobs." New York Times (4 July 1982).Abernathy; William J. and Rosenbloom; Richard S."The Institutional Climate for Innovation in In-dustry: The Role of Management Attitudes andPractices." In National Science Foundation. The 5-Year Outlook on Science and Technology. 1981.NSF 81.42. Washington, DC: U.S. GovemmentPrinting Office, 1982. pp._407-420.Strassmann. op. cit. p. 56. "Accepted economicindicators" may not be valid measures of white.collar productivity. See following text.Uttal. Bro. "What's Detaining the Office of theFuture." Fortune (3 May 19821: 176.Ginsberg. E.. "The Mechanizatioi. of Work." Scientific American (September 19827-75.Uttal. op. cit.

1 23

Fostering The Use Of AdvancedManufacturing Technology

Abstract

Manufacturing processes are a major factor in international economic competitiveness. Inrecent years, the United States has lagged behind other industrialized nations in the diffu _

Si on and implementation of advanced manufacturing systems: Many analyses have pointedout certain macroeconomic variables contributing to this lag; however; difficulties in imple-mentation of these systems at the factory level; while equally important; have not beenadequately addressed.

Several new technologies, including robotics, computer-aided manufacturing; grouptechnology, and flexible manufacturing systems, are particularly crucial to economiccompetitiveness. These systems must be seen as complex sociotechnical phenomena;involving major changes in corporate strategies, organizational design, and human resources.For example, their implementation involves interactions among functional units withinthe firm, with effects that are radical rather than incremental; full implementation oftentakes years. Especially relevant to decisions about manufacturing technology are anticipatedeffects on the skill and responsibility profiles of workers, including the need for retraining,involvement in operational decisionmaking, and job redundancy.

Given the difficulties of implementation, current technology transfer efforts are probablyinadequate: Shop floor involvement in implementation decisions is crucial, but often missing.Federal efforts in the transfer of advanced manufacturing technology have been unevenat best; particularly since much Federal R&D is mission-oriented with an emphasis ondefense requirements: Universities have b,-.1en reasonably active in research and develop=ment for new manufacturing technologies; but have not adequately addressed disseminationand implementation issues: The efforts of professional associations and societies havebeen considerable, but have been largely disaggregated: Given the need for widespreadimplementation of these technologies and the limits on current technology transfer,increased efforts reflecting a systems orientation by both government and private concernswill be required.

Introduction

One prominent economic and social issuewith significant science and technology im-plications is whether the United States'position as a major manufacturing powerwill continue to erode over the next decade.Evidence is mounting for the critical im-portance of manufacturing to economicrecovery and productivity. In the early 197Cvs,it was often argued that the decline of U.S.manufacturing reflected a "natural" move-ment toward a service economy. It was widelyexpected that service and knowledge-generation industries would supersedemanufacturing in importance, just as man-ufacturing had once superseded agriculture.This picture has not been sustained thus

far; the decline of productivity growth inmanufacturing has continued, and the serviceeconomy has not grown at a rate sufficientto compensate.

During the past 20 years, foreign com-petitionparticularly from Japan andWestern Europehas resulted in majorinroads into domestic markets traditionallydominated by U.S. manufacturers: Especiallyhard hit have been consumer electronics;automobiles, basic steel, textiles; and foot-wear. The impact is not evenly distributedgeographically; States with the highestcurrent rate of unemployment are thosewith the heavy concentrations of traditionalmanufacturing industries. But the effect isSpreading to other areas. Even in the semi-conductor industry, long considered thepreeminent preserve of American technol-

125113

ogy, the United States is being hard pressedby Japanese competition.

Many factorstechnological, economic,and politicalare involved in these events,including the growth of multinational com-panies and the consequent dispersion ofeconomic activity around the world, theincreasing costs of domestic capital, andthe use of "dumping" and other unfair tradepractices and even nontariff trade barriersimposed by foreign competitors. Differentindustries have been affected in differentways, but a common denominator is thatour manufacturing facilities are not beingreplaced by sociotechnically innovative andefficient operations. As one indicator of thisrelative disadvantage, a survey by theAmerican Machinist (1978) indicates thatof the seven industrial nations studied, theUnited States has the lowest percentage ofmachine tools under 10 years old; Japan,for example, has twice the U.S. percentageof newer tools.

The last decade has witnessed tremendousadvances in the development of new man-ufacturing technologies that could havecontributed to U.S. competitiveness evenin the face of wage and raw material costdifferentials, protective trade policies, anddifferences in the cost of capital. Theseinclude robotics, computer-aided designand manufacturing, group technology,flexible manufacturing, and various manage-ment and control systems associated withthese technologies. There are also technol-ogies on the horizon (for example, artificialintelligence) that will qualitatively extend thearray of ways to produce a product.

For various reasons, the United Stateshas been slow to install and use such tech-niques. While much of the early developmentof robotics was done in this country in the1960s by such companies as Unimation andAMF, Japan has come to be the undisputedworld leader in the deployment of roboticstechnology. Since the introduction of thefirst robots into Japan (by a U.S. firm) in1967, the use of robots has grown rapidly.By 1981 there were over 14,000 robots inuse in Japan; estimates for the deploymentof industrial robots in the United States andWestern Europe are approximately 5,000and 3,500 respectively (Robotics Institute

114

of America, 1981).* Japan's leadership inthe use of robots and related technologiesis not generally disputed. As an illustrationof radical possibilities offered by these tech-nologies, Fujitsu Fanuc recently built afactory that can be left essentially unmannedduring evening "ghost shifts." The plant has29 work stations, 7 of which are equippedwith robots, and 22 of which employ auto-matic pallet changers. The plant also hasautomatic warehouses, for both materialsand finished subassemblies (Yoshikawa,Rathmill, and Hatvany, 1981).

Given these rapid technological develop-ments and the increasingly frequent demon-strations of productivity gains in both foreignand domestic plants using these techniques,why have most U.S. manufacturing firmslagged in adopting and implementing them?One set of explanations currently offered islargely based on macroeconomic argumentsinvolving taxes and related incentives forinvestment in new technology. The macro-economic dimensions of R&D investmenthave been widely discussed (Annual Scienceand Technology Report to the Congress:1981), and many changes have already beenmade in the U.S. Tax Code that will pre-sumably address these and other invest-ment constraints.

But such remedies operate at a ratherhigh level of abstraction and aggregationand are by no means targeted at manufac-turing technology directly. In fact, the eco-nomics of advanced manufacturing systemsoperate at a very micro level and are notwell understood by most corporate decision-makers. The nations where these technol-ogies have enjoyed greater diffusion tendto have implemented much more focusedeconomic interventions based on promotingthe technology directly rather than throughthe manipulation of overall resource con-straints on firms. There is little empirical ortheoretical reason to assume that unfocusedincentives will more than marginally affectthe deployment of technology.

