RnD Stat Derek Bosworth Robert Wilson Alison Young

REVIEWS OF UNITED KINGDOM STATISTICAL SOURCES

Volume XXVII

RESEARCH ANDDEVELOPMENT


Other volumes in this series published by Chapman & Hall

xxmXXIVXXVXXVI

XXVIIxxvmXXIX

Agriculture, G. H. PetersLocal Govemment, J. M. GillespieFamily Planning, P. F. SelmanInternational Aspects 0/ UK Economic Activities, P. Bucldey and R. D.PearceResearch anti Development, D. L. Bosworth, R. A. Wilson and A. YoungThe Food Industries, J. Mark and R. StrangeDistribution, C. Moir and J. A. Dawson

Available from Chapman & Hall 2-6 Boundary Row, London SEI 8HNTelephone 071-522-9966

(No further titles will be published in this series)


Edited by M. C. FLEMINGProfessor of Economics ,Loughborough University

Volume xxvnRESEARCH ANDDEVELOPMENT

DEREK L. BOSWORTHManchester School of Management ,

UMISr

ROBERT A. WILSONInstitute for Employment Research,

University of Warwick

and

ALISON YOUNGOECD, Paris

Published for The Royal Statistical Society andThe Economic and Social Research Council

Springer-Science+Business Media, B.Y.

First edition 1993

ISBN 978-0-412-35640-7 ISBN 978-1-4899-2983-9 (eBook)DOI 10.1007/978-1-4899-2983-9

1993 Springer Science+Business Media DordrechtOriginally published by The Economic and Social Research Council in 1993.Softcover reprint of the hardcover 1st edition 1993

Typeset at the Oxford University Computing Service

Apart from any fair dealing for the purposes of research or private study, orcriticism or review, as permitted under the UK Copyright Designs and PatentsAet, 1988, this publication may not be reprodueed, stored, or transmitted, inany form or by any means, without the prior perrnission in writing of thepublishers, or in the case of reprographie reproduction only in accordancewith the terms of the Iicences issued by the Copyright Licensing Ageney inthe UK, or in aecordance with the terms of Iicences issued by the appropriateReproduetion Rights Organization outside the UK. Enquiries concemingreproduetion outside the terms stated here should be sent to the publishers atthe London address printed on this page.

The publisher makes no representation, express or implied, with regard tothe aeeuraey of the information eontained in this book and eannot aecept anylegal responsibility or liability for any errors or omissions that may be made.

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication data available

@J Printed on permanent acid-free text paper, manufaetured in aecordancewith the proposed ANSIINISO Z 39.48-199X and ANSI Z 39.48-1984

CONTENTS OF VOLUME XXVII

Foreword

Membership 0/ Joint Steering Committee

Editorial Introduction

Guide to the Series and How to Use it

Review No. 45: Research and Development Statistics

Subject Index

v

VB

IX

Xl

XV

1

219

FOREWORD

The Sources and Nature 0/ the Statistics 0/ the United Kingdom , produced under theauspices of the Royal Statistical Society and edited by Maurice Kendall, filled anotable gap on the library shelves when it made its appearance in the early post-waryears. Through aseries of critical reviews by many of the foremost national experts,it constituted a valuable contemporary guide to statisticians working in many fieldsas weil as a bench-mark to which historians of the development of statistics in thiscountry are likely to return again and again. The Social Science Research Council(now the Economic and Social Research Council) and the Society were bothdelighted when Professor Maunder came forward with the proposal that a revisedversion should be produced, indicating as weil his willingness to take on the oneroustask of editor (a task in which he was assisted from 1985 by Professor Fleming). Thetwo bodies were more than happy to act as co-sponsors of the project and to help inits planning through a joint steering committee. The result, we are confident, will bejudged a worthy successor to the previous vo1umes by the very much larger 'statisticspublic' that has come into being in the intervening years.

W. SOLESBURYSecretaryEconomic and Social Research Council

D. A. LIEVESLEYHonorary SecretaryRoyal Statistical Society

vii

MEMBERSHIP OF JOINT STEERING COMMITTEE

(December 1987)

Chairman: Miss S. V. Cunliffe

Representing the Royal Statistical Society:

Mr M. C. Fessey

Dr S. Rosenbaum

Mrs E. J . Snell

Representing the Economic and Social Research Council:

Mr I. Maclean

Miss J. Morris

Secretary: Mr D. E. Allen

ix

EDITORIAL INTRODUCTION

The series of Reviews 01 United Kingdom Statistical Sources is meant to serve a dualpurpose: to provide an authoritative guide to statistical sources in the UK (bothofficial and unofficial) and a critical appraisal of the nature and limitations of theavailable data. To maximise its usefulness as a source of reference, each volume inthe series folIows a standard format which incorporates a number of featuresdesigned to meet the varied needs of different users. A detailed guide for usersfollows this introduction but the key points to note for users in a hurry is that thetext, which pro vides a commentary on the sources and nature of the data, is followedby a Quick Reference List (QRL ) , which provides a quick and easy means ofidentifying what statistics are available, and a QRL Key to Publications which liststhe statistical source publications.

This volume may be regarded as a companion to volume 19 in the series onIntellectual Property Rights by D. L. Bosworth published in 1986. That volumecovered statistical sources relating to patents and other outputs associated with R&Dactivity. In contrast, this volume covers sources relating to the initial inputs ofresources into R&D. The importance of R&D as the mainspring of technologicalchange and economic growth requires no emphasis. Research on the determinants ofinvestment in R&D, on the interaction between such investment and economicactivities generally and international comparisons of R&D expenditure are acontinuing focus of attention. At the same time, the question of the appropriate roleof government in this field is a frequent topic of debate . But the lack of anauthoritative review of the sources and nature of the data available in the area hasbeen an important deficiency. This volume, therefore, is meant to fill a long-feltneed.

A few words must be said about the preparation and authorship of this volume.The work was initially undertaken by Alison Young of the Organisation forEconomic Co-operation and Development (OECD) but, regrettably, the pressure ofother duties denied her the time to complete the work and it remained dormant formany years . At the last moment, when the volume was about to be abandonded,Derek Bosworth and Robert Wilson agreed to take over and they have undertakenall the work required to complete it, including extensive redrafting and checking andthe preparation of much new material. As a consequence, the outcome is a volume inwhich it is not possible to attribute responsibility for the different parts of it toindividual authors and it appears, therefore, as the joint work of all three authors,each being named in alphabeticalorder.

Regrettably, this volume is one of the last in the series to appear under the jointsponsorship of the Royal Stat istical Society and the Economic and Social Research

xi

RESEARCH AND DEVELOPMENT STATISTICS

Council (ESRC). The work received the generous financial support of the ESRC(formerly the Social Science Research Council) from the start of the series in 1969until the end of 1987. During that time work on 25 volumes covering 43 reviewtopics was completed (although three of these volumes were not published until1988). Several reviews were then abandoned, but progress on four other volumes wassufficiently far advanced to make their completion seem feasible in a short space oftime without further financial support. Unfortunately, this optimism was misplaced.Although it has proved possible to avoid the abandonment of these volumes ,progress has been slow and has only been maintained with the aid of ad hoc fundingobtained from time to time to support either the authors' , or the editorial, work . Inthis respect we are particularly grateful to the Department of Trade and Industry formaking a small grant to help meet the editorial costs of this volume .

The original objective of the series was to try to cover every field of economic andsocial statistics in the Uni ted Kingdom; unfortunately its achievement proved to be amuch more lengthy task than was originally anticipated and the series is stillincomplete. However, thc need for such aseries and , moreover, aseries which isregularly updated remains as strong as ever. More than twenty areas are stilluncovered and many of the early volumes now stand in need of extensive revision.Although a valuable Guide to Official Statistics is now regularly published undergovernment auspices every two years or so, it does not provide a substitute for theseries because, valuable though it is, it is confined very largely to current sources ofofficial statistics and, perhaps more important, it ventures no appraisals to assist inthe use and interpretation of the available data. It is thus a complement to this seriesrather than a substitute for it. One must, therefore, hope that some further financialsponsorship may yet prove forthcoming which will enable the series to be continuedin one form or another.

Until the end of 1987, the series was directed by a Joint Steering Committee of theRoyal Statistical Society and the ESRC. It only remains here to express gratitude tothe members of the Committee, who directed the overall strategy with as admirable amixture of guidance and forbearance as any editors of such aseries could desire. Atthe same time they bear no responsibility for shortcomings in execution . Especialthanks are due to David Allen, the Secretary of the Committee, who was involvedwith the project almost as long and almost as closely as anybody. One must also paytribute to the work done by Professor W.F. Maunder as editor throughout the wholeperiod to 1987. A very great deal is owed to hirn both for the development of theoriginal idea and for his perseverance and determination in bringing so manyvolumes through to completion, as well as for helping to initiate work on so manyothers (including this volume) . Statistics users owe hirn a great debt.

