Global Models, World Futures, and Public Policy: A Critique

Global Models, World Futures, and PublicPolicy: A Critique

April 1982

NTIS order #PB82-251299

Library of Congress Catalog Card Number 82-600523

For sale by the Superintendent of Documents,U.S. Government Printing Office, Washington, D.C. 20402

Foreword

A number of long-range forecasts of international demographic and economic trendshave been published over the last decade, many of them based on the findings of globalmodels—computerized mathematical simulations of the world’s physical, economic, andpolitical systems. This report responds to a request by the Technology Assessment Boardfor an evaluation of the methodologies, findings, and implications of Global 2000 andother global modeling studies.

Computerized models are a useful way of handling a large amount of data at once and ofkeeping assumptions and calculations self-consistent. In this connection, our study notonly reviewed the findings and recommendations of five major global modeling studies,but also examined the underlying assumptions and data bases on which those results arebased.

The purpose of this report is neither to confirm nor to disprove the sometimes rosy butmore often dire predictions derived from global modeling studies, but rather to examinethe present use and potential usefulness of this rapidly developing technology as a powerfultool for long-range strategic analysis and policy development. OTA found significant andgrowing use of models by a variety of Federal agencies. In addressing the modelingcapabilit y of the Government, the report focuses not on whether models should beused—they already are, and have proven themselves valuable over a period of years—butrather on how to improve this capability and make its projections more useful to analysts,planners, decisionmakers, and the broader public.

We were greatly aided by the contributions of a number of contractors and by theguidance and expertise of the many individuals who generously provided information andreview comments. Their assistance is greatly appreciated.

Director

l l

Global Models, World Futures, and Public Policy Project Staff

John Andelin, Assistant Director, OTAScience, Information, and Natural Resources Division

William Mills, Project Director

Paul B. Phelps, Analyst

Marsha Fenn, Administrative Assistant

contractors

Jennifer Robinson, Fog’s Edge Research

The Futures Group, John Stover, principal investigator

Energy Information Administration of the U.S. Department of EnergyPaul Werbos, principal investigator

Claudia Lorge

OTA Publishing Staff

John C. Holmes, Publishing Officer

John Bergling Kathie S. Boss Debra M. Datcher Joe Henson

v

Acknowledgments

OTA wishes to thank the following people who took time to provide information or review part or allof this study.

David Barnhizer Dennis LittleWorld Wildlife Fund Congressional Research Service

Barbara Baum William LongU.S. Department of State U.S. Department of State

Anne Cheatham Marvin OttConfessional Clearinghouse on the Future OTA Director of Congressional and

CherylChristiansen -

U.S. Department of AgricultureBill Davis

OTA Materials ProgramGeorges Fauriol

Center for Strategic and International StudiesGeorgetown University

Katherine GilmanCouncil on Environmental Quality

Dolores GregoryU.S. Environmental Protection Agency

Khristine HallEnvironmental Defense Fund

Major Gary KnutsonCommand and Control Technical CenterU.S. Department of Defense

Barbara LauscheOTA Food and Renewable Resources Program

Donald LeshU.S. Association for the Club of Rome

and Teaching Laboratory

Institutional RelationsJudith Randal

OTA Health ProgramJohn M. Richardson, Jr.

Quantitative ResearchAmerican University

Bruce RossOTA Food and Renewable Resources Program

Dick RowbergOTA Energy Program

Dan SchottenfelfGeneral Accounting Office

Julie SullivanForeign Policy Association

Richard ThoresonOTA Energy Program

Louise WilliamsOTA Human Resources Program

Peters WillsonAlan Guttmacher Institute

vi

Contents

Chapter1$ Overview . . . . . . . . . . . . . . . . . . . . . .

2. Major Global Modeling Studies. . .

3. Findings of the Global Models. . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4. Global Models and Government Foresight. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5. Priorities and Strategies for Improving U. S. Government Foresight . . . . . . . . . . . . . .

AppendixesA. Population Projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

B. Agricultural Projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Page3

11

43

53

61

69

81

C. Energy Projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..................104

CHAPTER 1

Overview

.

ntFwesi@t ● * * * * * * * * * * * * * * * * * * * * * * * * * * * e k e * * * * * + 4

● * * * * * * * * * * * * e * * o * * * * * * * * , * * * * * * * * * * * .6

4 4 * * * * * * * ’ * * . * . * * * * * . * * * * * * . *

8 ,

1. Summary . ● e . . . * , 3

CHAPTER 1

Overview

The last 15 years have witnessed a growing inter-national effort to increase understanding andbroaden public awareness of the conditions, prob-lems, and opportunities that are likely to confrontthe world through the end of this century and intothe more distant future. This ongoing “futuresdebate” has been stimulated in part by the pub-lication of a series of long-range forecasts of globaltrends in population growth, resource availability,economic development, and environmental condi-tions. Many of these forecasts have been based onthe findings of “global models’ ’—computerizedmathematical simulations of the world’s physical,economic, and political systems. As tools of strate-gic analysis, these models have been used to study

the interactions and future implications of pastevents and current trends, As tools of policy for-mulation, global models have been used to evalu-ate or promote alternative actions and programs

that might bring about different or more favorableworld futures.

This report surveys the assumptions, findings,and recommendations of five major global model-ing studies (see table 1). It also considers the use ofglobal models within the U.S. Government, suchas the World Integrated Model (WIM) that is beingused by the U.S. Joint Chiefs of Staff (see pp.23-24), In addition, the report presents strategiesthat have been suggested for improving the qualityand relevance of the Government’s modeling ca-pability. Of great interest in this connection is thenewly created White House “national indicatorssystem” (see p. 65). The appendixes provide de-tailed comparative analyses of the models’ projec-tions of population, agriculture, and energytrends.

Table 1 .—Summary Description of the Five Global Modeling Studies Discussed in This Report

Historical Projection Geographical AlternativeModel Date base period time horizon regions scenarios

World 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1972 1900-1970 2100 1 (global) 11World Integrated Model (WI M). . . . . . . . . . . . . . . . . . . . . . . . . . 1974 1950-1975 2025 10 (later 14) 17Latin American World Model (LAWM). . . . . . . . . . . . . . . . . . . . 1976 1960-1970 2060 4 (later 15) 7United Nations Input-Output World Model (UNIOWM) . . . . . . 1977 1970b 2000 15 (3 blocs) 13Global 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1980 Not consistent 2000 5 to 28 12

(Not consistent)

SOURCE: Office of Technology Assessment.

Findings of the Global Models

Global modeling studies have varied widely intheir purposes, techniques, findings, and prescrip-tions. The results of some studies have beenguardedly optimistic, while others have beenhighly pessimistic. Their specific quantitativeresults have been different because they havemade different assumptions and have focused indifferent ways on different parts of the globalsystem. Nevertheless, they have generally iden-

tified the same problems and seem to have arrivedat roughly similar qualitative conclusions aboutthe present state of the world and its plausiblefutures:

● Population and physical capital cannot growindefinitely on a finite planet without eventu-ally causing widespread hunger and resourcescarcities. However, there is no physical or

3

4• Global Models, World Futures, and Public Policy

technical reason why basic human needscould not be supplied to all the world’s people,now and for the foreseeable future. Theseneeds are not now being met because of un-equal distribution of resources and consump-tion—not overall physical scarcities. The ab-sence of physical limits, however, does notnecessarily imply the existence of a practicalsolution.The continuation of current trends would re-sult in growing environmental, economic, andpolitical difficulties. As a result, “business asusual” is not a palatable future course. Tech-nological progress is expected (and in fact es-sential), but no set of purely technical changestested in the models was sufficient in itselfto bring about a completely satisfactory out-come. The models suggest that social, eco-nomic, and political changes will also benecessary.Over the next two or three decades, theworld’s socioeconomic system will be in a peri-od of transition to a state that will be signifi-cantly different from the present. However,the shape of this future state is not predeter-mined—it is a function of decisions andchanges being made now.Because of the complexity, momentum, andinterdependency inherent in the world’s phys-ical and socioeconomic systems, the full long-term effects of a given action are almost im-possible to predict with precision or certainty.However, actions taken soon are likely to bemore effective and less costly than the sameset of actions taken later, and cooperativelong-term approaches are likely to be more

●

●

beneficial for all parties than competitiveshort-term approaches.Many existing plans and agreements—particu-larly complex, long-term international de-velopment programs—are based on assump-tions about the world that are either mutuallyinconsistent or inconsistent with physicalreality.Pollution and resource availability may ormay not be problems on a global ‘scale,” butthere is general agreement that regional prob-lems of global concern—such as food short-ages in South Asia and perhaps Sub-SaharanAfrica-are far more likely than a global col-lapse.

In some cases individual global models havebeen used to support more dramatic conclusionsand more specific prescriptions than these, but itwould be a mistake to confuse global modeling us amethod of analysis with any particular prediction orrecommendation. As a tool of analysis, globalmodeling is in itself neutral, although like anycomplex tool a given model can be designed orused inappropriately. For instance, most globalmodels contain little or no representation of geo-politics; it would thus be inappropriate to usethem to predict short-term events that may in factbe more strongly affected by nonquantifiable polit-ical variables. Similarly, the findings that comeout of a model will also depend on the data andassumptions that go into it, the purposes to whichit is put, and the way it is interpreted. As a result,global modeling can be a useful technique in long-range analysis, but it should not be—nor is it likelyto become—the sole, or even the principal, basisfor decisionmaking.

Global Modeling and Government Foresight

Global modeling is used by a variety of organiza-tions and is by no means the exclusive preserve ofenvironmentalists or those who advocate a “newinternational economic order. ” A growing num-ber of large domestic and multinational corpora-tions routinely employ the projections of privateeconomic modeling services in their corporateplanning. Several foreign governments and inter-

national organizations support ongoing globalmodeling programs, and a variety of models—glo-bal and otherwise–are also in use throughout theU.S. Government in a wide range of forecastingapplications. The Joint Chiefs of Staff, for exam-ple, are developing a version of WIM for use intheir joint long-range strategic appraisal, and boththe Department of Agriculture and Bureau of

Ch.1—Overview 5

Mines have used WIM as well as other models.The Global 2000 Study found that numerous Fed-eral agencies (including the Central IntelligenceAgency and Department of Energy, as well as theAgency for International Development, Bureau ofthe Census, and Environmental Protection Agen-cy) routinely use regional or sectoral models in car-rying out their long-range analysis and planningfunctions. Similarly, the Members and committeesof Congress have access to long-term econometricmodels maintained by the Congressional BudgetOffice (as well as the findings of models main-tained by the executive agencies) for use in theiroversight, assessment, and legislative functions.This Government modeling capability existsbecause it is necessary, and it has shown itself tobe useful over the years.

The expanded and better coordinated use of glo-bal models could offer the U.S. Government anopportunity to improve its existing foresight capa-bility. “Foresight” relates to the ability to effective-ly address long-range issues by first anticipatingfuture developments, and then formulating pol-icies and programs that will minimize potentialproblems or exploit potential opportunities. Al-though global models cannot generate precise, de-tailed predictions of what will happen in thefuture, they can be used to produce conditionalforecasts of what is likely to happen or the probabil-

ity of different outcomes, given certain specificassumptions about trends, policies, and events.They can also be used to test the consistency ofassumptions and predicted outcomes for differentpolicy options. In addition, the models can gener-ate order-of-magnitude estimates of many demo-graphic, economic, and resource factors at the glo-bal, regional, and national levels.

This level of forecast accuracy and detail can beuseful for a wide range of applications in long-range assessment and policy-development activ-ities. Deficiencies do exist in the Government’scurrent capability, but if these deficiencies are cor-rected global models could become a more effec-tive input to policymaking in four specific areas:

●

●

●

●

assessing the future impacts of current trendsand existing policies;monitoring the national and internationalsituation to identify early signs of potentialproblems or opportunities;formulating and testing a wide range of alter-native policies and courses of action forachieving national goals, avoiding potentialproblems, and exploiting potential oppor-tunities; andproviding a framework to ensure consistencybetween short- and long-term analyses andacross agency jurisdictions.

Strengths and Weaknesses of Global Models

Global models offer a number of methodologicaladvantages over traditional techniques of long-range analysis and policy development:

● Longer time horizon. -Traditional methods are ●

used primarily for annual or short-term fore-casts, whereas global models typically havehorizons of 20 years or more. This allows glo-bal models to assess long-term effects andcumulative changes that might not otherwisebe anticipated.

● Comprehensiveness. —The computer can con-tain far more information about a system orprocess than any single mental or verbal mod-el, and it can keep track of far more interrela-tions and variables at the same time. Global

models can therefore enable the analyst toutilize substantially more information, and doso more meaningfully (e.g., with regional dis-aggregation) than could otherwise be done.Rigor and accessibility. -Modeling requires ex-plicit, precise, and complete statements of ob-jectives, assumptions, and procedures. Thesemust be written out before they can be run onthe computer, and this makes it easier for allsides to examine them and point out omis-sions and inconsistencies. Open communica-tion about the system and the model can leadto revisions and refinements even before anal-ysis begins, and it can also contribute to thedialog through which clear-cut goals are estab-lished.

6 ● Global Models, World Futures, and Public Policy

•Q Logic.—The computer can draw logically cor-rect and mathematically error-free conclu-sions from an extremely complicated set ofassumptions and data. This can lead to in-sights into unexpected or counterintuitivesystem behavior, reveal areas in which furtherresearch is needed, and expose assumptionsand objectives that are inconsistent, contra-dictory, or physically impossible.

● Flexibility. -It is possible to tailor globalmodels to fit particular problems or regions.By changing the magnitude of specific vari-ables and relations, global models can also beused to test a wide range of assumptions andpolicy alternatives. This can make the globalmodel a valuable tool for policy formulation,as well as a device with which planners andpolicy makers alike can sharpen their analyticskills and improve their intuitive “feel” for theprobable behavior of global systems.

Global models are, however, subject to a num-ber of limitations that can constrain their accu-racy, reliability, and usefulness:

● Methodological constraints. -The essence ofmodeling is a simplification that improvesunderstanding, but this means that a limited

●

●

set of discrete factors and relations are used todescribe the complexity and ambiguity of thereal world. There is little agreement, however,on the proper level of complexity or integra-tion. Similarly, there are no generally ac-cepted tests of model realism, making qualitycontrol and third-party validation importantconsiderations.Theoretical constraints.—Current understand-ing of some causal relationships is far fromadequate, particularly for environmental andsociopolitical processes, and this too can leadto inaccurate or invalid assumptions. As a re-sult, the theoretical biases of the modelers-orthe specific needs and assumptions of modelusers—can sometimes lead to oversimplifi-cation or distortion.Data constraints. –In many areas there is a lackof adequate, reliable, and consistent data.This, too, can be a source of forecast error, aswell as a constraint on the issues or regions towhich global models can be reliably applied.

Because of these limitations, it is vital to evaluatethe assumptions and uncertainties underlying theforecasts, if the results are to be understood andused by policy makers.

Institutional Barriers

Several assessments of the Government’s model-ing capability have concluded that the institu-tional context in which models are currently usedis as much of a constraint on their usefulness asthe above technical limitations. Frequently cited ●

institutional barriers include the following:

● poor communication between modelers andpotential model users, resulting in projections

●

that are unresponsive to the informationneeds of policy makers;

. narrow specialization of interests and respon-sibilities, at the expense of interactions among

sectors and cooperation among agencies, com-plicated by inadequate mechanisms for trans-ferring data and resolving problems betweenagencies;lack of understanding, confidence, or supportfor modeling among top-level policy makers,resulting in a failure to integrate forecastingand policymaking activities; andlack of interest in long-term global issues onthe part of the Federal agencies, Congress,and the general public.

Ch. I—Overview 7

Strengthening the Government’s Modeling Capability

Proposals for improving the Government’s mod-eling capabilities usually stress the need for a co-ordinated strategy involving complementary ef-forts at all levels. The proposed initiatives general-ly reflect four fundamental priorities:

● Correct existing deficiencies.—Relevant agenciesmight create internal advisory committees to:1) prepare an inventory of existing models,their uses, their deficiencies, and any plannedmodifications; 2) conduct a survey of currentor potential applications by analysts and pol-icymakers, with particular attention to theirspecific information needs; 3) evaluate existingdata bases to determine data needs and possi-ble ways of gathering data that are scarce; and4) improve communication between policy-makers and modelers in order to increase therelevance and responsiveness of forecasts.

● Coordinate existing capabilities and activi-ties.—Some form of interagency mechanismmight be established in order to: 1) identifyareas of compatibility and sources of inconsist-ency among models; 2) promote communica-tion and technical cooperation among agen-cies; 3) develop consistent standards andprotocols for the reliability, validation, anddocumentation of both models and data;4) provide a clearinghouse for easier access, ex-change, and integration of other agencies’data, assumptions, and projections; and 5) re-solve problems among agencies through nego-tiation or arbitration.

● Support technical improvements in the Govern-ment’s capability and the state of the art.—An in-dependent or “quasi-public” institute might

be created to promote research on global mod-eling and futures research. Its specific func-tions might be to: 1) encourage impartial,third-party validation and assessment of exist-ing models; 2) support nongovernmental re-search on global models and establish a “glo-

●

bal modeling forum” (analogous to the EnergyModeling Forum at Stanford University) atwhich modelers could exchange ideas and cri-tique one another’s work; 3) enlist the talentsand participation of the private sector in Gov-ernment foresight activities; and 4) assessmodeling work done outside the UnitedStates and maintain communication with in-ternational organizations such as U.N. agen-cies and the International Institute of AppliedSystems Analysis.Link foresight with Policymaking. -To ensurethat long-range global issues are routinelytaken into consideration in the formulationand implementation of U.S. policy, Congressmay wish to coordinate and upgrade the fore-sight capabilities of its legislative researchagencies and/or authorize the creation of anew unit in the Executive Office of the Presi-dent. The functions of this new unit might beto: 1) supervise and/or coordinate the strate-gies outlined above; 2) provide the Presidentand other top-level decisionmakers withthorough analyses and a broad range of policyoptions on global issues; 3) evaluate the long-term effects of agency goals and budget itemson global trends and U.S. strategic interests,for consideration by the Office of Manage-ment and Budget and Congress in the budget-ary process; 4) prepare a “policy statement onthe future” to be delivered by the President;5) issue periodic reports on specific globalissues; 6) conduct comprehensive, integratedstudies of long-range trends and problems atregular intervals; and 7) in conjunction with

the Department of State, encourage foreignnational assessments of long-range issues andsupport the data-gathering, analytic, and

problem-solving activities of the United Na-tions, international financial institutions, andnongovernmental organizations.


1.

2.

3.

Conclusions

Global modeling represents an important ana-lytic tool for exploring alternative world futuresand for testing the feasibility and long-term ef-fects of alternative policy actions.The current state of the art in global modelingoffers the U.S. Government a significant oppor-tunity to improve its foresight capability, if themodels are used judiciously and in combinationwith other techniques and inputs to strategicanalysis and policy development.If models are to be used properly within theirpresent limitations, it is critical to: 1) determineand state explicitly the purposes, assumptions,and theoretical biases of the model; 2) ascertainthe extent of uncertainty in a particular projec-

tion and its sensitivity to changes in the under-lying assumptions; and 3) differentiate betweendescriptive forecasts and those that are prescrip-tive or normative.

4. Improvements in socioeconomic theory, model-ing methodology, and data-gathering technol-ogies could substantially improve the usefulnessof the projections generated by global models.

5. Existing deficiencies in the Government’s mod-eling capability are institutional as well as tech-nical in nature, and any effort to correct thesedeficiencies will require better coordinationamong Federal agencies and increased attentionto the information needs of policymakers anddecisionmakers.

CHAPTER 2

Major Global Modeling Studies

Contents.

PageIntroduction . . . . . . . . . . . . . . . .. ..... 11Forecasts and Forecasting. . . . . . . . . . . . . . . . . . . . . . 11Models and Modeling . . . . . . . . . . . . . . . . . . . . . . . . . 11

The Trend Away From TechnologicalOptimism . . . . . . . . . . . .. .. ... ... .. 13

World 3—The Limits to Growth. . . . . . . . . . . . 15Origin and Purpose. . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Structure and Assumptions. . . . . . . . . . . . . . . . . . . . . 15Findings of World 3. . . . . . . . . . . . . . . . . . . . . . . . . . . 15Conclusions of World 3.... . . . . . . . . . . . . . . . . . . . 18

World Integrated Model—Mankind atthe Turning Point . . . . . . . . . . . . . . . . . . . . . 19

Origin and Purpose. . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Structure and Assumptions. . . . . . . . . . . . . . . . . . . . . 19Findings of WIM.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Conclusions of WIM. . . . . . . . . . . . . . . . . . . . . . . . . . 23U.S. Government Use of WIM. . . . . . . . . . . . . . . . . . 23

Latin American World Model–Catastrophe or New Society?. . . . . . . . . . . . . 24

Origin and Purpose. . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Structure and Assumptions . . . . . . . . . . . . . . . . . . . . . 25Findings of LAWM.. . . . . . . . . . . . . . . . . . . . . . . . . . 25Conclusions ofLAWM. . . . . . . . . . . . . . . . . . . . . . . . 29

United Nations Input-output World Model–The Future of the World Economy.. . . . . . . 30

Origin and Purpose. . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Structure and Assurnptions . . . . . . . . . . . . . . . . . . . . . 30

PageFindings of UNIOWM. . . . . . . . . . . . . . . . . . . . . . . . . 31Conclusions of UNIOWM. . . . . . . . . . . . . . . . . . . . . 34

Globa1 2000–Entering the 21st Century. . . . . 34Origin and Purposes. . . . . . . . . . . . . . . . . . . . . . . . . . . 35Structure and Assurnptions . . . . . . . . . . . . . . . . . . . . . 36Findings and Conclusions of Global 2000. . . . . . . . . 37

Figures

F i g u r e N o . Page1.2.3.4.5.

6.7.

8.9.

104

11<

Types of Models. . . . . . . . . . . . . . . . . . . . . . . . . . . 12Possible Global Behavior Modes . . . . . . . . . . . . . 16World 3 Standard Run... . . . . . . . . . . . . . . . . . . 17World 3 Equilibrium Run. . . . . . . . . . . . . . . . . . . 17World 3 Run With Stabilizing PoliciesIntroduced in the Year 2000. . . . . . . . . . . . . . . . . 18Regionalization of the World Integrated Model. 20Block Diagram of the Basic Elements of theWorld Integrated Model. . . . . . . . . . . . . . . . . . . . 21WIM Historical Scenario. . . . . . . . . . . . . . . . . . . . 22WIM Isolationist Scenario. . . . . . . . . . . . . . . . . . . 22Comparison of WIM Low- and Optimal-PriceOil Scenarios. . . . . . . . . . . . . . . . . . . . . .. .....+. 22LAWM Standard Run for Developed Nations. . 26

12. LAWM Standard Run for Latin America. . . . . . 2713. LAWM Standard Run for Asia..... . . . . . . . . . 2814. Regional Economic Growth Under

Three UNIOWM Scenarios . . . . . . . . . . . . . . . . . 32

CHAPTER 2

Major Global Modeling Studies

Introduction

Forecasts and Forecasting

Formal forecasting, which appeared in the early20th century, is based on the rigorous applicationof empirical inquiry and statistical analysis to theprediction of socioeconomic change. It insists oncareful monitoring, a firm data base, and the ju-dicious use of trend extrapolation, while rejectingunfounded optimism and utopianism as “wishfulthinking.” The first use of formal forecasting bythe U.S. Government came in 1929, when Presi-dent Hoover created a Presidential Research Com-mittee on Social Trends, and its techniques andfindings became linked with comprehensive plan-ning and decisionmaking during the New Deal.Further theoretical improvements and practicalapplications have emerged since World War 11through developments in econometrics, generalsystems theory, cybernetics, operations research,and input-output analysis. 1

For strategic analysis and policymaking pur-poses, three general types of forecasts can bedistinguished, based on their approach to foresee-ing the future:z

. unconditional forecasts, which determinethat certain events or trends will, in all pro-bability, occur in the future (these forecastsmight properly be called “predictions”);

● conditional or probabil ist ic forecastswhich determine that certain events or trendsare more or less likely to occur in the future,given certain limiting assumptions concerningpresent and future conditions and policy ac-tions (and that, given a different set ofassumptions, different events or trends aremore or less likely to occur); and

● exploratory forecasts, which examine a widerrange of policies and trends in an open-ended

exploration of possible future developments,with less emphasis on the plausibility ofassumptions or scenarios.

To these three types of descriptive forecasts, whichattempt to project what will or might happen inthe

●

future, a fourth could be added:

prescriptive or normative forecasts, whichidentify events or trends that should (orshould not) happen and determine the pol-icies and conditions that will promote the de-sired outcome.

Models and Modeling

A model is a simplified or generalized represen-tation of something else—an object, process, orsystem. The model need not resemble the originaland can in fact take many forms, depending onthe purpose it is to serve: as an aid to memory, asmall two-dimensional photograph can remind usof a large three-dimensional person or place wehave seen; as an aid to discocery, a 3-lb modelairplane can be tested in a wind tunnel to predictthe performance of a 30-ton airliner built on thesame design; and as an aid to explanation, a set ofgravitational equations can be used to elucidatethe intricate motion of planets orbiting a sun.

The model need not depict every detail of thething it represents. A good model reduces thecomplexity of the original by eliminating elementsand relations that are irrelevant to the purpose athand, retaining only the characteristics that areneeded for that purpose. Ingeniously simple mod-els may be described as “elegant, ” but in the end“a model can be made and judged only with respect toa clear purpose.”3

Models can be divided into three basic types:mental models, physical models, and symbolicmodels (see fig. 1). Mental models are the concep-

~ll>nell;~ H , \ft,dc)u ~, ]t)})n R.IC h,lrJw)n, < i n Li ( k’rh:lrt f3ruc km’lnn (c-J~ ~,

Gmpnq in tk [)Mk Tlw F[r,r [k~dc [)/ (;l~)hll kl( )d~,linq (>’ctl }’(irk W’111’\ ,(c,rr}l~ ,)mlr)u), ~.l> 1~- 3K: cmphail. t hcvr..

11

12 . Global Models, World Futures, and Public Policy

Figure I.—Types of Models

SOURCE: Arthur D. Little, Inc.

tual models people carry about in their heads anduse to think about the world. They are flexible,adaptable, creative, and contain rich stores of in-formation about such intangible factors as valuesand motivations. Some mental models are ex-tremely subtle and elaborate, even elegant. Butthey can also be vague, shifting, unverbalized, andimmune to objective criticism, and they are oftenbased on dubious but strongly held assumptions.Judgmental and qualitative forecasts (includingmany unconditional forecasts) are often based onmental models.4

Physical models are created from tangiblematerials, and the process of embodying the modelcan usually make it both more explicit and more

+Ibid., pp. 20-21, 37-38; and Arthur D. Little, Inc., op. cit., p. 11.6.

open to objective criticism. Iconic or schematicmodels, such as maps or diagrams, are physicallysimilar to their originals, although they may notbehave in the same manner. Analog models, suchas wind tunnels, reproduce the behavior of theiroriginals without necessarily resembling them inappearances Physical models can be a usefulmeans of communicating, clarifying, and correct-ing mental models.

Symbolic models make use of some system orlanguage of symbols to describe the relevant ele-ments and relations of the object, process, or sys-tem they represent. Verbal models, such as DasKapital or Wealth and Poverty, take the form of oralor written language. As a result, they can be more

‘Arthur D. Ltttle, Inc., op. ctt., p. 11.6.