*Since the operational definition of a robot in Japantends to include some devices excluded from othercountries' estimates, precise cross-national comparisonsare difficult.

ogy, the United States is being hard pressedby Japanese competition.

Many factorstechnological, economic;and politicalare involved in these events;including the growth of multinational com-panies and the consequent dispersion ofeconomic activity around the world, theincreasing costs of domestic capital, andthe use of "dumping" and other unfair tradepractices and even nontariff trade barriersimposed by foreign competitors. Differentindustries have been affected in differentways, but a common denominator is thatour manufacturing facilities are not beingreplaced by sociotechnically innovative andefficient operations. As one indicator of thisrelative disadvantage, a survey by theAmerican Machinist (1978) indicates thatof the seven industrial nations studied, theUnited States has the lowest percentage ofmachine tools under 10 years old; Japan,for example, has twice the U.S. percentageof newer tools.

The last decade has witnessed tremendousadvances in the development of new man-ufacturing technologies that could havecontributed to U.S. competitiveness evenin the face of wage and raw material costdifferentials, protective trade policies, anddifferences in the cost of capital. Theseinclude robotics, computer-aided designand manufacturing, group technology,flexible manufacturing, and various manage-ment and control systems associated withthese technologies. There are also technol-ogies on the horizon (for example, artificialintelligence) that will qualitatively extend thearray of ways to produce a product.

For various reasons, the United Stateshas been slow to install and use such tech-niques. While much of the early developmentof robotics was done in this country in the1960s by such companies as Unimation andAMF, Japan has come to be the undisputedworld leader in the deployment of roboticstechnology. Since the introduction of thefirst robots into Japan (by a U.S. firm) in1967, the use of robots has grown rapidly.By 1981 there were over 14,000 robots inuse in Japan; estimates for the deploymentof industrial robots in the United States andWestern Europe are approximately 5,000and 3,500 respectively (Robotics Institute

114

of America, 1981).* Japan's leadership inthe use of robots and related technologiesis not generally disputed: As an illustrationof radical possibilities offered by these tech-nologies; Fujitsu Fanuc recently built afactory that can be left essentially unmannedduring evening "ghost shifts." The plant has29 work stations, 7 of which are equippedwith robots, and 22 of which employ auto-matic pallet changers. The plant also hasautomatic warehouses, for both materialsand finished subassemblies (Yoshikawa,Rathmill, and Hatvany, 1981).

Given these rapid technological develop-ments and the increasingly frequent demon-strations of productivity gains in Goth foreignand domestic plants using these techniques,why have most U.S. manufacturing firmslagged in adopting and implementing them?One set of explanations currently offered islargely based on macroeconomic argumentsinvolving taxes and related incentives forinvestment in new technology. The macro-economic dimensions of R&D investmenthave been widely discussed (Annual Scienceand Technology Report to the Congress:1981), and many changes have already beenmade in the U.S. Tax Code that will pre-sumably address these and other invest-ment constraints.

But such remedies operate at a ratherhigh level of abstraction and aggregationand are by no means targeted at manufac-turing technology directly. In fact, the eco-nomics of advanced manufacturing systemsoperate at a very micro level and are notwell understood by most corporate decision-makers. The nations where these technol7ogies have enjoyed greater diffusion tendto have implemented much more focusedeconomic interventions based on promotingthe technology directly rather than throughthe manipulation of overall resource con-straints on firms. There is little empirical ortheoretical reason to assume that unfocusedincentives will more than marginally affectthe deployment of technology.

*Since the operational definition of a robot in Japantends to include some devices excluded from othercountries' estimates, precise cross-national comparisonsare difficult.

tools; custom printing, tailored clothing,and other- product lines where there arelimited numbers of potential purchasers witha particular set of product needs. Job shopstend to be small in size and employ a largeproportion of highly skilled workers.

Batch production involves runs rangingfrom 200 to 20,000 in size, although theabsolute size of the production run is lessimportant than the fact that the productionfacility needs to be changed at frequentintervals for different product lines. Forexample, a firm may set up a productionline to produce several hundred electricmotors of one size, and then in a matter ofdays or weeks convert that line to assemblinga motor of an entirely different Machinesand equipment tend to be relatively unspe-cialized, and a mix of skilled and unskilledworkers is required: Products typicallymanufactured in batches include general:purpose machine tools; major householdappliances; ready-to-wear clothing, books,furniture, and some types of industrial equip-ment There are also "batch-flow" processes,such as the production of ice cream andcosmetics. Batch production accounts forover 35 percent of U.S. manufacturing(Gerwin, 1982) and tends to involve mediumto large firms.

Mass production involves the continuousproduction of identical items, with high vol-ume and the use of single:purpose machinesand equipment. Demands on worker skillsare generally low; U.S. mass production in-dustries usually consider labor an easilyreplaceable commodity. Examples of massproduction industries include automobiles;electronics, small household appliances;light bulbs, and nails. Mass producers tendto be large industries.

Several types of new technology can fitinto this general pattern of manufacturing.Four types of advanced manufacturing sys-tems will be considered here: robotics,computer-aided manufacturing (CAM); grouptechnology; and flexible manufacturingsystems (FMS). Althoi..qh in practice thesecategories tend to ovei lap, (for example,robots are often a part of flexible manufactur-ing systems; and group technology is in manyways an essential ingredient of both CAMand FMS), they are discussed separately,since they employ rather different sets of

116

equipment and have different behavioral,economic, and organizational implications.

Robotics Technology

The Robotics Institute of America (1980)has defined a robot as:

A reprogrammable, multifunction manip-ulator designed to move material, parts,tools or specialized devices through van-able programmed motions for the per-formance of a variety of tasks.

Robots may range in complexity from sim-ple "pick-and-place" machines, designedprimarily for materials handling, to complexmachines possessing sufficient motion flex:ibility and sensing capability to emulate andoften surpass the performance of a humanworker at particular repetitious tasks. EarlyU.S. research and development of roboticswas done during the 1950s, and the firstindustrial robot was installed in a U.S. fac-tory (General Mo.)rs) in 1961. The majorU.S. producers of robots are Unimation andCincinnati Milacron, whose sales constituteapproximately 70 percent of the Americanmarket. However, several other major firmsare currently entering the field; and as notedearlier, the Japanese presently dominatethe world production of robots: There arenow between 130 and 140 firms in Japanmaking robots; as opposed to roughly thesame number in the rest of the world(Aron; 1982).

Robots range greatly in size and breadthof application. It is possible for robots tobe installed either at single work sites or atmany sites simultaneously; the "robotization"of a manufaduring plant can be approachedpiecemeal. Robots range in price fromapproximately $50,000 to $2(X),000. Accord-ing to an industry source, robots have beeninstalled in a plant employing as few as twodozen people. One impetus for the installa-tion of robots has been to displace workersin areas when there are significant healthand safety hazards and the cost of protectingthe workers is considered prohibitive.