The authors join me in thanking as weil all those who gave up their time to attendthe seminar which was held to discuss the first draft of the review and whichcontributed materially to improving the final version . We are most grateful toChapman and Hall Limited for their support and in particular to their productiondepartment, who put all the pieces together. The subject index entries have beencompiled by Mrs Marian Guest who acted as editorial assistant until October 1989.Special thanks are also due to Mr Ray Burnley of the Social Studies Data ProcessingUnit at the University of Exeter , who has given a great deal of help with the finaleditorial stages and again has masterminded our use of the Lasercomp System at

xii

EDITORIAL INTRODUCTION

Oxford University Computing Service, and to the latter for the use of this facility.Finally, we also wish to record our appreciation of the permission granted us toreproduce certain copyright material by the Controller of Her Majesty's StationeryOffice.

Michael FlemingLoughborough UniversityJune 1992

xiii

GUIDE TO THE SERIES AND HOW TO USE IT

The Scope and Nature of the Series

The purpose of the series is twofold . The primary aim is act as an authoritative workof reference to the sources of statistical material of all kinds, both official andunofficial, in the United Kingdom. The intention here is to enable the user todiscover what data are available on the subject in which he or she is interested andfrom where they may be obtained. The second aim is to provide a critical appraisalof the nature and limitations of the available data so that the user is able to interpretthem safely and avoid pitfalls in their use.

Data are regarded as available not only if published in the normal printed formatbut also if they are likely to be released to a bona fide enquirer in any other form,such as duplicated documents, computer print-out or even magnetic tape. On theother hand, no reference is made to material which, even if it is known to exist, isnot accessible to the general run of potential users. The distinction, of course, is notclear-cut and mention of a source is not to be regarded as a guarantee that data willbe released; in particular cases it may well be a matter for negotiation. The lattercaution applies with particular force to the question of obtaining computer print-outsof custom-specified tabulations.

The intention is that the source for each topic should be reviewed in detail, andthe brief supplied to authors calls for comprehensive coverage at the level of'national interest' . This term does not denote any necessary restriction to statisticscollected on a national basis (stilliess, of course, to national aggregates) but it meansthat sources of a purely local character, without wider interest in either content ormethodology, are excluded. Indeed, the mere task of identifying all material of thislatter kind is an impossibility. The interpretation of the brief has obviously involveddiscretion and it is up to the users of these reviews to say what unreasonable gapsbecome apparent to them. They are cordially invited to do so by communicatingwith the Editor.

The Format and Content of Reviews

To facilitate the use of the series as a work of reference, a common format isadopted for each review. This involves the incorporation of six standard features :

I. Text2. Quick Reference List (QRL)3. QRL Key to Publications4. Bibliography

xv

RESEARCH AND DEVELOPMENT STATISTICS

5. Subject Index6. Specimen Forms and Questionnaires

Each of these is described in turn below.

1. The Text . This is designed , in so far as varying subject matter permits, to folIow astandard form of arrangement, covering introductory material on the activity coveredand its organisation, core material on the available sources and a discussion ofdesirable improvements.

The introductory material is meant to give a c1ear background understanding ofhow data are colIected, what is being measured, the stage at which measurements aremade , what the reporting units are, the channels through which returns are routedand where they are processed . Coupled with this is a discussion of specific problemsof definition and measurement. This is folIowed by core sections or chapters onavailable sources. Wherever possible these are arranged according to subject (ratherthan source) . But in practicc thcy may be arrangcd according to the author'sdiscrction - by origin, by subject subdivision, or by type of data - as there is toomuch heterogeneity between topics to permit any imposition of complete uniformityon alI authors. A final chapter is devoted to a discussion of general shortcomingsand desirable improvements. In case a contrary expectation should be aroused, itshould be said that authors have not been asked to produce a comprehensive planfor the reform of statistical reporting in the whole of their field. However , a reviewof existing sources is a natural opportunity to make some suggestions for futurepolicy on the colIection and publication of statistics within the scope of their reviewand authors have been encouraged to take fulI advantage of it.

2. The Quick Reference List (QRL) . This provides a detailed list of alI the series andcategories of data that are available and, again, is generalIy arranged according tosubject. It also includes cross-references to the sections of the text in which the dataare discussed as welI as publication references. Each publication shown as a da tasource is given aserial number and the prefix 'QRL'

3. The QRL Key to Publications. This gives fulI details of the publications shown asdata sources in the QRL.

4. The Bibliography. This gives references to works discussing wider aspects of theactivity and the statistics including methodology. These publications are identified byaserial number and the prefix 'B'.

5. The Subject Index. This acts as conventional line of inquiry on textual references.But an important feature is that it is a computerised system. For an individualreview this introduces the possibility of easy permutation of entries and this is fulIyutilised in order to facilitate search by giving as many variants as possible . The indexis prepared in the editorial office, not by the author.

6. Specimen Forms and Questionnaires. FinalIy, specimen copies of the moreimportant returns or forms used in data colIection are reproduced, as appropriate, asappendices so that it may be seen what tabulations it is possible to make as welI ashelping to c1arify the basis of those actualIy available.

XVI

GUIDE TO THE SERIES AND HOW TO USE IT

How to Use the Series

As we have indicated, the standard format adopted for each review in the series isdesigned expressly to facilitate its use for reference purposes. The features which itincorporates provide three possible 'ways in' for the user. These are :

I . The Contents List2. The Quick Reference List3. The Subject Index

For users most interested in discovering qu ickly whether or not a particular sort ofdata is available and where it is published, the Quick Reference List should be themost useful feature to consult first. Inc1uded within the list are cross-references todiscussions and descriptions of the data in the text and references to the publicationsources for the data (QRL reference) . To facilitate its speedy use, it is arranged bysubject as far as possible and the list is itself preceded by its own summary contentslist.

The contents list provides, of course, a summary indication of the subject matterof each chapter but also shows the main sub-divisions into which it is divided. It isuseful, therefore, for anyone interested in locating material in the book on broadsubject areas, rather than more specific statistical sources.

The subject index provides an alternative means of locating information but onewhich is intermediate between the Quick Reference List and the contents list. It givesreferences to the subject matter at a more detailed level than the contents list, but itdoes not duplicate the degree of detail of the entries, relating to particular categoriesof statistical data, given in the QRL.

xvii

45: RESEARCH AND DEVELOPMENTSTATISTICS

D. L. BOSWORTH

Professor of EconomicsSchool of Management

University of ManchesterInstitute of Science

and Technology

R. A. WILSON

Principal Research FellowInstitute of Employment Research

University of Warwiek

ALISON J. YOUNG

Organisation for EconomicCo-operation and Development

Paris

REFERENCE DATE OF SOURCES REVIEWED

This review is believed to represent the position, broadly speaking, at December1987. Major changes since then are noted in the Addendum at the end of thisvolume .

2

CONTENTS OF REVIEW 45

Abbreviations and Acronyms 9Acknowledgements 12

1 Introduction and General Concepts 131.1 Coverage of the Review 131.2 Sources and Methodology 141.3 Definition of Research and Development 15

1.3.1 Definitions 151.3.2 Distinctions Between R&D and Other Activities 161.3.3 Distinction Between Research and Higher Education and Training 171.3.4 Distinction Between R&D and Other Scientific and 18

Technical Activities1.4 Scientific and Social R&D 191.5 Expenditure and Employment 19

1.5.1 Expenditures 201.5.2 Manpower 20

1.6 R&D Content and "M ainly R&D" Accounting 201.6.1 Applied to Performing Units 201.6.2 Calculation of Full-time Equivalents 21

1.7 Performance and Source of Funds 211.8 Occupation versus Level and Type 0/ Qualification 22

1.8.1 Occupation 221.8.2 Level and Type of Qualification 231.8.3 Occupation crossed with Level and Type of Occupation 23

1.9 Problems of Comparison 241.9.1 Between Different Sources 241.9.2 Problems of Comparison over Time 251.9.3 Problems of International Comparison 27

2 Tbe National Survey2.1 Origins2.2 Basic Characteristics

2.2.1 The National R&D Effort2.2.2 Major Fields of Science2.2.3 Method of Accounting for R&D

3

292930303030

4 RESEARCH AND DEVELOPMENT STATISTICS

2.3 System 0/ Sec tors and Sub-Sectors 312.3.1 Industry 312.3.2 General Government Sector 332.3.3 Higher Education 352.3.4 Other 362.3.5 Abroad 37

2.4 Surveys and Publications 382.4.1 Introduction 382.4.2 Frequency and Time Period 382.4.3 Survey Method 392.4.4 Publications 40

2.5 Expenditure Data 412.5.1 Total Intramural Expenditure on R&D 412.5.2 Total Intramural Expenditure by Type of Cost 412.5.3 Current R&D Expenditure by Type of Activity 422.5.4 Sources of Funds 432.5.5 Extramural Expenditure 462.5.6 Gross Expenditure in R&D 47

2.6 Employment Data 472.6.1 Total R&D Manpower 472.6.2 Scientists and Engineers 492.6.3 Technicians 502.6.4 Other Supporting Staff 50

2.7 Price Indices 512.7.1 Industry 512.7.2 Public Sector 522.7.3 Government 522.7.4 Higher Education and Other 53

2.8 International Comparisons 532.8.1 National Sources 532.8.2 International Sources 53

3 Net Government Expenditure on R&D 553.1 Origins 553.2 Basic Characteristics 56

3.2.1 General Budgetary Procedures and Publications 563.2.2 Estimates and Outturn 573.2.3 General Definition of Net Expenditure 583.2.4 Difference Between Net Expenditure and Gross Expenditure 583.2.5 Major Fields of Science 603.2.6 Coverage and Method of Identifying R&D 60