Ch. 2—Major Global Modeling Studies ● 1 3

explicit and precise than mental models, but at thesame time they are potentially diffuse, impres-s ionis t i c , ambiguous, and rhetoric al.Mathematical models, on the other hand, canrepresent the relevant elements and relations of areal object, process, or system in mathematicalsymbols and equations. This allows them to ex-press complex operations concisely, precisely, andexplicitly in a rigorous and consistent language.This in turn makes them more open to objectivecriticism and correction, but mathematical modelsremain susceptible to omissions, distortions, andmisinterpretations like those that afflict mentaland verbal models. They can be no more valid orreliable than the theoretical understanding onwhich they are based and the mathematical formin which they are expressed.

Computerized models are mathematical mod-els that have been rewritten in a programinglanguage that can be run on a computer. They canbe used to investigate a process, system, or theorythat is too large or too complicated to model ade-quately (or manipulate conveniently) in words or afew simple equations. 6 Such models can containmore elements (variables), more relations (equa-tions), and far more empirical data than simplermodels. The computer can keep track of all ofthese factors simultaneously, manipulate themvery rapidly, and produce results that are free fromcomputational error. However, human judgment

fi\lc,IJ<)u ~, RI< h’irdwln, (Ind ku~ km,lnn, op \ It , p 2(I

The Trend Away From

Until about 1970, most long-range forecastswere characterized by generalized optimism aboutthe benefits of continued economic growth andconfidence in the ability of technology to over-come any barriers. The most influential of theforecasts was The Year 2000, by Herman Kahn andAnthony Wiener of the Hudson Institute, whichoffered a set of alternative “scenarios” as a“framework for speculation” on the future. Its cen-tral finding was “that economic trends will pro-ceed more or less smoothly through the next thirtyyears and beyond,” and that “we are entering aperiod of general political and economic stability

is still required to determine what factors to in-clude, hem’ to represent them, what data to use,and how to interpret the numerical findings. Con-sequently, the results that come out of the com-puter are only as reliable as the general assump-tions, structural decisions, and data that go into it,and even the best results are subject to biased ormistaken interpretations.

Global models are simply computerized mathe-matical models whose purpose is to investigatesystems, theories, and issues of a global scale andcomplexity, usually with a relatively long timehorizon:

Global modeling is distinguishable from othertypes of modeling of social systems only by thequestions it asks. Its methods, strengths, andweaknesses are identical to those of all policy-oriented computer models. It draws from the samebase of theory and data. Therefore, if there are anydistinct properties of global modeling, they followdirectly from the characteristics of globalproblems. i

The following survey will therefore focus not onlyon the modeling techniques that have been usedand the findings that have resulted, but also onthe global problems that have been addressed inthe models and the purposes to which their find-ings have been put.

-It.ld , p 45

Technological Optimism

at least so far as the frontiers and economies ofmost of the old nations are concerned.”8 This“surprise-free” scenario was based on exponentialextrapolations of postwar demographic and eco--nomic trends, but it was also influenced by the au--thors’ underlying assumption of “continuity” inglobal affairs, particularly the increased rate oftechnological innovation, and by their confidencethat society would be able to find “physically non-harmful methods of over-indulging.”9 Kahn and


Wiener do caution that “increasing discrepanciesbetween rich and poor” could lead to resentmentand instability, and that the “problems of develop-ment constitute a serious economic and moralconcern. ”1° Nevertheless, according to one critic,“they simply refuse to be overawed by themagnitude of the problems posed. ”l 1 Althoughsharply criticized in recent years, this view of thefuture has remained influential in both Govern-ment and corporate policy, as well as public opin-ion, in the United States. 12

Since the late 1960’s–and particularly since the1973 oil embargo–a less optimistic view of thefuture has gained currency, a view characterizedby increased concern for the feasibility and envi-ronmental consequences of unrestrained eco-nomic growth and by criticism of the social andpolitical institutions that have supported suchgrowth, This new mood, which has been charac-terized as “neo-Malthusian pessimism, ” was influ-enced in part by the projections of economistJoseph Spengler and by the popular success of sev-eral books by Anne and Paul Ehrlich, who arguedthat the world is already over-populated and over-developed in terms of its ecological resources.13 Byfar the largest stimulus to public debate over theseissues came from the activities of the Club of Rome(an international group of businessmen, academ-ics, and civil servants) that was organized in 1968by Italian management consultant Aurelio Peccei,

The Club of Rome’s “Project on the Predica-ment of Mankind” focuses on the complex inter-

acting socioeconomic problems that make up theso-called “world problematique:”14

● poverty in the midst of plenty;● degradation of the environment;● loss of faith in institutions;● u n c o n r t o l l e d u r b a n s p r e a d ;● insecurity of employment;● a l i e n a t i o n o f y o u t h ;. rejection of traditional values; and. inflation and other monetary and economic

disruptions.

The predicament of mankind, according to theClub, is to be able to perceive this problematiquebut to be unable to understand its origins or oper-ation and, therefore, unable to respond to it effec-tivelv.

The Club’s continuing program, consequently,has two objectives: 1) to gain a better understand-ing of the limits of the world system, the interac-tion of its dominant elements, and the constraintsit puts on human numbers and activities; l 5 and 2)to encourage appropriate sociopolitical reforms by

bringing the world problematique to the attentionof the general public and (more pointedly) theworld’s leaders and decisionmakers. The Club “hiton the idea of using a computer to advertise theircause, ” as one critic puts it, not only because “thefield of Systems Dynamics had created a bodyof expertise uniquely suited to the research de-mands,” but also because the resulting reportmight prove to be “a vehicle to move the heartsand minds of men out of their ingrained habits. ”16

These dual purposes led to the first true globalmodel, which also remains the best known and themost controversial.


World 3—The Limits to Growth

Origin and Purpose

In June 1970, when the Club of Rome was seek-ing a suitable methodology for their investimationof the global system, Jay For-rester of the Massa-chusetts Institute of Technology (MIT) invited thegroup to Cambridge for a demonstration of the ca-pabilities of systems dynamics. Within 3 weeks hedesigned and documented a simple global model- World l - as the basis for presentations and dis-cussions at the end of July 1970. (A revised ver-sion, World 2, was the subject of Forrester’s subse-quent World Dynamics (197l).) Impressed, theClub obtained a $250,000 grant from the Volks-wagen Foundation to fund For-rester’s colleagueDennis Meadow’s and a team from MIT indeveloping a full-scale model—World 3—based onthe systems dynamics approach. Under theauspices and direction of the Club of Rome, theMIT team produced both an elaborated modeland a popularized presentation of it in less than 2years—perhaps too quickly, in the view of onecritic:

The Club only relinquished control then the ex-ercise had produced their desired product, as evi-denced by the fact that client pressure drove themodelers to violate their scientific values by pub-lishing The Limits of Growth beforc the technicaldocumentation for World 3 was completed.

Structure and Assumptions

The World 3 model describes the global systemin terms of five interacting subsystems—popula-tion, natural resources, capital, agriculture, andpollution-which are averaged on a global basis.Its most important conceptual contribution is the incor-poration of “feedback” relations between these vari-

] I.r,r,,tcr R(hIII.,111 III ~lL(IJI\\\ RI Il<lr.!.rli <Ir)d Ihu km<ll~n, p Lit ,

;> +’+

ables; due to these relations, attempts to solve oneproblem may unintentionally exacerbate another.The model also introduced the concept of “carry--ing capacity ’’—the level of population and produc-tion that could be sustained indefinitely by theprevailing physical, political, and biological sys-tems of the world-and posited four possible “be-havior modes” that a growing population couldexhibit with regard to this carrying capacity (seefig. 2). None of these behavior modes reflects thepotential ability of technology to expand the carry-ing capacity, primarily because the model assumesnonsubstitutibility between technology and resources.

The purposes of the model, according to theauthors, were “to determine which of [these] be-havior modes . . . is most characteristic of theglobe’s population and material outputs under dif-ferent conditions and to identify the future policiesthat may lead to a stable rather than an unstablebehavior mode,”18 According to one critic, how-ever, given the authors’ “ specific motivating con-cern with limits, the broad conclusions thatemerged from the model are, perhaps, not surpris-ing’’—they assumed that limits exist and wouldeventually be reached; “[the] question was whenand how.”19

Findings of World 3

The standard or “reference run” of World 3,based on a continuation of the trends that havecharacterized the world system since 1900, resultsin the model output that has given The Limits ofGrowth its reputation for “gloom and doom” (seefig. 3). In this case the collapse of the system iscaused by rapidly expanding population and in-

———-———1 fi[)t.l, r,,< [., Alc:lJ(l\\ s, C’r ‘II ! [),,MI1lL \ ,)/ L ;?, It, [h I* I ,i f l~llr, U’ r),l(l \( ‘<111 ) -

hrldgt, hl,I\s \Y’rlgl~r Allc,n, 1 Q;+), II ~‘g~l~:ldtni i, [{I( h{lrdw~l]. <id f’mu~ km(llll~, ~~p ( 11., ;> [> I


Figure 2.— Possible Global Behavior Modes

Organic growth

Oscillationabout stablecarryingcapacity

Time

TimeTime

Erosion ofcarryingcapacity

Time

dustrial output and a diminishing resource base.Essentially the same results are achieved in addi-tional runs with increasingly optimistic assump-tions about the five system variables:

● doubling the nonrenewable resource base;● “unlimited” nuclear power and extensive re-

cycling;. strict and effective pollution control;● doubling the average agricultural yield; and● “perfectly effective” but voluntary birth con-

trol.

In some of the runs population and industrial pro-duction climb to higher levels before collapsing,but–according to this analysis–no single techno-logical change can avert the final catastrophe, norcan any combination of them delay the collapsebeyond the year 2100. In some runs the collapse iscaused by a resource crisis, in others by a pollutioncrisis or a food crisis; but no matter what the as-

sumptions, say the authors, “The basic behaviormode of the world system is exponential growth ofpopulation and capital, followed by collapse.”20

In keeping with their second objective, the MITteam also used the World 3 model to identify con-ditions and policies that would avoid these prob-lems and lead to a stable behavior mode like one ofthose in figure 2. Continuous growth was ruledout by the basic assumptions of the model; theywere looking for an output that represented a “sus-tainable” world system that would avoid collapseand would also be “capable of satisfying the basicmaterial requirements of all of its people. ”21

By working backward from the desired outcometo the conditions that would produce it, the au-thors were able to find a combination of “realistic”policy changes that, implemented simultaneously


Figure 3. —World 3 Standard Run

1900 2100

The “standard” world model run assumes no major changein the physical, economic, or social relationships that havehistorically governed the development of the world system.All variables plotted here follow historical values from 1900to 1970. Food, industrial output, and population grow ex-ponentially until the rapidly diminishing resource baseforces a slowdown in industrial growth. Because of naturaldelays in the system, both population and pollution con-tinue to increase for some time after the peak of industri-alization. Population growth is finally halted by a rise in thedeath rate due to decreased food and medical services.SOURCE: Limits to Growth.

in 1975, would lead to an “equilibrium state” (seefig.

●

●

●

4): 22

restrict population growth by reducing aver-age desired family size to two children andmaking “perfect” birth control universally ac-cessible (population stabilizes at about 6 bil-lion in 2050, after a delay inherent in the agestructure of the current population);restrict capital growth by maintaining aver-age industrial output per capita at the 1975level and holding the capital investment rateequal to the depreciation rate (excess capacityis used to produce consumer goods and serv-ices);reduce resource consumption and pollu-tion generation per unit ‘of industrial andagricultural output to one-fourth of their 1970

1900 2100

Technological policies and growth-regulating policies pro-duce an equilibrium state sustainable far into the future.Technological policies include resource recycling, pollu-tion control devices, increased lifetime of all forms of capi-tal, and methods to restore eroded and infertile soil. Valuechanges include increased emphasis on food and servicesrather than on industrial production. Births are set equal todeaths and industrial capital investment equal to capitaldepreciation. Equilibrium value of industrial output percapita is three times the 1970 world average.

SOURCE: Limits to Growth,

●

●

●

levels (largely through recycling and advancedabatement technologies);divert capital to agricultural production inorder to produce sufficient food for all people,even if such an investment would be consid-ered “uneconomic;”prevent soil depletion and erosion by usingsome of the agricultural capital for enrich-ment and preservation (e.g., comporting ur-ban organic wastes and returning them to theland); andextend the average lifetime of industrialcapital stock through improved durabilityand maintenance, in order to reduce obsoles-cence and make more capital and resourcesavailable for other sectors.

The authors recognized that different combina-tions of the above policies might be adopted by dif-ferent societies, and that “[a] society choosing sta-


bility as a goal certainly must approach that goalgradually.” 23 However, they hastened to add thataction must be taken soon: if the implementationof these policies were to be delayed by 25 years, forexample, they would not result in an “equilibriumstate” (see fig. 5); this implicitly suggests that farmore severe measures would be required after thattime.

These findings led the authors to call for a “con-trolled, orderly transition from growth to globalequilibrium, ” but they were vague about the spe-cific actions and tradeoffs this transition would re-quire, explaining that “much more information isneeded to manage the transition. ”24 Some criticsfeel that the model’s “no growth” bias “can be seenas supporting the interests of the materially well-

25 However, others Pointoff” and the rich nations.out that the equilibrium state necessarily implies a“world-wide radical egalitarian levelling of in-comes and property, ”26 yet the MIT team has“almost nothing to say about what should ormight happen to poor nations . . . under thepolicy of no growth. ”27 Because of “their deliberateself-restriction to physical properties of the world,”according to another critic, “they have chosen tobe unconcerned with politics [and] social struc-ture; ”28 The Limits to Growth speaks instead of thegreater demands that will be placed on“humanity’s moral resources. ”29 Above all,however, the model’s simplification and global ag-gregation of imperfectly understood factors makeit unsuited for generating specific, detailed policyrecommendations. This is a limitation shared byother global models:

The breadth of focus and coherent conceptualdevelopment of the world models ensure their util-ity for clarifying the nature of long-term globalproblems. However, their limitations render themunsuitable as primary tools of analysis or as toolsfor detailed analysis of global problems and theirsolutions. 30

“ibid., p. 167.“11md., p. 180.J5Keith L. R. Pawtt, “Malthus and C)ther E c o n o m i s t s : S o m e Doomsdavs

Rewslted,” In IMocfek oj Doom A CrltIque of [he Llmm to Grouth, H. S. D. Cole(eel.) (New York: Unlveme Books, 1973), pp. 154-157.

2sHarvey S i m m o n s , “Svstems Dynamtcs a n d T e c h n o c r a c y ) ” [n ,Models of

Doom, pp. 206-207.27c(11e ,n W’Or/d Futures T h e Grear Debate, P. 29.28 Marie Jahoda, “Postscript on Social Change,” In Models of Doom, p. 212 .Z~Meadc)w,s, et al., T h e Ltmtts to Grouth, P. 179.

~O]ennifer Robinson, “worlds 2 and 3,” [n The Glofra/ 2(XN.I ~eporc [0 lb Presl-

dem (Washington, D. C.: U.S. Council on Environmental Qualltv and Depart-m e n t of State, 1980), Iol. 2, ~, 608,

Figure 5.—World 3 Run With Stabilizing PoliciesIntroduced in the Year 2000

1900 2100

If all the policies instituted in 1975 in the previous figure aredelayed until the year 2000, the equilibrium state is nolonger sustainable. Population and industrial capital reachlevels high enough to create food and resource shortagesbefore the year 2100.

SOURCE: Limits to Growth,

Conclusions of World 3

Within these limits, World 3 arrives at three cen-tral conclusions that have been influential in thesubsequent “futures debate:”31

10

2.

3.

If present growth trends in global population,industrialization, resource depletion, pollu-tion, and food production are allowed to con-tinue unchanged, the limits to growth on thisplanet will be reached sometime within thenext 100 years, resulting in a catastrophicdecline in both population and industrial ca-pacity.These growth trends can be altered in such away as to establish economic stability at levelsthat are both sustainable into the foreseeablefuture and capable of satisfying the basic ma-terial needs of all the world’s people.If the people and nations of the world decideto strive for this equilibrium state, the soonerthey start working to attain it, the greatertheir chances for success will be.

1JILleadous, Rl~hardson, a n d Bruckmann, op. cit., PP. ~7-68.

Ch. 2—Major Global Modeling Studies • 19

Similarly, the technical limitations that restrict creasingly have been to achieve both greaterthe utility of World 3 have proven to be a stimulus specificity and greater relevance to the needs offor subsequent global models, whose purposes in- policy makers.

World Integrated Model —Mankind at the Turning Point

Origin and Purpose

The World 3 model achieved most of the objec-tives set for it by the MIT team and by the Club ofRome, but system dynamics was still viewed withskepticism in traditional scientific and policy cir-cles. As a result, the popularized report on themodel, The Limits to Growth, was the subject ofconsiderable debate and controversy because of itsmethods—primarily its radical aggregation ofglobal factors—and because of its vagueness onpolicy issues. When the club began planning a fol-lowup in 1972, therefore, it sought a modeling ap-proach that would accomplish three goals:32

●

●

●

to represent the world as a system of interde-pendent regions, rather than a single homoge-neous unit, and to represent those regions ingreater sectoral detail;to develop recommendations that would be ofmore direct relevance to policy makers; andto gain greater acceptance from the scientificcommunity by incorporating more “harddata” and, wherever possible, by explicitlyemploying state-of-the-art theories and meth-odologies from the relevant academic disci-plines.

The model the club chose to support, again withfunds from the Volkswagen Foundation, was theWorld Integrated Model (WIM). This model wasdeveloped in parallel by two teams, one led byMihajlo Mesarovic at Case Western Reserve Uni-versity in Cleveland and the other by EduardPestel (a member of the executive committee of theClub of Rome) at the Technical University inHannover, West Germany.

The authors first presented their model at a con-ference for high-level policy makers sponsored bythe Woodrow Wilson International Center forScholars in Washington, D. C., then at the firstglobal modeling conference of the International

‘>lhd , pp. ;~-?+.

Institute for Applied Systems Analysis (IIASA) inAustria, and finally at a series of scientific meet-ings and congresses throughout the world. Onlyafter these formal presentations—and the distribu-tion of technical documentation to selected ex-perts—did they release the popular description ofthe model in the fall of 1974.


The WIM methodology is based on Mesarovic’s“multilevel hierarchical systems theory, ” whichviews the world in terms of five interrelated planesor strata:

●

●

●

●

●

the environmental stratum, which com-bines geophysical and ecological factors andcorresponds roughly to the natural “carryingcapacity,” although perhaps too superficiallyto satisfy some environmentalists;the technology stratum, which embraces ac-tivities whose biological, chemical, or physicalterms involve mass and energy transfer;the demographic-economic stratum, whichcombines the human population and indus-trial capital of World 3 and, with the environ-ment stratum, makes up most of the model’scontent;the group stratum, made up of sociopoliticalinstitutions, policies, and decisions, which areusually represented as sets of alternative sce-narios among which the model user chooses;andthe individual stratum, reflecting personalattitudes and values, again represented by al-ternative scenarios to be selected by the modeluser.

According to the theory, these levels ordinarilyoperate with a fair degree of independence, al-though they can become highly interactive under“crisis” conditions. The authors therefore feel thattheir model can help us to understand and predict


the system’s behavior in both present and futurecrises. 33

The major improvement in the WIM representa-tion of the world system, however, is its greatergeographic and economic detail (regionalizationand disaggregation). Instead of a single homogene-ous world, the model contains 10 regions made upof similar countries (see fig. 6), although in someruns they are grouped in three or four blocs. As aresult, WIM can represent varying levels of devel-opment and resource endowment, as well as cul-tural and environmental differences; and it cantherefore be used to investigate potential regional(as opposed to global) problems and crises. In addi-tion, because these regions are connected by atrade network, WIM can be used to investigate thepotentially mitigating effects of international trade(see fig. 7). Within each region, physical and eco-nomic sectors are differentiated into numeroussubcategories—85 age groups for population, 19

‘JJennLfer Rolxnson, “Mesaro\w-Pestel W o r l d M o d e l , ” In The Global 2cXXIReport to the Presderrt, vol. 2, p. 6 1 6 .

categories for industrial capital, five for energy cap-ital, two for agricultural capital, and so on. Fromthe point of view of an economist, in fact, WIM isa collection of regional economic models. Theresulting mathematical model is quite large: World1 contained only.40 equations, and World 3 about200, while WIM (according to its creators) con-tains over 100,000.34

Another improvement, one that is more directlyrelevant to policy applications, is the model’s in-teractive design. The model user is allowed to esti-mate social and political behavior by selectingamong alternative scenarios in the individual andgroup strata, thereby manipulating certain vari-ables in such a way as to test a wide range of policyassumptions about energy prices, food exports,capital investments, and development aid. In addi-tion, WIM’S various submodels can be used inde-pendently to generate and test alternative policiesfor specific countries and regions. This capability

~+cole in World Futures: The Great Debate, P . 34

Figure 6.—Regionaiization of the Worid integrated Modei

SOURCE: Mankind at the Turning Point.


Figure 7.—Block Diagram of the Basic Elements of the World Integrated Model

. . ‘ .

Trade and payments

Agriculture

SOURCE: Command and Control Technical Center.

was in fact one of the stated objectives of the mod-elers:

We hoped thus to furnish political and economicdecisionmakers in various parts of the world with acomprehensive global planning tool, which couldhelp them to act in anticipation of the crises at ourdoorstep and of those that loom increasingly largein the distance, instead of reacting in the spirit ofshort-term pragmatism .35

In keeping with this objective, which is shared bythe Club of Rome, the WIM team at Case WesternReserve has actively marketed their model, usingsatellite-telephone patches to make presentationsto prime ministers and other officials in at least 18different nations.

Findings of WIM

The WIM model has shown its versatility in ex-tensive use for policy testing, both to evaluate al-ternative scenarios within its own assumptionsand to test the scenarios and assumptions of other

~j~~ihallo D, Mesaroi,lc and Ecjuard Pestel, hfankmd at the T[/mw Pomr (New

Y o r k : Dutton, 1974), p. Ix.

modelers and futurists. 36 Because the purpose andoutput of the model vary significantly from user touser and run to run, however, it is difficult toisolate any particular “results, ” although severaltest runs are illustrative. The reference or “histor-ical scenario” run of the model, based on a con-tinuation of present trends, results in the modeloutput shown in figure 8: a steady increase in thereal cost of food on the world market, whichwould also drive up domestic prices in the UnitedStates, and a catastrophic increase in the numberof deaths caused by starvation in South Asia. Thealternative “isolationist scenario” (fig. 9) indicatesthat, should the United States act to keep domes-tic food prices down by restricting exports, starva-tion in South Asia comes sooner and is even morewidespread. In another pair of policy tests (fig. 10),the model output suggests that a policy of low,fixed oil prices leads to a catastrophic economic de-cline in the developed world when the resource isexhausted, whereas “optimal” price increases (per-

Msee for example Barry B, Hughes and Mlha]lo D. Mesarowc, “Tes t ing the

Hudson Institute %enarlos” (~’ashlngton, D. C.: U.S. Assmlatmn for the Club

of Rome, Sept. 1979, mimeograph .


1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025

SOURCE: Systems Research Center, Case Western Reserve University.

Figure 10.–Comparison of WIM(A)

0

1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025

SOURCE: Systems Research Center, Case Western Reserve University,

1975 1980 1985 1990 1995 2000 2005 2010 2015 20202025

Year

Cheap energy in the form of oil has been a prime fuel for theunprecedented growth of the world economy in the 1950’sand 1960’s. The dramatic increase in oil prices in 1973 wasviewed as a catastrophe. However, computer analysis ofour world system model indicates that the continuation ofwhat amounts to overexploitation of oil, spurred by anunreasonably low price, would lead to major dislocationsbecause of the exhaustion of reserves and the lack ofmotivation to develop substitutes in time. Pursuance ofshort term objectives would lead to major dislocations in


Low” and Optimal”Price Oil Scenarios

. . . ‘ 14

01975 1980 1985 1990 1995 2000 2005 2010 2015 20202025

Year

the long run (see A). A much more beneficial developmentfor all concerned results from the “optimal price scenario”in which the price is gradually increased up to the “op-timum” level. Such a policy would bring in the substitutesin a more regular fashion while prolonging the reserves.Both exporting and importing regions would fare better(see B). It is only by taking a global and long term view thatsuch a course of development, most beneficial to all con-cerned, can be identified.


mitting gradual adaptation and substitution) bene-fits both oil producers and consumers; a third oil-price scenario (not shown) suggests that increasesabove the “optimal” level leave all regions worseoff. 3’

Conclusions of WIM

From these and numerous additional interac-tions with the model, the authors arrive at severalconclusions about the nature of the world systemand

●

●

●

the management of future development:38

the current crises in agriculture, energy, etc.,are not transient but persistent, and representthe first signs of an “oncoming era ofscarcity ;“the solutions to these crises cannot be foundthrough isolated, short-term, or narrowly na-tionalistic strategies, but only through an inte-grated global context and “in the spirit oftruly global” cooperation . . . guided by a ra-tional master plan for long-term organicgrowth;” and“the time that can be wasted before develop-ing such a global world system is runningout .“

The model indicates that oil, substitutes for oil,and agricultural land will be the greatest con-straints on growth. To address these problems, theauthors recommend a policy of “organic growth,”based on a recognition that the world system is a“collect ion of functionally interdependentparts, ”39 This policy, which would encouragegrowth where needed and discourage it where it“threatens not only that part but the whole aswell, ” include such specific steps as the following:40

●

●

●

●

encourage worldwide diversification of indus-try to achieve a truly global economic system;build up the economic base and especially theexport potential of the poorest countries sothey can pay for food imports;give food aid to the poorest countries, but giveinvestment aid only in the form of “intermedi-ate” or appropriate technology; andcarry out effective social and institutional re-forms, because the required economic trans-

fers are impossible under prevailing inter-national economic arrangements.

The authors suggest that unless such steps aretaken, spreading regional collapse and interna-tional tension will, like falling dominoes, eventu-ally reach the developed world. If these and othersteps are taken, on the other hand, “the worldgrowth rates implied by (WIM’s) computer resultsare much closer to those of Kahn and Wiener thanto those of Meadows and Forrester. ”41

U.S. Government Use of WIM

More recently, former members of the modelingteam have designed a specially tailored version ofWIM for the U.S. Department of Defense.42 Themodel, which is fully operational, is maintainedand operated by the Command and ControlTechnical Center (CCTC) in support of the Plansand Policy Directorate (J-5) of the Joint Chiefs ofStaff. J-5 is currently creating a new division, spe-cifically devoted to long-range analysis, which willuse the model to develop projections of global sys-tems behavior for use in long-range nationalsecurity planning. At present, the model is beingdeveloped to provide data on political, economic,and demographic conditions under varioussubcontingencies of four basic scenarios or “futureworlds” defined by J-5:

●

●

●

●

“A-muted bipolarity,” essentially a referencerun based on current trends and internationalrelations;“B-superpower dominance (conflict mode),”including contingencies representing differentlevels of East-West conflict;“C-superpower dominance (cooperationmode);” and“D-devolution of power,” representing a fu-ture in which the superpowers must shareworld power with other groups of nations,and including contingencies for potentialNorth-South conflicts, such as an oil em-bargo.