Computer-Aided Manufacturing

The initial move toward CAM was numericalcontrol of machine tools, developed in the

128

1950s largely at the Massachusetts Instituteof Technology (MIT) (Ettlie, 1971,_andGroover, 1980), where the U.S. Air Forcefunded a project to improve precision injet aircraft fabrication. The complexity ofmachining necessary for modern combataircraft had put tremendous strains onexisting control technology. MIT's solutioninvolved automatic control of machine toolsby a punched paper tape; airplane wingpanels were milled to specifications on a pre-programmed, high-speed machine.

The development of small computers andmicroprocessors led eventually to the re-placement of control by paper tape to controlby electronics. The first phase in this devel-opment involved direct numerical control(DNC) in which a single large computercontrolled a number of machine tools byhard-wired connections. As computer tech-nology evolved, particularly with greaterminiaturization; it became possible to havea single microcomputer in control of eachmachine. This led to computer numericalcontrol (CNC) and enabled even greatercontrol over the production process byallowing feedback on tool wear, greater useof sensors, etc. CAM has often involvedcombining these techniques by linking singlemachines' dedicated computers to a largercentral control computer to schedule opera-tions and related intermachine flows. Inaddition, it is often linked more or lessdirectly to computer-aided design (CAD)systems.

Group Technology

In many ways; group technology can beconsidered an organizational or technicalmanagement precursor to flexible manu-facturing and CAM systems. Group tech-nology is more an organizational constructthan a set of hardware. Its purpose is tobring the economies and benefits of massproduction to small-scale manufacturing;down to the level of the job shop. Pro-cedurally, it involves the categorization orgrouping of parts on the basis of design andmanufacturing similarities. Once a group ofparts has been identified, a corresponding"cell" of machines performing interrelatedfunctions can be dedicated to the productionof that family of parts. Advantages in de-

129

creased set-up time are considerable, andproductivity increases of major proportionshave been reported (Gettleman, 1975;Gallagher and Knight, 1973; Wilson andHenry, 1977). Group technology in itspresent form began in Europe, includingextensive work by the Russians, but manyof the concepts and techniques of grouptechnology derive from the work of anAmerican Frederick Taylorin the 1920s.Group technology has many and variedapplications in the machine tool industry,but these methods have not been extensivelyimplemented in U.S. firms.

Flexible Manufacturing Systems

Flexible manufacturing systems are in anywas outgrowths or extensions of develop-ments in group technology, computer-aidedmanufacturing, and robotics. As indicatedabove, early numerical control applicationsusually involved a single function such asdrilling or shaping. These systems evolvedso that tools could be changed and differenttypes of machines could be linked togethersequentially by materials-handling capabil-ities (for example, pick-and-place robots),all controlled, monitored, and servicedby computers.

One example of a flexible manufacturingsystem in the process of development is theAutomated Programmable Assembly System(APAS) created by Westinghouse Corpora-tion. The APAS demonstration is focusedon the assembling of small electric motors;which historically have been batch-processedthroughout the world: The APAS assemblyline involves the linking of robotics; roboticworkstations; advanced optical sensors; auto-mated materials handling; and overall com-puter control into a completely automatedassembly system. Most importantly, the sys-tem is "reprogrammable," such that motorsof different sizes can be assembled literallyat the flick of a switch. Another example ofan FMS developed under Governmentsponsorship is the Automated ManufacturingResearch Facility being developed by theNational Bureau of Standards. This demon:stration will involve an integrated series ofmachining and inspection stations with par-ticular applications for small shops.

117

FMS exists at present more in theory thanin widely deployed and developed applica-tions: According to one industry analyst.there are less than 20 FMS systems in opera-tion in the United States: and planning andinstallation currently requires something onthe order of 2 to 5 years.

Problems of Use

AS we have suggested, the availability oftechnical knowledge about advanced man-ufacturing systems has not resulted in theirwidespread adoption and use in the UnitedStates. It is necessary to look at the decisionprocesses of individual firms to understandsome of the problems involved and howthey might be addressed. The social andorganizational dimensions of technologicalchange are as critical to its adoption anduse as its technical dimensions and deserveas careful an evaluation.

In recent years, a body of knowledge hasemerged about the process of innovation.and this can be applied to decisions involvingtechnology. To understand innovation,technologies must be considered not onlyas collections of hardware: but as knowledgeembodied both in the rr .'chines themselvesand in the software, cont -ol, and organiza-tional systems necessary to operate thetechnology effectively. Defined thus. thedeployment of manufacturing technologycan be seen partly as a knowledge dissem-in'tion and utilization issue.

The innovation process literature hasrecently begun to view technology transferas moving through several separable stages,from initial awareness. to evaluation anddeciding, to adoption. aiid, finally, to imple-mentation. The transition from adoption toimplementation is likely to be as critical forsuccess as the dissemination of knowledge,but it has received much less attention. Upto the commitment point, knowledge dis-semination is primarily a cognitive andintellectual activity that involves learningabout the innovation: Implementation, bycontrast, involves the expenditure of humanand material resources and behavioralchanges at many levels of the organization:Thus two separable knowledge transfer issuesare involved: transmitting information that

118

might lead to adoption and providing de-tailed operational information about puttingthe innovation in place.

An adoption decision is usually premisedon management's perception that an inno-vation is more profitable or effective thanexisting practice, and thus there is a reciprocalrole for vendors to make potential usersaware of new technologies. The opportunityfor users to observe real-time installationsand demonstrations (essential for technol-ogies of this complexity) is quite limited.particularly for smaller firms: Since thepotential for marketing is usually greaterwith large firms, smaller firms may get lessaccess to information about new technol-ogies. While an adoption decision is usuallyfairly easily identified in space and time,implementation is not the result of a singledecision but rather a whole series of decisionsfrequently made by different people in dif-ferent places.

A key concept in understanding imple-mentation is to recognize the impossibilityof separating decisions about hardware andeconomics from their implications for thesocial behavior of those using the hardware.Engineers and technical designers oftenmake critical decisions about social/organi-zational issues, such as how people will bedivided into groups. how many people willbe in these groups, and where individualworkers will be located (Davis and Taylor,1976), although often they are not awareof the implications of their choices: Thus,the adoption and implementation of even asimple pick-and-place robot is a set of com-plex sociotechnical processes involving manydifferent people from many different groupsin the firm and. thus. subject to evaluationfrom many points of view.