3.3 Surveys and Publications 643.3.1 Surveys 643.3.2 Publications 64

3.4 Data A vai/able 653.4.1 Institutional Classifications 653.4.2 Functional Classifications 66

CONTENTS 5

3.5 Derived Series 673.5.1 Price Indices 673.5.2 Comparison with Other Series 67

3.6 International Comparisons 673.6.1 EEC 673.6.2 OECD 683.6.3 Special Problems 68

4 Tbe Science Budget 714.1 Origins 714.2 Basic Characteristics 71

4.2.1 Coverage 714.2.2 Main Fields of Science 724.2.3 Method of Accounting for R&D 72

4.3 Surveys and Publications 734.3.1 The Science Budget Proper 734.3.2 Additional Sources 73

4.4 Data A vailable 734.4.1 Expenditure: Science Budget Proper 734.4.2 Tables in the CSO and ACSP Reports 744.4.3 Tables in the Annual Review 744.4.4 Research Council Annual Reports 74

4.5 Miscellaneous Sources 784.6 Sources 0/ International Comparisons 78

5 Individual Departmental R&D Reports 815.1 Origins 815.2 Basic Characteristics 83

5.2.1 General Characteristics 835.2.2 Methods of Accounting for R&D 845.2.3 Net and Gross Expenditure 845.2.4 Classification Units for Central Government R&D Expenditure 84

5.3 Surveys and Publications 855.3.1 List of Reports 855.3.2 Reports of the Major R&D Spending Departments 85

5.4 Data Available 855.5 Department 0/ Energy 86

5.5.1 General Pattern of the Department's R&D 865.5.2 Tables in the Annual Report 865.5.3 Comparison with Other Series 86

5.6 Department of Industry 885.6.1 General Pattern of the Department's R&D 885.6.2 Tables in the Annual Report 88

5.7 Ministry 0/ Agriculture, Fisheries and Food 935.7.1 General Pattern of the Ministry R&D 935.7.2 Tables in the Annual Report 93


5.8 Departments 0/ the Environment and Transport5.8.1 General Pattern of the Departments' R&D5.8.2 Tables in the Annual Reports

5.9 Department 0/ Health and Social Security5.9.1 General Pattern of the Department's R&D5.9.2 Tables Available in the Departmental Report

5.10 Home Office5.10.1 Department's R&D Activities5.10.2 Data Available from Other Sources

5.11 Overseas Development Administration5.11.1 General Pattern of the Department's R&D5.11.2 Tables in the Annual Report5.11.3 Comparison with Other Sources

5.12 Department 0/ Agriculture and Fisheries for Scotland (DAFS)5.12.1 DAFS R&D Activities and Related Reports5.12.2 Data Available from Other Sources

5.13 Information on Individual Departments from the Annual Review

6 Triennial Scientific and Technological Manpower Surveys6.1 Origins6.2 Basic Characteristics

6.2.1 Types of Manpower Included6.2.2 Major Fields of Science6.2.3 Method of Accounting for R&D

6.3 System 0/ Sectors and Sub-Sectors6.3.1 Industry6.3.2 Government6.3.3 Higher Education Sector6.3.4 Summary of Unsurveyed Sectors6.3.5 National Territory

6.4 Surveys and Publications6.4.1 General Remarks6.4.2 Surveys

6.5 Data Available and Comparison with Other Sources6.5.1 Industry6.5.2 Government6.5.3 Higher Education and Other

6.6 Comparisons with Other Series6.7 International Comparisons

7 Miscellaneous Sources7.1 Introduction7.2 Industry

7.2.1 Private Industry7.2.2 Nationalised Industries7.2.3 Research Associations

959595989898

100100100100100101101101101102103

105105106106107107107108108108109109109109109111111III112112113

115115115115117117

CONTENTS

7.3 Government7.3.1 The Estimates7.3.2 Annual Reports7.3.3 Special Reports

7.4 Higher Education Sector7.4.1 Annual Statistics7.4.2 Special Reports

7.5 Other Bodies

8 Conclusions8.1 Introdu ction8.2 Industry8.3 Government and Higher Education8.4 Future Developments8.5 Conclusions

Quick Reference List Table of ContentsQuick Reference ListQuick Reference List Key to Publications

Bibliography

Appendix: Specimen Forms and Questionnaires

Addendum

Subject Index

7

118118118119120120120120

123123123124125126

127128170

191

197

215

219

LISTS OF TABLES AND FIGURES

List of Tables in the Text

3.1 Coverage of Net outturn data 1961/62-1965/66 compared withprogramme headings 1966/67-1970/71 62

5.1 Relation between PESC headings and departmental R&Dspending - 1979/80 82

5.2 Functional Classification for R&D Expenditure of the Departmentof Industry 89

5.3 Functional Classification, Departments of the Environmentand Transport 96

5.4 Functional Classification of R&D used by the Department of Healthand Social Security 99

5.5 Functional Classifications used by the Overseas Development Agency 102

5.6 Civil 'Oriented' R&D by Central Government Departments 1979/80 103

List of Figures in the Text

3.1 Flows of Funds: Net and Gross Expenditure on R&D

8

59

ABRCACARDACSPAEAAFRCARCBECBGBSCBSOBTGCBICECCEGBCEICERN

CMEACNAACNRCNRSCOICRESTCSECSIICSMCSOCSPDAFSDEDESDOEDHSSDIDSIRDTIEEC

ABBREVIATIONS AND ACRONYMS

Advisory Board for the Research CouncilsAdvisory Council for Applied Research and DevelopmentAdvisory Council on Scientific PolicySee UKAEAAgricultural and Food Research CouncilAgricultural Research CouncilBusiness Education CouncilBritish GasBritish Steel CorporationBusiness Statistics OfficeBritish Technology GroupConfederation of British IndustryCommission for the European CommunitiesCentral Electricity Generating BoardCouncil of Engineering InstitutionsLe Centre European pour Recherche Nucleaire (European Commissionfor Nuclear Research)Council for Mutual Economic AssistanceCouncil for National Academic AwardsConsiztio Nazionale delle Ricerche (Italy)Conseil National pour la Recherche ScientifiqueCentral Office of InformationCornite de Recherche Scientifique et Technique (EEC)Certificate of Secondary EducationCentre for the Study of Industrial InnovationMemorandum by the Chief Secretary to the TreasuryCentral Statistical OfficeCouncil for Scientific PolicyDepartment of Agriculture and Fisheries ScotlandDepartment of EmploymentDepartment of Education and ScienceDepartment of the EnvironmentDepartment of Health and Social SecurityDepartment of IndustryDepartment of Scientific and Industrial ResearchDepartment of Trade and IndustryEuropean Economic Community

9

10

ELDOESAESFESRCESROEVAFBIFEFSMFTEGCEGCSEGDPGERDGNERDGPOGVFHEHMSOHNCHNDHSEICLISYJCOMAFFMinTechMLHMODMPBWMRCMSCNABS

NCNCBNCCNERCn.e.sNHSNIRNSNRDCNSENSFODAOECDOEEC

RESEARCH AND DEV ELOPMENT STATISTICS

European Launeher Development OrganisationEuropean Space AgencyEuropean Science FoundationEconomic and Social Research CouncilEuropean Space Research OrganisationEuropean Stat istical Vnit of AccountFederation of British IndustryFurther EducationMemorandum by the Financial Secretary to the TreasuryFulI Time EquivalentsGeneral Certificate of EducationGeneral Certificate of Secondary EducationGross Domestic ProductGross Domestic Expenditure on Research and DevelopmentGro ss National Expenditure on Research and DevelopmentGeneral Post OfficeGeneral Vniversities FundHigher EducationHer (His) Majesty's Stationery OfficeHigher National CertificateHigher National DiplomaHealth and Safety ExecutiveInternational Computers UdInternational Statistical YearsJoint Consultative OrganisationMinistry of Agriculture , Fisheries and FoodMinistry of TechnologyMinimum List HeadingMinistry of DefenceMinistry of Public Building and WorksMedical Research CouncilManpower Services CommissionNomenclature for Analysis and Comparison of Scientific Programmesand BudgetsNature ConservancyNational Coal BoardNature Conservancy CouncilNatural Environment Research Councilnot elsewhere specifiedNational Health ServiceNational Institute for Research in Nuclear ScienceNational Research and Development CorporationNatural Science and EngineeringNational Science FoundationOverseas Development Agency (Overseas Development Administration)Organisation for Economic Cooperation and DevelopmentOrganisation for European Economic Cooperation

ONCONDPESCPNPPSIQSERAR&DRCRCSRD&D

RICRRCRSCRSES&ES&TSERCSHHDSICSOECSRCSSHSSRCSSTSTIUUGCUKUKAEAUNUNESCOWHO

ABBREVIATIONS AND AC RONYMS

Ordinary National Cert ificateOrdinary National DiplomaPublic Expenditure Survey CommitteePrivate Non-ProfitPolicy Studi es InstituteQualified Scientists and Engineers(Industrial) Research AssociationsResearch and DevelopmentResearch CouncilRo yal College of ScienceResearch Development and Design (Research, Development andDemonstration)Royal Institute of ChemistryResearch Requirement CommitteeRoy al Society of ChemistryResearch and Development Scientists and EngineersScientists and EngineersScientists and TechniciansScience and Engineering Research CouncilScottish Horne and Health DepartmentStandard Industrial ClassificationStatistical Office of the European CommunitiesScience Research CouncilSocial Science and HumanitiesSocial Science Research CouncilSelect Committee on Science and TechnologyScience and Technology Indicators Unit (OECD)University Grants CommitteeUnited KingdomUnited Kingdom Atomic Energy AuthorityUnited NationsUnited Nations Educational, Scientific and Cultural OrganisationWorld Health Organisation

11

ACKNOWLEDGEMENTS

This contribution to the ESRC/RSS series on UK Statistical Sources has evolvedthrough a number of stages. Alison Young began the first draft which was presentedat a seminar in January 1981. Derek Bosworth and Rob Wilson have built on thisearly draft, updating and adding to the material, and take responsibility for anyremaining errors.