These WIM projections will become an input tothe Joint Long-Range Strategic Appraisal begin-ning with its 1982 revision. CCTC also plans acomplete update of its data base (facilitated by anew software package) and further refinements inWIM itself that may make it a more flexible toolfor determining future military requirements. Forinstance, CCTC's version of WIM contains 12geographical regions rather than 10 and will soonbe expanded to 14, with the further capability ofsubdividing each new region into five subregions;it also contains 87 rather than 85 age groups and(for the United States and Soviet Union) a labor-skills submodel that further divides the population

into male or female and urban or rural; and therehave been similar refinements in the agriculturaland materials submodels. These improvementscreate data problems, however, since reliable dataare not available for many subregions and sectors.CCTC is working with the Bureau of Mines to up-date and expand the data base for the materialssector; in addition, J-5 has instructed CCTC tocontact other Federal agencies about possible coor-dination of global modeling and strategic assess-ment activities. Such coordination might be facil-itated in at least two cases by the fact that the De-partment of Agriculture, as well as the Bureau ofMines, is already using a version of WIM.

Latin American World Model—Catastropheor New Society?

Origin and Purpose

When the Club of Rome presented the prelim-inary results of World 3 at a 1970 meeting in Riode Janeiro, the reaction of the mostly Latin Ameri-can audience was strongly negative. They felt thatpredictions of global crises, based on extrapolationof present trends and arrangements, reflected a pa-rochial developed-world perspective; for two-thirds of the world’s people such crises are alreadyat hand. The audience refused to accept scenariosthat implicitly curbed development and widenedthe income gap, and they felt that policies aimed atachieving a state of global equilibrium wouldmerely ensure that the present disparities and in-equities in the world system are perpetuated. Theyresolved, therefore, to design a model of an egali-tarian “ideal society” in which basic human needs(not profits) would be the basis for resource alloca-tion. The purpose of the model is to demonstratethe material viability of such a society, and there-by to demonstrate “that the different countriesand regions of the world (particularly the poorest)could reach the goals we advocate in a reasonableperiod of time,” relying primarily on their ownhuman and economic resources.43

This global modeling effort was carried out atthe Fundacion Bariloche in Argentina, with prin-cipal support from the International DevelopmentResearch Center in Ottawa, and was first pre-sented at the second IIASA modeling conferencein Berlin in October 1974. An expanded version ofthe model, developed for the International LaborOrganization (ILO), was warmly received at the1976 World Employment Conference in Geneva,where “basic needs” were formally adopted as amajor target of development.44 The model contin-ues to have considerable impact through UnitedNations organizations, including the United Na-tions Educational, Scientific, and CulturalOrganization (UNESCO) and the United NationsIndustrial Development Organization (UNIDO) aswell as ILO, and it is by far the most popularglobal model among scientists and decisionmakersin the Third World.45

Unlike World 3 and WIM, which provide condi-tional descriptive forecasts of global trends, theLatin American World Model (LAWM) is openlyand insistently normative—it is not concernedwith “predicting what will occur if the contem-porary tendencies of mankind continue, butrather with sketching a way of arriving at the final

+JAmllCar 0. Herrera, et al., catu.s[~ophe or ~eu, %et>? A Lutm A m e r i c a n

World ,Model (Ottawa: Internatmnal Development Research Center, 1976), p. 8.++cole In WOTM Futures T h e G r e a t D e b a t e , P . 39.+5 Meadow5, Richar&n, and Bruckmann, op. at., P. 92.


goals of a world liberated from underdevelopmentand misery. ”4b Its purpose is not to demonstratethat certain changes might bring the present worldsystem into equilibrium, but rather “to show thefeasibility of solving the fundamental problemsthrough deep socio-political changes.” 47 T h emodel is also distinguished by its emphasis on theideological issues involved in global modeling: themodeling team was composed of humanistic social-ists, 48and on questions of development and redis-

tribution they “explicitly take up a stance favour-able to the Third World in general, and to LatinAmerica in particular.”49


The LAWM team’s goals and purposes leadthem to include some rather unusual assumptionsin the structure of their model. For instance, themodel divides the world into four regions (later15), each of which is treated as an economic unit,which “presupposes total collaboration betweenthe countries forming it.”50 The model contains“simplistic” trade linkages: the contribution of in-ternational trade relative to regional gross na-tional product (GNP) is held constant at the 1970level, and all trade deficits are eliminated by2000. 51 Instead, each region satisfied basic needsthrough “autarchy, ” using almost exclusively localeconomic resources. However, these resources areassumed to be available in unlimited quantitiesand at constant cost: after a static analysis of cur-rent resource data, the modelers conclude that“the environment and its natural resources willnot impose barriers of absolute physical limits onthe attainment of [an ideal] society,” at least notwithin a “historically significant time-scale. ”52

AS

a result, they do not include these physical factorsin the computer model, and in this specific theirmodel reflects a technological optimism akin toHerman Kahn’s.

+ 6 Amllcar 0. Herrera, Procwd[ngs of the Zil IIAS,4 Global l+foddmg Con/ercnce

(Beriln, 1~74), quoted h}, hleaclow$, RIC hardwn, and Bruckmann, op. cit.,p, 9],

‘THURO t). Sc(,lnlk, et al., “The Barlloche !tfodel,” In Meadow, R icha rdson ,

a n d Bruc kmann, op. cit., p. 247.‘S]enntfcr Robinson, “The Latin American V’c,rld Model,” in The G/obul ?000

Report co (h,’ PrLw&n[, \wl. 2, p. 647.~~~ole ,n ~’or/~ Fu[14rcI Tht, GTCU[ Dc/w[c, P. %.5 0 Herrera, et al., op. L i t . , P. ‘~.

‘lRohlnwn, “ T h e L a t i n Amcrlcan V’orld klociel,” p. 642‘ : Herrera, ct al., (Jp. cit., p. 8.

However, the authors do assume a radicalchange in the sociopolitical factors that controlpatterns of resource use—i.e., an equal distributionof consumption between regions and a total, egali-tarian redistribution of income within regions. Inaddition, the model includes no assumptionsabout population policies, although it does includeseveral untested assumptions about the effect ofliving conditions on demographic change. LAWMalso appears to assume automatic growth in pro-ductivity through technological progress, at nocost, at rates between 0.5 and 1.5 percent annuallydepending on the sector.53

LAWM is essentially an economic model thatoperates through optimization procedures; i.e., ithas five production sectors representing basicneeds—food, housing, education, capital goods,and other goods and services—to which labor andcapital are allocated through optimal control tech-niques in such a way as to maximize life expect-ancy at birth, which is taken to be the best indi-cator of general living conditions. These calcula-tions proceed independently for each region from1960 to 2060, but all countries are assumed to fol-low optimal policies after 1980.

Findings of LAWM

The standard or reference run of this optimiza-tion model indicates that all regions except Asiacan reach their basic needs targets within 30 years.Developed nations (including the Communistworld) “can reach high levels of well-being even iftheir economic growth rate is reduced drasticallyin the future” (fig. 11); in reality, economic growthis restricted to between 1 and 2 percent—far belowthe developed region’s capacity for growth—whichthe authors acknowledge “assumes a political deci-sion. ”54

Latin America could fulfill basic needs by theearly 1990’s by maintaining a relatively high in-vestment rate, particularly in housing and educa-tion (fig. 12). The output for Africa closely resem-bles that for Latin America, although basic needsare not met until 2008 and some shortfalls occur inthe housing sector.

[Page Omitted]

This page was originally printed on a gray background.The scanned version of the page is almost entirely black and is unusable.

It has been intentionally omitted.If a replacement page image of higher quality

becomes available, it will be posted within the copy of this report

found on one of the OTA websites.


Figure 12.— LAWM Standard Run for Latin America

1960 1970 1980 1990 2000 2010 2020 2030

Key;

1 (B) Birthrate 7 (C) Total calories2 (5) Percentage of GNP allocated to sector 5 8 (E) Life expectancy3 (4) Percentage of GNP allocated to sector 4 9 ($) GNP per capita in4 (A) Population growth rate 10 (P) Total population5 (M) Enrollment 11 (U) Urbanization6 (V) Houses per family

SOURCE Catastrophe or New Society?

2040 2050 2060

1960 dollars

[Page Omitted]

This page was originally printed on a gray background.The scanned version of the page is almost entirely black and is unusable.

It has been intentionally omitted.If a replacement page image of higher quality

becomes available, it will be posted within the copy of this report

found on one of the OTA websites.

Ch. 2—Major Global Modeling Studies Ž 29

not keep up with population growth; daily food in-take peaks at less than 3,000 calories per capita in2008 and declines steadily thereafter. (These in-vestments would also divert resources from the sat-isfaction of other basic needs, and they wouldprobably prevent economic growth if investmentin capital goods were not fixed at 25 percent.)Only an effective population policy and the use ofnonconventional foodstuffs, both of which the au-thors advocate, could avoid catastrophe in Asiaunder these conditions.

The policy tests conducted with LAWM indi-cate that capital transfers from the industrializedcountries (in isolation from other measures) wouldhave little effect on the above outcomes, but theyalso reveal that both technological progress andinternal income redistribution are vital to achiev-ing regional goals.

●

●

●

In the “international solidarity” run, the devel-oped nations transfer capital aid to Africa andAsia at a rate that rises from 0.2 percent of GNPin 1980 to 2.0 percent in 1990 and thereafter.The result is higher investment rates and fastereconomic growth in the industrialized nations(in order to compensate for the aid), but very lit-tle effect on the time needed to satisfy basicneeds elsewhere and a negligible impact on thefood shortage in Asia.In the “technological stagnation” run, on theother hand, growth in economic productiondue to technical progress falls to zero between1980 and 2000 and remains there. The outcomeis disastrous in every region except the devel-oped nations. Latin America requires a longerperiod of time to satisfy basic needs, particularlyfood and housing, and in Africa and Asia “theeconomic system finally collapses” sometime be-tween 1990 and 2020 as population steadily out-strips production.By far the greatest difference in results, however,comes from the “historical” run, in which theassumption of egalitarian intraregional redistri-bution is replaced by a pattern of consumptionthat reflects current income distributions andsocioeconomic structures. To satisfy basic needsin the same period of time under these condi-tions would require economic growth rates of 10to 12 percent in the developing countries, rates

which “are in fact impossible to attain. ” The au-thors conclude that “at the very best” theirgoals would be delayed by two or three genera-tions, and would require 3 to 5 times more re-sources, under these conditions.55

Conclusions of LAWM

The conclusions the LAWM team draws fromits interactions with the model do not always re-flect the above results (apparent discrepancies arenoted in parentheses):56

●

●

●

●

●

“it is possible to control population growth tothe point of equilibrium by raising the generalstandard of living” (population stabilizes onlyin the developed regions, and is still growingglobally at a rate of 1.1 percent per year in2040);“if the policies proposed here are applied, allof humanity could attain an adequate stand-ard of living within a period little longer thanone generation” (this is true for Asia onlywith an effective population policy and con-siderable development aid);“this equilibrium could be achieved on aglobal scale well before the earth’s capacity toproduce food—the only foreseeable physicallimitation within the time horizon of the mod-el—is fully exploited even if food productioncontinues to be based on currently availabletechnology” (the model assumes considerabletechnical progress at no cost in agricultureand all other sectors, and fails to achieve itsgoals if technology stagnates);“[the] obstacles that stand in the way of theharmonious development of humanity are notphysical or economic in the strict sense, butare essentially sociopolitical;” and“[the] goals are therefore [to be] achieved . . .by a reduction of nonessential consumption;increased investment; . . . the rational use ofland . . . the egalitarian distribution of basicgoods and services; and . . . the implementa-tion of an active policy to eliminate deficits ininternational trade.”

~jlbld., pp. 106-107.

Wbd.t p. 107.‘ L .-

30 • Global Models, World Futures, and Public Policy

United Nations Input-Output World Model—The Future of the World Economy

Origin and Purpose

Like LAWM, the United Nations Input-OutputWorld Model (UNIOWM) represents a ThirdWorld reaction to the unpalatable conclusions ofThe Limits to Growth. However, its central concernis not the satisfaction of basic needs, but rather thenarrowing of the income gap between the rich andpoor nations. The model was commissioned in late1972 by the Centre for Development Planning,Projections, and Policies (CDPPP), a U.N. agencyresponsible for long-range integrated planning. In-itial financial support came from the Governmentof the Netherlands, and subsequent funding wasobtained from the U. N., the Ford Foundation,and the National Science Foundation. WassilyLeontief, the project director, outlined the con-cepts behind the model in his acceptance speechfor the Nobel Prize in economics in 1973; the workof collecting data and building the model was car-ried out by Anne Carter, Peter Petri, and others atBrandeis University. Petri presented the model atthe Fifth IIASA modeling conference in Septem-ber 1977, shortly after The Future of the WorldEconomy was released in New York.

The study was conducted under U.N. auspicesand direction. Although its findings did not repre-sent official U.N. recommendations, the model’sprimary purpose was to determine whether phys-ical or environmental limits would pose a signifi-cant barrier to the economic growth targets set bythe U.N.’S International Development Strategy,which had been proposed by the General Assem-bly in 1970 as the basis for the Second Develop-ment Decade. As modified and expanded by var-ious U.N. agencies, these targets include the fol-lowing:

●

●

●

●

reducing the average income ratio betweenthe developing and developed countries byalmost 50 percent, from 12:1 to 7:1;improving internal income distribution toeradicate mass poverty;creating 1 billion new jobs in the developingworld by 2000;satisfying the basic needs of all people;

●

●

increasing food production in developingcountries by at least 4 percent per year;increasing the developing nations’ share of theworld market in manufacturing to 14.3 per-cent by 1985 and 25 percent by 2000; andachieving a new international economic order(NIEO), including stabilized commodityprices, increased financial and technologytransfers, open markets for the less developedcountries’ (LDC) exports, and a code of con-duct of translational enterprises.


The authors describe UNIOWM as “basically ageneral-purpose economic model and thus applic-able to the analysis of the evolution of the worldeconomy from other points of view, ”5’ notably theenvironmental. However, Leontief has cautionedthat:

We cannot predict the future of the worldeconomy. However, we can rule out of our expecta-tions future scenarios that are internally inconsis-tent and thus impossible.

To rule out internally inconsistent expectationswe need to construct a model that guarantees inter-nal consistency . . . by visualizing the world as asystem of interdependent process in which eachprocess . . . generates certain output and absorbs aspecific combination of input.58

The rigorous accounting required by this input-output analysis forces the model to balance thegrowth of one economic sector against its effect onother sectors; similarly, imports and exports in oneregion must be balanced against the imports andexports of other regions. This technique also per-mits “an unusual degree of detail” in representingparticular industries or regions, which is “advan-tageous” because of its “relatively specific policysignificance. “59 On the other hand, critics have

STW~~SllV LeOnrlef, Anne Carter, and Peter Petri, The Future Of t~ wor~

Econom>: A Uruted Nutlom Stud} (New York: Oxford, 1977), p. 7.58wa5511y Leontlef, “s t ruc ture of the V’orld E c o n o m y : Outline o f a S imple

I n p u t - O u t p u t F o r m u l a t i o n ” ( N o b e l Memorial Lecture), Amencun Econom~cRetwu, D e c e m b e r 1977, p. 823.

5gLeonttef, Carter, and Petr], op. cit., p. 8.


questioned whether this level of explicit detail isworthwhile or justifiable, particularly since it de-mands an enormous amount of data, much ofwhich had to be adapted from other informationand data bases.

60 UNIOWM’s population sector,for example, merely assumes the projectionsprepared by the U.N. Population Division in 1973(see app. A).

The model divides the world into 15 regionscomposed of fairly homogeneous economies, al-though for purposes of interpretation and presen-tation of results these regions are further aggre-gated in three categories: developed nations (eightregions), resource-rich LDCs (three regions), andresource-poor nations (four regions). Each region’seconomy contains 45 sectors of economic activity,described by 175 equations with 229 variables.Prices are calculated in a separate submodel, and(as in LAWM) the representation of internationaltrade has been kept “almost artificially simple. ”61

The environmental sector includes eight pollut-ants and five types of abatement activities, but themodel does not reflect the effect of developmenton ecological systems, nor does it contain anyother feedback loops; “it cannot, in any sense, beviewed as a dynamic model. ”62 The model’s equa-tions are solved simultaneously, usually at 5-yearintervals, in order to provide “snapshots” of theworld in 1980, 1990, and 2000.

The model can be applied to a wide variety oftasks, but its utility is limited by its large data re-quirements and by the many controversial as-sumptions that have been included. This has ledone critic to conclude that:

The huge number of assumptions made inestimating time trends for input-output matricesmakes for confusion when it comes to consideringthe model as a whole. There are so many assump-tions that one is hard put to evaluate the reason-ableness of the total picture. 63

Nevertheless, the model could be and has beenused for a wide range of policy tests that reflect the

S o sam Cole, G/oba[ LIo&/s and [he Inw-nattonal Economc ~d~ A pQPH ~or t~

L’N’ITAR ProJecr on the Future (New York: Pergamon Press, 19?7), p. 22; and .len-nlfer Rc>hlnqc>n, “(J, N. World )wfodel,” In The G1oIxI1 2(XXI Report to the Presdenc,

, 0 1 . 2 , p. 64961Rc)b1n\On, “IJ,N. World Nlodel,” p. 6 5 2 .

‘~ifeadows, RIL hardson, and Bruckmann, op. c[t., p . 167; a slmllar crltlclsmcould he made of ‘W’Ihf.

b~Rohlnson, “~1.N. N’orld Model,” p. 652.

interests of the modelers and the organizationsthat commissioned them.

Findings of UNIOWM

The study’s optimistic findings, particularly that“no insurmountable physical barriers exist withinthis century to the accelerated development of thedeveloping regions” and that “pollution . . . is atechnologically manageable problem,”64 receivedwidespread attention in the media, where it wassometimes reported that UNIOWM “discredited”The Limits to Growth. However, the authors havecautioned that the model “cannot settle, and wasnot designed to settle, the many fundamentalquestions raised in The Limits-to-Growth debate.”65

And in fact the optimism of these general state-ments is not supported by the specific results ofmost of UNIOWM scenarios.

Policy tests conducted for the UN have includeda number of different scenarios relating to eco-nomic growth rates and per capita income gaps.

●

●

The “old economic order” scenario is basedon historical trends in internal and externalinvestment and existing international eco-nomic arrangements. Income per capita growsin all three categories of nations but incomegaps increase, despite decelerating growth inthe developed regions after 1990, and someLDCs would face an absolute decline in livingstandards (see fig. 14). This scenario “turnsout to be rather pessimistic, ” according to theauthors, and because of its dim economicprospects for the developing regions it is“downplayed in the UN documentation. ”66

The standard run, based on the minimumgrowth targets of the International Develop-ment Strategy (IDS), also turns out to be rela-tively pessimistic. Because of their higher ratesof population growth, accelerated economicgrowth in the LDCs does not lead to corre-sponding gains in GNP per capita. The in-come gap between developed and less devel-oped regions remains at the current 12:1 ratio.

6+ LeOntle(, Carter, and Petri, OP. Cit., P P. 48-49.

“Peter Petri , “An Introductmn to the Structure and Applications of the~lnlred N a t i o n s ~’or[d hfodel,” App/,ed hfa[hm[tcal ,Mode/lng, vol. 1, No. 5,

-June 1977, p. ?62.66 Leontief, Carter, and Petri, op. cit., p. 73; Robinson, “L1. N. World Model,”

p. 653.

32 . Global /Mode/s, Wor/d Futures, and Public Policy

Fig

7,000

6,000

5,000

4,000

3,000

2,000

1,000900

800

700

600

500


● Four additional runs were then conducted byaltering the standard or IDS scenario in sucha way as to reflect more optimistic assump-tions about: 1) resource endowments; 2) in-creased foreign aid from the developed re-gions; 3) fewer constraints on balance-of-payment deficits; and 4) faster agricultural in-vestments to achieve food self-sufficiency inlow-income Asia.

None of these scenarios, however, was capable ofproducing the desired reduction in per capita in-come ratios. In two final scenarios, therefore, theauthors preset the model in such a way as toroughly halve the income gap by 2000 and close itcompletely by about 2050, and then solved itsequations to determine the investment andgrowth rates that would produce the desiredresults:

● Scenario “C,” based on the U.N.’s “low”population growth projections (see app. A),requires a 6.9 percent GNP growth rate in theLDCs to reduce the income ratio to 7.15:1 by2000.

● Scenario “X, ” based on the U.N.’s “medium”population projections, requires an evenhigher GNP growth rate of 7.2 percent in theLDCs and reduces the income ratio only to7.69:1. Scenario “X” also requires a fivefoldincrease in overall agricultural output in thedeveloping regions, including a nearly tenfoldincrease in resource-poor Latin America.

In subsequent policy tests, UNIOWM has beenused to examine the economic consequences ofmineral- and energy-conservation strategies. Astudy of the future production and consumptionof nonfuel minerals, based on the resource conser-vation strategies of the Economic Council of Can-ada, is nearing completion but has yet to be pub-licly documented.b7 Another study, conducted forthe U.S. Department of Commerce, compares the“old economic order” scenario with an “energyconservation” scenario based on the maximumreasonable reduction in fossil fuel consumptionover the next 20 years through the substitution oflabor and capital for energy. It revealed that ener-gy conservation could reduce the balance-of-pay-ments deficits of both developed regions and

b~kleadows, Richardson, and Bruckmann, op. cit., p. 175; Rohlnson, “U.N.

‘ik’orld ktodel,” p. 653.

resource-poor LDCs and allow increased GNPgrowth in the latter, but that the capital require-ments for conservation would require a 17- to23-percent increase in the savings rate.ba

More recently, Leontief has conducted twostudies of the economic implications of the NIEOand of worldwide military spending, with fundingfrom the U.N. and from American disarmamentorganizations.b9 In the NIEO scenario, theresource-poor LDCs are allowed to import what-ever quantities of goods and services are requiredto reduce income ratios by 50 percent by the year2000, with their balance-of-payments deficits–upto 75 percent of their imports—to be financed by“extraordinary credits” from the developed re-gions and resource-rich LDCs, carrying a nominal5-percent interest rate. The model output for thisscenario (see fig. 1.4) shows that the developedregions, w h i c h w o u l d h a v e t o ‘‘workovertime . . . to provide [these] huge amounts ofeconomic aid,” would have a higher GNP butlower per capita consumption in 2000, at whichtime they would be allocating 3.1 percent of theirtotal GNP to development assistance. Leontiefhimself doubts that such a plan could be imple-mented:

On the whole this projection of the future devel-opment of the world economy under the new eco-nomic order suggests that the practical possibilityof carrying out such an optimistic program must beseriously questioned.70

As an alternative, Leontief proposes an “armslimitation” scenario, noting that the current half-trillion-dollar annual worldwide defense spendingrepresents “the largest existing economic reservethat might be utilized to accelerate the growth ofthe resource-poor less developed regions. ” Wherethe “old economic order” scenario assumed thatall regions would continue to devote the same per-centage of their respective GNPs to defense thatthey had in 1970, the “arms limitation” scenarioassumes that by 2000 the defense expenditures ofthe United States and the Soviet Union would bereduced by one-third, and that all other regions

6B.Anne P. carter and Alan K. Sin, “An Energy Ccmser~wt[on Scenario for

t h e V’orlcl kfodel,” prepared for the Bureau of Internatmnal and Econornlc

POIICY, L1.S. Department of Commerce, No~ember 1977, p. 2.s~L~eadc>ws, R i c h a r d s o n , a n d Bruckmann, op. clt., P. 175.

~OM’assll}, W’, Leontlef, “ T h e V’orld E~onomv o f t h e Y e a r 2030,” Sclentlfic

A m e m x m , jol. 243, No. 3, Septemher 1980, p. 230.


would reduce defense spending 25 percent by 1990and 40 percent by 2000. The savings realized ineach region would first be used to satisfy its owncivilian needs, but the developed regions would al-locate 15 percent of their savings to developmentaid by 1990 and 25 percent by 2000. The modeloutput for this scenario (see fig. 14) indicates thatper capita income and consumption in the re-source-poor LDCs would increase far faster thanunder the old economic order. Since developed-region defense savings would be given to LDCs inthe form of direct aid, their balance-of-paymentsdeficits would also be far smaller. Based on a com-parison of these scenarios, Leontief concludesthat:

. $ . the reallocation of economic resources arisingfrom the kind of international arms-limitationagreement that has been suggested repeatedly, bothformally and informally by individuals and organi-zations inside and outside the U. N., is by far themost promising of the three schemes for world eco-nomic development .71

Conclusions of UNIOWM

The results of these numerous UNIOWM sce-narios suggest, in general, that the economic pros-pects of the resource-poor LDCs are not very opti-mistic. The growth rate targets of the U.N.’s Sec-ond Development Decade are insufficient to beginclosing thedevelopingtaken into

7 Ilbld., p. 231.

income gaps between developed andregions when population increases areconsideration. The limits imposed by

mineral resources, agriculture, and the environ-ment are not insurmountable and could be over-come through appropriate policies and invest-ments; but “the principal limits to sustained eco-nomic growth and accelerated development arepolitical, social and institutional in character.”TzTotwo

●

●

achieve accelerated development, therefore,general conditions are considered necessary:

far-reaching internal reforms in the LDCs in-cluding often drastic changes in sociopoliticalinstitutions and economic policies—between30 and 40 percent of GNP ‘must be used forcapital investment, particularly in the agricul-tural and export sectors, and both equitableincome redistribution and increased public-sector participation are needed to increase theeffectiveness of these investments; andsignificant reforms in the international eco-nomic order, aimed at reducing the potential-ly large balance-of-payments deficits in thedeveloping regions—stabilizing commoditymarkets, stimulating exports of manufacturedgoods from the LDCs, and increasing finan-cial transfers from the developed regions andresource-rich LDCs.

Neither of these conditions, taken separately, issufficient to ensure a favorable outcome: 4’Ac-celerated development leading to a substantialreduction of the income gap between the develop-ing and the developed countries can only beachieved through a combination of both of theseconditions. ”73

TILeontief, Carter, and Petri , op. cit., p. 4~.

TIIbld., p. 11.

Global 2000—Entering the 21st Century

Global 2000 is a global modeling study, rather plus a number of projections based on analyticalthan a “global model” in the same sense that techniques other than computerized simulationWorld 3, WIM, LAWM, or UNIOWM are. Its models. Despite the very limited degree of interac-projections result not from a single, integrated tion among the sectors and agencies, however, andmodel but rather from a collection of sectoral despite the frequent lack of consistency in theirmodels, independently developed or adopted by various assumption and data bases, these sectoralvarious Federal agencies and other organizations, “submodels” collectively provide the U.S. Govern-


ment with the same type of projections that theother more integrated global models produce fortheir users.74

Origin and Purposes

Global 2000 was carried out by an interagencytask force of the U.S. Government in response to adirective issued by former President Carter in hisenvironmental message to Congress on May 23,1977:

Environmental problems do not stop at nationalboundaries. In the past decade, we and other na-tions have come to recognize the urgency of inter-national efforts to protect our common environ-ment.