The empirical literature dealing with theproblems of either dissemination or imple-mentation of major technological systems islimited. Much of the traditional literatureon innovation processes has assumed thatimplementation of technology followsnaturally from decisions to adopt or pur-chase. The literature pertaining to the im-plementation of industrial innovations isespecially meager. and studies of the imple-mentation of manufacturing processes arerare: For a recent review of implementationstudies (Scheirer and Rezmovic, 1982), only

130

FMS exists at present more in theory than,videly deployed and developed applica-is: According to one industry analyst.re are less than 20 FMS systems in opera-

in the United States, and planning and:illation currently requires something onorder of 2 to 5 years.

Problems of Use

we have suggested, the availability ofinical knowledge about advanced man-cturing systems has not resulted in theirespread adoption and use in the Unitedtes. It is necessary to look at the decisioncesses of individual firms to understandle of the problems involved and how

might be addressed. The social andanizational dimensions of technologicalnge are as critical to its adoption andas its technical dimensions and deserve:ireful an evaluation.-1 recent years, a body of knowledge has?,rged about the process of innovation.this can be applied to decisions involving

inology. To understand innovation.inologies must be considered not onlyollections of hardware: but as knowledge)odied both in the rr3chines themselvesin the software, cont -ol, and organiza-

al systems necessary to operate themology effectively. Defined thus. theloyment of manufacturing technOlogybe seen partly as a knowledge dissem-on and utilization issue.he innovation process literature hasntly begun to view techndlogy transferloving through several separable stages,

initial awareness. to evaluation andding, to adoption, al finally, to ittiple-tation. The transition from adoption rciementation is likely to be as critical foress as the dissemination of knowledge,t has received much less attention. Uple commitment point, knowledge dis-i nation is primarily a cognitive andlectual activity that involves learningit the innovation: Implementation. byrast: involves the expenditure of humanmaterial resources and behavioralges at many levels of the organization:two separable knowledge transfer issues

nvolved: transmitting information that

might lead to adoption and providing de-tailed operational information about puttingthe innovation in place.

An adoption decision is usually premisedon management's perception that an inno-vation is more profitable or effective thanexisting practice, and thus there is a reciprocalrole for vendors to make potential usersaware of new technologies. The opportunityfor users to observe real-time installationsand demonstrations (essential for technol-ogies of this complexity) is quite limited,particularly for smaller firms: Since thepotential for marketing is usually greaterwith large firms; smaller firms may get lessaccess to information about new technol-ogies. While an adoption decision is usuallyfairly easily identified in space and time.implementation is not the result of a singledecision but rather a whole series of decisionsfrequently made by different people in dif-ferent places.

A key concept in understanding imple-mentation is to recognize the impossibilityof separating deciSionS about hardware andeconomics from their implicationS fbr thesocial behavior of those using the hardware.Engineers and technical designers oftenmake critical decisions about social/organi-zational issues, such as how people will bedivided into groups, how many people willbe in these groups, and where individualworkers will be located (Davis and Taylor.1976). although often they are not awareof the implications of their choices: Thus,the adoption and implementation of even asimple pick-and-place robot is a set of com-plex sociotechnical processes involving manydifferent people from many different groupsin the firm and. thus. subject to evaluationfrom many points of view.

The empirical literature dealing with theproblems of either dissemination or imple-mentation of major technological systems islimited. Much of the traditional literatureon innovation processes has assumed thatimplementation of technology followsnaturally from decisions to adopt or pur-chase. The literature pertaining to the im-plementation of industrial innovations isespecially meager, and studies of the imple-mentation of manufacturing processes arerare. For a recent review of implementationstudies (Scheirer and Rezmovic, 1982), only

130

of advanced manufacturing systems is thatto be fully effective these systems must tietogether disparate functional units withinthe firm in new and different ways. Unfor-tunately, in the typical large industrial firm,manufacturing, distribution, and accountingprocedures are much better integrated witheach other than is manufacturing with eithermarketing or engineering. One of theadvantages that Japanese companies mayhave is that over a career their managerswill work in many areas of the companyand, as a result, are less protective of func-tional fiefdoms (Aron, 1982). Often, newmanagerial techniques are required bytechnical decisions, and organizationalstructures adapted to change rather thanrigidity have a substantial advantage.

The pervasiveness of organizational effectsattributable to manufacturing systems alsohas implications for leadership and decision-making pertaining to implementation. Sincethe implementation of such systems as CAMcuts across different functional units, it isprobably unwise to give leadership of thatimplementation to any one function. Oneanalyst has called for the assignment of a"process champion," who would have leader.ship responsibilities for implementation ofsuch new technology across various functionsin the company (Gerwin, 1982).

A process champion cannot operate in aunilateral or authoritarian manner. One ofthe more consistent findings in the innovationprocess literature is that adoption and im-plementation of complex technologies isfacilitated by participative decisionmaking(see, for example, Tornatzky, et al., ! 980).For the implementation of complex manu-facturing systems to "stick," individualsfrom many different levels and functionswithin the firm will need to be involved inplanning and decisionmaking. Often thiswill go against the prevailing organizationalclimate and may actually lengthen theimplementation process.

However, it should be noted that ourforeign competitors have become quite awareof the role of worker involvement in theimprovement of process technology In manyways, the quality circle movement in Japanhas been an important factor in the ad-vancement of manufacturing processes.Analogously; European nations have been

120

acutely aware of the relationship betweenthe new technologies and worker involvementin their use (Norges Offentlige Utredninger,1980). Worker involvement can be used toassist implementation planning, and indeedmany initiatives for installation of advancedtechnology such as robots have come fromthe shop floor. Many U.S. companies areinvolved in Quality of Work Life (QWL)efforts, which may have similar repercus-sions on the improvement of process tech-nology through participative decisionmakingand planning. For any of these workerinvolvement programs to be effective, theparticipation must be substantial and sig-nificant on the part of both managementand labor. To the extent that they are seenas ways to manipulate workers rather thanto take them seriously, they will fail.

Longitudinal and Strategic Aspects ofManufacturing Technology

As we have suggested; the implementationof such systems as CAM, group technology;or robotics cannot be considered as merelya routine capital investment decision. Fullimplementation may take years (such as inthe case of FMS) and must be well integratedinto a strategic vision for the firm. In addition,the costs of adoption and implementationand the benefits from use of manufacturingtechnologies are not easily quantifiable inadvance. Decisions may require as muchan act of faith as a thorough economicanalysis.