We would like to note the special role that the Department of Trade and Industryhas played in the collection of R&D survey data in the UK. In addition, our thanksgo to John Bowles from the DTI for his help on a number of occasions.

This book would not have been possible without the dedication of the team ofworkers at Exeter, particularly Marian Guest and Professor Maunder until 1987 andthen Linda Lilburne (working under Professor Fleming's direction atLoughborough). Thanks also go to Peter Miller and Samantha Wilson for the librarysearches that they undertook and the collation of information about many of thesources that appear in this volume, and to Joan Bosworth for her efforts in helpingto organise the QRL.

12

CHAPTER I

INTRODUCTION AND GENERAL CONCEPTS

1.1 Coverage of the Review

Research and Development (R&D) is an activity which is carried out over a widerange of sectors and sub-sectors of the economy. In consequence there are amultitude of statistical sources which contain some data on resources (money andpersonnel) used for R&D. These sources may be divided into four broad classes. Thefirst class includes regular official surveys designed specifically to collect R&Dstatistics for a whole economic sector (government, industry, the universities, etc.)using standard instructions and classifications. These 'R&D surveys proper' arediscu ssed in Chapters 2 and 3. The second group of sources comprises regular(usually annual) reports by official bodies (for example the Research Councils) ontheir R&D activities which include statistical tables . These 'annual reports' arediscussed in Chapters 4 and 5. The third group of sources covers other regularofficial statistical surveys and publications which yield some R&D data (for examplethe Defence estimates, and the triennial surveys of qualified manpower) and they arediscussed in Chapters 6 and 7. Chapter 7 also includes a fourth group of sourceswhich comprises ad hoc official reports quoting R&D statistics (for example reportsof the Select Committee on Science and Technology) and various non-official studiesand reports. The coverage of this fourth class, in particular, is illustrative rather thanexhaustive.

The underlying idea of the review is to use the specific R&D surveys described inChapters 2 and 3 as a framework and in later chapters to try and explain where thedata from other sources fit into this framework and how far their specifications andcharacteristics vary from those of 'R&D da ta proper' as defined by the relevantnational and international standards. This normative approach means that data inthe later chapters are evaluated using criteria which were not necessarily relevant tothe aims of the authors of the sources concerned. Such evaluations should not beread as criticisms.

At the time of writing, full national R&D surveys had been undertaken regularlyfrom 1955 (with a triennial or greater frequency). The review thus covers data forthe period 1955 to 1986 but with more detail for the period after 1966 than for theearlier years .

It is perhaps useful at this stage to point the reader to one or two other reviewswhich cover R&D statistical sources. Particularly useful in this respect are theCabinet Office guide to sources of information about UK Government R&D:Government Research and Development: A Guide to Sources of Information [B.49] andAnnual Review of Government Funded R&D 1985 [QRL.51] Annexe D. In addition ,there is the Guide to Ojjicial Statistics, which is regularly updated (e.g. Guide to

13


Official Statistics No. 4 [B.50] pp. 340-342 and Guide to Official Statistics No. 5[B.51] p. 111). In particular, Guide No. 4 sets out the general sources of R&D data,sources relating to government funded R&D, Research Councils' activities andindustrial research, in four separate sections.

This Review of statistical sources focuses primarily on R&D inputs (i.e.expenditure, personne1, etc.), rather than R&D outputs. The latter are partly dealtwith e1sewhere, in the companion Review of UK Intellectual Property Statistics byBosworth [B.7].

1.2 Sources and Methodology

There is no handbook of methodology for R&D statistics in the United Kingdom .Information has to be gleaned from the notes and appendices to the various studiesand, where these are available, from the questionnaires used in the major surveysdiscussed in Chapters 2 and 3. The basic source for R&D methodology in general isthe 'Proposed Standard Practice for Surveys of Research and ExperimentalDevelopment' [B.56], otherwise referred to as the 'Frascati Manual' , published byOECD. The relevant UK authorities have played an important role in thequinquennial revisions of this Manual and with one or two exceptions UK R&Ddata are collected in line with 'Frascati Standards' . Unless elsewhere specified,references in the present review are to the 1981 version of the Manual, [B.56].

Other international organisations have issued complementary methodologiesnotably the various European Economic Community standards for government R&Dbudgets, which will be discussed in Chapter 3 and the 'Nordic Manual' issued in1974 [B.52], which, although principally relevant to Scandinavia, contains someinteresting general methodological discussion.

UNESCO has also published a number of norms and methodological documentsdealing with the measurement of resources devoted to R&D [B.66] and Manual onthe National Budgeting 0/ Scientific and Technological Activities [B.55]. UNESCO isinterested in the whole area of scientific and technical activities. See also TheMeasurement 0/ Scientific and Technological Activities [B.16] and Measurement 0/Output 0/ Research and Experimental Development [QRL.13]. The basic definitionsof R&D are compatible with those used by the CEC and by OECD but thec1assifications have been specially designed to permit comparison between marketand planned economies and also between developed and developing countries. Onthe whole they appear to have had little effect on the way data are actually collectedin the United Kingdom.

The rest of the present chapter is, therefore, based mainly on the internationalstandards outlined in the Frascati Manual [B.56] either as defined in the manualitself or as paraphrased in national survey questionnaires. The primary aim is toidentify the special problems of measuring resources devoted to R&D in the UnitedKingdom.

INTRODUCTION AND GENERAL CONCEPTS

1.3 Definition of Research and Development

15

1.3.1 Definitions

The official international definition of R&D is currently as follows: 'Research andExperimental Development comprise creative work undertaken on a systematic basisin order to increase the stock of knowledge , including knowledge of man, cultureand society, and the use of this stock of knowledge to devise new applications'(Frascati Manual, [B.56] p. 15), However, the UK authorities have generallypreferred to retain an earlier international definition of 'creative work undertakenon a systematic basis to increase the stock of scientific and technical knowledge andto use this stock of knowledge to devise new practical applications.' [QRL.153] p. 3.

Furthermore R&D is, in principle, generally defined as being the sum of threeexhaustive and mutually exclusive activities : basic research, applied research andexperimental development, described in the UK surveys as follows:

i) Basic Research is original investigation undertaken in order to gain newscientific knowledge and understanding. It is not primarily directed towardsany specific practical aim or application.

ii) Applied Research is also original investigation undertaken in order to gain newscientific or technical knowledge . It is, however, directed towards a specificpractical aim or objective.

iii) Experimental Development is the use of scientific knowledge in order toproduce new or substantially improved materials, devices, products, processes,systems or services.' [QRL.153]

The aims of government R&D expenditure are set out in Annual Review ofGovernment Funded R&D [QRL.51], p. 5 :i) Advancement ofScience. Work funded primarily in order to increase human

knowledge, i.e. to advance scientific understanding of natural phenomena.This research contributes to the scientific base of the nation and, although,originally funded with no specific application in view, much of it eventuallyresults in lang term benefit through the eventual application of knowledgegained. The heading is equivalent to the OECD term 'basic research' .

ii) Support for Policy Formulation and Implementation. Applied research andexperimental development carried out in order to meet the Government's ownneeds for knowledge of improved products or processes.

iii) Improvement oJ Technology. Applied R&D, funded by GovernmentDepartments but often carried out within industry, to advance the technologyof different sectors of the UK economy.

iv) Support for Procurement Decisions. Applied R&D which contributes to thespecification and development of goods and services required by Departmentsand to equip the purchasing Department to act as an informed buyer.

v) Support for Statutory Duties . Applied R&D which assists Departments tocarry out statutory responsibilities or other obligations.

vi) Support for Other Activities. Applied R&D which cannot be classified underthe other headings.

An enormaus amount of time and effort has been spent in the Uni ted Kingdomand elsewhere on discussing and re-drafting all these theoretical definitions . What is


more important from a practical point of view is to ask how R&D should becircumscribed for the purposes of measurement. Which activities does it include andwhich does it exclude?