As part of this process, I am directing the Coun-cil on Environmental Quality and the Departmentof State, working in cooperation with . . . other ap-propriate agencies, to make a one-year study of theprobable changes in the world’s population, natu-ral resources, and environment through the end ofthe century. This study will serve as the foundationof our longer-term planning. ’s

This mandate, as interpreted by the Global 2000task force, imposed dual objectives on the study:its purpose was not only to “identify [future] prob-lems to which world attention must be directed,”but also “to identify and strengthen the Govern-ment’s capability for longer-term planning andanalysis. ”7b The resulting report, released in July1980, addressed both of these goals, although rely-ing on the Government’s existing capability mayhave detracted from the accuracy and usefulness ofthe resulting projections. According to a Scienceeditorial:

A reading of portions of the report producedafter 3 years reveals more about the functioning ofthe federal government than it conveys new reli-able information about the future of the world. ”

‘ + G e r a l d 0. Barney, stud} director, The Globul 2000 Report m the Pre.wdetr[:

Enwmg the Tu mt>-fmt Cenrum (V’a~hlngrnn, D. C.: U.S. C o u n c i l on En-~vronmental Qualitv a n d D e p a r t m e n t o f S t a t e , 1980, ~’01. ~, PP. ~’-vI.

75Jlmmv C a r t e r , The Pre~ldmrr’~ En{ [mnmental P r o g r a m 1977 (W’ashlngton,

D. C.: U.S. Government Prlntlng Offlce, 1977), p. hi-l 1.?SBarnel,, The G/o~a/ ~~] ~ePf)7[ tO [~, pr~s&nt, km!. 1 , p . 6.

~~rhdlp H , A b e l s o n , “The Global 2@W Reporr,” Scwnce, vol. Z@9, No. 4458,

A u g . 1 5 , 1980, p. 76].

Various Federal agencies are already conductinga considerable amount of long-term analysis andplanning, and a number of them have the capabil-ity to produce projections based on extensive databases and sophisticated sectoral models, many ofthem computerized. These existing tools and pro-cedures (and the skilled personnel who use them)represent the “present foundation” of the Govern-ment’s long-range global planning—they embodythe assumptions on which current analysis isbased, and they are actually being used as at least apartial basis for current planning and decision-making. As a result, the study plan chosen by thetask force was “to develop trend projections using,to the fullest extent possible, the long-term globaldata and models routinely employed by the Feder-al agencies. ”78

However, they found that “each agency has itsown idiosyncratic way of projecting the future, ”based on its individual planning requirements andarea of responsibility .79 As a result, each agency’sprojections tend to focus on a single factor (such aspopulation, food, or energy) without adequatelyconsidering the feedback involved in a systemwhere these factors are interacting variables. Fur-thermore, although these separate projections“have generally been used by the Government andothers as though they had been calculated on amutually consistent basis, ” the different agencies’models “were never designed to be used as part ofan integrated, self-consistent system. ”8° This leadsto one of the study’s basic findings:

To put it more simply, the analysis shows thatthe executive agencies of the U.S. Government arenot now capable of presenting the President withinternally consistent projections of world trends inpopulation, resources, and the environment for thenext two decades.81

Despite these deficiencies, Global 2000 presentsthe most comprehensive and consistent set of pro-jections yet produced by the U.S. Government,and it represents the first attempt to make such

7L3T~, ~~oba/ 2000 Rcport to the Pr~~[dent, fw1. 1 , p. 6.

Wbid., J,oI. 2, p. 454.‘OIhld.8’Ihld.


projections on a coordinated, integrated basis.The task force has been disarmingly frank andf o r t h c o m i n g i n t h e i r a n a l y s i s o f “ t h eGovernment’s global model,” and their discussionof its current weaknesses points to a number ofways in which existing long-range analysis andplanning tools can be improved. Several of themodels have in fact been modified or expanded inthe last 3 years, often in response to problemsidentified by the task force, although many prob-lems still remain.

82 T he task force cautions, how-ever, that “in the absence of ongoing institutionalincentives to address cross-sectional interactions,the present form of the government’s global modelis not likely to change significantly in the foresee-able future. ”83


For the purposes of Global 2000, the study teamimposed a “special limited discipline, ” within thetime and resource constraints of the study, underwhich: 1) the assumptions, structures, and projec-tions of the agencies’ sectoral submodels weremade “more mutually consistent” and 2) the out-put from one sector was used as the input foranother “whenever this was readily feasible. ”84 De-spite these efforts, however, the Government’smodel has almost no feedback loops and remainsat best “quasi-integrated,” with the result that, “ifanything, the severity of the effects of these basictrends may be understated. ”8S In addition, thesubmodels employ different patterns of regionali-zation, varying from as few as five to as many as 28regions, with a similar variation in the degree ofdetail provided.86 This not only makes coordina-tion difficult but also leaves the projections with-out a consistent geographic reference for policy

analysis.

Furthermore, there are numerous major incon-sistencies in the values assigned to the same vari-able in different sectoral submodels, reflecting mu-tually contradictory agency assumptions about the

behavior of crucial factors.87 The population sec-tor, for example, assumes that birth rates in LDCswill decline because of continued moderate socio-economic development, whereas the agriculturaland economic projections indicate only marginalincreases in global food and GNP per capita, withreal declines in some LDCs—hardly reflectingmoderate socioeconomic development.88 Similar-ly, the GNP submodel assumes that the real priceof wheat will decrease by 0.6 percent per year andthat the real price of oil will remain constant dur-ing the early 1980’s, whereas the food and energyprojections indicate real price increases of 2.1 and5.0 percent per year, respectively, over roughly thesame period.

Far more serious, however, is the absence of anyconsistent accounting of capital or resource alloca-tions in any of the sectoral submodels. This leadsto what the report calls “significant omissions anddouble-counting’ –in effect, the model recognizesno conflicts from competing demands or uses, andit places no constraints on the amount of capitaland resources available to each sector.89 Underthese conditions, the same acre-foot of water is as-sumed to be available for both irrigation andenergy development, just as the same barrel of oilis assumed to be available for transportation, ener-gy generation, and petrochemical feedstock.

Different sectors also contain contrasting or con-tradictory assumptions about the course of publicpolicy, despite the report’s frequent repetition ofthe caveat that its trend projections are made “un-der the assumption that present policies and policytrends continue without major change. “9° The mostsignificant exception to this rule comes in the pop-ulation projections, which are based on the as-sumptions of: 1) continued socioeconomic progressdespite marginal gains in food and GNP per capitain the LDCs; and 2) adoption of family planningpolicies in all countries and major extensions of ex-isting programs, especially in rural areas. Theother sectoral projections, however, are based onequally significant policy assumptions, includingthe following:91

ezlbid., vol. 2, p. 46J) n. 2.

a]lbld., VO[. 2, p. 461.*’Ibid., vol. 2, p. 457.e>[hid., J*oI. 2, pp. 456, 481.s’$Ibid., VOI. 2, pp. 485 n. 1, 478; see also table 14-3, p. 479, and the n-d-md-

ological maps following p. 442.

eTIbld ~,ol. 2, pp. 46 I -476; see also tahle 14-2 and pp. 470-475 for an extensive.!discussion of “selected contrasrtng assumptmns,”

ssIbld., vol. 2, pp. 481-482.

aglbld., vol. 2, p. 467.~OIb,d, ~,oI, 2, p. 3; emphasis theirs.$IIIM,l \(~I, 2, table 14-2 and ~Ip. 470-475; see also pp. 485-4~.

Ch. 2—Major Global Modeling Studies . 37

●

●

●

●

GNP.—Implementation of “prudent” policiesto maximize export earnings, with GNPgrowth in the LDCs largely dependent onGNP growth in the developed regions;Food.—Major public and private investmentsin land development; a worldwide shift to-ward more fossil-fuel-intensive agriculturaltechniques and inputs; and (implicitly) im-proved resource management to protect fish-eries and to prevent overgrazing, erosion, andfarmland degradation;Energy.–Widespread deployment of light-water nuclear electric powerplants; implemen-tation of more effective energy conservationprograms in OECD countries; willingness ofOPEC countries to meet oil demand up totheir maximum production capacity; and ma-jor public and private investment in air pollu-tion abatement so that by 1985 all countries’energy facilities are retrofitted to meet U.S.new-source performance standards for CO,S OX, NOX, and particulate; andTechnology.–Major technological progressin almost ‘all sectors, with no ‘technologicalsetbacks or adverse side effects; and extensiveworldwide transfer and deployment of family-planning, yield-enhancing, nuclear-power,and pollution-abatement technologies.

In addition, the “no-policy-change” assumptionitself explicitly excludes the possibility of eitherplanned change or sudden upheaval in the world’sexisting political institutions and economic ar-rangements:

. . . the Study assumes that there will be no majordisruptions of international trade as a result of war,disturbance of the international monetary system,or political disruption. The findings of the Studydo, however, point to increasing potential for inter-national conflict and increasing stress on interna-tional financial arrangements. Should wars or a sig-nificant disturbance of the international monetarysystem occur, the projected trends would be alteredin unpredictable ways.92

Findings and Conclusionsof Global 2000

In the absence of more extensive policy testing,and because of the presence of contradictory and

often controversial policy assumptions, Global2000 does not provide an adequate basis for coor-dinated analysis or detailed policy recommenda-tions. Furthermore, because of the omissions andinconsistencies outlined above, the study teamconcludes “that it is impossible to assign a highprobability to any of the specific numeric projec-tions presented” for its different sectors. 9 3

However, the current weaknesses and deficienciesof “the Government’s global model” do not neces-sarily or completely invalidate its overall results,and the study team concludes that “these basicfindings are qualitatively correct,” since they arein general agreement with past projections by thesame agencies, are supported collaterally by thealternative sectoral projections of outside organiza-tions, and correspond “in . . . their most basicthrusts” with projections generated by “less com-plex but more highly integrated global models. ”9q

Global 2000’s often-quoted general conclusionsabout the future are as follows:

If present trends continue, the world in 2000 willbe more crowded, more polluted, less stable ecolog-ically, and more vulnerable to disruption than theworld we live in now. Serious stresses involvingpopulation, resources, and environment are clearlyvisible ahead. Despite greater material output, theworld’s people will be poorer in many ways thanthey are today.

For hundreds of millions of the desperately poor,the outlook for food and other necessities of lifewill be no better. For many it will be worse. Barringrevolutionary advances in technology, life for mostpeople on earth will be more precarious in 2000than it is now—unless the nations of the world actdecisively to alter current trends.95

The principal sectoral findings on which thesegeneral conclusions are based, briefly outlined, areas follows:

● Population.—Global population growth rateswill not decline significantly by 2000 and, inabsolute terms, net population growth will befaster than it is today. The world’s total popu-lation will increase by 55 percent, from 4.1billion in 1975 to 6.35 billion in 2000, with 92percent of the growth occurring in the LDCs,particularly in Africa and Latin America.

9’Ihld. , \ [)1. 2, p. +31.‘+Ihd.

‘51hd. , \ 01, 1, p. 1.

.

38 “ Global Models, World Futures, and Public Policy

● GNP.—Worldwide GNP is projected to in-crease 145 percent during the 1975-2000 per-iod, with faster annual growth in the LDCs(4.5 percent) than in the developed nations(3.3 percent), although growth rates in allregions will decline after 1985. Due to dif-ferential population growth, however, GNPper capita will grow much more slowly—anoverall increase of only 53 percent worldwide,with marginal improvements or actual de-clines in a number of LDCs in Africa andSouth Asia. Existing income disparities be-tween the richest and poorest nations willwiden, and “dramatically different rates ofchange would be needed to reduce the gap sig-nificantly by the end of the century. ”9b

● Food.—Worldwide food production is pro-jected to increase by 2.2 percent annuallyfrom 1970 to 2000, a rate approximating therecord increases of the Green Revolution.Since most of the good arable land is alreadyunder cultivation, most of this increase willcome from more intensive use of energy-inten-sive inputs and technologies, resulting in anincreased dependence on oil and gas and atleast a doubling of real food prices by the endof the century. Since food production growsmore rapidly than population, average percapita consumption will increase 15 percentworldwide by 2000, but with significant re-gional variations—increases of 21 percent inthe developed regions but only 9 percent inthe LDCs, with smaller increases in NorthAfrica, the Middle East, and South Asia; anda “calamitous” 19. l-percent decline in CentralAfrica, where average caloric intake is alreadywell below the minimum requirements set bythe U.N. Food and Agriculture Organization.These projections suggest the need for foodimports and food assistance will continue togrow in the developing regions, particularly inthe poorest countries.

● Energy .-The energy projections, made inlate 1977, indicate that world energy demandwill increase 58 percent over the 1975-90 peri-od. However, petroleum production capacityis not increasing as rapidly as demand, andthe rate of reserve additions per unit of ex-

~~[b,~, ~,c)l 1 ~,, 13; ‘c,~l,il~c tl,lth rhe Jlic u+ion ()( L ‘h’IIJ\~’hl, Jk)\ c.

●

●

ploratory effort appears to be declining. As aresult, technical considerations indicate thatpetroleum production will peak before the endof the century, although political and eco-nomic decisions by OPEC countries couldcause production to level off even earlier. Theresulting transition away from petroleum de-pendence is projected to be led by nuclear andrenewable sources (primarily nuclear, but in-cluding hydro, solar, and geothermal), whichare forecast to increase 226 percent by 1990;production of petroleum, natural gas, andcoal is projected to increase by 58, 43, and 13percent, respectively, over the same period.The projections also indicate considerable po-tential for reducing energy consumption perunit of economic production.Resource prices.–Global 2000 finds that thereal prices of food, fish, lumber, water, and en-ergy will increase significantly by 2000, withthe steepest increases occurring after 1985.However, this finding shows how the noninte-grated model can lead to an economic para-dox:

If the real prices of these commodities in-crease as projected, for what correspondingcommodities will real prices decrease? If nocompensating real-price decreases are pro-jected, what do these “real” price increasesmean theoretically—or even semantically?Unfortunately, even attempting to developanswers to these difficult questions wouldhave exceeded the time and resource con-straints of the study .97

Environment.–Major strains will be placedon ecological systems throughout the world,leading to significant deterioration in terres-trial, aquatic, and atmospheric resources thatwould have adverse impacts on agriculturalproductivity, human mortality, and economicdevelopment. There are already signs of manyof these effects, which will be felt more strong-ly, particularly in the LDCs, toward the endof the century. The projected increase in fossilfuel combustion could be expected to doublethe COZ content of the atmosphere by 2050,leading to a 2° to 3° C rise in temperaturesand significant alterations in weather and pre-cipitation patterns in the temperate zones,


●

where most of the world’s food-exporting na-tions are located.Species extinctions.—Between 0.5 millionand 2.0 million species of plants and animalscould be extinguished by 2000, mainlythrough the loss of wild habitats or throughpollution. This threat is particularly great inthe tropical forests, which are an importantpotential source of new foods, pharmaceuti-cals, and building materials. An equally im-portant threat is posed by the possible loss oflocal and wild varieties that are needed tobreed pest-and disease-resistant traits intohigh-yield cereal grains.

Updates of these projections, developed on thebasis of subsequent events or improvements in theforecasting tools, generally support the initial find-ings of Global 2000 and, if anything, provide evenless

●

●

●

reason for optimism:

Fertility rates have declined more rapidly thanexpected in some areas, but world populationin 2000 will be only 3 percent lower than orig-inally projected.GNP projections are somewhat lower, due toincreased petroleum prices and efforts to con-trol inflation in OECD countries, with a con-sequent drop in LDC growth rates.Agricultural projections have also been re-

●

vised downward, due to rapid increases in en-ergy-related production costs and diminishingreturns for other yield-enhancing inputs. Inaddition, increased concern with the conse-quences of intensive cultivation has led (in theUnited States and other developed nations) topressure for resource-management policiesthat would prevent further erosion and soildeterioration. Some LDC governments are in-tervening in domestic markets to keep foodprices low, often to the detriment of ruraldevelopment and production capacity.The greatest differences are found in theenergy sector: updated projections, reflectingthe sudden large increase in oil prices in 1979,show that demand will be lower due to higherprices and slower economic growth caused byenergy impacts in other sectors. Estimates ofmaximum OPEC production levels are lower,reflecting the cartel’s resource-conservationpolicies. Estimates of future OECD nuclearcapacity are also lower, reflecting constructiondelays and public concern as well as the U.S.licensing moratorium, and coal is projected toprovide a larger share of energy supplies.Higher prices are also expected to encouragethe adoption of alternative sources (includingsolar) and conservation measures.

CHAPTER 3


Contents

PageIntroduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Qualitative Conclusions About the World and Its Future. . . . . . . . . . . . . . . . . . . . . . . . . 43

Population Projections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Agricultural Projections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Technical Progress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48Uncertainties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

Energy Projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

T a b l e s

Table No. Page2. Comparison of Short- and Long-Term Population Projections . . . . . . . . . . . . . . . . . . . . . .......453. Percentage Increases in Projected Global Food Production, Food Prices, and Food per

Capita, 1970-2000. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

CHAPTER 3


Introduction

The five global modeling studies addressed inthis report demonstrate at least three fundamen-tally different “predictive styles’ ’—World 3 andGlobal 2000 examine what might happen if pres-ent trends continue, while the Latin Americanand United Nations (U. N.) world models examinethe goals that might be achieved through broadchanges in those trends, and the World IntegratedModel (WIM) examines the policies and actionsthat might bring those changes about.1 The mod-els also vary significantly in their more specificpurposes, assumptions, and methodologies. In ad-dition, they focus on different parts of the worldsystem, at different levels of detail, and over dif-ferent spans of time. These fundamental differ-

ISee S t u a r t 13remer, “Test)ng ~fndel~,” in Donella H. Nfeadows, John Rtch-

ardscln, and Gerhart Bruckmann (edi. ), (%oplng ~n the Dark. The First Decude Of

Global Nfodd[ng (New Y o r k : N’llev, forth~(~mlng), pp. 376-~77.

ences, as well as the more specific differences inpatterns of regionalization and degrees of aggrega-tion, make it difficult to compare their projectionsin any extensive or sustained manner.

The five models nevertheless display a limitedconsensus about the nature of the world systemand the identity of the problems facing it, as wellas some of the steps that might be taken to addressthem. The following discussion will examine theareas of general agreement or disagreement thatemerge from these five studies, first in theirqualitative conclusions about the general prob-lems of the world future, and then in their quan-titative findings in three key sectors: population,food, and energy. Extended technical analysis oftheir projections in these three key sectors and ofthe structural differences between the models isprovided in the appendixes.

Qualitative Conclusions About the World and Its Future

Despite the many differences between these fiveglobal models, it is possible to draw a number ofcommon themes from them about the presentstate of the world and the possible paths it mightfollow in the future. The following statements arebased on a list compiled by Donella Meadows,John Richardson, and Gerhart Bruckmann for aforthcoming review’ of the first decade of globalmodeling. 2 The statements reflect a number ofqualitative findings with which (according to theauthors) almost all global modelers would agree,and they are arranged in such a way as to form aloose logical argument:

● Population and physical (material) capital can-not grow indefinitely on a finite planet.

● There is, however, no reliable or complete in-formation about the planet’s ultimate carrying

—1hicadcm S, R i c h a r d s o n , a n d 13ruckmann, op CI[., pp 1S- 19, +9-50; see also

John Nf. Rlchardw)n, jr., “Tmvardi Effcctlie F o r e s i g h t In t h e U n i t e d State<(jc>,crnment” (prepared (or rhe LI. S. Department of Srsre, June 1~7~), pp 5-6and app. C.

●

●

capacity. * There is a great deal of partial infor-mation, which optimists read optimisticallyand pessimists read pessimistically.Nevertheless, there is no known physical ortechnical reason why the basic human needs*of all the world’s people cannot be suppliednow and into the foreseeable future, Thesebasic needs are not now being met because ofpolitical, economic, and social factors, not be-cause of overall physical scarcities.Continuing “business as usual” policies overthe next few decades will not lead to the bestpossible outcome, nor to a desirable outcome,nor even to the satisfaction of basic humanneeds. It would result instead in an increasinggap between the rich and the poor, worseningeconomic conditions, growing internationaltension, problems of resource availability, andenvironmental degradation.

*The terms “carrying capac ltv ” a n d “haslc h u m a n need” are h]ghlv value-laden and therefore ln&capablv lead to dehare.

43


●

●

●

●

●

●

●

Because of these difficulties, the continuationof current trends is not a likely future course.Over the next three decades, therefore, theworld’s socioeconomic system will be in a peri-od of transition to some new state that will beboth quantitatively and qualitatively differentfrom the present.The exact nature of this future state, andwhether it will be better or worse than thepresent, is not predetermined—it is a functionof decisions and changes being made now.Because of the momentum inherent in theworld’s physical and socioeconomic processes,policy changes that are made soon are likelyto have more impact with less effort and costthan the same set of changes made later; andif the changes are put off for too long, theymay not work at all.Changes in technology are expected and in-deed essential: even the most optimistic sce-narios might fail if technological progress is in-adequate. However, no set of purely technicalchanges tested in any of the models was suffi-cient in itself to bring about a desirable out-come. The models suggest that restructuringsocial, economic, and political systems willalso be necessary and may in fact be more ef-fective.The interdependencies among peoples and na-tions, over time and space, are far greater thancommonly imagined: actions taken at onetime in one part of the world have far--reaching consequences that are often difficultto anticipate intuitively and are probably im-possible to predict (totally, precisely, perhapsat all) even with computer models.Because of these interdependencies, isolatedmeasures intended to reach narrowly definedshort-term goals are likely to be less effectivethan anticipated. Decisions should thereforebe made within the broadest possible context,across space and time and intellectual disci-plines.As a further consequence of these inter-dependencies, cooperative, long-term ap-proaches to achieving individual or nationalg;oals often turn out to be more beneficial toall parties than short-term, competitive ap-proaches.

● Many existing plans, programs, and agree-ments—particularly complex internationalones like the U.N."s International Develop-ment Strategy—are based on assumptionsabout the world that are either mutually in-consistent or inconsistent with physical real-ity. Much time and effort have thus beenspent in designing and debating policies thatare in fact simply impossible.

In short, according to the authors, the modelersgenerally agree that the world system is going tochange in the near future, and that a continuationof current trends and policies will lead to a changefor the worse. They also agree that changes for thebetter are possible, although they disagree sharply

on what those changes should be and which poli-cies would bring them about. In the authors’words, the models indicate that “we should dosomething, [but] we can’t be sure what we shoulddoo”

3 Environmentalism and the Club of Rome’s“world problematique” (see below) seem to haveinfluenced the earlier models, which stressed thelimits of the present system and called for an “equi-librium state” or “organic growth.” The more re-cent models, which stress the inequities of the pres-ent system and call for internal and/or interna-tional redistribution of growth and consumption,seem to have been influenced more directly by theissues surrounding the “new international eco-nomic order. ”

In spite of these differences in emphasis and pre-scription, however, general agreement does emergeabout the fundamental issues or “problem nexus”for which projections must be made and solutionsfound.4 The following sections, therefore, sum-marize projections made by the five global modelsin three of these crucial areas:

●

●

●

population, which is addressed by all of themodels and is the most fundamental drivingvariable in most of them;food supply, the most basic of human needsand the most promising basis for comparisonsbetween the models; andenergy, which reflects the more general prob-lems of resource depletion but has a uniquely

important impact on agriculture and econom-ic activity.

‘Meadows, Richardson, and Bruckmann, op. cit., p. 51,

4Bremer, op. cit., p. 375.

Ch. 3—Findings of the Global Models ● 4 5

The purpose of these summaries is not to arrive ata “consensus projection, ” but rather to illustratethe similarities and differences among the models.

More detailed information and technical analysescan be found in the historical survey and appen-dixes.

Population Projections

Table 2 shows the results of a number of studiesof future population growth, including three thatdid not employ global models. Population projec-tions play a key role in any assessment of futureworld conditions, since the size of the populationwill determine the number of consumers of goodsand services and the number of people available toproduce those goods and services. In some globalmodels, the future size of the population is pro-—

5The following dlscussmn N htghly condensed; see app. A for a more detailedcomparative analysis of population projections. For further Information on thistopic, see OTA’s assessment, Wodd Populutlon and Fertd~t? Pkmnmg Techndog[es:77w ,Nexr 20 Yetm, OTA-HR-157 (Washington, D. C.: U.S. Government Prlnt-tng Off]ce, February 1982).

jected without regard to changes in other condi-tions; in others, population growth projections areaffected by other factors such as technology oreconomic development. These structural differ-ences, combined with uncertainties about thepresent size and future behavior of the world’spopulation, lead to variations in the projectionsthemselves and differences in their reliability andusefulness to the policy maker. Reliability is also af-fected by time horizon, which in turn reflects oneof two basic goals:

● to provide an accurate short-term (25 years orless) forecast of world conditions; or

Table 2.—Comparison of Short- and Long-Term Population Projections

Scenario or Population in 2000Model or source projection (billions) Longer term projections

World 3 . . . . . . . . . . . . . . . . . . . . . . . . .

World Integrated Model . . . . . . . . . . .

Latin American World Model . . . . . . .

United Nations. . . . . . . . . . . . . . . . . . .

Global 2000 (Census Bureau) . . . . . . .

CFSC . . . . . . . . . . . . . . . . . . . . . . . . . . .

World Bank. . . . . . . . . . . . . . . . . . . . . .

Harvard . . . . . . . . . . . . . . . . . . . . . . . . .

Standard run

Equilibrium run

Standard run

Standard run

Second run (improvedconditions in Asia)

1978 assessment(provisional)

High, medium, low

High, medium, low

Standard

Standard

6.0

NA

6.4

6.4

NA

5.9 to 6.5

5.8 to 6.5

5.8 to 6.0

6.0

5.9

Population increases to 7.0billion by 2025, then decreases to4.0 billion by 2100

Population stabilizes at 6.0billion by 2050

Population stabilizes at justunder 7.0 billion by 2015. Deathrates due to starvation are highin South Asia

Population reaches 11.0 billionby 2040 and is still growing at 1.1percent/yr. Death rates due tostarvation rising rapidly in Asia

Population reaches almost 11.0billion by 2060 and is growing atless than 0.5 percent/yr

Population reaches 8.0 to 12.0billion by 2050 and stabilizes at8.0 to 14.0 billion by 2150

NA

Population reaches 7.8 to 8.1billion by 2050 and is virtuallystationary

Population stabilizes at 9.8billion by 2175

Population reaches 8.4 billion by2075 and is virtually stationary

SOURCE: The Futures Group.


● to describe the long-term behavior of theglobal system.

Findings

There is relatively little variation (plus or minusabout 5 percent) in the population projections for2000, which range from a low of 5.8 billion to ahigh of 6.5 billion. This reflects the higher degreeof certainty inherent in population projections forperiods under 25 years: there is relatively little un-certainty about the number of reproductive-age fe-males between now and 2000, although there ismore uncertainty about the number of childreneach will bear. The global models that aim for thissort of accurate, short-term forecast (the UnitedNations Input-Output World Model (UNIOWM)and Global 2000) are based primarily on expertjudgments about changes in fertility and mortal-ity. As a result, population is linked to other con-ditions only to the extent that the experts considerthe rest of the world system when they make thesejudgments. This short-term approach strives foraccuracy and usefulness by making separate pro-jections for individual countries, which can besummed to produce a world total.

The long-term global modeling studies (World 3,WIM, and LAWM) attempt to describe the gener-a! behavior of the entire global system over thenext 50 to 125 years, and they consider populationas only one of many factors in dynamic, integratedsystem behavior. The level of accuracy that is suffi-cient for this purpose is quite different from thatsought in the U.N. or Census Bureau projections.However,, the difficulty with this approach is thatthe relationships between fertility, life expectancy,and the factors that affect them—such as food pro-duction, pollution, and economic and social devel-opment—are not known, nor is the historical evi-dence rich enough to allow these relationships tobe estimated with any degree of confidence.