Unfortunately, the application of suchstrategic vision of manufacturing technologyhas been distinctly rare in American industryof late. Rapidly accumulating empirical andobservational evidence (Hayes and Aber-nathy, 1980) suggests that a lack of strategicplanning has serious implications for eitheradopting complex manufacturing systemsor their successful implementation: Severalfactors have been identified as contributingto this lack of vision. In the last few decades;corporate managers in the United Stateshave been more likely to have financial andlegal rather than technical backgrounds. Ithas been suggested that this trend hasstemmed from the greater emphasis in thelast 20-30 years on matters external t o t hefirm, particularly interactions with govern-

132

mentinteractions focused more on lawand money than on technology. Thus, thetrend may well have had short-term sury valvalue. But it has also imposed costs. Thereis a greater preoccupation with short-termprofits and decisionmaking; a dominanceof a marketing orientation; and an increas-ing tendency toward corporate mergers. Allof these trends have exacerbated a splitbetween shop floor manufacturing technol-ogy and corporate strategy, and this splithas made it difficult for managers to under-stand the implementation impacts of theirstrategic choicesor the opportunity costsof their tack of choiceS.

Sociotechnical Aspects ofImplementationIt is increasingly apparent that implemen-tation of advanced manufacturing systemswill permanently change the nature of workand of the workforce in industry. Althoughthe extent to which these technologies willproduce net job losses or gains has yet tobe determined; it is quite clear that thenature of job skills needed in manufacturingwill change dramatically: Traditional massproduction has treated workers as low-skilledoperatives who are easily replaceable; the"second industrial revolution" will demandworkers with multiple skills who will be heavilyinvolved in maintenance of the new tech-nologies, information transmission, andtechnologically demanding tasks. It will alsoprobably require fewer of them. The shrinkingsize of the primary labor market is a factwith rather profound social policy impli-cations, which are as yet largely unexplored.

Confronted with this reality, U.S. corpora-tions will have to make a basic choice betweenretraining employees or simply replacingthem. Interestingly, much of the high-technology manufacturing component ofJapanese industry has thus far kept its com-mitment for lifetime employment and hasinvested heavily in worker retraining andskills upgrading: This high job security mayactually contribute to innovativeness, asworkers not confronted with job loss maymore readily accept changes: In turn, man-agement faced with a permanent workforcewill need to explore ways to maintain andincrease productivity.

133

Changing the skill and responsibility pro-files of workers will likely have implicationsfor the distribution of power and decision-_making within the company. The center ofinfluence may shift to those areas Wherethe new sophisticated procesS technology isbeing employed on the shop floor. Americanmanufacturers have tended to utilize workorganizations based on hierarchical leader-ship and specialized job classifications inconjunction with advanced manufacturingsystems. This probably reflects an attemptto apply management principles that aretraditional in batch manufacturing. However,the nature of the new equipment may besufficiently different from typical stand-alonemachine tools that an alternative workorganization may be in order. An integratedsystem is characterized by relatively sharpboundaries at either end and a continuousflow of material within the borders. It istherefore more akin to process manufac-turing than batch manufacturing. This sug-gests that a work organization baSed on thegroup as the fundamental unit may be moreappropriate than one based on the individual.

One possibility to consider, according tosociotechnical systems theory, is a joint workgroup of operators and loaders that doesaway with their separate designations. Eachparticipant would have an opportunity toshare in all or most tasks. The group wouldbe responsible for the complete cycle ofloading; monitoring, and unloading as wellas some repairs, tooling, and supervisoryfunctions. Consequently, task identity,meaningfulness, and feedback on per-formance from coworkers would be in-creased. Job rotations would augmentautonomy and participation in work-relateddecisions. European manufacturers appearto be more willing to experiment with newapproaches to work organization that mightprovide a better fit with advanced manufac-turing technology (see Taylor, 1977, forillustrations).

Efforts to undeittand new job skill require-ments will go hand in hand with implemen-tation of the technology per se if maximumbenefits' are to be realized. Like technicalchoices, social and organizational choice§related to implementation are frequentlydriven by the technology (for example, size,structure, and composition of work group),

121

but always with the potential to shape itreciprocally. The more this interaction isunderstood and planned for, the less willbe the stress on the system and the betterthe results (Reinke. 1982).

Technology Transfer Strategyand TacticsIt should be clear that the disseminationand implementation of advanced manufac=turing systems will not just happen, but willlikely be a painful and stressful situation formost organizations. Some of this stress canbe reduced through strategic planning forimplementation. For example, the relation-ship between the equipment vendor and theimplementii.g firm is much more intensiveand extensive than in other capital investmentprogramS. The vendor should be willing towork closely with the adopting companyand in turn the adopting firm must be willingto devote resources to understanding thetechnology and its implications. A simpleturnkey approach to the technology is notlikely to be workable for either vendor oradopter. There is evidence that the morecomplex the technological system beingimplemented, the less viable the turnkeyapproach (Cooke and Malcolm. 1981).

Worker participation is crucial in imple-mentation strategy. Some programs haveheavily involved lower level staff in new plantdesign and the implementation of advancedmanufacturing systems (Gustayson andTaylor, 1982). More of this will probablyoccur as management structures orientedtoward traditional linear and hierarchicaltechnology are replaced by managerialpractices oriented toward advanced manu-facturing systems and the lower level dis-cretion they require. Given the extensiveinvolvement of lower level staff in successfulimplementation, acceptance of informationabout new technologies might be enhancedby focusing dissemination efforts at thatlevel. For example, in an automobile plantbeing established in Tennessee; a largepercentage of the new workforce was flownto the home plant in Japan to becomefamiliar with the manufacturing processesto be used there.

The major argument for explicit attentionto implementation in technology decisions

122

is that it will help both managers and workersavoid some surprises: While new technologyis inherently uncertain in its effects; thereare many potential contributions of imple-mentation analysis that can guide tech-nological strategy. In general, the Japanesehave been better at applying these principlesthan have Americans, for reasons partlycultural and partly deliberate. Japaneseemphasis on consensus-building has in-corporated delayS and trials into the process.explicitly recognizing that multiple decisionsare involved. By contrast, American man-agert frequently treat technolbgy decisionsas simple choices and set artificially shorttimetables for implementation. The almostinevitable disappointment with the resultstends to be interpreted more as a reflectionon the technology than on the inadequateimplementation process. There is no doubtthat many managers have interpreted astechnical deficiencies problems that in factderive from the shortcomings of their ownstrategic organizational choices:

Programs to Assist TechnologyTransfer

Efforts to assist firms in understanding andusing advanced manufacturing technologyare emerging from three sources:_ the Fed-eral Govemment, university research centers,and private nonprofit educational andprofessional associations. The nature andscope of these efforts will be described below.Few if any of these activities appear torecognize the systematic properties of eitherthe technologies or the client systems withwhich they work. In many ways, these groupsare more concerned with the hardware thanwith the software; organizational; or humanresource aspects of manufacturing technol-ogies: As Noble (1979) indicates; the devel-opment of advanced manufacturing systemshas tended to be left to "industrial tech-nocrats...and self-serving computer jocks."

The Federal Role in ManufacturingTechnology

There are approximately one dozen Federalprograms focused on advanced manufac-turing systems. These programs are scattered

134

always with the potential to shape itiprocally. The more this interaction islerstood and planned for, the less willthe stress on the system and the betterresults (Reinke. 1982).