R&D needs to be distinguished i) from post-R&D innovative activities andproduction in industry, ii) from education and training in the higher education sectorand iii) from a number of other scientific and technological activities. In all thesecases a useful criterion, first used in the United States, but later quoted in the'Frascati Manual' and paraphrased in the instruct ion to several UK surveys, is tha t:"The guiding line to distinguish R&D activity from non-research activity is thepresence or absence of an element of novelty or innovation. Insofar as the activityfollows an established pattern, it is not R&D. Insofar as the activity departs fromroutine and breaks new ground, it qualifies as R&D.' (See notes to the variousquestionnaires). The word 'innovation' is perhaps misplaced as we shall see below,but otherwise this is a useful guide.

1.3.2 Distinction Between R&D and Other Innovative ActivitiesScientific and technological innovation consists of all those scientific, technical,commercial and financial steps needed for the successful development and marketingof new or improved processes and equipment (or the introduction of a new approachto a social service). R&D is only one of these steps and one of the main problemswhen surveying R&D in industry is to decide where it ends and where the next stageof the process begins. For example, should the costs of prototypes and pilot plant beincluded in R&D? What about the operating costs incurred during pilot productionruns? A theoretical discussion of these points will be found in the Frascati Manual[B.56] (pp. 69-77).

In the United Kingdom virtually all the industrial R&D data discussed in thereview comes from the same source, i.e. aseries of surveys made by the Departmentof Trade and Industry (and its predecessors). In practice, therefore, the instructionsfrom this survey (which will be discussed at length in Chapter 2) effectively definethe borderline.

They specify that R&D includes 'the prototype or pilot-plant stage and all workdone on development contracts with government departments, the Atomic EnergyAuthority and similar public bodies. Firms in the aerospace industry should includeexpenditure on development batches'. Respondents are also told to exclude:i) Routine testing and analysis of all kinds, whether for control of materials ,

components or products, and whether for control of quantity or quality .ii) Market research, operational research, work study, cost analysis, management

science, surveying, 'trouble shooting'.iii) Royalty payments for the use of the results of research and development

unless required as an essential part of the research and developmentprogramme within the unit.

iv) Trial production runs where the primary objective is not further improvementof the product.

v) Design costs to meet changes of fashion and artistic design work.vi) Legal and administrative work in connection with patent applications, records

and litigation . Work involved in the sale of patents and licensing

INTRODUCTION AND GENERAL CONCEPTS 17

arrangements. Experimental work performed solely for the purposes of patentlitigation.

The specifications used for surveys of government R&D which yield informationabout government R&D contracts placed with industry indicate the position of theline drawn between development and production:

i) If the primary objective is to make further improvements on the product orprocess, the work comes within the definition of R&D. If, on the other hand,the product or process is substantially 'set' and the primary objective is todevelop markets or to do pre-production planning, or to get the productionprocess going smoothly, then the work is no longer R&D.

ii) The design, construction and testing of prototypes normally falls within thescope of R&D.

iii) Development inc1udes the engineering activity required to advance the designof a product or process to the point where it meets specific functional oreconomic requirements and can be turned over to the manufacturing units.The costs of trial production runs should not normally be inc1uded as R&Dunless problems that are encountered require further R&D work.

1.3.3 Distinction Between Research and Higher Education and TrainingResearch is c1early one of the prime functions of universities (and, to a lesser extent,of other institutions of higher education) but it is sometimes extremely difficult todistinguish resources devoted to research from those devoted to education andtraining, especially in the cases of 'General University Funds and postgraduatestudies' (see Frascati Manual [8.56] p. 59-64 and 174-176).

1.3.3.1 General University Funds. Some university R&D projects are financed byspecific contracts, grants or other earmarked funds, but a large proportion of suchresearch is undertaken by university staff interspersed with their teaching andadministration activities and is financed from 'General University Funds', i.e. in theUnited Kingdom the grant from the University Grants Committee (UGC) . Suchresearch is not separately accounted for in university statistics and has to beestimated on the basis of the share of time which staff spend on research as opposedto their other activities, (see [8.68] and Bowles in Economic Trends [QRL.8], p. 95).The basis and coverage of such estimates has varied in the period under review andalso between different series.

1.3.3.2 Postgraduate Studies . Some postgraduate studies are based almost entirelyon course work and have a negligible research content. However, high levelpostgraduate courses nearly always require the student to undertake a substantialpiece of independent and novel work. According to international standards (FrascatiManual [B.56] p. 60-64) such work is bona-fide research and should be inc1uded intotal university R&D. However, there are two schools of thought in the UnitedKingdom, one of which agrees that resources devoted to research by postgraduatesshould as far as possible be inc1uded, and a second which believes that the activitiesconcerned are training in research and not research proper and thus the resourcesinvolved should not be credited to R&D.

Prior to the middle 1960s this question does not seem to have been examined and,on the whole, postgraduate studies were inc1uded in official UK R&D data rather


than excluded. This type of problem is mentioned, for example, in [QRL.153] , p. 6.Since the mid-1960s two sets of series of data have existed, one of which includespostgraduate studies as far as possible and the other which excludes them.

The resources devoted to postgraduate research should include time spent by thestudents and by their superiors and by supporting staff (technici ans, typists, etc.). Inexpenditure terms they comprise the relevant labour costs of student, supervisor andsupporting staff plus the other costs of the research project (materials, overheadsetc .). Some of these costs may be met from earmarked research contracts and grants,but others will be met from 'General University Funds' . It is for this reason thatthere are two series of estimates for 'scientific' university R&D financed from theTreasury Grant which appear in different publications in the Uni ted Kingdom. Theupper one contains an estimate for postgraduate research, the lower one does not.However, this does not 'include the labour costs of the postgraduates themselves asin the United Kingdom they are usually supported by means of individual grants. '(They are rarely, as is the general case in continental Europe, on the universitypayroll.) The Research Councils are the main providers of such grants although thereare other public and private sources. Both the Research Councils and the othersources pay these grants direct to the students concerned. The Research Councilsalso pay a 'Research Training Support Grant' to the university in respect of thestudents they finance. It follows that there are also two series of data for researchsupport by the Research Councils, one induding payments to , and in respect of,postgraduate students and the other excluding them . In the cases of the Agricultural,Medical and Natural Environment Research Councils the difference is small but ismueh larger for the Scienee and Engineering Research Council (SERC) and theEconomic and Social Research Council (ESRC), the latter was formerly the SocialScience Research Council (SSRC). For an early description of the structure of theResearch Councils and their links with other government agencies, see Rose andRose , Science and Society [B.29], p. 101. A more recent listing can be found inAnnual Review 0/ Government Funded R&D, 1985 [QRL.51].

1.3.4 Distinction Between R&D and Other Scientific and Technical ActivitiesThe following scientific and technical activities, some of which have already beenmentioned, are also formally excluded from R&D :

i) All legal administrative work in connection with patent applications, recordsand litigation. Work involved in the sale of patents and licensingarrangements. Experimental work performed solely for the purpose of patentlitigation.

ii) Routine testing and analysis of all kinds, whether for control of materials,components or products, and whether for control of quantity or quality.

iii) Scientific information activity which comprises all aspects of communicationamongst scientists (e.g. the publication, dissemination and translation ofinformation resulting from research and development) . Generallibrary andtechnical advisory services are included under this heading.

iv) General or broad-purpose data eollection (e.g. geologieal and geophysiealsurvey work, mapping and exploration activities, including those of oil andmineral eompanies; hydrographie and oceanographic survey work of a routine


nature not specifically re1ated to the development of new knowledge ortheories; daily meteorological records, monthly production statistics, thecollection and arrangement of specimens for museums, zoological gardens andso forth).

v) Testing and standardisation inc1uding such public and quasi-public functionsas the establishment of standards, calibration of secondary standards andnon-routine quality testing which are separately identifiable from research anddeve1opment.

vi) Market researchThe above definitions are taken from various UK questionnaires. They cover

approximately the same ground as the current international standards except that nomention is made of the exc1usion of 'specialised medical care' and 'policy-relatedstudies' and the inc1usion of 'routine space exploration', etc. (Frascati Manual [B.56]p.52, 54 and 67), although two of them are mentioned in the 1980 governmentquestionnaire .

1.4 Scientific and Social R&D

It is usual when collecting R&D statistics to make a distinction between R&D in thenatural sciences and engineering (NSE) and in the social sciences and humanities(SSH). Until relatively recently in the UK, data were collected for the NSE only.This is usually referred to as 'scientific' R&D. The collection of data for R&D in the'social sciences' dates from the 1970s.

It is almost certainly more difficult to measure R&D in the social sciences than inthe natural sciences and engineering, but the UK authorities seem to have sufferedparticularly serious qualms about doing so. The problems concerned are dealt within Annexe 1 to the Frascati Manual [B.56], which was written by Jeremy M .MitcheII, a British consultant, and at greater length in a mimeo by the same authoravailable from OECD.

'Scientific research' covers 'rnedical sciences' and other science, agriculture,engineering and technology. Adefinition of 'social sciences' is given in the officialsurveys discussed in Chapters 2 and 3. It c1early inc1udes economics and c1earlyexc1udes the humanities. However, the CSO ([QRL.152], [QRL.153], p. 130) suggeststhat it is a residual, i.e . all R&D which is not covered by the 'disciplines covered inscientific R&D' including multi-disciplinary work. Note that the medical scienceshave not always been inc1uded in 'science and technology', especially in earliermanpower series (Chapter 6).