As a result, long-range models are more specula-tive and their population projections show consid-erably more variation. Two differences in table 2are particularly notable. In World 3, populationactually begins to decline due to increasing deathrates after 2025, and presumably would do so forAsia in WIM and LAWM if they were extendedbeyond 2060. All of the other projections show

population growing more slowly until it reachessome stationary level. However, there are immensevariations in the size of that stationary population,which ranges from a low of 8 billion to a high of 14billion.

Lim i tat ions

The reliability of the projections, and their use-fulness to the policy maker, are also influenced by anumber of theoretical and data constraints and bythe policy assumptions that have been built intothe global models. As mentioned above, there isneither theoretical agreement nor sufficient histor-ical evidence about the relationships between pop-ulation variables and conditions in the rest of theworld system. There is also considerable uncer-tainty in the base-year data for the initial popula-tion figures—estimates of China’s population vary

by as much as 14 percent, for instance, and thecurrent population of Nigeria has been estimatedat anywhere between 65 million and 85 million.

These differences in base-year estimates tend tocancel out when they are summed at the globallevel, and the recent round of censuses has sub-stantially improved the information available formany countries, particularly in Africa. However,there is still uncertainty about present rates andfuture changes in fertility and life expectancy,resulting from at least four factors:

●

●

●

●

uncertainty about how much birth rates havealready declined;uncertainty about the contribution of existingfamily planning efforts and technologies topast declines in birth rates;uncertainty about how many countries willadopt family planning programs, and uncer-tainty about how strong or effective these ef-forts will be; anduncertainty about the relevance of past ex-perience with family planning to those coun-tries that have little experience with such pro-grams, notably in Africa.

Different global models contain different as-sumptions about the above factors as well as aboutother policy decisions, all of which may have someeffect on future changes in population growth.The short-range models that use exogenous popu-

Ch. 3—Findings of the Global Models ● 4 7

lation figures also seem to assume that currenttrends will continue unchanged (at least until2000), thereby excluding such disasters as interna-tional conflict and massive starvation. The long-range models, on the other hand, suggest that re-gional or even global disasters are increasinglylikely in the longer term. World 3 and LAWMpoint out the dangers of inaction and delay andsuggest alternatives to present trends, but neither

model has adequate mechanisms for testing specif-ic policy options. WIM, which was designed as apolicy tool, is both more flexible and more disag-gregate. The more detailed stand-alone projec-tions, like those used in UNIOWM and Global2000, can become a valuable input to further anal-ysis for developing policy options, testing develop-ment goals, and planning broad strategies for theworld future.

Agricultural Projections

The world food problem has been a major con-cern for global modelers. All of the well-knownmodels have one or more agricultural sectors; allconsider measures of food availability to be majorindices of system performance; and all indicatethat the performance of the global agricultural sys-tem over the next 20 to 100 years is a matter of ma-jor concern. Table 3 compares the projections forkey agricultural variables in 2000 generated byfour global models and by two large-scaleagricultural models, the Model of InternationalRelations in Agriculture (MOIRA) and the Grain-Oilseed-Livestock model (GOL) used for Global2000.

Findings

In general, the most optimistic food supply pro-jections come from assumptions of rapid economicgrowth and technical progress, slow populationgrowth, and large reserves of easily developed agri-

cultural land. There is far greater variation in re-gional projections than in global projections; themost severe problems are foreseen in South Asiaand Sub-Saharan Africa. The results are also high-ly dependent on time horizon—longer time hori-zons generally lead to more pessimistic findings.

All of the models except UNIOWM indicatethat there will be problems in supplying food to atleast some of the world’s people over the next 20years. The reason for this finding is fairly straight-forward: all of the models except UNIOWMassume diminishing marginal returns to agricul-tural inputs and increasing costs for land develop-ment as the amount of undeveloped land de-creases; in short, the models show agriculturalproblems because they include agricultural limits.Similarly, WIM’s relatively pessimistic estimates ofpotentially arable land, which are 25 to 30 percentlower than the other models, undoubtedly con-tribute to its dire predictions of impending faminein the developing regions.

Price projections (for the models that makethem) vary far more than supply projections, with

Table 3.—Percentage Increases in Projected Global Food Production, Food Prices,and Food per Capita, 1970-2000

World Latin U.N.Integrated American Input-Output Global 2000

World 3 Model World Model World Model MOIRA GOL Model


real increases ranging from 13 to 422 percent by2000. MOIRA, which assumes that farmers willincrease production to maximize their profits at agiven price for agricultural inputs, concludes thatmeasures that drive food prices up will be a suc-cessful means of reducing world hunger. However,profit maximization may be a better approxima-tion of the behavior of rich farmers than that ofpoor farmers, who may not be able to borrowfunds to expand production, or who may resistgiving up their traditional agricultural practices.For this reason, MOIRA may overestimate the re-sponse to price incentives in the developing coun-tries (see app. B). World 3, LAWM, andUNIOWM all lack price mechanisms, and none ofthe six models takes the international monetarysystem into account.

Technical Progress

World 3 assumes that most increases in agricul-tural production will come from increases in landunder cultivation; Global 2000 assumes that theywill come from increased yields per acre; andUNIOWM, the most optimistic of the models, as-sumes both increased cultivation and increasedyields. All of the models except World 3 also in-clude some form of “disembodied technologicalprogress’ ’—income growth not attributable to in-creases in capital, labor, or other inputs—whichamounts to an assumption that agricultural pro-ductivity will increase automatically at no cost.The rate of such progress is 1.0 percent per year in

LAWM but is unreported for the other models,despite the fact that model results are highly sensi-tive to its presence and magnitude. When techno-logical progress is eliminated from LAWM, for ex-ample, Africa as well as Asia faces land constraintsand economic collapse. In World 3, on the otherhand, sufficiently strong assumptions about tech-nological progress can eliminate the overshoot-and-collapse mode entirely.

Uncertainties

Population and income growth are calculatedindependently from food supply in MOIRA,UNIOWM, and Global 2000. The effects of pollu-tion on agricultural yields is omitted in all of themodels except World 3 and UNIOWM. These fac-tors have an important influence on model results,as do assumptions about the availability and priceof inputs such as fertilizer, irrigation, and farmmachinery, but there is little agreement among themodels on the values that should be assigned tothem. Nor does any of the six models account forseveral trends that are likely to affect agriculture inthe coming decades:

●

●

●

●

regional or sectoral competition for water sup-plies;unusually bad weather, including adverselong-term climatic changes;Increased productivity due to advances ingenetic engineering; andpotential shrinkage or destabilization of oiland gas supplies.

Energy Projections

The future availability and price of energy re-sources are crucial variables in long-term projec-tions of world economic development. However,some of the models do not address energy specific-ally or in detail, and the findings of those t-hat doare significantly influenced by their assumptionsabout the global energy system and future energy

The following discussion is highly condensed; see app. C for a more detailedcomparative analysis of energy projections. For further information on thistopic, see OTA’s technical memorandum, Wodd Petroleum Availability f980-

2000, OTA-TM-E-5 (Washington, D. C.: U.S. Government Printing Oflice.October 1980), and the OTA assessment, Technology and Soviet Energy Avadabd-~ty, OTA-ISC-153 (November 1981).

trends. In general, those models that include afinite resource stock tend to show that depletionwill raise prices, slow industrial production, anddampen global economic growth. In short, as withagricultural projections, they predict energy prob-lems because they include energy constraints.

Findings

Collectively, the models indicate that the worldfaces a near-term transition away from depend-ence on conventional sources of petroleum and

Ch. 3—Findings of the Global Models . 49

natural gas, and that the major alternatives amongwhich future energy choices must be made arecoal, nuclear, and solar power. The models whoseprojections extend farthest into the next centuryindicate that coal, conservation, and conventionalnuclear power may not be enough to sustain con-tinued economic growth. Breeder reactors, fusion,and large-scale solar power may therefore be neces-sary.

In World 3, rising extraction costs are the prin-cipal cause of economic collapse. As an increasingpercentage of capital is allocated to obtaining non-renewable resources, investment and productivitydecline in agriculture and other sectors. New min-ing technologies and nuclear power can delay butcannot prevent a global economic collapse.

WIM, on the other hand, finds that collapse dueto costly resources is less likely than a future basedon widespread deployment of breeder reactors, butit also examines an alternative based on large “so-lar farms” in the deserts of the Middle East. Thismodel is perhaps the most flexible in dealing withenergy choices, and it has been used in a variety ofpolicy tests focusing on energy prices and the be-havior of both producers and consumers.

LAWM explicitly excludes the problem ofenergy and other resources. The authors assumethat conventional fission and the potential devel-opment of fusion power will eliminate the globalenergy crisis without significantly raising prices.

UN1OWM, because of its restricted time hori-zon, foresees “[no] problem of absolute scarcity in

the present century. ” However, the model doesproject a 77-percent depletion of conventional oilreserves by 2000. Its optimism about future energysupplies is based primarily on abundant coal sup-plies (it projects a decline in the real price of coaldespite a 400-percent increase in demand) and onthe assumption that nuclear power will generatean increasingly large share of the world’s elec-tricity.

Global 2000’s energy projections indicate asupply-constrained oil market before 1990, withproduction declining thereafter. As a result, “aworld transition away from petroleum dependencemust take place. ” Global 2000 examines severalpotential energy systems, but foresees nuclearpower and coal as the most likely alternatives. Italso foresees considerable potential for “conserva-tion-induced reductions in energy consumption. ”

Limitations

The accuracy and reliability of these projectionsare affected by their assumptions about a numberof physical, technical, and economic factors:

●

●

●

●

●

the total reserves of each resource;the rates of population and GNP growth;the degree to which energy demand growthwill be moderated by conservation or substitu-tion among sources;the bottlenecks involved in mobilizing addi-tional or alternative energy sources; andpotential energy breakthroughs such as fusionand solar power.

CHAPTER 4

Global Models andGovernment Foresight

Contents

PageIntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Benefits of Global Models. ......,.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . * * 54

Limitations o fGlobal Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Institutional Opportunities and Barriers . . . . . . . . . . . . . .: .....,... . . . . . . . . . . . . . . 55

CHAPTER 4

Global Models and Government Foresight

Introduction

The “futures debate” of the last 15 years hastaken place against a background of rapid andsometimes unexpected changes in national andworld affairs, including:

●

●

●

●

●

a general slowdown in the rapid economicgrowth that had characterized the world econ-omy since World War II, leading to stagnationin the industrialized nations and stalleddevelopment in the less developed countries(LDCs);rapid population growth and urbanization inmuch of the Third World, leading in somecases to widespread hunger, social unrest, andpolitical instability;structural changes in the world’s political andeconomic systems, typified by the emergenceof OPEC and new nuclear powers and byThird World demands for a “new interna-tional economic order;”growing apprehension about the cost and con-tinued availability of natural resources, bestexemplified by the energy crisis; andincreasing concern for the regional and global—environmental consequences of continued in-dustrialization.

These and similar developments have shownthat long-term global trends can have serious im-plications for the economic and national securityinterests of the United States. This in turn has ledto proposals in Congress and elsewhere that theU.S. Government should improve its “foresightcapability ’’—its institutional capacity to projectlong-range global trends and their consequences,and to use these projections as inputs in the proc-ess of strategic assessment, policy development,and decisionmaking. A number of Federal agen-cies are already using global models and other

computerized models as tools of long-range anal-ysis and planning, but the Government’s presentcapability is limited by unevenness of data, incon-sistency of assumptions, and lack of proper coordi-nation. * If existing deficiencies are corrected, glo-bal models could become a more effective tool infour specific areas:

●

●

●

●

assessing the potential future impacts of cur-rent trends, policies, and decisions;monitoring the national and internationalsituation to identify early signs of potentialproblems or opportunities;formulating and evaluating a wide range ofalternative policies and courses of action forachieving national goals, avoiding potentialproblems, and exploiting potential opportu-nities; andproviding a framework to ensure consistencybetween short- and long-term analyses andacross agency jurisdictions.

●

●

●

●

●

●

53


Benefits of Global Models

The essential claim for global models is that theyrepresent the behavior of the real global system, orat least some of its components, in ways that aresuperior to the less formal mental models and fore-casting techniques currently being used by deci-sionmakers, policy analysts, and the general pub-lic.2 Their benefits are those of mathematical mod-els in general and computerized models in particu-lar, the difference being that global models are spe-cifically designed to address problems and issuesof a global scale and importance. This can makethem a valuable tool of analysis and a valuable ad-ditional input to policy development and decision-making. Specific benefits of global models includethe following.3

● Longer time horizon.—Many current fore-casting techniques are used primarily for an-nual or short-term projections, whereas globalmodels typically have time horizons of 20years or more. This allows them to assess long-term effects and cumulative changes, howevercritical, that might not otherwise be detected.

● Comprehensiveness.—A computerized mod-el can contain far more information about theworld system than any single mental model. Itcan also keep track of many more variablesand interrelations at the same time and, ac-cording to some modelers, it is far more sensi-tive to subtle, remote, or counterintuitive ef-fects and outcomes.

● Rigor.—They impose a logical discipline byrequiring the modelers—and the model users—to make explicit, precise, and completestatements of their objectives, assumptions,and procedures. The system or process beingmodeled must be clearly divided into its majorcomponents, and the relations between thosecomponents must be specified. This proceduremay lead modelers and model users to revise

—.——2L. M. N’arci, et al., “N’orld Modellng: Some Crlttcal F o u n d a t i o n s , ” Beha~I-

ioral Scwnce, }ToI, 23, hlo. 3, Mat, 1978 , p . 138.

‘See Donella H. Meadows, John Richardson, and Gerhart Bruckmann (eds. ),Gopmg m the Dmk: The Fwst Decade oj Global )vlodehng (New York: Wdey,

forthcoming), pp. 20-21, 42-43, 203, 3 10-312; and Denms L. Little, et al., Long-

Range Pkmmng (Washington, DC.: U.S. Library of Congress CongressionalResearch %rvlce, 1976), prepared for the Subcommittee on the Enwronmentand the Atmosphere of the House Comrnlttee on .%lence and Technology, pp.456-457.

●

●

●

●

their mental models—or to refine them byidentifying previously ignored componentsand relations—even before computer analysisbegins.Accessibility.-The assumptions and struc-ture of the model must be written out beforethey can be run on the computer. This allowsall sides to examine them, point out omissionsor inconsistencies, and suggest improvements.Open communication about both the systemand the model can lead in turn to the incor-poration of fresh insights and differing view-points.Logic.-If properly designed and programed,a computerized model will draw logically cor-rect and mathematically error-free conclu-sions from an extremely complicated set ofassumptions and data. This can lead to novelinsights into unexpected or counterintuitivesystem behavior, reveal areas in which furtherresearch is needed, and expose assumptionsthat are inconsistent or contradictory.Flexibility and range.–By making smallchanges in the magnitude and relations of var-iables, it is possible to examine the implica-tions of a wide range of alternate assumptionsand to test the sensitivity of the outcome tochanges in different parameters. The modelscan also be tailored to “fit” particular prob-lems, regions, or issues. The model can there-fore become a powerful planning tool, andproperly updated runs can be used to monitorprogram progress.Instructiveness.–The flexibility of globalmodels also makes them a valuable tool foranalysts, planners, and policy makers alike, al-lowing them to examine a broad range of pos-sible outcomes, responses, and policy options.This can allow them to reject physically im-possible options, clarify the nature of variousrisks, and evaluate the adequacy of differentoptions for minimizing those risks. Thus, evenwhen the models cannot give precise quantita-tive answers, they allow the users to sharpentheir analytic skills and improve their intui-tive “feel” for the operation of the system.

Ch. 4—Globa/ Models and Government Foresight ● 55

Limitations of Global Models

Despite these benefits, global models remainsubject to a number of limitations that may con-strain their accuracy, reliability, and usefulness forpolicy making:4

●

●

� ✎

Theoretical limitations.–The structures ofdifferent global models are based on under-lying assumptions adapted from systems anal-ysis and several other disciplines, including

●

engineering, economics, and the natural sci-ences. There is no general agreement on therelative validity of their competing explana-tions of socioeconomic phenomena, nor isthere any evidence that one approach pro-duces results that are consistently more reli-able than the others. In addition, theoreticalunderstanding of a number of important proc-esses (for example, the effect of socioeconomicconditions on fertility rates) is too weak toallow adequate modeling, although projec-

●

tions based on alternative assumptions canstill be useful and instructive.Methodological limitations.–The essence ofmodeling is a simplification that improvesunderstanding, but this means that a limited

-.———‘See Nfeadows, R]c hardwm, and Bruckmann, op. clt., pp. 22, 43-44, 312.3 13;

Little, et al., op. cit., pp. 458-459; V’arcl, et al., op. cit., and The Global 20(xl Re-p o r t m the Pmwfcn[, \ ol. 2, ch. 14.

number of discrete factors and relations mustbe used to describe the dynamic complexityand ambiguity of the real world. Theoreticalbias and data constraints often determinewhich variables and relations are included oromitted, but no model could include everyfactor without becoming as complicated as thereal world.Data limitations.—Data vital to proper fore-casting are often nonexistent, inaccessible, orunreliable. This is particularly true for envi-ronmental data and for most data on theLDCs. This situation has improved somewhatin recent years, but data limitations remain aserious constraint on reliability and on thesectoral and regional coverage that modelscan achieve. In some cases there are inade-quate empirical data on which to base im-provements in theoretical understanding.Practical limitations.–An effective globalmodeling effort requires considerable time, aninterdisciplinary team of modelers and techni-cians, a large and continuously updated database, access to computers of sufficient capaci-ty, support services, and money. Skimping onany of these requirements greatly increasesthe risk of error and unreliability in the result-ing forecasts.

Institutional Opportunities and Barriers

Since President Theodore Roosevelt created the cies of theNational Conservation Commission in 1908, nu- presentingmerous presidential and congressional commit- ent set oftees, commissions, task forces, and studies haverecommended in one way or another that the U.S.Government should improve and/or institution-alize its long-range analysis, planning, and policy-making capability. The most recent study dis-cussed in this report, The Global 2000 Report,reveals that, individually, the executive agenciespossess an impressive, if uneven, capability forlong-range analysis and forecasting within theirseparate areas of responsibility and interest; but italso reveals “that, collectively, the executive agen-

government are currently incapable ofthe President with a mutually consist-projections of world trends in popula-

tion, resources, and the environment. ”5 Neverthe-less, according to the report, “Important deci-sions—involving billion-dollar federal programsand even the national security—are partially basedon these projections, ” which “have generally beenused by the government and others as though theyhad been calculated on a mutually consistentbasis.” 6

5The Global 2(xX) R@m to the President, Ywl. 2, p. 3‘Itxd,, ~ml. 2, p. 45-I.


The interagency followup report, Global Future:Time to Act, carries these conclusions a step fur-ther:

If there is one clear lesson from the exercise ofputting the Global 2000 Report together, it is thatthe U.S. government currently lacks the capacityto anticipate and respond effectively to these globalissues. . . . As of today, the government still doesnot adequately: 1) project and evaluate futuretrends; 2) take global population, resource, and en-vironmental considerations into account in its pro-grams and decisionmaking; and 3) work with othercountries to solve these problems. T

The deficiencies in the Government’s currentforesight capability appear to be institutionalrather than technical. In a recent analysis of thepotential Government policy applications of com-puter models, A. D. Little found modeling to bepotentially quite useful:

Current state-of-the-art techniques in long-rangeforecasting of population, resources, and the en-vironment present significant opportunities for theState Department to enhance its capabilities foranalysis of the long-run future socioeconomic andpolitical consequences of foreign national demo-graphic, resource, and environmental conditions.8

After considering the various limitations of com-puterized modeling (see above), the A. D. Littlereport concludes that “Institutional conditions areoften much more of a constraint to the effectiveuse of forecasting models than methodological ordata considerations.”9 These findings, althoughaddressed to the Department of State, would ap-pear to be equally relevant to the modeling ac-tivities of other agencies and of the Government asa whole.

The Global 2000 Report includes similar find-ings—the discrepancies and lack of integrationamong its forecasts arise:

. . . essentially because of the institutional contextin which the elements of the model were developedand are being used, This context emphasizes sec-toral concerns at the expense of interactions amongthe sectors and leads to distorted and mutually in-consistent projections. 10

TNlcholas Yost, staff director, Global Future: Time to Act, Report to the president

on Global Resources, En(uronment and Popufmon (Washington, D. C.: U.S. Coun-cd on Enwronmental Quahty and Department of State, January 1981), p. 206.

‘Arthur D. Little, Inc., op. cit., pp. IX-X.

‘Ib]d., p. xii.l’JThe G/obu/ 2~ R e p o r t [O the Pres[derrt, VO1. .2, p. 454.

This design is no institutional accident but is inconformit y with the bureaucratic division ofresponsibility within the executive agencies [andwithin the committee structure of the U.S. Con-gress]. . . . Furthermore, in the absence of ongoinginstitutional incentives to address cross-sectionalinteractions, the present form of the government’sglobal model is not likely to change significantly inthe foreseeable future. 11

* * *

Moreover, it would be naive not to recognize thatprojections and the procedures used to producethem have frequently been criticized by Congres-sional committees and others as subject to influ-ences not purely analytical in origin. Each agencyhas its own responsibilities and interests, its ownconstituencies, and its own pet projects. Often, anagency finds it helpful to use advanced analytictechniques (and associated projections) as weaponsin the adversary process of initiating, justifying, anddefending its programs. As a result, there have beenmany occasions in which the elements (and associ-ated projections) of the government’s global modelhave been used in support of (or in opposition to)highly controversial programs, and the credibility

of the projections has become a subject for debate.This has been especially true in recent times, asboth the issues and the advanced analytic pro-cedures used for examining the issues have becomeincreasingl y complex and, in a sense, incomprehen-sible to many nonexperts. 12

Another analysis of the Government’s currentforesight capability points to these and similar in-stitutional barriers. The following list of majorobstacles presents 10 frequently cited reasons forthe Government’s failure to correct the perceiveddeficiencies in its existing foresight capability.13

1<

2.

3.

4.

There is little or no top-level support for fore-sight.The “best talent” has never worked onbroad, long-term issues.Bureaucratic rigidity, compartmentalization,and specialization have frustrated attemptsto promote cooperation among departmentsand to take a broad, long-term view.Time pressures restrict vision to the shortrun.

I IIbld., Vol. 2, p. 461 and footnote.

IZIbid., vol. 2, pp. 478-480.i 3John M. R1chard50n, Jr., “Tov,ard5 Effective Foresight in the United States

Government” (prepared for the U.S. Department of State, June 1979), pp.13-18.

Ch. 4—Global Models and Government Foresight ● 57

5.

6.

7.

By the time models or forecasts are devel- 8.oped, policy-level officials have either movedon or lost interest.Policy-level officials lack the knowledge andexperience to properly use models. 9.The products of modelers’ efforts are incom-prehensible or irrelevant [to practical policy 10.issues], or both.

There is poor communication among thosewho contract for models and” forecasts, thosewho develop them, and those who are sup-posed to use them.Congress doesn’t care about the long-termfuture.The public doesn’t care about the long-termfuture.

CHAPTER 5

Priorities and Strategies forImproving U.S. Government

Foresight

Contents

Page, * . . . . . * * . * , 4 0 , . . . . . 4 * e e . * . . e . * e e . * . * * * * * * + * . * . * 61Introduction . ... .. . . . .. .

.. cnON i.-Correct the Existing Deficiencies Identified by Global 2000 and ::lther Assessments ... ...... . .... . ........ .. ........ ... ........ . . .. ..... 61

.. crION 2.-Coordinate the Government's Current Modeling Capabilities ond Activities . . ... . . . .. .. ...... .. . . . . .... . . .. .. .. .... . .. .... . . . . ... ... 62

.. cnON 3.-Support and Technical Improvements in the Current Capability and "dvance the State of the Art ...... .... ..... ..... ........ . .. ..... .. . .. ...... 63

.. crION 4.-Link the Government's Foresight Capability With It. Policymaking and Management Activities . . .. . .. .. .... . . .. . ... . . . .. . .. . .. . .. .... ... .. . . ... . . 64

CHAPTER 5

Priorities and Strategies for ImprovingU.S. Government Foresight

Introduction

Numerous proposals for improving Governmentforesight have been put forward, both in the pastand in response to the Global 2000 Report. 1 Somehave been modest and limited, others sweeping;some would require major legislation from Con-gress, while others could be encouraged throughoversight or carried out through executive orderor agency initiative. The following discussion pre-sents the most frequent and representative pro-posals, which generally reflect four fundamentalpriorities:

ISee especially Denms L. Litt le, et al . , Long-Range Plunrung (Washington,D. C.: U.S. Library of Congress Congressional Research Service, 1976), pre-pared for the Subcommittee on the Enwronment and the Atmosphere of the

House Committee on Science and Technology, pp. 384-390; John M. Richard-son, jr., “Tow.arcls Effectlte Forestght In the United States Grnernment” (pre-

pared for the L’.S. Department of State, June 1979), pp. 5-6 and app. C; ArthurD. Ltttlc, Inc., LongRange Fo~eca~[Lng Nfode[s Oj Population, Narmd Resources, andrhe Ent tmnment Their Use in F o r e i g n Po[[o Aswssments at [kc ,N’utional Let el(prepared for the US. Department of State, Not. 1979), PP. 11.2-3; and G~ohu~

f%we T u n e co Act

ACTIONIdentified

●

●

●

●

correct the existing deficiencies in Govern-ment models, as identified by Global 2000 andother assessments;coordinate the Government’s current predic-tive capability and activities;support technical improvements in the cur-rent capability and advance the state of theart; andlink the Government’s foresight capabilitywith its policymaking and management activi-ties.

These priorities and the various strategies forcarrying them out do not represent “options” inthe normal sense of the term. They are comple-mentary and mutually reinforcing parts of a larger,integrated effort to make the Government’s fore-sight capabilities more reliable, more coordinated,and more useful to both analysts and policy-makers.

1. —Correct the Existing Deficienciesby Global 2000 and Other Assessments

Global 2000’s authors, reviewers, and criticshave identified numerous deficiencies in each of itscomponent submodels, many of which are notedin chapter 2 and the appendixes to this report.The followup report by the interagency TaskGroup on Data and Modeling Capability identi-fied the correction of these existing deficiencies asthe first priority in improving the Government’sability to analyze and address global problems. z

This action could be taken as a matter of course byindividual agencies, many of which are alreadyplanning or carrying out evaluations and modifi-

1 (’ Repc~rt ~)f the Glohal 2CKY2 Task Grou p on Data and Nfodellng Capahlllty, ”

prepared for the Pres&nt’s Task Force on Global Resources and En~wonment,N o v . 7, 1980.

cations of their present capabilities. However, suchactions might be given higher priority, higher levelattention, and greater coherence by the agencies ifthey were encouraged by a presidential directiveand/or congressional oversight.