:hnology Transfer StrategyTactics

hould be clear that the disseminationI implementation of advanced rnanufac:ng systems will not just happen, but willly be a painful and stressful situation forst organizations. Some of this stress canreduced through strategic planning for)Iementation. For example, the relation-) between the equipment vendor and theilementii.g firm is much more intensiveI extensive than in other capital investmentgramS. The vendor should be willing to.k closely with the adopting companyI in turn the adopting firm must be willinglevote resources to understanding thelnology and its implications. A simpleikey approach to the technology is notly to be workable for either vendor orpter. There is evidence that the morenplex the technological system beingrlemented, the less viable the turnkeyroach (Cooke and Malcolm, 1981).Vorker participation is crucial in imple-ntation strategy. Some programs havevily involved lower level staff in new plantign and the implementation of advancedn u fa ct uri ng systems (Gustayson andriot. 1982). More of this will probablyur as management structures oriented'ard traditional linear and hierarchicalnnology are replaced by managerialctices oriented toward advanced manu-uring systems and the lower level elis-ion they require. Given the extensive)1vernent of lower level staff in successfullementation, acceptance of informationut new technologies might be enhanced'ocusing dissemination efforts at that!I. For example, in an automobile plantlg established in Tennessee, a largetentage of the new workforce was flownhe home plant in Japan to becomeiliar with the manufacturing processes,e used there.he major argument for explicit attention

nplementation in technology decisions

is that it will help both managers and workersavoid some surprises. While new technologyis inherently uncertain in its effects, thereare many potential contributions of imple-mentation analysis that can guide tech-nological strategy. In general, the Japanesehave been better at applying these principlesthan have Americans, for reasons partlycultural and partly deliberate. Japaneseemphasis on consensus-building has in-corporated delays and trials into the process.explicitly recognizing that multiple decisionsare involved. By contrast, American man-agerS frequently treat technology decisionsas simple choices and set artificially shorttimetables for implementation. The almostinevitable disappointment with the resultstends to be interpreted more as a reflectionon the technology than on the inadequateimplementation process. There is no doubtthat many managers have interpreted astechnical deficiencies problems that in factderive from the shortcomings of their ownstrategic organizational choices.

Programs to Assist TechnologyTransfer

Efforts to assist firms in understanding andusing advanced manufacturing technologyare emerging from three sources:_ the Fed-eral Government. university research centers,and private nonprofit educational andprofessional associations. The nature andscope of these efforts will be described below.Few if any of these activities appear torecognize the systematic properties of eitherthe technologies or the client systems withwhich they work fn many ways, these groupsare more concerned with the hardware thanwith the software, organizational, or humanresource aspects of manufacturing technol-ogies. As Noble (1979) indicates; the devel-opment of advanced manufacturing systemshas tended to be left to "industrial tech-nocrats...and self-serving computer jocks."

The Federal Role in ManufacturingTechnologyThere are approximately one dozen Federalprograms focused on advanced manufac=turing systems. These programs are scattered

134

members and nonmemberS. Similarly, theRobotics Institute of America Supports con-ferences and workShopS oriented aroundrobotics technology. At a more general level,the Society of Manufacturing Engineers hasa heavy investment in these new technologiesand has been quite active in organizingsymposia, training, and general knowledgetransfer.

One problem with these training/dissem-ination efforts has been that theyhaveaddressed general technical issues ratherthan specific operations at particular sites.They also tend to treat only the hardwareaspects of implementation; which may bethe smaller (if more manageable) part of theproblem. Implementation issues are oftenlikely to be idiosyncratic and unpredictablein nature and need to be addressed on theshop floor. This aspect of assistance hasbeen largely left to vendors of equipment,and their performance haS been uneven.Marketers, like manufacturerS, tend to takea short-term view of their client. What needsto be determined is how they might beinduced to take a longer view, and if thereis an appropriate dissemination and imple-mentation assistance role for Governmentor Government-supported activities.

National Implications andPolicy Options

In some senses the issues raised in this paperillustrate the limitations of traditional Gov-ernment policy levers. If adoption andimplementation of advanced manufacturingsystems are phenomena played out in thecontext of the firm itself, it is only partiallyaffected by interventions at the industry orvector level. Management practices andstrategies seem heavily implicated in thesuccessful implementation Of these tech-nologies, as does knowledge transfer re-garding the technologieS themselves. It is

unlikely that Government actions in suchareas as taxation could substantially alterthe incentive structure and consequentlythe strategic vision operating in Americanmanufacturing within anything like the timehorizons required. This would require a muchmore detailed taxation package than haspreviously been enacted, one heavily oriented

124

toward rewarding productivity enhancementand implementation of advanced processtechnologies: This is an extremely microlevelmanipulation of economic policy, and it islikely to be looked on with disfavor as anexcessive intrusion i ito the operations ofmarket forces.

There does seem to be a legitimate Gov:ernment role in two particular areas beyondthe traditional (and still needed) support ofbasic and applied research leading to tech-nology development. The first is in knowl-edge transfer and dissemination of tech-nological informatiOn. Current technologytransfer and implementation assistance isextremely scattered, and Federal agenciesinvolved in the development of manufac-turing technologies have not been giveneither a clear mandate or sufficient resourcesto promote civilian implementation;

The barriers to such an effort are moreconceptual and ideological than practical.Programs of thistype would radically alterthe traditional Government posture towardtechnology development. In the past, theGovernment has restricted itself to supportof research and early development and hasdepended upon market forces to enhancedissemination. This is the classic "demand-pull" approach. In contrast, an aggressivecoordinated technology transfer program(somewhat akin to agricultural extension)would imply that Government had made astrategic choice to "push" a family of tech-

-

nologies. While this would be unprecedentedin regard to manufacturing technology forthe United States, it would more closelyapproximate the posture taken by suchcountries as Japan and West Germany.

An essential corollary of a "technologypush" posture would be explicit attentionto issues of worker displacement andretraining. While the extent of workerredundancy produced by the new tech-nologies is unclear; it is virtually certain thatdislocations of labor will occur: Such optionsas guaranteed employment; incentives forretraining; and employee gain-sharing ofproductivity growth should be consideredin this context (Business Week, 1982).Moreover; policy should recognize that jobcreation and new industry are priorities.

Another area of modest Federal activitywould be research to improve understand-

1 3 6

ing of the dissemination and implementationprocesses: As noted several times, the proc-esses by which complex technologies areadopted and implemented are not wellanalyzed: nor have they been translated intoaction programs in any systematic way. ftshould again be emphasized that advancedmanufacturing systems constitute verycomplex innovations for adopting organi-zations, and we could understand thatinnovation process itself considerably betterthan we do now.