1.5 Expenditure and Employment

Resources devoted to R&D can be measured both in terms of expenditure andemployment. Most of the data available for the United Kingdom are forexpenditure. Very little in the way of employment data is available before the 1970sand even now whole sectors are excluded.


1.5.1 ExpendituresTheoretically, R&D expenditure should cover both current and capital expenditurebut should exclude depreciation allowances. This generally applies to most R&Dexpenditure da ta collected in the Uni ted Kingdom. However, a special problemoccurs in the government sector with respect to materials, services and facilities usedin the performance of R&D by government establishments, but paid for on thegeneral budget of the Department of Environment (earlier the Ministry of PublicBuildings and Works). In R&D surveys proper, imputed sums are generally includedfor these , but such sums are often excluded in ancillary sources.

1.5.2 ManpowerTheoretically, manpower indicators should include all persons employed directly onR&D plus R&D managers and staff providing direct services, such as clerical staff.All the above should be included, regardless of their level of responsibility or of thetype or level of education they have received . Those providing indirect services, suchas canteen staff, commissionaires, office cleaners, etc., should be excluded. However,such ' total R&D manpower' data are rarely available before the 1970s. What dataare available are usually only for scientists, engineers and technicians.

1.6 R&D Content and 'Mainly R&D' Accounting

1.6.1 Applied to Performing Units

R&D is not just what research institutes or laboratories do. It is both less and morethan this , since it is unlikely that many of the units included in an R&D survey willhave only one activity. Most perform a mixture of R&D and other scientific andtechnological activities or R&D and production activity.

Thus the identification of R&D may involve three stages (Frascati Manual [B.561,p. 147).

i) The 'identification and measurement of the total activity of all specialisedR&D units.

ii) Subtraction from this total of their activity which is non-R&D.iii) Addition to this of any R&D activity performed outside research units , e.g. in

production units, education departments or hospitals.'If all these stages are followed through then only the 'R&D content' of the unit's

activities will be included. If, however, only the first stage is attempted, then 'mainlyR&D' accounting has been used . For example, both approaches will be found in thecase of the Research Councils. Sometimes all their activities can be considered asR&D (thus, 'mainly R&D' accounting). Sometimes certain of them, notably supportfor postgraduate education, are excluded (R&D content accounting). In theory,'R&D content' accounting has been used throughout our period in the official R&Dsurveys, although its actual application has tightened up considerably. However, thedata in many of the ancillary sources, such as the reports of various committees andannual reports, generally use the 'mainly R&D' approach.


1.6.2 Calculation 0/ Full-Time EquivalentsThe text above deals with the activities of whole units such as laboratories andresearch establishments. The same distinction between 'mainly R&D' and 'R&Dcontent' can be made for individuals working in these units. Thus it is possible tocollect da ta for all persons working mainly on R&D or to collect a full-timeequivalent (FTE) or even 'person-years' on R&D, which corresponds to R&Dcontent accounting.

1.7 Performance and Source of Funds

R&D activities are often financed by one unit and performed by another and theseunits may be in different parts of the same sector (i.e, payments by the DHSS forR&D performed by the Medical Research Council, or a sub-contract passed from anaerospace firm to an electronic firm) or even in different sectors of the economy, (i.e.research grants from the MRC to universities, or an R&D contract from the MODto the aerospace industry).

The term 'sectors' can be used to refer to a variety of different levels ofdisaggregation from product groups upwards to industrial Divisions (i.e. the previousSIC Orders). In broad terms, however, the Government chooses to break down thestatistics by seven sectors (see, for example, [QRL.152] and [QRL.153] as folIows:

Central Government . This 'is defined as in the government sector of the nationalaccounts for the period in question. Thus it inc1udes the UKAEA andResearch Councils' .

Local Authorities. These 'are also defined as in the local authorities sector in thenational accounts, except that R&D work done or financed by their furthereducation establishments falls under higher education' , below .

Private Industry. "This inc1udes firms in private manufacturing, construction, miningand quarrying industries, private water companies and the distributive andwholesale trades. Agriculture is not inc1uded, nor are any of the privateservice industries not listed above'.

Public Corporations . These 'are also defined in the national accounts and are thoselisted in the descriptive handbook for the national accounts figures' (NationalAccounts Statistics: Sources and Methods [B.60]). The list of publiccorporations has been amended from time to time, particularly in the recentperiods of privatisation. A listing can be found in the footnotes of NationalIncome and Expenditure (now UK National Accounts) [QRL.109] in successiveyears. Arecent statement of the privatisation of previously publiccorporations can be found in an artic1e on 'Privatisation' in The Observer[B.63], p. 67.

Industrial Research Associations . These inc1ude 'co-operative research associationsand other organisations grant-aided by the DTI and also the other industrialresearch associations' .

Higher Educat ion. This 'covers the universities and the spending of the localeducation authorities on the polytechnics'.

Private Non-Profit-Making Bodies. These inc1ude 'such research performing and/orfinancing charitable bodies as the Cancer Research Campaign and the majorgrant-dispensing private foundations'.


Tracking these flows of funds down is one of the main problems of work withR&D statistics. Furthermore, each flow may be reported in two ways, as a receiptfrom an external source by the performer and as an extramural payment by thesource unit. In principle, the performers' reports are to be preferred as they aregenerally better placed to measure the ' R&D content' of expenditure, whereasfunding sources are more likely to report on a 'mainly R&D' basis . Although'performer-based' reporting is theorctically best, a good deal of the data discussed inthis review are 'source-based' .

1.8 Occupation Versus Level and Type of Qualification

In theory, people working on R&D may be analysed either in terms of their currentoccupation or in terms of their level and type of qualification. Again, in theory, thebest approach to use depends on the type of question being asked. In practice, mostUK sources use a mixture of the two approaches which tends to produce eitheroverlapping categories or gaps . The following, rather theoretical, discussion isprovided as a background to the descriptions of the data provided in Chapters 2 and6.

1.8.1 OccupationFour occupation levels may be identified: R&D scientists and engineers; technicians;clerical and administrative staff; and other supporting staff.

1.8.1.1 R&D scientists and engineers (RSEs) . These are also sometimes referred toas 'researchers', 'scientists and engineers' or 'research staff'. The latter term is used inthe questionnaire for the only official UK survey to collect such data ([QRL.153] andAppendix 1). This questionnaire does not offer any further explanations. The officialinternational definit ion is: 'Persons actually engaged in the conception andforcreation of new knowledge, products, processes, methods and systems' [B.56].

1.8.1.2 Technicians. The question of how to define the job of a technician hasbeen more widely discussed. The questionnaire for the 1968 Survey of Engineers andTechnological and Scientific Manpower, quoted in [QRL.194] (this survey will bediscussed in Chapters 6 and 7) indicates that "Technicians and other technicalsupporting staff occupy positions between that of the qualified engineer, technologistor scientist on the one hand and the skilled foreman, craftsman or operative on theother. In most cases they are in jobs-v-in which they are either subject to thedirection of an engineer, technologist or scientist or are engaged primarily in theapplication of proven techniques. Within these prescribed limits their education andaccumulated specialised skills enable them to exercise technical judgement. By this ismeant an understanding, by reference to general principles, of the reasons for andthe purposes of their work and an ability to select the appropriate establishedtechniques and skills to carry it out. '

The principal border-line problem between research staff and technicians occurs inthe case of R&D which involves the use of extremely sophisticated equipment. Thepersons who operate and maintain this equipment may not themselves be directlyengaged in the conception and creation of new knowledge, but their job may take

INTRODUCTIO N AN D G ENERAL CONCEPTS 23

equivalent intellectual demand s. The French, for example, identify such personsseparate ly in cert ain studies as 'ingenieurs non-chercheurs.'

1.8.1.3 Clerical and Administrative Stajf. A rath er similar problem arises in thecase of the supervisor of the person who is actually engaged in the conception andcreat ion of new kno wledge, particularly when the former is an ex-researcher.Accord ing to internation al standards, based on original United States practice,persons who supervise the scientific or intellectual aspects of a researcher's workshould be included with research ers, whereas persons principally engaged in work onbudgets, social security payment s, etc., should be included with clerical andadm inistrative staff. UK sources have generally preferred to include all R&Dadministration with clerical work in the clerical and admini strative class.

1.8.1.4 Other supporting stajf. This group comprises industr ial staff of variouskinds, and broadly the same problems outlined abov e are relevant for this group.

1.8.2 Level and Type 0/ QualificationBoth these criteria are significant in the context of comparing the various series ofR&D employment data available in the United Kingdom.

1.8.2.1 QSEs. These are persons with university degrees or equivalent in science,technology and engineering, but not in the medical and allied sciences or in the socialsciences and hum anitie s. They include persons with university degrees, withequ ivalent technical awards and/or membership of selected professional institutes.(Further deta ils can be found in Chapters 6 and 7).

1.8.2.2 Other university graduates or equivalent. This class covers: persons withuniversity degrees or equivalent in medicine, pharm acy, dentistry or veterinarysciences; persons with university degrees or equ ivalent in the social sciences orhumanities.