The Department of Energy’s (DOE) Energy In-formation Agency has institutionalized this evalu-ative function in its Office of Energy and Informa-tion Validation, which might provide a model forother agencies. Another possible strategy for car-rying out this priority would be the creation ofhigh-level advisory committees within each of therelevant agencies. The functions of these agencyadvisory committees might include the following:

61


●

●

●

prepare an inventory of existing models and ●

data bases, including the purposes for whichthey were originally developed and the uses towhich they are currently put, with particular ●

attention to their scope, complexity, and as-sumptions;identify existing deficiencies and evaluate any

plans to modify existing models or obtain newmodels, preferably through independent as-sessment by outsiders; ●

evaluate current and potential applications ofmodels and projections by agency analysts,planners, and policy makers, with particularattention to the specific information needs ofpotential users;

encourage expanded use of models and projec-tions through educational and training pro-grams for agency personnel;improve communication between those whouse models and those who develop or main-tain them, with particular attention to in-creasing the relevance and responsiveness ofmodel outputs to the needs of potential users;andidentify likely future problems and issueswithin the agency’s area of responsibility andinterest, with the goal of developing problem-oriented and policy-relevant models and databases.

ACTION 2.—Coordinate the Government’s CurrentModeling Capabilities and Activities

While the creation of agency advisory commit-tees might be a useful and necessary foundationfor improving the Government’s foresight capabil-ity, it fails to address the equally important prob-lem of linking and coordinating these agency capa-bilities, which currently focus on relatively nar-row, mission-oriented sectoral concerns. Problem-oriented models would most appropriately bedeveloped by the agencies that would use them,but overall analysis and policymaking would re-quire consistent, integrated forecasts that incor-porate data and projections from several agencies.Global 2000 shows that it is in this area that cur-rent Government efforts have been most unsuc-cessful: the delays in completing the study itselfwere caused in part by problems involving com-puter compatibility and tape transfer.

The simplest strategy for carrying out this actionwould be a process of interagency negotiation andarbitration to bring about greater consistency andcompatibility between the separate agencies’ databases, assumptions, and projections. Ultimately,however, the effectiveness of such a process mightrequire the creation of an interagency task force toprovide a focal point and to resolve conflicts. Thecoordinatin g functions of such a task force mightinclude the following:

●

●

●

●

●

●

prepare an inventory of models and data basescurrentl y maintained by individual agencies(including their respective strengths andweaknesses), and identify any areas of compat-ibility, overlap, or redundancy;identify gaps, sources of inconsistency, andpoints of conflict among existing agency capa-bilities and suggest possible solutions;promote greater understanding,. communica-tion, and technical cooperation among agen-cies (simply getting the Government’s mod-elers together was one of the Global 2000study’s major accomplishments);review and coordinate agency plans to modifyor obtain new models or data bases, in orderto prevent redundancy, ensure greater com-patibility, and identify software or hardwareneeds;develop consistent procedures and protocolsfor data collection, standards of reliability,and validation, as well as for model documen-tation and validation;establish a central clearinghouse to provideinformation on the location of models anddata bases and to permit easier access. ex-change, and integration of data, assumptions,and projections;

Ch. 5—Priorities and Strategies for Improving U.S. Government Foresight ● 63

. identify key future problems and issues, desig- grated Model (WIM) maintained by the Jointnate lead agencies to gather information and Chiefs of Staff and other agencies, the Grain-monitor trends in each area, and ensure the Oilseed-Livestock model maintained by thepublication of timely projections; and Department of Agriculture, and the IEES/

● link existing models, such as the World Inte- LEAP models maintained by DOE.

ACTION 3.—Support Technical Improvements in theCurrent Capability and Advance the State of the Art

Technical coordination among agencies andbetween the executive and legislative brancheswould eventually require some form of third-partymediation. Creation of an institutional mecha-nism for this purpose would also present an oppor-tunity for carrying out research and developmentaimed at technical improvements in existing agen-cy capabilities, the Government’s capability as awhole, and the state of the art in global modelinggenerally. To carry out these functions, however,this institutional mechanism would have to be in-sulated from the day-to-day concerns of the lineagencies and, thus, able to take a long-term viewand incorporate broader and more diverse per-spectives. Because its mission would be only indi-rectly linked to policy-related concerns, however,it would seem that such an organization should becreated only in conjunction with other initiativesthat are directly relevant to policy developmentand coordination (see action 4). Technical im-provements are nevertheless a necessary prelude topolicy applications,

Several strategies have been suggested for carry-ing out this priority. In the short term, an ad hoccommission or research advisory panel might beappointed to identify key technical problems andestablish research priorities. To be effective, how-ever, such an effort would have to be both openand ongoing. A frequently encountered proposalfor the long term is the creation of a “hybrid” or“quasi-public” institute devoted to long-rangeanalysis, global modeling, and futures research.

The primary goal of such an institute would be toencourage private-sector understanding of, sup-port for, and participation in Government fore-sight activities. Specific functions might includethe

●

●

●

●

●

●

●

following:

solicit the thoughts and enlist the creative tal-ents of the private sector, particularly thebusiness and educational communities, to“cross-fertilize” Government ideas and initia-tives;support research by nongovernmental organi-zations to create and/or improve global mod-els and other analytic tools, especially thosebased on paradigms other than economics;encourage impartial, third-party validationand assessment of existing or proposed Gov-ernment models;establish a “global modeling forum, ” pat-terned on the Energy Modeling Forum, atwhich modelers could exchange ideas and cri-tique one another’s work;assess work done outside the Government oroutside the United States and, where appro-priate, suggest its incorporation into the Gov-ernment’s capability;support data-gathering efforts and the devel-opment of needed data-gathering technologiesand systems; andestablish and maintain communication withsimilar organizations in other countriesthrough such organizations as the Interna-tional Institute of Applied Systems Analysis.


ACTION 4.—Link the Government’s ForesightCapability With Its Policymaking and

Management Activities

The above-described institute, although it couldhelp to broaden the dialog on global problems andadvance the state of the art in global modeling,could not by itself ensure that these concernswould be translated into coordinated Federal pol-icy. If the U.S. Government is at present giving in-sufficient attention to long-range global problems,it could be in part because no single agency has themandate or the ability to look at these problemson an integrated, ongoing basis. A final priority,therefore, might be to create an institutional focusthat could coordinate the various elements of theGovernment’s foresight capability and ensure thatlong-range global concerns and national prioritiesare routinely taken into consideration in the for-mulation, selection, and implementation of U.S.policy at all levels.

In Congress, this would require continuing ef-forts to ensure that legislative proposals are evalu-ated in terms of their long-term global impacts andimplications. One rationale for such evaluationmight be provided by House Rule X, which directsin part that each standing committee (other thanBudget and Appropriations):

, . . shall review and study any conditions or cir-cumstances which may indicate the necessity or de-sirability of enacting new legislation within the ju-risdiction of that committee . . . and shall on acontinuing basis undertake futures research andforecasting on matters within the jurisdiction ofthat committee (2(b)(l)).

The long-range analytic capabilities of the legisla-tive support agencies might also be coordinatedand brought to bear on such issues. In addition,Congress might also encourage appropriate initia-tives in the executive branch through oversighthearings, personal appeals, or directed research.For example, several committees have already heldor plan to hold hearings on long-range demo-graphic issues, Global 2000, and Governmentforesight. In addition, Sen. Charles McC. Mathiashas written a letter to the President, cosigned by84 other Members, strongly urging that he give the

Global Future: Time to Act report his thoughtfulconsideration and that he “put into motion themachinery that will translate these recommenda-tions into action. ”3 Another example is the re-quest by the Technology Assessment Board forthis OTA study.

Proposals for creating an institutional focus forlong-range global policymaking in the executivebranch usually suggest that–to ensure supportfrom and access to high-level decisionmakers—itshould be located in the Executive Office of thePresident (EOP). The current administration has

Ch. 5—Priorities and Strategies for Improving U.S. Government Foresight • 65

apparently taken a step in this direction by creat-ing a new “national indicator system” directed bya special assistant to the President. He describesthe project as “a system for providing social anddemographic information to the policy people in asystematic and regular way, in advance of policydebates, “ in order to give high-level policy makers“a view of [the] changing world” and a sense ofhow “everything has its cross-impacts on every-thing else in society. “4 The system, which will leadto twice-monthly briefings for the President, Vice-President, Cabinet, and senior EOP staff, will fo-cus primarily on national trends but will also ex-amine international trends if there is an obviousconnection or a special request from the “long-range policy group” that previews the briefings.

Other proposals for structuring and housing anEOP foresight capability have included the follow-ing

●

●

●

options:

Create a new office in EOP devoted exclu-sively to long-range global issues.—Such anentity would give the issues greatest emphasis,ensure access to the President, avoid compet-ing responsibilities, and facilitate coordina-tion of agency capabilities and activities.Assign responsibility for these issues to anexisting EOP office.—This would avoid theneed to create a new EOP unit. Several exist-ing offices (e.g., the Office of Management andBudget, National Security Council, or Coun-cil on Environmental Quality (CEQ)) haveanalogous or complementary missions and ex-pertise, and CEQ has already begun planninga followup to the Global 2000 study. Addi-tional staff and resources would have to bemade available, and there is a possibility thatlong-range global issues might not receive fullattention because of the unit’s existing func-tions and responsibilities.Create an interagency coordinating com-mittee on long-range global issues.—Such acommittee might be chaired by the Vice Presi-dent (in conjunction with his current duties aschairman of the Crisis Management Team) orby the head of an existing EOP unit, but even

4Rlchard Flea], Spec]al Asslwant to the Pres]clent and director of planmng andevaluatmn; quoted by Phlllp J. Hilts, “U’hlte House Uses %clal Sciences ButCuts Fundtng for Research,” LY’ushlngton Po~t, June 29, 1981, p. A8.

●

with a staff of its own it would probably be lessefficient and less effective than a dedicated of-fice.Assign responsibility for policy develop”ment on long-range global issues to a Spe-cial Assistant to the President.—Such an in-dividual, with the help of a small staff, couldhave access to the President and could ensurea somewhat better degree of interagency coor-dination, but this office would have to de-pend, in turn, on modeling and analytic ex-pertise from other sources.

The functions and objectives that have beensuggested for this new EOP office include the ini-tiation, supervision, and coordination of all of thefunctions outlined for the preceding priorities, plusthe

●

●

●

●

●

●

●

following:

ensure that the President and other top-leveldecisionmakers are presented with the bestpossible analyses and broadest possible rangeof policy options on long-range global issues;use global models (in combination with otheranalytic techniques) to determine the effect ofvarious agency goals and budget items onlong-range global trends and strategic inter-ests;encourage an open and vigorous national dia-logue on long-range global problems and is-sues, with the goal of developing a clear defini-tion of U.S. national goals and strategic inter-ests in these areas;prepare a “policy statement on the future,” tobe presented by the President, as a means offocusing attention and forcing action on long-range global issues;issue periodic reports, similar to an executiveagency’s annual report to Congress, on majorglobal issues and, at longer intervals, conductcomprehensive, integrated studies of long-range global trends and problems;issue periodic reports on the state of the art inglobal modeling and the state of the Govern-ment’s foresight capability; andin conjunction with the Department of State,encourage similar assessments of long-range is-sues by foreign governments and cooperate inthe data-gathering and analytic activities ofthe various international organizations.

Appendixes

APPENDIX A

Population Projections

Introduction

Population projections play a key role in any assess-ment of future world conditions, since the size of thepopulation determines the number of consumers ofgoods and services and the number of people availableto produce those goods and services. The rate and dis-tribution of population growth will also determine theshort-term strains that are put on the world’s socioeco-nomic system, just as the ultimate size of population willdetermine the long-term strains that are put on theplanet’s physical and biological systems. Three of theglobal models–World 3, World Integrated Model(WIM), and Latin American World Model (LAWM)—try to describe the general long-term behavior of allcomponents of the global system, including population.In these models, the determinants of populationchange—fertility and life expectancy—are linked toother sectors through feedback loops that simulate theeffect of changing physical and socioeconomic condi-tions on the rate of population growth. The other mod-els try to provide a detailed, short-term prediction ofworld conditions; in these models the future size of thepopulation is projected in a more straightforward way,unrelated to the projection of other conditions.

These differences in purpose and technique, com-bined with differences in time horizon and the uncer-tainties surrounding the present size and future beha-vior of the world’s population, lead to variations in theprojections themselves and in their relative reliability.The projections are also influenced by policy assump-tions and by the models’ degree and pattern of geo-graphical regionalization. All of these factors affect theusefulness of the models and projects for the policy-maker.

Purposes,

World 3

Structures, and Projections

This highly aggregated global model was designed todisplay the long-term interactions among major worldsystems—population, nonrenewable resources, capital,agriculture, and pollution—in order to investigate theconsequences of five major trends at the global level: ac-celerating industrialization, rapid population growth,widespread malnutrition, depletion of nonrenewable re-—..—

IThe follow Ing material IS based on an OTA umrklng paper prepared bv j.

Stover of the Futures ~JHNIp, Glastonhurv, Corm. For further Information onthts sub]ect, s e e O T A ’ s a s s e s s m e n t , Wrodd Popukxtlon and Ferti(lty-Pfunnmg

Technologies The Next 20 Years, OTA-HR-I 57, Fehruary !982.

sources, and environmental deterioration. In order toexamine the long-term effects of these trends, the modelwas designed to simulate interactions over a 200-yearperiod, from 1900 to 2100. The system dynamics model-ing technique is particularly useful for simulating twoimportant kinds of system behavior—feedback and de-lays–and is thus a powerful tool for understandingcomplex system behavior. This makes it a suitable ap-proach for exploring the general shape of long-term glo-bal system behavior, but it is generally less useful thanother techniques for producing accurate short-termforecasts of individual variables.

The population sector of World 3 is highly aggre-gated: it has only one geographic region and only fourdifferent age groupings: O to 14, 15 to 44, 45 to 64, andover 64. Global population is determined solely by theeffects of changes in birth rates and death rates, but thedetermination of these rates depends heavily on inter-action with the rest of the model. The death rate is cal-culated from average world life expectancy, which inturn is determined by four influences that the authors,based on historical evidence and expert judgment, pos-tulate in the following relationships:

●

●

●

●

Life expectancy ‘increases with increasing foodavailability, reaching a maximum when food percapita is about four times the present world aver-age.Life expectancy increases with increasing healthservices per capita, although with about a 20-yeardelay.Life expectancy initially increases with crowdingbecause of increased industrialization and urbanservices, but beyond a certain level crowding has anegative influence on life expectancy due to the ef-fect of local pollution and stress-related diseases.Life expectancy decreases as the amount of persist-ent pollution increases.

These relationships are intended to capture the generaldirection and magnitude of the four influences on lifeexpectancy. Although exact equations must be enteredinto the model, these equations are not meant to be rig-orous quantifications of the actual relationships.

The birth rate is calculated from the fertility rate,which in turn is influenced by two factors: desired fertil-ity, and fertility-control effectiveness. Fertility-controleffectiveness is expressed as a function of level of devel-opment, as measured by industrial output per capita,and reaches 100 percent at a level of development threetimes higher than the current average world level. De-sired family size is assumed to be affected by two furtherinfluences-the social norm and the individual response

69

.

70 ● Global Models, Word Futures, and Public Policy

to the social norm—in such a way that desired familysize ranges from a high of 5 children when income is lowto a low of 1.5 children when income is high. As withlife expectancy, however, the specification of these rela-tionships is based on expert judgments aided by thesmall amount of available historical evidence.

The entire World 3 model was calibrated by simulat-ing the period from 1900 to 1970, during which periodthe model correctly represents the broad dynamics ofworld population.

In the standard run of the World 3 model, world pop-ulation increases to about 6 billion by 2000 and reachesa maximum level of about 7 billion by 2025, but thendeclines rapidly to only 4 billion by 2100 (see fig. A-l).This decline occurs because resource depletion causesan increasing fraction of capital to be used in extractingraw materials, thereby reducing the amount availablefor other investments, especially agricultural produc-tion. As investment in agriculture lags, food per capitabegins to decline, causing a higher death rate, a decline

Figure A-1 .—World 3 Standard Run

The “standard” world model run assumes no major changein the physical, economic, or social relationships that havehistorically governed the development of the world system.All variables plotted here follow historical values from 1900to 1970. Food, industrial output, and population grow ex-ponentially until the rapidly diminishing resource baseforces a slowdown in industrial growth. Because of naturaldelays in the system, both population and pollution con-tinue to increase for some time after the peak of industri-alization. Population growth is finally halted by a rise in thedeath rate due to decreased food and medical services.

SOURCE: Limits to Growth.

in life expectancy, and ultimately a decline in popula-tion. This model run is meant to represent the mostlikely mode of system behavior if there is no change inpast trends and policies, and the authors argue that itdemonstrates the need for action to ensure that such aresult does not occur.

In a series of sensitivity and policy tests, the authorsshow that neither increased resources nor improvedtechnologies have any significant effect on this basic be-havior mode: population invariably peaks before 2050and declines thereafter, although the maximum popula-tion varies from 7 billion to 9 billion and the populationin 2100 varies from 3 billion to 5 billion. The reasonsfor the collapse do change, however: for example, whenthe resource limits to growth are removed, increases inpollution become the factor that eventually causes thedeath rate to increase.

Only two scenarios were able to avoid the basic pat-tern of overshoot and collapse. In the first, based on thehighly unrealistic condition of removing all the physicallimits to growth, population continued to grow andreaches about 14 billion by 2100. The second is the“global equilibrium” scenario, based on an integratedset of policies designed to bring about a gradual transi-tion to a stable, nongrowth world (fig. A-2). These pol-icy changes, all of which are implemented beginning in1975, include the availability of “perfectly effective”birth control and a worldwide reduction of desired fam-ily size to 2 children. Under these conditions, the popu-lation grows to only 5 billion by 2000 and stabilizes ataround 6 billion by 2050. If the needed policy changesare delayed until 2000, however, the equilibrium state isnot sustainable (fig. A-3).

The usefulness of this model to policy makers liesmainly in pointing out the potential dangers ahead, thecosts of delaying action, and the need for consideringthe whole system when thinking about one aspect ofthe problem. However, since the model does not con-tain adequate policy levers (i.e., policymakers cannotdirectly control the variables in the model), its useful-ness for policy testing and evaluation is limited.

World Integrated Model (WIM)

This global model, the first of several to be built in re-sponse to what many saw as the inadequacies of World3, differs from the latter in one major aspect—disaggre-gation. WIM represents a world composed of many dif-ferent subsystems that interact with one another hierar-chically on five different planes or strata: individual,group, demographic-economic, technological, and envi-ronmental. It also divides the world into 10 or moregeographic regions; each region is a complete model initself, but interacts with the other regions through in-ternational trade. The model is not an explicit system

Figure A-2.—World 3 Equilibrium Run

I

/

1900 2100

Technological policies and growth-regulating policies pro-duce an equilibrium state sustainable far into the future.Technological policies include resource recycling, pollu-tion control devices, increased lifetime of all forms ofcapital, and methods to restore eroded and infertile soil.Value changes include increased emphasis on food andservices rather than on industrial production. Births are setequal to deaths and industrial capital investment equal tocapital depreciation. Equilibrium value of industrial outputper capita is three times the 1970 world average.

SOURCE: LImIts to Growth.

dynamics model, but it does incorporate some of thefeedback relationships and delays that are contained inWorld 3, as well as many new ones.

The population sector of each region divides the pop-ulation into 85 l-year age cohorts. Fertility is deter-mined by the level of socioeconomic development ineach region, including both the reduction in desiredfamily size and the increased effectiveness of contracep-tive use. Increasing levels of education also lead to de-clining fertility.

Mortality is also linked to the rest of the model. Lifeexpectancy increases with increasing level of develop-ment and (in some versions of the model) it is also af-fected by food availability: the model calculates theamount of calories and protein available in each region,and when these amounts fall below sufficient levels, ad-ditional deaths due to starvation result. All of these re-lationships are judgmentally determined–as in World3, they are meant to capture the general shape of thereal-world relationships, without attempting to quan-tify them rigorously.

Figure A-3.—World 3 Run With Stabilizing PoliciesIntroduced in the Year 2000

1900 2100

If all the policies instituted in 1975 in the previous figure aredelayed until the year 2000, the equilibrium state is nolonger sustainable. Population and industrial capital reachlevels high enough to create food and resource shortagesbefore the year 2100.

SOURCE: Limits to Growth.

The standard run of the model shows the result ofcontinuing historical patterns of development. In thisrun, however, the relationships between education, de-velopment, and fertility are replaced by the “optimisticthough reasonable” assumption that successful popula-tion policies will cause fertility rates in all regions todrop to the replacement level (i.e., births equal deaths)within 35 years. The authors did this in order to “avoidbiases that might be introduced by the predominance ofpopulation growth factors. ”2 This standard run showsworld population increasing from 4 billion in 1975 toabout 6.4 billion by 2000 and stabilizing at about 6.8billion by 2015.

In other runs the authors have examined the effectson world population of delays in starting a worldwidepolicy to achieve replacement fertility rates. Assumingthat it takes 35 years from the start of such an effort un-til replacement level fertility is reached, they found thatworld population would stabilize at about 7 billion ifthe effort is started in 1975, 10 billion if it is started in1985, and 12 billion if it is started in 1995.

Because of the differences in regional detail, the con-clusions of the WIM study are different from those of

‘M. D, Mesaro\lc and E. Pestel, ,Lfanklnd At the Turntng PoInr (New York: Dut-

ton, 1974), p. 57.


World 3. The authors conclude that worldwide collapseis not likely to result even if past trends continue, butthat catastrophes could well occur at the regional levelwell before 2050. These regional collapses, occurring atdifferent times and for different reasons, will neverthe-less profoundly affect the entire global system. The solu-tion to this problem lies not in stopping all growth butrather in achieving what the authors call “organicgrowth’ ’-i.e., different kinds of growth in different re-gions, such as continued industrial growth in the LDCsbut service-oriented growth in the developed countries.Without early action to ensure a transition to this kindof balanced, differentiated growth, the authors say thatregional and finally global disaster is inevitable.

The model was designed to be employed by policy-makers. Extensive work has gone into making the mod-el interactive so that policy makers can experiment withit and, thereby, increase their understanding of the fu-ture effects of current policy decisions. Since the ori-ginal model is not disaggregated to the country level,policymakers cannot see the effect of specific policiesthat the y might implement in their own countries.However, subregional and country models based on thesame concepts have been developed by former membersof the modeling team in response to the needs of specificclients.

Latin American World Model (LAWM)

This model was also constructed in response to limi-tations seen in the World 3 model, and in particular toits disturbing implications for the developing world, Ifeconomic growth must stop soon in order to avertworld collapse, what hope do the poor countries haveof ever alleviating their poverty? The Latin Americanauthors of this study concluded that the proper ques-tion to ask is: “What changes in the structure of thepresent socioeconomic system would be required in or-der to bring about a better life for all?”

Since the model attempts to answer a different ques-tion, its modeling technique is also different. LAWM isbased on optimization techniques, and the model isstructured in such a way as to determine the allocationof labor and capital that will ultimately maximize lifeexpectancy—its measure of the satisfaction of basic hu-man needs. The model divides the world into four re-gions—the developed countries, Latin America, Africa,and Asia—each of which functions more or less inde-pendently. The model does contain some physical con-straints, but its structure reflects a general underlyingassumption that physical limits to growth will not be asimportant as social and political limits. As a result, itsprojections do not reflect how the world will be, butrather how it could be if the indicated allocation policieswere followed.

The population sector of LAWM, which is designedto identify the social and economic factors that influ-ence population growth and life expectancy, is bothmore detailed and more complicated than those ofWorld 3 and WIM. The complex web of variables andinfluences (shown in fig. A-4) can, however, be reducedto a single, fairly straightforward hypothesis: “The onlytruly adequate way of controlling population growth isby improving basic living conditions.”3 Thus, the allo-cation of capital and labor to any of the basic needs sec-tors—food, education, housing, and other consumergoods and services—will result in lower birth ratesand/or higher life expectancy. These influences are infact compounded, since lower birth rates lead to higherlife expectancy, which in turn leads to a further de-crease in birth rates.

Education and agriculture seem to have been givenparticular weight in LAWM, but all of the relationswere estimated through descriptive analysis, using scat-tered data and rather esoteric mathematical techniques,and they should not be confused with causal relation-ships, although the formulation of the model implicitlyassumes them to be causal.4 The correlations betweenestimated and actual birth rate and life expectancy forthe test period (1960-70) are very high (+0.90 and+0.95, respectively), but this technique gives no indica-tion of whether or not the postulated functions are cor-rect. If demographic change in the real world followsdifferent rules, this optimization procedure will not gen-erate reliable projections.

The results of the LAWM standard run are mildly op-timistic for all regions except Asia, where the only basicneed that is satisfied is education. By 2010 all availableland in Asia is being cultivated, which means that thegrowth of food production is unable to keep up withpopulation growth; although progress continues inhousing and income, food availability per capita falls todangerous levels, and the delay in achieving basic needscauses a delay in reducing population growth. The pop-ulation projections show a total world population of 6.4billion by 2000, 1,2 billion in the developed world and5.2 billion in the developing world. By 2040 the worldpopulation reaches 11 billion and is still growing at 1.1percent per year, due largely to the 1.4 percent growthrate in Asia.

The usefulness of this model for policymakers lacksmainly in its ability to display the effects of adopting aparticular broad strategy—that of allocating labor andcapital in a manner that corresponds statistically to

‘A. O. Herrera, et al., Cutu.strophe or Neu Soaerj ? A Latin Amemzm WorldModel (Ottawa: International Development Research Center, 1976), p. 8.

4See J. Robinson, “The Latin American World Model,” In The Global 2000Report to the Pres&nt (Washington, D. C.: U.S. Councd on EnwronmentalQuahty and Department of State, ]980), vol. 2, pp. 638 and 641; and D. H.Meadows, J. Richardson, and G. Bruckmann (eds.), Gro#ung m the Dark. TheF[rst Demde of Globul ,Modelmg (New York: Wiley, forthcornlng), p, 97.

App. A —Population Projections ● 7 3

Figure A-4.— Basic Structure of LAWM

( Foodproduction )

I Health t PopulationI

I

●

Expectancyof life at birth

m

Capital Labor

74 Ž Global Models, world Futures, and Public Policy

maximized life expectancy. LAWM does not containpolicy levers that might be exercised by an individualplanner, but the model has nevertheless contributed tothe debate on broad strategy by describing an alterna-tive to the present world order.

The United Nations Input-OutputWorld Model (UNIOWM)

The particular goal of this study was to assess the im-pact of various economic issues and policies on the In-ternational Development Strategy for the SecondUnited Nations Development Decade. Specifically, themodel was to address the question of whether the exist-ing policies and development targets were consistentwith the availability of world resources. Because it usesan input-output approach, the model can be used eitherto investigate the rates of economic growth achievablegiven certain constraints (resources, balance of pay-ments, etc. ) or to investigate investment and consump-tion levels that would be consistent with given rates ofeconomic growth.

Unlike the other global models, however, populationgrowth is entirely exogenous in UNIOWM. Althoughthe size and composition of the population affect therest of the model, there is no feedback from the rest ofthe model to population. In fact, the model has no pop-ulation sector as such: it simply uses the populationprojections prepared by the United Nations PopulationDivision for the 1973 assessment of world population,(This assessment was updated for all countries in 1978,and the discussion below deals with the more recent fig-ures.) These projections, include four variants-high,medium, low, and constant-the constant case is mere-ly a reference projection assuming no change in a coun-try’s fertility rate. The medium variant is designed torepresent the most likely future demographic trends,based on past and present demographic trends, ex-pected social and economic progress, ongoing govern-ment policies, and prevailing public attitudes towardpopulation issues. The high and low variants are in-tended to reflect plausible variations on these factors.