The bottom line is that one is unlikely tobe able to implement advanced manufac7turing technology adequately if it is viewedas just another machine or tool for doingwhat one is doing in the same way as atpresent. While these systems are potentiallyextremely productive-and probably essentialto the survival of the American economy-they are radically different from presenttechnologies in crucial ways. Taking advan-tage of advanced manufacturing capabilitiesis a process that will require considerablymore; and more systematic, attention to thephenomenon of deployment than hasheretofore been generally in evidence inU.S. industry. It will also require a significantchange in how managers view work andworkers. The types of technical and man-agerial control required by advanced man-ufacturing systems open up all kinds of newpossibilities for developing organizationalstructures and processes with both economicand human benefits. This is an area wherethe U.S. Government must think about itsrole in enhancing such transitions, with atleast as clear a vision as that of its intema-tional economic competitors.

References

1. Annual Science and Technology Report to the Con-gress: 1981. Washington-DC: Office of Scienceand Technology Policy, 1982.

2. American Machinist (December 1978): 133-148.3. Aron. Paul. "Robots Revisited: One Year Later."

In Exploratory Workshop on the Social Impactsof Robotics. Congress cf the United States, Officeof Technology Assessment. February 1982.

4. Battelle. "The Manufacturing Engineer-Past.Present and Future." Report to the Society ofLlanufecturing Engineers, 1979.

5. Business Week. "Retraining Displaced Workers:Too_Little, Too Late?" Business Week (19 July1982): 178-185.

13

6

7

. Cooke. R.A. and Malcolm, B. "Choosing and Im-plementing Resource Recovery Systems." NationalCenter for Resource Recovery Bulletin. vol. 11,no. 2 (1981), 35.39.

. Davis L.E. and Taylor, J.C. "Technology. Organi-zation, and Job Structure." In R. Dubin (ed.).Handbook of Work, Organization; cnd Society.Chicago, IL: Rand McNally, 1976.Ettlie, J.E., "Technology Transferin the MachineTool Industry: The Utilization of NC Machine Toolsin Captive Shops." Unpublished Master's Thesis.Evanston, IL: Northwestern University, 1971.Fisk, John D. Industrial Robots in the UnitedStates: issues and Perspectives. CongressionalResearch Service Report No. 81-78E, March 30,1981:Gallagher, C. C., and Knight, W.A. Group Tech-nology. London: Butterworth, 1973.Gerwin, Donald. "Do's and Don'ts of ComputerizedManufacturing7_Harvard Business Review (March-April 1982): 107-116.Gettleman. K. M. "Group Technology is a Three-Way Winner." Modern Machine Shop, vol. 48(July 1975).Giuliano, V.E.. "The Mechanization of Office Work"Scientific American. vol. 257, no. 3 (September1982): 149-164.Groover, M.P. Automation. Production Systems,and Computer-Aided Manufacturing. EnglewoodCliffi, NJ: Prentice-Hall, 1980.Gunn, T.G. "The_Mechanizatiori of Design andManufacturing." Scientific American, vol. 257,no. 3 (September 1982L 107-116.Gustayson, Paul and Taylor, James. "Socio-Technical Design and New Forms of WorkOrganization: integrated Circuit Fabrication."Unpublished paper, January 6, 1982.Hayes, Robert H. and Abernathy, William J."Managing Our Way to Economic Decline." HarvardBusiness Review (July-August 1980): 67-77.Hetzner, W. A., Tornatzky, Louis G., and Klein,Katherine: _"Manufacturing Technology in the1980's: A Survey of Federal Programs and Prac-tices." Management Science; 1982 (in press):Noble, David. "Social Choice in Machine Design."In Case Studies on the Lollar Process. New York:Monthly Review Press, 1979.Norges Offentlige Utredninger. Employment andWorking Conditions in the 1980', 1980, p. 33(English translation).Pelt, D.C. and Munson, D. "The Innovation Process:A Conceptual Framework." Working paper. AnnArbor: University of Michigan, Center for Researchon Utilization of Scientific Knowledge, 1980.Reinke, Debra: "Implementation of a Shop FloorControl System." Proceedings of the 19th AnnualMeeting _and Technical Conference. NumericalControl Society, April 18.21, 1982.Robotics Institute of America. "Robotics Today,"RTA News (Spring 1980),Robotics Institute of America. Worldwide RoboticsSurvey and Directory, 1981,Scheirer, Mary Ann and Rezmovic, Eva L. Meas7using the implementation of innovations. FinalReport to the National Science Foundation, Grant

8.

9.

10.

11.

12.

17.

18.

19.

20.

21.

22.

23.

24.

25.

125

No. PRA 80-22612. Annandale, VA: AmericanResearch Institute, 1982.

26. Taylor, J.C. "Experiments in Work System Design:Economic and Human Results." Personnel Review.vol. 6. nos: 2 and 3 (1977).

27. Tornatzky. L.G. and Klein, K.J. "Innovation Char-acteristics and Innovation Adoption-Implementa-tion: A Meta-Analysis of Finding." IEEE Trans-actions on Engineering Management. vol. EM29(1982): 28=45.

28. Tomatzky; Louis a, Fergus. Esther; Ave liar, Joseph.and Fairweather, George W. Innovation and SocialProcess. New York: Pergamon, 1980.

126

29. TornatZky, L.G., et al. The Process of Innovation:Analyzing the Literature. Washington. DC: NationalScience Foundation, 1982:

30. Wilson, R. C. and Henry, R. A. Introduction toGroup Technology in Manufacturing and _Engi-neering. Ann Arbor: University of Michigan, 1977.

31: Yoshikawa; Hiroyuki; Rathmill; Keith; and Hatyany,Jozsef. Computer-Aided Manufacturing: An Inter-national Comparison, Washington. DC: NationalAcademy Press. 1981.

1 38

Acknowledgements

Kenneth W. AbbottNorthwestern University

Gail AlbrittonFlorida State House of Representatives

Marietta BabaWayne State University

Leona BachrachUniversity of Maryland

Dale & BakerChemical Abstracts Service

Richard BalzhiserElectric Power Research Institute

Darryl BanksThe Rand Corporation

Donald BeilmanMicroelectronics Center of

North Carolina

Tora K. BiksonThe Rand Corporation

Eu la BinghamUniversity of Cincinnati Medical School

David G. Black, Jr.Research Corporation

Erich BlochInternational Business Machines Corp.

Richard BoltBolt, Beranck and Newman

Sidney BorowitzNew York University

Jane BortnickCongressional Research Service

Haivey BrooksHarvard University

Donald BuchwaldKansas Technical Institute

Elaine Bunten-MinesNational Bureau of Standards

John BurnsSt. Regis Paper Company

Wayne BurwellUnited Technologies Corp.