1.8.2.3 All other qualifications. This is a residual class which covers a wide rangeof qualifications, includ ing: other post-secondary qualifications, such as HNC,HND, HTD and equivalent Business Education Council (BEC) qualifications;Higher Secondary qualifications, such as GCE A Level, ONC, OND, OTD and theequivalent BEC qual ification; Lower Secondary qualifications, such as GCE 0Level, CSE , GCSE, City & Guilds, etc.

1.8.3 Occupation Crossed with Level and Type 0/ QualificationNo UK survey of R&D employment has ever publi shed a full cross-classification ofqualification and occupation, although such a cross-tabul ation would be extremelyinteresting. Most surveys have used a mixed approach where the highest level hasbeen defined in terms of occupation. Thus it assumes that, on the one hand there areno RSEs who do not have a university degree or equivalent and on the other thatthere are no technicians or administra tive staff with university degrees or equivalent.The first of those assumptions is probably fairly accurate in recent years, butprobably not for the earlier years of the period covered by this review, before themajor expans ion of university education in the 1960s. (It is still far below the 100 percent figure). It is quite possible that in the 1950s there were persons working as RSEswhose highest formal qualification was HNC. (In France in 1963, nearly a third of


all industrial 'chercheurs' did not have a university degree or equivalent). Since theuniversity expansion we can expect significant numbers of graduates to be working atvirtually all levels. By 1968, for example, approximately 29 per cent of all QSEsworking on R&D in the government sector were employed as technicians[QRL.113] . Similarly, one might currently expect to find significant numbers of SSHgraduates in administrative and clerical posts .

1.9 Problems of Comparison

1.9.1 Between Different Sources

This type of difficulty is, of course, not peculiar to R&D statistics. It is, however, aparticularly common problem when assembling R&D data for a number of reasons.First, R&D is an activity which is undertaken in virtually all sectors and sub-sectorsof the economy. The reader interested in compiling data for a given field or area ofR&D will be using da ta reported by a wide range of institutions whose practicalapplication of the theoretical concepts of R&D accounting may vary, particularlybetween funders and performers, or who may indeed have been asked differentquestions. For example , in the case of health-oriented R&D one must try to fittogether R&D on drugs, medicine and medical equipment performed in industry,R&D by health departments in government, biomedical R&D financed or performedby the Medical Research Council and medical research in private non-profitinstitutes and in the universities. A second reason for problems in matehing differentsources for R&D data within the United Kingdom in the 1950s and 1960s is thatR&D expenditure data were collected quite separately from R&D employment data.For example, the two types of data used different sampling frames for industry.

The third reason reflects recent policy attitudes towards R&D . Immediately afterthe implementation of A Framework for Government Research and Development[QRL.83], more data were issued for publicly funded R&D , but at the cost ofexcessive fragmentation. All the important R&D data available for the period1955/56 - 1972/73 could be found in about a dozen publications, supplemented bythe annual Memoranda of the Financial Secretary to the Treasury and the AnnualReports of the Research Councils. The result was about two dozen reports issuedannually containing some R&D data, all using different classifications and someusing different definitions. These problems of comparability occur for many kinds ofR&D data. They are, however, at their worst for military and civil aerospace R&Dand also for governrnent-funded R&D for other high technology programmes, inparticular, in the earlier years, for nuclear energy and, latterly, for electronics .

The diversity of government reporting of R&D was criticised in the Report by theHouse of Lords Select Committee on Science and Technology Science andGovernment [QRL.164]. The Government's response was to introduce 'AnnualReviews of Research' in 1983. The main concern of the Annual Review was toprovide information about central Government expenditure on R&D, which is notcontained in a single separate R&D budget, Annual Review of Government FundedR&D . 1985 [QRL.51] p. 1. The 1983 Review attempted to establish a consistent database for the period 1977/78 to 1982/3, along with consistent procedures for the


regular reporting of information. The surveys have evolved, over time, with betterinformation about R&D employment, a clearer set of definitions (i.e. on 'strategic'rather than 'seedcorn' research), more information on nationalised industries , etc.

1.9.2 Problems 0/ Comparison over Time1.9.2.1 Changes in concepts and definitions. The quality of United Kingdom R&Ddata has certainly improved during the period under review. In the main this is notthe result of changes in concepts and definitions, but rather of the improvedapplicat ion of those already existing. This is particularly true of the mid-I960s, whena number of improvements and revisions were introduced. Two types of data have,however, suffered considerably from changes in methodology: information aboutexpenditure in the higher education sector and statistics on R&D employment.

1.9.2.2 Institutional changes. During the period under review the institutionalarrangements for the funding and performance of R&D by government haveundergone two major upheavals ; the first as a result of the Science and TechnologyAct of 1965 [QRL.166]; the second followed the application of the White Paper onthe organisation of R&D [QRL.83] in the early 1970s. Furthermore there has beenan almost steady stream of individual institutional changes as industries have beennationalised/privatised, public corporations set up, old ministries broken up and newministries or departments formed, etc. These changes have nearly all affected theR&D series in one way or another. The major cases will be discussed in the chapterswhich follow. (See also the footnotes to UK National Accounts - formerly NationalIncome and Expenditure [QRL.109]. A discussion of recent privatisation can be foundin [B.63].

1.9.2.3 Changes in the price, quality and mix 0/ factors devoted to R&D. Analysesof trends in the amount of resources devoted to R&D generally stress the importanceof one factor, i.e. the research scientists and engineers (RSEs) who actually generatenew knowledge. However, their work depends on the support of auxiliary staff(technicians, clerical and administrative staff etc.) and on the availability of suitablematerials, equipment, buildings and other supporting services. Conceptually , forthere to be an increase in the volume of R&D in an institution, sector or nation ,there must be an increase in the number of RSEs with no change in their 'R&Denvironment'. The first UK work on this topic was The Sophistication Factor inScience Expenditure [B.ll] by Cohen and Ivens, although they built on previousAmerican studies of expenditure per scientist and engineer, notably Arnow [B.l],Brunner [B.9], Milton [B.23] and Searle [B.31].

Obviously the first element to be excluded when examining trends in this way isthe effect of inflation, particularly for the second half of the 1970s when the rapidrate of inflation made comparisons over time at current prices meaningless. Onepossibility is to deflate the R&D data using a general price index such as the implicitdeflator of the Gross Domestic Product. However, over half of R&D expenditure ison labour costs and only about one tenth is in the form of capital expenditure . Thusthe rate of inflation for R&D can be expected to be different from, and indeedprobably higher than, that in the economy at large.

The calculation of such special 'R&D price indices' raises all the classic problemsof price index work . (See, for example, [B.71] and [B.6]). The main practical problem


is, however, to obtain the price series. The most accurate approach would be toestablish special price series specific to the various types of resources devoted toR&D in different sectors , fields, industries, etc. This has been attempted in theUnited States for defence R&D [B.47] and in Germany [B.38], but never in theUnited Kingdom, as the cost and the difficulty of obtaining the basic data is toogreat. The most common approach, therefore, is to use proxy indices for the maintypes of resources, which are either applied direct to the R&D data by type of cost(e.g. [QRL.21O] and [B.6]) or used to produce an index based on the cost pattern in aset year (e.g. [B.72]). In the more sophisticated cases, different price series are usedfor each industry or field whereas in the simpler models one index is caIculated for awhole sector. A general discussion of the topic will be found in Frascati Manual[B.56], p. 269-295. Further details of the main 'R&D price indices' available will begiven in the chapters dealing with the type of da ta they are designed to deflate.

Once allowance has been made for general inflation, i.e. the effect of changes inprice of a fixed basket of R&D goods, one has a first approximation of changes inthe volume of R&D activities, either in total or per RSE . Then comes the questionof changes in the quality of the inputs which cannot be dealt with by deflation basedon proxy price series. Cohen and Ivens [B.ll] coined the term 'sophistication' todescribe residual changes in the amount of resources per RSE after deflation. Sincethen it has come to be used mainly in connection with equipment (Nicholson [B.25]pp. 512-30), linked to the increase in costs brought about by the installation andmaintenance of more complex equipment, and often associated with the underlyingidea that such equipment is necessary to maintain RSE's 'environment' or even their'productivity'.

Equipment is not the only factor devoted to R&D whose quality has changed overthe period under review. The R&D labour force has undergone major changes. Theaverage level of education of supporting staff in general and of technicians inparticular has certainly risen. The case of the RSEs themselves is open to discussion .Cohen and Ivens identified what they called the 'youth factor' which has sincebecome known as the 'ageing factor'. Until recently, R&D has generally been anactivity for young graduates. The number of senior R&D posts has traditionally beenfairly small and staff were expected to move on to other activities after aperiod asRSEs. For example, the amount of time spent on R&D was known to decline withrank in the universities [B.68]. However, in recent years, at least up to the early1980s with the economy sluggish and the university boom over, the degree of upwardor outward mobility of RSEs has declined considerably, not only in the UnitedKingdom but also in a number of other OECD countries. RSEs are staying put and'ageing' and so research teams are not being renewed by the entry of young RSEs .Thus, there is some feeling that the overall quality of the RSE labour force may bedeclining. On the other hand, the average cost of employing RSEs is rising not onlybecause of increases in salary rates for the various grades , but also because in manysectors staff receive regular increments within their grade . Furthermore, establishedstaff may cost more in terms of social benefits, such as family allowances, thanrecent graduates. It is this increase in cost which Cohen and Ivens referred to as the'youth factor' and, at that time, put at 2.5 per cent p.a. Either this increase has to betreated as a special form of inflation, or it has to be taken to reflect an increase inquality, incremental or grading drift. Treating this as an increase in quality may be


inappropriate in the light of the arguments about the higher contributions of youngerRSEs, although the precise relationship between productivity and age has never beenfirmly established.