The fertiltiy assumptions underlying the projectionsare based on past and present fertility trends in eachcountry, adjusted judgmentally by experts in the U.N.regional population offices around the world and re-viewed in the Population Division at U.N. headquar-ters in New York. These judgments were guided bythree general principals:5

● Fertility rates will decline as economic and socialdevelopment takes place.

● Existing or anticipated government policies andprograms, as well as nongovernmental activities

aimed at such a goal, will expedite the process offertility decline.

● Once initiated, fertility decline will begin slowly,gain momentum, and then slow down again. Forthose countries with fertility levels near or belowthe replacement level, it is assumed that fertilitywill begin to converge on replacement level beforethe end of this century.

Judgmental estimates of life expectancy were also pre-pared for each country, following the general rule thatlife expectancy would increase by 2.5 years every 5 yearsuntil it reaches 55, after which the rate of increasewould be less. (For some countries where developmenthas been slow, this rule was changed to allow slower in-creases. ) The projections also take into account assump-tions about internal migration (urbanization) in eachcountry.

In the projections prepared by the U.N. in 1978, thepopulation of the world increases from 4 billion in 1975to 5.9 billion, 6.2 billion, and 6.5 billion by 2000 for thelow, medium, and high variants, respectively. Most ofthis growth occurs in the developing world: the popula-tion of the more developed countries increases by only13 to 21 percent from 1975 to 2000, while the increasefor the less developed world is 57 percent in the low var-iant to 77 percent in the high projections.

The most rapid rate of growth is in Africa, wherepopulation increases by 114 percent from 1975 to 2000,and in Central America and the Caribbean, where itincreases by 115 percent. China’s population is pro-jected to increase by 26 to 37 percent; this is only halfthe average rate for all developing countries, although itmeans an additional 235 million to 335 million peopleby 2000. India will add even more people to its popula-tion, an additional 360 million to 480 million people.

These population projections are used by UNIOWMto explore the consequences of the goals adopted for theSecond Development Decade. This model has greatutility in displaying the results that would occur if thesegoals were achieved, and as such it has been useful tothose planning the broad strategy of the developmentdecade. Unfortunately, the goals and prescriptions ofthe development decade do not always have a signifi-cant influence on actions taken by individual countries.

Global 2000

This study, conducted b y an interagency task forcethat drew on the expertise and models of the entireU.S. Government, is one of the most comprehensivemodels in terms of the number of variables it examines.However, the separate submodels comprising “the Gov-ernment’s global model” are not linked to each other,and as a result there is no feedback or interaction be-tween sectors. As a result, the population figures do not

App. A—Population Projections ● 7 5

reflect the model’s pessimistic projections of GNP orfood production.

The population projections that were used in mostsectors of Global 2000 were prepared by the U.S. Bu-reau of the Census in 1977. Like the U.N. projections,they are presented with high, medium, and low variantsfor all regions and selected countries. Underlying allthree variants are the assumptions that: 1) fertility willdecline more or less continuously throughout the peri-od; 2) all countries will have adopted some kind of fam-ily planning program by 2000; and 3) the effectivenessand coverage of such programs will increase. The differ-ent fertility rates used in the high, medium, and lowvariants are based on the judgment of experts and re-flect the range of uncertainty about current fertility, fu-ture development patterns, and government familyplanning policies. Specific individual judgments are notreported, however.

The final Census Bureau projections extend to theyear 2000 and are substantially the same as the U.N.projections: global population grows to 5.8 billion inthe low variant, 6.2 billion in the medium variant, and6.5 billion in the high variant. The population of thedeveloped regions grows to between 1.3 billion and 1.4billion by 2000, while the less developed regions grow tobetween 4.5 billion and 5.1 billion.

An alternative set of projections, prepared by theCommunity and Family Study Center (CFSC) at theUniversity of Chicago under the sponsorship of theAgency for International Development (AID), is alsopresented in the Global 2000 Report but is not used inother sectors. These projections, too, include high, me-dium, and low variants, but in this case the projectedfertility rates are based on a model of fertility declinethat relates the rate of decline to the current level of fer-tility and the strength of family planning efforts withineach country. The variants therefore reflect more ex-plicit assumptions about family planning efforts: thehigh variant assumes that each country maintains itspresent level of family planning; the medium variant as-sumes that nations will implement strong family-plan-ning programs by 2000; and the low variant assumesthat strong programs will be in place by 1995 and thatthe effectiveness of these efforts will increase. These as-sumptions were developed through regression analysisof data for 1968-75, which showed that the strength offamily-planning efforts had as much effect on fertilityrates as all the measures of socioeconomic developmentcombined.

Because they assume policy changes that would in-crease the efficacy of family-planning services, t h eCFSC projections are generally lower than those of theU.N. and Census Bureau: world population reaches 5.8billion, 5.9 billion, and 6.0 billion by 2000 in the low,

medium, and high variants, with the developed regionsgrowing to between 1.2 billion and 1.3 billion and theless developed regions to between 4.5 billion and 4.7 bil-lion. These projections have also been extended to2050, at which date they show a global population ofbetween 7.8 billion and 8.1 billion people.

The Census Bureau projections, those prepared byCFSC for AID, and all other “stand-alone” projectionsare useful to the policy maker mainly because they pro-vide an input describing the population conditionsunder which policy must operate. Such projections areof little direct use to policy makers, but they oftenbecome key elements in the analyses that are preparedduring the development of policy options.

Other Population Projections

Projections of world population have also been pre-pared, without the use of a global model, by the WorldBank and Harvard University. The World Bank projec-tions, last revised in 1979, were prepared by estimating,for each country, the year in which fertility will reachreplacement Ievel. b (As defined by the World Bank, “re-placement-level fertility” refers to the level at which thenumber of births equals the number of deaths at theprevailing level of mortality. For countries with highmortality rates the replacement fertility level is consid-erably higher than 2 children per couple; as mortalitydeclines, fertility must also continue to decline. Thus, astationary population might not be achieved for morethan 100 years after the achievement of replacement-level fertility.) The World Bank projections assume thatfertility decline toward the replacement level hadstarted by 1975 in all regions except Sub-Saharan Afri-ca, where the decline was expected to start between1980 and 1985. These projections show population in-creasing to 6.0 billion by 2000, with 1.3 billion in thedeveloped world and 4.7 billion in the developingworld; the population of the entire world becomes sta-tionary around the year 2175 at a level of 9.8 billion.

In 1977, the Center for Population Studies at Har-vard University prepared a single population projectionfor all countries and regions of the world to the year2075.7 In these projections, fertility is assumed to de-cline to replacement levels by 1990-95 for the developedcountries and by 2000-05 for the developing countries;they show the population of the world increasing to 5,9billion by 2000 and 8.4 billion by 2075.

‘K. C. Zachartah and My Thl \’u, Popukmon Pro]ecrton~, J 975-1000 and Long-Term (Stutmruq Pop&tmn) (V’ashlngton, D. C.: Nrorld Bank, 1979).

TG. Llttman and N. Keyfitz, The .Nexc Hundred Years (Cambrtdge, Mass.: Har-vard Umverslty, 1977).

76 ● Global Mode/s, World Futures, and Public Policy

Strengths and Weaknesses ofPopulation Projection Techniques

Among the studies discussed here there are at leasttwo different goals:

● to describe the long-term behavior of the global sys-tem; or

● to provide an accurate, short-term (25 years or less)picture of the world situation.

The long-term global modeling studies (World 3, WIM,LAWM) are clearly concerned with the first goal only.They examine the behavior of the entire global systemover the next 50 to 125 years, but they make no attemptto provide accurate forecasts of world population (orany other variable) for 2000. All three studies take greatpains to point out that accuracy of individual numbersis not a concern. The studies that limit themselves toprojecting population only (World Bank, Harvard) aregenerally trying to satisfy both goals; they present coun-try-by-country forecasts of population in 2000, but theyalso present longer views of population growth. Thetwo modeling studies that limit themselves to 2000(UNIOWM and Global 2000) are clearly interested pri-marily in the second goal.

These two different goals require the use of two fun-damentally different modeling techniques. The first ap-proach (long-term system behavior) considers popula-tion as only one of the many elements involved in theintegrated behavior of the system: it affects the othersectors and, in turn, is affected by them. With this ap-proach the most important aspects of the populationsector are the effects of the rest of the model on popula-tion growth. The second technique (accurate short-term forecasts) considers population alone, without in-tegrating it into the world system, It strives for accuracyby making separate projections of the individual coun-tries, which are useful in themselves and can also besummed to produce a world total.

Long-Term Integrated Projections

The first technique has obvious advantages for pro-jecting long-term system behavior. Population really i saffected by the rest of the global model—by food pro-duction, pollution, and economic and social develop-ment—and it is essential to take these factors into ac-count in projecting population growth over a long timeperiod. The difficulty with this approach, however, isthat the relationships between fertility, life expectancy,and other variables are not known precisely, nor is thehistorical evidence rich enough to allow these relation-ships to be estimated with any degree of confidence. Inorder to include the relationships in their model, there-fore, the modelers are often forced to rely on rough esti-mates and informed guesses. The sensitivity of the mod-

el’s behavior to these judgments can be tested, ofcourse, but the reliance on judgment does make themodel speculative.

Within this first general approach the models dis-cussed here have adopted three variations. World 3 ag-gregates everything to the level of the world, with nogeographical divisions. This was certainly appropriatefor the preliminary versions of the model, which weremeant to be simple, easily explained outlines of theeventual model. The large effort that went into con-structing and documenting World 3 might have beenbetter served if some regional disaggregation had beenemployed, but this would have at least doubled theamount of work involved in developing, testing, anddocumenting the model.

The development gap between the developed and lessdeveloped world is so large, however, that the two re-gions cannot properly be considered as one. WIM di-vides the world into 10 or more regions, and it thereforesupplies a much more detailed representation of real-world behavior. In most of the initial WIM runs, how-ever, the link between fertility and the rest of the modelwas broken and an optimistic fertility assumption im-posed, although mortality rates did respond to the restof the model.

LAWM takes the same general approach to the popu-lation sector as the other two models, with one majordifference: the model is not designed to simulate actualworld conditions, but rather to simulate what wouldtake place if resources were allocated optimally in orderto maximize life expectancy. Since this is not the waythe world actually operates, LAWM’s usefulness is indescribing the world that would evolve, given a particu-lar new mode of behavior, and not in describing what islikel y to happen.

Short-Term Stand-Alone Projections

The second method of projecting population is usedby each of the other studies. In this approach the popu-lation projections are linked to the rest of the worldonly insofar as the rest of the world is considered by theexperts when they make their judgments about fertilityand mortalit y decline. When the U.N. experts madetheir estimates, for example, they had in mind someview of the future world which influenced their judg-ments. Their worldview obviously excluded such disas-ters as nuclear war and massive famine, or their life ex-pectancy assumptions would not have shown steady in-creases; but it is impossible to describe just what thatworldview was. Indeed, since many different peoplewere involved in this exercise, many different (and pos-s ib ly conflicting) views of the world were used. It isimpossible to say whether these worldviews were intern-ally or mutually consistent, or to what extent any of


them reflect the feedback effects between populationand development. The most that can be said is that theexperts generally assumed that the forces at work in thepast would continue to operate, with the future largelygrowing out of a continuation of past trends.

Is this assumption of a “surprise-free” future reason-able? All of the global studies discussed here (includingthe long-term studies as well as Global 2000 and theUNIOWM) apparently consider this assumption rea-sonable in the short term. None of the studies showlarge-scale effects on population due to famine or re-source shortages before 2000, although WIM shows re-gional starvation in Asia. In the longer term, however,the situation is different. The three long-term modelingstudies show a high likelihood of significant changesbeyond 2000: World 3 shows a collapse of populationafter 2030, WIM shows rapidly increasing deaths due tostarvation in South Asia before 2025, and LAWMshows food availability in Asia dropping to starvationlevels by 2040 in its standard run. Other studies notbased on mathematical models, such as Herman Kahn’sThe Next 200 Years,a do not foresee such problems, but— —

‘H. Kuhn and A. Vrlener, The Year 2(W A Frameu ork ~c~~ Specdutmn on the?Y’ext ThmwThree Ye~rs (New York. M o r r o w , 1967).

the assumption of a long-term continuation of pasttrends is certainly questionable. The likelihood of sucha future actually developing can only be addressedthrough a comprehensive, integrated study of all factorsaffecting population growth.

Factors Affecting the Reliability andAccuracy of Population Projections

Differences in purpose and time horizon also affectthe reliability and accuracy of the resulting populationprojections. World 3, for example, attempts to describethe general behavior of the global system to the year2100. The precision sufficient for such a model is quitedifferent from that desired in the U.N. or Census Bu-reau population projections, which are used to predictthe populations of individual countries in 2000.

Table A-1 shows the results of the studies that havemade long-term population projections. There are twonotable differences. Due to increasing death rates, pop-ulation actually begins to decline in the World 3 modelafter 2025 and presumably would do so for Asia inWIM and LAWM if they had continued beyond 2060;the other projections all show steadily growing popula-

Table A-l.—Comparison of Long. Range Population Projections

Projection source Projection Results

World 3 . . . . . . . . . . . . . . . . . . . . . .

World Integrated Model . . . . . . . .

Latin American World Model . . . .

United Nations. . . . . . . . . . . . . . . .

CFSC . . . . . . . . . . . . . . . . . . . . . . . .

World Bank. . . . . . . . . . . . . . . . . . .

Harvard . . . . . . . . . . . . . . . . . . . . . .

Standard run

Stationary world

Standard run

Standard run

Second run—improvedconditions in Asia

1978 populationassessment

Medium and lowproject ions

Standard

Standard

Population increases to 7 billionby 2025 then decreases to 4billion by 2100

Population stabilizes at 6 billionby 2050

Population stabilizes at justunder 7 billion by 2015. Deathrates due to starvation are highin Southeast Asia

Population reaches 11 billion by2040 and is still growing at 1.1percent/yr. Death rates are ris-ing rapidly in Asia

Population reaches almost 11billion by 2060 and is growing atless than 0.5 percent/yr.

‘Population reaches 8 to 12billion by 2050 and stabilizes at8 to 14 billion by 2150

Population reaches 7.8 to 8.1billion by 2050 and is virtuallystationary

Population stabilizes at 9.8billion by the year 2175

Population reaches 8.4 billion by2075 and is virtually stationary -

aThe~e are unpublished, provisloflal estimates that are not official.



tion until some stationary level is reached. The secondmajor difference is in the size of that stationary worldpopulation, which ranges from a low 8 billion (U.N.low, CFSC, Harvard) to a high of 14 billion (U.N.high). (The lower figures of World 3 and WIM are basedon exceptional population programs begun in 1975,and are thus not comparable.)

There is much less variation in the population projec-tions for 2000, which are compared in table A-2 and fig-ure A-5. They vary from a low of 5.8 billion (CensusBureau, CFSC) to a high of 6.5 billion (U.N. and Cen-sus Bureau). This much smaller variation indicates themuch higher degree of certainty inherent in populationprojection for periods under 25 years: there is relativelylittle uncertainty about the number of reproductive-agefemales between now and 2000, although there is muchgreater uncertainty with respect to the number of chil-dren that each woman will have.

The reasons for the differences among these differentprojections and their relative accuracy and reliability

depend on several factors, among which are:Ž projection technique;Ž data base for initial population figures; andŽ estimates of future fertility and life expectancy.

Table A.2.—Comparison of Population Projectionsfor 2000

Populationin 2000

Projection source Projection (billions)

World 3 . . . . . . . . . . . . . . . . . . . . . . . . . Standard run About 6World Integrated Model . . . . . . . . . . . Standard run 6.4Latin American World Model . . . . . . . Standard run 6.4United Nations. . . . . . . . . . . . . . . . . . . High 6.5

Medium 6.2Low 5.9

U.S. Bureau of the Census . . . . . . . . . High 6.5Medium 6.2

Low 5.8CFSC. . . . . . . . . . . . . . . . . . . . . ... , . . High 6.0

Medium 5.9Low 5.8

World Bank. . . . . . . . . . . . . . . . . . . . . . Standard 6.0Harvard ., . . . . . . . . . . . . . . . . . . . . . . . Standard 5.9


Projection Technique

The integrated models include the effects on popula-tion from other parts of the global system, and this re-sults in the projection of increasing death rates due tofood shortages in at least some parts of the world. The

Figure A-5.-Aiternative Projections of World Population in 2000Billions7.0

6.0

5.0

4.0

3.0

2.01960 1970 1980 1990

Year2000


projections that did not employ global models assumedthat such disasters would be prevented by technicalchange, but when forecasters are forced to specify theirassumptions mathematically in a model, they were alldrawn to similar conclusions: it is unlikely that techno-logical advances alone will allow the world sufficienttime to solve the problem of rapidly growing popula-tion. The technology assumptions or, indeed, the verystructure of the models could well be wrong; but thosewho develop long-term population projections withoutthe use of models have certainly side-stepped this issue.The problems identified by the integrated global modelsmight be solved through improved technolog y or differ-ent social responses, but none of the projections pre-pared without models explicitly addresses these ques-tions.

Data Base for Initial Population Figures

All of the recent projections use essentially the samedata base for base-year population. Information fromnational sources is collected by the U. N., the WorldBank, and the U.S. Census Bureau, and for most coun-tries all three organizations report the same populationtotals. However, there is very little available informa-tion for some countries; it is therefore necessar y t oprepare estimates based on incomplete data, and theseestimates may vary.

Uncertainty about base-year data is not expressed inthe ranges reported by the U. N., and the Census Bu-reau projections incorporate different base-year dataonly for China. The recent round of censuses has sub-stantially improved the information available for many

countries, particularly in Africa, but for many countriesthe variation in estimates can still be quite large. China,for example, has not had a census since 1953, and thereis a 14-percent difference between the U.N. low andCensus Bureau high estimates for China in 1980, a dif-ference about equal to the population of the UnitedStates. Similarly, the current population of Nigeria hasbeen estimated at anywhere between 65 million and 85million, and the most recent census in India found 12million more people than expected. Even though thesedifferences may be quite large for individual countries,however, they are fairly small once they are summed tothe global level: the difference between the highestestimate for world population in 1975 (Census Bureauhigh series of 4,134 million) and the lowest estimate(World Bank estimate of 4,014 million) is only 3 per-cent.

Estimates of Future Fertility andLife Expectancy

It has generally been true that projection of popula-tions for countries and regions is relatively accurate inthe short term (15 to 20 years) and fairly inaccurate in

the long term (over 25 years). The reason for this lies inthe nature of population change. Population sizechanges through three mechanisms: births, deaths, andmigration. Except for certain special cases, migration isusually not a major factor in population growth, al-though movements from rural to urban areas can havesignificant social, economic, and political conse-quences. Uncertainty in birth and death rates, on theother hand, can be quite large.

Changes in the birth rate have occurred quite rapidlyin the past, and since the reasons for changes in thebirth rate are poorly understood, projections of birthrates have tended to err on the side of assuming lesschange than actually takes place. However, even if theprojection of the birth rate misses the mark by a signifi-cant amount, the short-term projection of the size oftotal population may still be reasonably accurate.

After 15 or 20 years, however, errors in projectingbirth rates begin to have a much larger effect. By thattime children born during the first years of the projec-tions would begin to have children of their own, andany overestimation or underestimation of births willbecome compounded, leading to exponentially increas-ing errors in the projection. Since changing death ratesaffect the mortality of children much more than olderage groups, this delayed compounding effect also oper-ates on errors in death-rate projections. For this reasonthe record of population projections beyond 20 yearshas not been especially good.

Another crucial factor is the accuracy of the base-yeardata for birth and death rates. If they are not knownwith much accuracy, then the projection will have abuilt-in error and, even if the future change in theserates is correctly forecast, the total error of the projec-tion can be substantial. These factors most seriously af-fect projections for developing countries, where popula-tion data are often scant and unreliable.

The Accuracy of Past andPresent Projections

The 1950’s and 1960’s were a period of rising popula-tion growth rates, and U.N. projections made duringthis period generally underestimated future population.Past U.N. projections of world population for 1980 il-lustrate this point: the latest assessment in 1978 pro-jected the 1980 population at 4.41 billion; in 1957 it wasprojected at 4.22 billion (4.3 percent less than the latestestimate); in 1963 at 4.33 billion (1.8 percent less); andin 1973 at 4.37 billion (0.9 percent less) (see fig. A-6).

The major factor affecting these estimates was inac-curate base-year data on death rates, which were 10 or12 percent lower than estimated. Birth rates, whichwere estimated more accurately for the base year, havedeclined more rapidly than predicted. The current esti-


Figure A-6.—United Nations Population Projectionsfor 1960 Prepared Over the Past 25 Years

Billions Billions4.6

4.4

4 . 2

4 . 0

3.8

3.6

4.6

4.4

4.2

4.0

3.8

3.61957 1963 1973 1978

Year in which assessment was made

SOURCES: 1957 and 1963: World Population Prospects as Assessed in 1963(New York: United Nations, 1966), 1973: Se/ecfed Wor/dDemographic Indicators by Countries, 1950-2000 (New York:United Nations, 1975). 1978: Wor/d Popu/atlon Trends and Pro-spects by Country, 1950-2000, Summary of the 1978 Assessment(New York: United Nations, 1979),

mates of population size, birth rates, and death rates aremuch improved over what they were in the 1950’s and1960’s, but large uncertainties still remain, particularlywith the LDCs.

The uncertainty over present birth rates also contrib-utes to uncertainties about future birth rates. In the1980-2000 period birth rates will probably decline foralmost all ‘countries of the world, but projecting theamount of that decline is made difficult by at least fourfactors:

●

●

●

●

uncertainty about how much birth rates have al-ready declined up to the present;uncertainty about how many countries will adoptserious family planning policies and how strongthose efforts will be;uncertainty about how much family planning ef-forts have contributed to past declines in birthrates; anduncertainty about how relevant the experience offamily planning effectiveness in countries that havealready adopted such efforts will be in those coun-tries which have not yet done so.

Most of the uncertainty is about birth rates in the de-veloping countries, notably in Africa where there is al-most no experience with strong family-planning pro-grams. It is simply impossible to know whether the fam-ily-planning experience of countries in Asia and LatinAmerica will be repeated in Africa.

Michael Stoto of Harvard University has calculatedthe average error in individual country projectionsmade by the U.N. in 1957, based on the difference be-tween predicted and actual growth rates for 40 coun-tries for four time periods: 1955-60, 1955-65, 1955-70,and 1955-75. He found that the absolute average errorin annual growth rates for all four time periods wasabout 0.46 percent, or about 25 percent of the absoluteaverage annual growth rate for the world (about 1.9percent). 9 The Futures Group has repeated these calcu-lations for the period 1970-75, using projections madeby the U.N. in 1973 and actual growth rates taken fromits 1978 assessment, Using the same 40 countries, buteliminating those for which no new data are availablesince 1972, they found an absolute average error of 0.27percent. Most of the improvement is probably due tothe improved quality of base-year data available for the1973 forecasts, which were used in UNIOWM, If theseerrors are representative of the kind of errors we can ex-pect in the future, projections of population size 20years into the future should have an uncertainty rangeof 10 to 20 percent.

In the longer term the accuracy is much lower. This isdue to the greater uncertainty about fertility and mor-tality trends, the compounding of errors made in theshort-term forecasts, and uncertainty about the effectsof the rest of the world system (food availability,economic development, pollution) on populationgrowth. The range of projections shown in the studiesdiscussed here–from a projection of steadily growingpopulation reaching a level of 8 to 14 billion, to a pro-jection of population growth and collapse—is represen-tative of the kind of uncertainty that exists in trying toproject 50 to 100 years into the future, even if we arewilling to exclude such disasters as nuclear war (see fig,A-7).

.‘J. Stover, op. cit.

Figure A.7.—lncreasing Range of Uncertainty inLonger Range Projections of World Population

1 9 8 0 1 9 8 0 2000 2020 2040 2080 2080 2100

NOTE: For illustrative purposes only; see table A-1 and text for explanation.


APPENDIX B

Agricultural Projections

Introduction

World food supply has been a major concern for glo-bal modelers. All of the well-known global models haveone or more agricultural sectors; all of them considermeasures of food availability as major indices of systemperformance; and all indicate that performance of theglobal agricultural system over the next 20 to 100 yearsis a matter of concern. The apparent discrepancies intheir findings result from differences in time horizons,model structure, assumed rates of population and eco-nomic growth, the pace of technological progress, andassumptions about the quantity of agricultural landavailable.

In general, the most optimistic findings come fromassumptions of high income growth, low populationgrowth, continued rapid technological progress, andlarge reserves of agricultural land. Longer time hori-zons, however, lead to more pessimistic results; andthose models that include diminishing returns to landand agricultural inputs show the situation in the globalagricultural system getting tighter as time progresses,with South Asia and Sub-Saharan Africa the most se-verely affected regions.

Purposes, Structures, and Findings

World 3

The World 3 model was intended to examine the in-teractions between a set of global trends and to identifythe long-term impact of their interdependent evolution.as a result, the World 3 model assumes that agriculturemust compete with the industrial and service sectors forinvestment capital and natural resources, and it con-tains a mechanism by which capital and resources flowto sectors that show signs of supply shortfalls. Themodel also assumes diminishing returns for investmentsin land development and agricultural inputs, such asfertilizer and farm machinery; but it excludes both theprice mechanism and “disembodied technological prog-ress” (see below). The model does assume, however,that soil degradation and pollution will have negativeeffects on yields.

In the standard run (see fig. B-1), all indices ofagricultural performance improve until the seconddecade of the next century. Yields and land under

IThe followlng material IS based on an OTA working paper prepared by, Jen-mfer Rolmnson, a fellow of the International Institute of Applled Systems Anal-vsls, For further Information on this sub]ect, see OTA’s forthcoming assess-m e n t , T h e Impact Oj Technolog> on [he PmducrN [tv of the L u n d

cultivation both make considerable gains between 1980and 2000, and food per capita increases 10 percentdespite rapid population growth. Around 2015, how-ever, at about the same time that the limits of arableland are reached, industrial growth so depletes theresource base that investment must be shifted awayfrom agriculture in order to compensate for the increas-ing costs of resource extraction (see fig. B-2). Industrialoutput declines, as does the use of agricultural inputs,causing both yields and total production to declinemore rapidly in the 21st century than they had ex-panded in the 20th century (see fig. B-3). Since popula-tion continues to increase, the result is widespreadhunger, mass starvation, and a delayed but catastrophicdecline in global population.

World 3 produces different projections when plausi-ble changes are made in its assumptions, but althoughthe timing of events may be changed by a few decadesthe net result is the same. For example, given more op-timistic assumptions about industrial resources and/ormore pessimistic assumptions about agricultural re-sources, agricultural decline causes investments andresources to be drawn away from industry rather thanvice versa; but decline feeds on decline and mass starva-tion ensues. However, one group of critics reports thatthey have been able to move the physical limits to agri-cultural production beyond the time horizon of themodel by assuming continuing technical progress inboth land-development techniques and high-yield plantvarieties, as well as a more rational use of agriculturalresources. 2 The structure of World 3 is not sufficientlydetailed or flexible, however, to determine what specificp o l i cy actions this might entail. Its sensitivity to re-source depletion in other policy tests also suggests thatthe world agricultural crisis, though moved beyond2100, might still occur.


WIM’S projections have greater relevance for foodpolicy analysis because its structure allows a region-by-region investigation of the interaction between popula-tion, agriculture, and industrialization. WIM goes intomore detail than World 3 and also includes both theprice mechanism and a simulation of food and othertrade between regions. Environmental effects have beenexcluded, but the shortened time horizon—2025—is stilllong enough to encompass the crises foreseen by World3. WIM also includes numerous “policy levers,” making

‘H. S, D. Cole (cd.), Models OJ Doom (New York: Universe Books, 1973), p. 64.