139

Arthur A: BushkinTelemation Associates, Inc.

Hershell CantorInstitute for Defense Analyses

Jaime G. CarbonellCarnegie-Mellon University

Roger CanickCalifornia Governor's Office

Stephen CheshierSouthern Technical Institute

Patrick ChoateTRW Corporation

Dahl E. ChubinGeorgia Institute of Technology

Edward ClarkWorcester Polytechnic Institute

Joseph ClarkDepartment of Commerce

Albert ClogstonBell Laboratories

Donald W. CollierBorg-Warner Corp.

Benjamin CompaineHarvard University

Arnold C. CooperPurdue University

Floyd Culler, Jr.Electric Power Research Institute

Lee L. DavenportGreenwich, CT

William T. DavisPfizer, Inc.

Stephen DeanFusion Power Associates

Robert DiMarcoGeneral Foods Corp.

Paul DoiganGeneral Electric

Richard D'OnofrioFranklin Institute of Boston

Herbert DordickUniversity of Southern California

127

Donald DunnStanford University

Leon DunningUniversity of Tennessee

James DurigUniversity of South Carolina

Tait ElderAllied Corporation

Robert M. EntmanDuke University

Thomas E. EverhartCornell University

Yvan Fabian0.E.C.D.

Gary FraserSUNY

Ellen L. FrostWestinghouse Electric Corporation

Robert FullerJohnson & Johnson

Herbert I. FusfeldNew York University

E. C. GallowayStauffer Chemical Company

Patrick GarveyRockefeller University

Cyrelle GersonAmerican Chemical Society

Anthony GiordanoPolytechnic Institute of New York

Vincent GiulianoArthur D. Little, Inc.

J. E. GoldmanXerox Corporation

Frank A. GourleyCalifornia Power and Light Company

Ruth L. GreensteinNational Science Foundation

Jerrier A. HaddadNew York, NY

Milton HarrisWashington, DC

Robert M. HayesUCLA

128

Robert HaymanConsolidated Edison Company of

New York

Wayne HodgesGeorgia Institute of Technology

Charles M. HugginsGeneral Electric Company

George HupmanGeneral Electric Company

Gregory JonesR&D Associates

Rodney JonesCenter for Strategic and International

Studies

David KalergisDewey, Ballantine, Bushby, Palmer and

Wood

Samuel J. KalowInternational Business Machines Corp.

Francis B. KapperDepartment of Defense

Lucian KasprzakInternational Business Machines Corp.

James KellattUnited States Air Force

Charles A. KieslerCarnegie-Mellon University

Donald D. KingNorth American Philips Corp.

Lawrence KravitzBendix Advanced Technology Center

George KrumbhaarWashington, DC

Ron KunkelNational Bureau of Standards

W. Henry LambrightSyracuse Research Corporation

Gordon LawOffice of Technology Assessment

Dorothy Leonard-BartonMassachusetts Institute of Technology

Gerald J. LiebermanStanford University

140

Quentin LindseyNorth Carolina Governor's Office

Thomas M. LodahlGray Judson, Inc.

John M. LogsdonGeorge Washington University

Thomas V. LongTechnology 2000

Henry C. Lucas, Jr.NeW York University

John LyonsNational Bureau of Standards

Allen. McClellandE. I. du Pont NernourS & Co.

John McElroyNational Oceanic and Atmospheric

AdministrationJames McEvoy

Council for Chemical Research

Edward McGaffiganOffice of the Honorable Jeffrey Bingaman

Patricia McGarryHoffmann-LaRoche, Inc.

Thomas McGuireBoston University

W. M. McKeemanWang Institute

Paul MakowskiGould; Inc:

Carol MannDigital Equipment Corp.

David MechanicRutgers University

Richard MeserveCovington & Burling

Michael MichaelisPartners in Enterprise, Inc

Mary MogeeAmerican Council on Education

Joseph MoroneGeneral Electric Company

Loren MosherUniformed_Services University of the

Health Science

14f

Mitchell MossNew York University

Richard P. NelsonYale University

Scott H. NelsonPennsylvania Department of Public

Welfare

Frederick NewmanEastern Pennsylvania Psychiatric Institute

Rodney NicholsRockefeller University

Eric NimkeBrooklyn Union Gas Company

Alfred NissanScarSdale, NY

David PadwaAgrogenetics Company

Gail PesynaDu Pont

Don I. PhillipsDuke University

Herman PollackGeorge Washington University

Joel PopkinJoel Popkin & Company

Thomas PagelNew York University

Raj RahejaInternational Business Machines Corp.

William P. RaneyNational Aeronautics and Space

AdminiStration

Gerald A. RathWichita State University

Philip L. Ray, Jr.Department of Commerce

Darrel RegierNational Institute of Mental Health

Neils ReimerStanford University

Peter ReimoldScarsdale, NY

Harold C. RelyeaCongressional Researce Service

129

Richard RettigIllinois Institute of Technology

Robert RichCarnegie-Mellon University

William RoeASARCO. Inc.

J. David RoessnerGeorgia Institute: of Technology

Gene RowlandNational Bureau of Standards

Robert W. RycroftGeorge Washington University

H. Charles SchulbergWestern Psychiatric Institute/Clinic

Perry SchwartzGeorgia Institute of Technology

Willis ShapleyWashington. DC

Steven S. SharfsteinNational Institute of Mental Health

Marvin SirbuMassachusetts Institute of Technology

Eugene B. SkolnikoffMassachusetts Institute of Technology

Joseph M. SnarponisNational Society of Professional Engineers

Roger W. StaehleUniversity of MinneSota

Stuart G. StearnsMerck & Co.. Inc.

Eric StoverAmerican Association for the

Advancement of Science

David SwanKennecott Corp.

Leonard SwernSperry Corp.

Ben Sykes ..Exxon Enterprises

Vernal TaylorFlorida A&M University

Albert H. TeichAmerican Association for the

Advancement of Science

130

John F. TormeyRockwell International

Louis G. TornatzkyNational Science Foundation

Elizabeth UseemNewton. MA

Gary VandenBosAmerican Psychological Association

Jon VeigelAlternative Energy Corporation

Betty VetterScientific Manpower Commission

William WebsterRCA Labs

William WellsAmerican Association for the

Advancement of Science

N. Richard WerthamerBecton-Dickerson Development

Corporation

John F. WilhelmInstitute of Electrical and Electronic

Engineers

F. Karl WillenbrockSouthern Methodist University

David WilltamSonGeorgetown Center for Strategic and

International Studies

C. R. WischmeyerBell Laboratories

Lawrence J. WolfHouston University

Theodore WoodCornell University

Christopher WrightCarnegie Institution

Craig WyvillGeorgia Institute of Technology

Jonathan YorkNortheast Guidance

Shoshana ZuboffHarvard Business School

142