Some degree of substitution is possible between the types of resources devoted toR&D and has certainly taken place over the period under review. Cohen and Ivensfound an increase both in the number of supporting staff and in the volume ofcapital equipment per RSE in the public sector during the period 1955-1964 [B.ll].More recent studies of industrial R&D, 1967-1975 [QRL.167] suggest that there hasbeen a decline both in supporting staff and capital expenditure per RSE, but anincrease in the balance of expenditure on materials and bought-in services.

1.9.3 Problems 0/ International ComparisonAlthough this is too vast a field to examine thoroughly in this review, it is worthidentifying the main types of problems which arise.

1.9.3.1 Divergences between national and international standards. Most countriescollect R&D data broadly in line with the standards laid down by internationalorganisations. However, there are a number of divergences from these standards, forexample: in the United States industrial R&D data include depreciation instead ofcapital expenditure; in Japan R&D employment da ta in the higher education sectorare over-estimated because all academics who claim to work regularlyon R&D areincluded (and very few academics admit to not working 'regularly' on R&D), inaddition, Japanese data include all disciplines whereas other countries generally donot. Such divergences are usually known even if they cannot be accuratelyquantified.

1.9.3.2 Divergences between dijferent sets 0/ international standards. The UnitedKingdom collects data in line with the 'Frascati Standards' of the OECD area . Thecountries of Eastern Europe have their own system of 'CMEA Indicators on Scienceand Technology.' A satisfactory explanation of exactly what the latter comprise doesnot currently exist in English, but they are clearly different from the OECD series ina number of respects. Theoretically, a link between the two is provided by theUNESCO system for statistics of science and technology.

1.9.3.3 Institutional dijferences. Even where a country responding to aninternational organisation does its best to observe all the standards concerned, therewill always be differences caused by the cultural and organisational history of thecountry concerned. Those differences can obviously be very large when comparisonsare made with countries with a totally different economic system, e.g. the CMEAcountries. But such variations can and do occur even between the relatively similarcountries of Western Europe. For example, in the United Kingdom the highereducation sector is narrowly defined to include only teaching establishments. All theResearch Council units, including those situated in universities, are allocated to theGovernment sector. In France, however, the Conseil National pour la RechercheScientifique (CNRS), which corresponds to the Research Councils , is officially partof the higher education sector and all its research units are included along withuniversities. When making international comparisons of R&D data, especiallyoutside industry, itis always worthwhile to check on the institutional background.

Social and cultural factors mayaiso affect other types of R&D data, notably thesub-classification of R&D manpower by qualification and /or function and the


breakdown by type of activity.1.9.3.4 Exchange rates. The classic method of making direct international

comparisons is to convert the expenditure of the country concerned into a commoncurrency, such as the dollar. However, in the case of the Uni ted Kingdom, thenormal exchange rate has historically over-estimated the internal cost of R&Dcompared with a number of countries, notably the United States and, more recently,the continental countries, and even Japan, although this conclusion may be changedsomewhat by recent realignments of currencies.

The earliest efforts made to calculate purchasing power parities for R&D weremade by Freeman round about 1960. These calculations compared the industry inthe United Kingdom and the United States [B.17]. They were extended to a widerrange of countries for 1961 [B.18]. This approach established the most popularmethod for estimating such rates, i.e. weights based on a type of cost breakdown,specific parities for salaries of R&D manpower and proxy parities drawn from moregeneral studies (in Freeman's case, the Gilbert and Kravis studies [QRL.69] and[B.19]). Rather a different approach was used by Brunner in a study of the basicresearch costs in Europe and the United States in the middle 1960s [B.43] and [B.9].McDonald [B.21] concentrated on a novel method of deriving equipment costs fromforeign trade data. A gap followed until the results of a new round of internationalpurchasing power pari ties for national accounts became available in the middle1970s, see Kravis , Heston and Summers International Comparisons of Real ProductPurchasing Power [QRL.16] and Kravis, Kenessy, Heston and Summers A System ofInternational Comparisons of Gross Product Purchasing Power [B.19]. The UnitedKingdom is included in both volumes of the UN exercise and the 1975 EEC study[B.53]. The OECD secretariat has used both these EEC and UN detailed sources toestimate special weighted R&D purchasing power ratios for industrial R&D for the1970s Trends in Industrial R&D [QRL.209]. It has used the ordinary GDP purchasingpower parities now published by the CEC and OECD for later work. A generaldiscussion of constructing R&D exchange rates will be found in Frascati Manual[B.56] p. 296-306.

CHAPTER 2

THE NATIONAL SURVEY

2.1 Origins

The first official attempt to calculate the total amount of national resources devotedto scientific R&D was made for the Advisory Council on Scientific Policy (ACSP)for the year 1955, and the results were published in their annual report for 1956/57[QRL.39] . The ACSP had a mandate to advise on alt aspects of science andtechnology and their interest in R&D was strictly economic. They argued that, 'Anyexpansion of the economy and any increase in our standard of living will dependlargely on the development of new processes and new products which we can selt inoverseas markets. For a country so placed the importance of our efforts in scientificresearch and development can hardly be over-estimated' [QRL.39]. ACSP's estimatesof private industry's expenditure on R&D were based on the preliminary results ofan enquiry conducted by the Department of Scientific and Industrial Research(DSIR) in 1955. The value of Government R&D contracts with private industry wasestimated separately by ACSP. Both sets of data were published in the 1955/56annual report. However, the complete findings of the 1955 survey were laterpublished by DSIR in Estimates 0/ Resources Devoted to Scientific and EngineeringResearch and Development in British Manufacturing Industry . 1955 [QRL.76].

Thi s first survey was modelled largely on those already undertaken by theNational Science Foundation in the United States. It is worth noting that there wasalread y some experience of related unofficial surveys in the UK, primarily those ofthe FRI Scientifi c and Technical Research in British Industry [QRL.173]. We return tothe se in Chapter 7 below. The basic characteristics ha ve varied little in subsequentsurveys although the quantity and quality of the data coltected have improvedconsiderably (see section 2.4.2). Over this period similar concepts have been adoptedand refined in most of the advanced market economies.

In fact it is not strictly correct to speak of the 'na tional R&D survey' but rather of'national R&D totals' as the latter are built up from the results of several surveyseach with its own questionnaire and slightly different specifications (see, for example,Appendix I). The two largest ones deal with R&D financed and/or performed: (i) byindustry; (ii) by Central Government. Less detailed surveys of R&D are undertakenin other sub-sectors (private non-profit-rnaking (PNP) institutes, local authority(LA), and further education (FE)) and special estimates are made for R&Dperformed by uni versities.

Some data are available about government R&D from the inter-war period[QRL.153], p . 135. This ha s become increasingly comprehensive over time, althoughit was not systematised during the early post-war period. This culminated in thecriticisms contained in the report on Science and Government produced by the

29


House of Lords Select Committee on Science and Technology [QRL.I64], para 20. Inresponse, the British Government decided to introduce a system of annual reviews ofGovernment Research. The first Annual Review of Government funded R&D wasfor 1982/83, published in January 1984 [QRL.53]. Subsequent reviews have followed[QRL.54] and [QRL.51). We return to this in detail below.

2.2 Basic Characteristics

2.2.1 The National R&D Effort

2.2.1.1 GERD. The most common R&D expenditure aggregate for a country is the'Gross Domestic Expenditure on R&D', or GERD, which is defined as 'Totalintramural expenditure for R&D performed on national territory during a givenperiod ' (Frascati Manual, [B.56]). Thus, it includes foreign-financed R&D performedon national territory, but largely excludes horne financed R&D performed abroad.GERD is usually presented in the form of a matrix showing the flows of funds forR&D performance between the sectors .

This aggregate is sometimes referred to in older UK reports as the Gross NationalExpenditure on R&D. However in official international nomenclature this term(shortened to GNERD) is reserved for another national aggregate , i.e. total R&Dfinanced by a country excluding foreign funded R&D at horne, but includingpayments for R&D abroad. In the case of purely UK data there need be noconfusion as all the surveys except the first were designed to collect Gross DomesticExpenditure on R&D. However some foreign sources , notably for France, containdata for GNERD.

2.2.1.2 Total R&D manpower. According to the international standard, theemployment aggregate corresponding to GERD should comprise total employmentworking directly on R&D on national territory during a given period expressed asfull-time equivalents (F.T.E.)

RnD Stat Derek Bosworth Robert Wilson Alison Young

Documents

Transcript of RnD Stat Derek Bosworth Robert Wilson Alison Young