81


II

0 0 0 0 00 0 0 0 0

Simulation of World

1900 1950 2000 2050 2100

NOTE: Around 2015 the rapid decline of nonrenewable resources forces the industrial sector to shift in-vestments away from industrial and agricultural production to compensate for increasingresource extraction costs. This causes declines in industrial output and agricultural production.The latter, in turn, causes massive starvation and decline of global population by nearly 3 biillonover the period 2030-2100.

SOURCE: Dynamics of Growth in a Firtite Wor/d, pp. 501-s03.

it a flexible tool for testing different combinations of ac-tions that could be taken to address potential food sup-ply problems.

The WIM standard or “historical” scenario is basedon a continuation of present trends, but it neverthelessmakes some rather optimistic assumptions about agri-cultural progress in the developing regions. (The reportfocuses on South and Southeast Asia because of thisregion’s existing food problems and the number of peo-ple involved, but the authors assert that their conclu-sions “are applicable to Tropical Africa and to anyother needy region. “ 3) It assumes, for instance, that allavailable arable land is quickly brought under cultiva-tion and that all technological inputs, such as irrigationsystems and farm machinery, will be available asneeded. It also assumes “quite optimistically” that theaverage use of fertilizer per hectare in the region willsurpass the present North American level b y 2025, atwhich time South Asia alone will be consuming morefertilizer per year than the entire world consumed in

‘M. D. Mesarovlc and E, Pestel, IMunkind U[ the Turning Pmnt (New York: Dut-

ton, 19 T-1), p. 121.

1960. These factors increase yields by about 1,000 kgper hectare, approximately the same increase achievedby the Green Revolution on the best land before fer-tilizer prices began to soar. Finally, the standard runassumes that other regions make enough food availableto cover any production shortfall in South Asia.4

Despite these assumptions, however, and despite theassumption that population will stabilize by 2025, thefood supply projections for South Asia are grim (see fig.B-4). The region’s annual protein production increasesby two-thirds, but the population almost triples and theannual protein deficit grows from 12 million to 50 mil-lion tons–an amount equal to the region’s own pro-duction. Deficits amounting to half the region’s proteinneeds “could never be closed by imports, ” according tothe authors: paying for them would require one-third ofSouth Asia’s total economic output and three times itsexport earnings, and “the physical problems of han-dling those quantities of food would be incredible.”5

Even if the needed imports were available, the annual

‘Ibd., p. 121.51bd., p. 122.

(

1I<

1

B-2.–Output From Standard Simulation of World 3

output allocated to●

/● — ●

output

Agriculturalinput perhectare

1900 1950 2000 2050 2100

Around 2015 the rapid decline of nonrenewable resources forces the industrial sector to shift in-vestments away from industrial and agricultural production to compensate for increasingresource extraction costs. This causes declines in industrial output and agricultural production.The latter, in turn, causes massive starvation and decline of global population by nearly 3 billionover the period 2030-2100.

SOURCE: Dynamics of Growth in a F)rwle World, pp. 501-503.

number of child deaths caused by malnutrition willdouble by 2005 (see fig. B-5).

Policy tests conducted with WIM indicate that only acombination of food aid, population policies, and bal-anced development can avert tragedy in South Asia. Inthe “isolationist” or “tragic” scenario, in which food im-ports are not available because of balance-of-paymentsconstraints, annual child mortality is twice as high as inthe standard run; it rises sharply after 1985, peaks in2010, and declines thereafter only because of the de-layed impact of earlier deaths on the later number offertile women (see fig. B-5). A third scenario, designedto investigate policies aimed at food self-sufficiency forSouth Asia, assumes that virtually all regional invest-ment is shifted from industrial development to agricul-ture; but the results indicate that yields per hectare,after initially rising faster than in the standard run,would peak around 2000 and decline thereafter. Thisdecline occurs because the agricultural sector would notbe able to maintain its growth without the industrialbase that must supply it with fertilizer and machinery.By 2025, gross regional product is only half what it was

in the standard run, and the region is left with evenfewer means of paying for food imports.

In further policy tests, WIM shows that populationpolicies aimed at achieving an equilibrium fertility ratecould have a significant effect on food deficits and childmortality, even in the absence of imports, if they are im-plemented quickly enough. Such policies, if initiated in1995, would not reduce the number of child deaths inthe “isolationist” scenario; if initiated in 1990, however,the same policies-might save more than 150 million lives(see fig. B-5). If initiated in 1975, these policies couldavoid more than 500 million child deaths. b The needfor food imports would be significantly reduced andwould come later in the period, but the cost would stillbe prohibitive. In a final scenario, therefore, it is as-sumed that the developed regions provide South Asiawith sufficient investment aid to develop “its ownexportable and competitive industrial specialization, ”whose exports could pay for most of its food imports.7

‘Ihlcl., p. 124 and fig. %9.Vtd., p. ] ~T.


iII

0 0 0 0 00 0 0 0 0 1900 1950 2000 2050 2100

NOTE: Around 2015 the rapid decline of nonrenewable resources forces the industrial sector to shift in-vestments away from industrial and agricultural production to compensate for increasingresource extraction costs. This causes declines in industrial output and agricultural production.The latter, in turn, causes massive starvation and decline of global population by nearly 3 billionover the period 2030-2100.

SOURCE: Dynamics of Growth in a Finite Wor/d, pp. 501-503.


LAWM was developed to show that, given optimalresource allocation and the universal objective of satis-fying basic human needs, the global system need not betroubled by physical limits. Because development pro-ceeds “autarchically” in each of its four regions, interna-tional food trade is unimportant and is largely excludedfrom the model. Environmental effects are also omitted,as are food prices; and although the model assumesdiminishing returns on land development and yield-enhancing inputs, it also assumes “disembodied techno-logical progress” (see below) in the form of an automaticI. O-percent annual increase in the productivity of thefood and agriculture sector.

The food and agriculture sector is by far the mostcomplicated in LAWM, containing three subsectors—agriculture, livestock, and fisheries–among whichcapital and labor are allocated in patterns that shiftover simulated time as the return on investment dimin-ishes in each. Each of these subsectors contains at leastone optimistic assumption. The agriculture subsector,

for instance, assumes that fertilizer will be available inunlimited quantities and at constant prices throughoutthe simulation period, and that processing losses in thedeveloping regions will decrease automatically eachyear until they reach the levels currently found in thedeveloped region. Similarly, the livestock subsectorassumes that agricultural wastes and excess agriculturalproducts will be used for animal fodder once humanneeds are met, thereby transforming food wastes into ameasure of meat consumption. The fisheries subsectorassumes a maximum sustainable catch of 120 millionmetric tons per year, considerably higher than the levelindicated by more recent reports.

The model assumes no policies to limit populationgrowth, other than the general improvement of livingconditions. However, it does assume a radical, egalitar-ian redistribution of income and consumption withinregions, which greatly increases the effective demandand relative benefit for the lowest socioeconomic strata.

Given these optimistic assumptions, the standard runof this optimization model indicates that all regions ex-cept Asia will be able to satisfy their own food needs

App. B—Agricultural Projections ● 8 5

Figure B-4.—World Integrated Model Standard Run for South and Southeast Asia

South Asia

L /Annual protein needs

Population

protein production

Annual proteindeficit

I I I I I I J

Year

f

within 30 years. The agriculturally relevant variable inthe simulation outputs is the total daily caloric intakeper capita: in the developed nations, it rises to 3,200 cal-ories by 1980 and equilibrates at that level thereafter(fig. B-6); in Latin America it rises to 3,000 calories by1990 and stabilizes at that level, which is lower due todifferences in climate and diet (fig. B-7); Africa achievesand stabilizes at a similar level around 2008. In Asia,however, the only need that is met is education; foodper capita peaks at less than 3,000 calories per day in2010 and declines steadily thereafter (fig. B-8). This agri-

cultural collapse is similar to the catastrophe foreseenby WIM in both its standard and self-sufficiency runs:

The problem in Asia arises in the food sector. By 2010,all available land is being cultivated. Thereafter, eco-nomic effort in the sector is devoted to increasing live-stock and fisheries. This, however, is not enough to feedthe growing population adequately, and consumptiondrops rapidly to below the minimum needed for survival.

The rapid increase in the cost of producing food, due tothe development of new land for agriculture, takes re-sources from the rest of the economy, causing backward-


Figure B-5.—Child Mortality in Four Scenarios of the World Integrated Model

Scenario 1 is the standard scenario. Scenario 2 shows the consequences for Scenario 1 if im-ports are not available to cover the protein deficiency gap. Scenario 3 shows the reduction inchild mortality achieved over Scenario 2 by a population policy instituted in 1990. Scenario 4shows the consequences of implementing the same population policy 5 years later. Com-posite of Figures 9-2 and 9-4, Mankind at the Turning Point, pp. 122 and 128. The projection ofScenarios 2 and 4 are essentially identical in the original.

SOURCE: Mankind at the Turning Point,

ness and also hindering the satisfaction of the other basicneeds. In summary, the delay in reaching adequate levelsof well-being leads to a sustained high population growthrate, and a vicious circle develops: increased populationand the increased cost of producing food make it moreand more difficult to satisfy basic needs.8

Rather than show the full details of this catastrophe,the modelers have truncated the Asia run at 2040, 20years before simulations for other regions are ter-minated. The authors advocate effective populationpolicies and the use of nonconventional foodstuffs toavoid mass starvation in Africa, but then present nei-ther specific details nor policy tests to support these rec-ommendations.

The policy tests that were conducted by the LAWMteam indicate that capital transfers from the developedregion would have a negligible impact on the foodshortage in Asia. They also show that technologicalstagnation after 1980 would lead to a similar collapse in

8 A . 0. Herrera, et al,, Cutustrophe m .\Teu SO C ie ty 7 A Lar[n Amerlcun WorkfModel (Ottawa: international Development Research Center, 1976), p. ,?9.

Africa as well as Asia and that, in the absence ofregional income redistribution, the satisfaction of basicneeds (though possible) would require three to fivetimes more resources and two to three more generationsof human suffering.

United Nations Input-Output World Model(UNIOWM)

UNIOWM is the only model that does not indicatepotential problems in supplying food to all the world’speople over the next two decades. Its optimistic find-ings, however, result in part from its purpose and struc-ture: its projections do not show what is likely to hap-pen in the agricultural sector, but rather what trendswould be required in order to achieve the goals of theU.N.’s Second Development Decade. The input-outputapproach is well suited for consistent accounting of in-tersectoral flows, but it is not particularly well adaptedfor agricultural analysis because it is totally linear.Many biological processes, on th e other hand, are

App. B—Agricultural Projections • 87

Figure B-6.—Time Period and Conditions Required for Developed Countries toSatisfy Basic Needs to Given Levels

I I I I 1 1 I 1 i 1 I

1960 1970 1980 1990 2000 2010 2020

1 (5) Percent GNP allocated to sector 57 (v)

2 (B) Birthrate 8 (C)

3 (4) Percent GNP to other goods and services 9 (E)

4 (U) Urbanization 10 ($)

5 (A) Population growth rate 11 (P)

6 (M) EnrollmentSOURCE: Catastrophe or New Societyq

2030 2040 2050 2060

Key:

Houses per familyTotal calories per capitaLife expectancyGNP per capitaTotal population


Figure B-7.—Latin American World Model Simulation for Latin America

1 (B) Birthrate 7 (C) Total calories2 (5) Percentage of GNP allocated to sector 5 8 (E) Life expectancy3 (4) Percentage of GNP allocated to sector 4 9 ($) GNP per capita in 1960 dollars4 (A) Population growth rate 10 (P) Total population5 (M) Enrollment 11 (U) Urbanization6 (V) Houses per family

App. B—Agricultural Projections Ž 89

Figure B-8.— Latin American World Model Simulation for Asia

1960 1970 1980 1990 2000 2010 2020

Key:

1 (A) Population growth rate 7 (M)2 (V) Houses per-family 8 (U)3 (5) Percentage of GNP allocated to sector 5 9 ($)4 (B) Birthrate 10 (E)5 (4) Percentage of GNP to other goods and services 11 (P)6 (C) Total calories per capita

SOURCE. Catastrophe or New Soc/ety? p. 92.

EnrollmentUrbanizationGNP per capita inLife expectancyTotal population

1960 dollars

90 “ Global Models, World Futures, and Public Policy

highly nonlinear, and many critical agricultural flows(such as the externalities associated with overgrazingand deforestation, or the changing probabilities of pestdamage under different cropping systems) are difficultto include in an input-output framework. In addition,UNIOWM shows linear returns on investments in agri-cultural inputs and must depend on off-line analysis todetermine the amount of land under cultivation. Thesefeatures produce odd results in some places, such as a169-percent increase in the productivity of Japanesefarmlands, which are already intensively cultivated,and a 387-percent increase in the Middle East.

In general, the projections show rapid increases foralmost all agricultural variables in almost all regions,with the most dramatic gains being made in the devel-oping countries. Over the 30 years of the simulation(1970-2000), global grain production almost triples andglobal production of animal products more than triples(table B-1). Developing regions achieve astounding in-

creases in both land productivity and total agriculturalproduction (table B-2), and by 2000 all regions havereached an average daily per capita consumption ofover 2,400 calories and 66 grams of protein (table B-3).These results, however, do not seem to be accompaniedby any symptoms of economic or financial stress. Agri-cultural prices, relative to general price levels, increaseonly 14 percent over 30 years. In no region does invest-ment in irrigation or land development grow by morethan a few percent per year, and in man y r e g i o n sagricultural investments actually decline. If anything,the pressure on the agricultural system appears to beeasing in 2000: rates of agricultural demand growthdecrease slightl y in the last decade of the simulation,largely because incomes have risen to a point whereconsumers spend less of each additional dollar of in-come on food.

It should be repeated that these projections are in-tended only to prove the technical and physical feasibility

Table B- I.—Global Agricultural Output in UNIOWM Standard Scenario

1970 1980 1990 2000


Table B-2.–Land Requirements and Yields in 2000in UNIOWM Standard Scenario

(Index, 1970 = 100 percent)

AgriculturalRegion output

Arable Landland productivity

Developed market:North America . . . . . . . . . . . 196Western Europe

(high-income) . . . . . . . . . . 130Japan . . . . . . . . . . . . . . . . . . . 176Oceania. . . . . . . . . . . . . . . . . 192

Centrally planned:Soviet Union . . . . . . . . . . . . . 164Eastern Europe . . . . . . . . . . . 143Asia (centrally planned) . . . . 468

Developing market:Latin America

(medium-income) . . . . . . . 495Latin America

(low-income) . . . . . . . . . . . 532Middle East . . . . . . . . . . . . . . 950Asia (low-income) . . . . . . . . . 506Africa (arid) . . . . . . . . . . . . . . 409Africa (tropical). . . . . . . . . . . 438

SOURCE: The Future of the Wor/d Economy, p. 40.

111

100100183

100100120

166

140126113131152

194

162269162

2151862 7 8

311

328487331282324

Table B-3.—Regional Daily per Capita FoodConsumption in 1970 and 2000, UNIOWM Standard

Scenario

Kilo-calories Proteins(thousands) (grams)

Region 1970 2000 1970 2000

Developed market:North America. . . . . . . . . . . . . . . . . . . . 3.2 3.2 96 100Western Europe (high-income) . . . . . . 3.0 3.2 91 105Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 3.2 71 117

Centrally planned:Soviet Union . . . . . . . . . . . . . . . . . . . . 3.2 3.2 92 108Eastern Europe . . . . . . . . . . . . . . . . . . . 3.1 3.2 93 108Asia (centrally planned) . . . . . . . . . . . . 2.1 2.5 59 79

Developing market:Latin America (medium-income) . . . . . 2.4 3.0 60 86Latin America (low-income) . . . . . . . . . 2.2 2.9 50 74Middle East . . . . . . . . . . . . . . . . . . . . . . 2.0 2.9 53 92Asia (low-income) . . . . . . . . . . . . . . . . . 2.0 2.4 52 66Africa (arid) . . . . . . . . . . . . . . . . . . . . 2.5 2.5 72 78Africa (tropical) . . . . . . . . . . . . . . . . . . . 2.2 2.8 62 87

SOURCE: The Fufure of the World Economy, p. 39.

of certain U.N. development goals The modelers in-volved in the UNIOWM have tended to merely statetheir findings and allow readers to draw their ownpolicy conclusions. Like WIM and LAWM, however,their model points to the need for increased food self-sufficiency and export earnings in the LDCs. Findingsrelevant to food and trade policy include the following:

The most pressing problem of feeding the rapidly in-creasing population of the developing regions can besolved by bringing under cultivation large areas of cur-rently unexploited arable land and by doubling and treb-ling land productivity. Both tasks are technically feasible

but are contingent on drastic measures of public policy fa-vorable to such development and on social and institu-tional changes in the developing countries.9

Self-sufficiency in food is a promising kind of “importsubstitution” for reducing balance of payments deficits indeveloping countries.10

A relatively stable increase in the prices of minerals andagricultural goods exported by the developing countries,as compared to the prices of manufactured goods, is oneway of increasing the export earnings of these countriesand closing their balance of payments deficit . . . . Fordeveloping regions which are not large net exporters ofminerals or agricultural goods, the main way to reducethe potential trade imbalance is to significantly decreasetheir import dependence on manufactured products . . .while . , . increasing their share of world exports of somemanufactured products, particularly those emanatingfrom light industry. . . . Increase in aid; measures tocreate a more favorable climate for a better mix of capitalinvestment flows to these regions; [and] . . . reduction inthe financial burden arising from foreign investment areimportant, but . . . secondary . . . compared to . . .changes in the commodity markets and trade in manufac-tured products.

To ensure accelerated development two general condi-tions are necessary: first, far reaching internal changes ofa social, political and institutional character in the devel-oping countries, and second, significant changes in theworld economic order.l 1

Global 2000

The agricultural projections in Global 2000 were gen-erated by the grain-oilseed-livestock (GOL) model thatwas developed in 1974 by the U.S. Department of Agri-culture (USDA) to assist in the formulation and execu-tion of U.S. agricultural and trade policy. Maintainedby the Foreign Demand and Competition Division ofUSDA’s Economics, Statistics, and Cooperative Serv-ice, GOL is a computer-based static equilibrium econo-metric model that was specifically designed to capturethe interaction between the largely cereal-oriented foodeconomies of the developing regions and the livestock-oriented food economies of the industrialized regions.For the purposes of Global 2000, GOL was supple-mented with three independently developed submodelsthat project the availability of ‘arable land, the totalfood supply (including fisheries and other miscellaneoussources), and the use of fertilizer in each region. GOLhas been used to analyze the potential impact of U.S.parity pricing policies on international food trade andto analyze the potential impact of alternative assistanceprograms for the U.S. Agency for In terna t ionalDevelopment.

‘AI’. Leontlef, et al., The Future oj the W o d d Econom? (New, York : Oxford ,1977), p. 21.

*OIbld., p. 22.1 IIbld., p, ,?3.

92 ● Global Models, Worlds Futures, and Public Policy

The greatest advantage of the GOL is its scope anddetail: it consists of 28 interactive regional submodelscontaining equations for the supply, demand, trade,and prices of 16 different food commodities. USDAanalysts claim that the model represents 70 to 80 per-cent of world production and consumption and aneven larger percentage of food trade. Its greatestweakness is that, as a static equilibrium model, it is in-capable of representing market behavior that is in dis-equilibrium. In addition, dynamic factors such as popu-lation and income growth must be calculated exoge-nously in advance and thus are not affected by themodel’s operation. These factors were adjusted to bemore consistent with other sectors of Global 2000, buta number of minor discrepancies exist between theGNP and population projections and the correspond-ing GOL assumptions. Other critical assumptions in-corporated into GOL include the following:

●

●

●

no major wars, changes in the international eco-nomic order, or natural disasters such as climaticchange or large-scale land degradation;no increase in world grain reserves to keep pacewith population growth, no change in Western Eu-rope’s somewhat protectionist agricultural andtrade policies, and no major increase in U.S. foodtrade with the Soviet Union, Eastern Europe, orthe People’s Republic of China; andcontinued technological progress (measured inyields) comparable to: the rapid growth of the last20 years, with the industrialized nations and (to alesser extent) the LDCs taking advantage of tech-nology according to the incentives provided bychanges in the prices of production factors andfood commodities.

GOL generated four alternative sets of agriculturalprojections for Global 2000, using different assumptionsabout population, income, weather, and energy prices:

Alternative 1, the standard or “baseline” projec-tion, assumes medium rates of population and in-come growth, constant weather, and constant realenergy prices at 1974-76 levels.Alternative IA, a variant of the standard run, as-sumes a doubling of real energy prices by 2000.Alternative II, the optimistic upper-bound projec-tion, also assumes constant real energy prices, butassumes lower population growth and higher percapita income growth, as well as more favorableweather conditions than over the last 25 years.Alternative 111, the pessimistic lower bound pro-jection, assumes a doubling of real energy prices,higher population growth, lower income growth,and less favorable weather conditions.

Detailed regional projections for 1985 and 2000 oftotal and per capita grain and food production, con-sumption, and trade are presented in tables B-4 and

B-5. “Other African LDCs” are included in order toshow the model’s most problematic region, and SouthAsia is included to allow comparison with the results ofother global models. World grain production and re-gional per capita consumption figures are compared infigures B-9, and B-10, which illustrate the range of un-certainty that results from different exogenous assump-tions. In case of energy variables, according to thereport, “[the] range reflects not so much uncertaintyabout petroleum price increases as uncertainty about. . . the ability of farmers to maintain or expand pro-duction while shifting away from energy-intensive in-puts .” l2

Within this range, the results indicate a near dou-bling of global food supply between 1970 and 2000.Roughly speaking, this comes from a 50-percent in-crease in the developed regions and a 150-percent in-crease in the LDCs. In both cases the increase comesfrom fertilizer use rather than land development: globalcultivated land increases less than 5 percent by 2000,whereas the application of fertilizer per hectare in-creases 160 percent, doubling in the developed regionsand quadrupling in the LDCs. However, because popu-lation growth is more rapid in the LDCs than in thedeveloped regions, LDC consumption generally in-creases more rapidly than production. As a result, inter-national food trade will expand ,briskly, with theUnited States and Argentina benefiting most from thelarger markets (see table B-6). Gains in per capita con-sumption are small and unevenly distributed in theLDCs: Tropical Africa shows net declines in per capitafood consumption even in the most optimistic scenario,and gains in South Asia are less than 10 percent at best;on the other hand, per capita increases of 10 to 30 per-cent are projected for Latin America and East Asia.The real price of food on the world market is projectedto increase by between 30 and 115 percent, dependingon the scenario; under the higher figure, the poorestLDC importers could find themselves priced out of themarket as they were in 1973-75.13

These findings lead to several conclusions relevant tofood policy. The world has the physical and economiccapacity to meet substantially increased food demandthrough 2000, but to do so it must maintain the near-record growth rates of the 1960’s and 1970’s. Significantincreases in food trade will be needed to balance excessdemand in food-deficit Western Europe, Japan, and thecentrally planned economies, as well as parts of devel-oping Africa and Asia. Variations in supply will be-come more important as the world’s productive capa-city is used at higher levels, particularly if there is no in-crease in world grain reserves. This suggests, according

lq~ G&d 2~ Re@rt co che Pres&nr (Washington, D. C.: U.S. Council on

Enwronmental Quality and Department of State, 1980), vol. 2, p. 85.I] Ibid., vol. 2, p. 556.


Table B-4.–Grain and Total Food Production, Consumption, and Trade (Alternatives 1, II, Ill)

568.1515.7

+ 52.4

536.2455.9

+ 60.3

126.6- 118.1121.0 -116.6

127.3 118.2127. ? 114.6

730.0 683.3848.4- 610.8 687.6 590.2

+ 93.1

157.0- 143.7155.8- 147.7

157.1 143.5165.7 143.6

297.5229.5

+ 68.0

309.7194.4

+ 115.3

137.8 -134.9119.6 -114.0

135.1 140.2129.2 111.2

416.0- 402.0 409.8 414.0290.2- 272.4 +325.0 +256.8

+ 126.0- + 129.6 + 84.6 + 157.2

184.3 -178.5160.3 -151.3

181.8 183.5178.3 143.2

589.5597.5– 6.0

485.3529.7

-44.4

534.0578.5

– 44.5

138.2143.3

174.0179.9

179.5 166.1179.2 173.2

470.5506.3

-35.8

154.4 -161.4163.4 -162.8

244.5- 247.7247.8 -242.6

266.2 246.4261.2 249.0

163.6 168.4169.2 159.2

279.5-284.4269.7 -265.3

298.4 288.1275.4 257.0

104.3103.7

+ .6

107,697.2

+ 10.4

158.7 -174.8162.7 -160.7

57.360.9

-23.6

48.651.5

– 2.9

53.079.9

-26.9

146.3- 148.1167,4 -165.1

149.6 136.9168.1 165.8

92.2- 89.0 94.5 88.1127.5- 123.7 125.5 132.5

-35.3- -29.7 -31.0 -44.4

252.5 -257.8276.1 -267.3

259.3 240.4271.4 267.7

45.548.5

– 3.0

150.7 -160.2161.2- 160.0

155.6 145.5160.0 150.5

61.3- 63.7 63.1 61.563.3- 63.0 60.7 62.0

- 2 . 0 - + . 7 + 2.4 – .5

197.1 -204.9196.4 -196.4

203.0 197.8189.1 193.2

190.0200.0

-10.0

178.6186.3– 7.7

154.0-155.5158.7- 158.2

158.9 149.3159.0 148,0

265.0- 259.0 269.0 271.0284.3- 275.7 290.7 293.9

-19.3- -16.7 -21.7 -22.9

221,6-216.8226.2 -219.4

225.2 226.9231.3 233.9

38.629.9

+ 8.7

39.630.7

+ 8.9

43.261.3

– 18.1

179.1 -194.3168.0- 168.0

180.6 185.5164.5 169.1

62.0- 65.0 62.6 64.147.9- 47.0 48.0 49.9

+ 14.1- + 18.0 + 16.6 + 14.2

71.9- 73.0 72.4 72.298.0- 97.0 95.0 100.5

-26.1- -24.0 -22.6 -26.3

295.3 -310.0268.8 -263.6

298.3 305.6257.8 280.4

46.563.7

-17.2

155.8- 154.7173.1 -172.3

161.4 149.7172.9 166.2

251.4 -255.3267.7 -264.9

253.1 252.4259.4 274.6

1,642.91,642.9

141.5 -140.5141.5 -140.5

144.5 137.0144.5 137.0

2,196,7 -2,141.7 2,233.0 2,119.62,196 .7-2,141.7 2,233.0 2,119.6

194.0 -191.0194.0 -191.0

198.0 191.5198.0 191.5

1,540.71,540.7

NOTE: In trade figures, + indicates export; minus sign indicates import.

SOURCE: The Global 2000 Report to the President, vol. 2, pp. 91-92.

to the report, that “[the] agricultural and trade policiesof a small number of importers and exporters will playan increasingly dominant role in determining the quan-tities and prices of food traded on the world market.”The United States is projected to play an increasinglydominant role in balancing world supply and demandby expanding or contracting production in order tomoderate price fluctuations. *4

The Global 2000 environmental projections relatedto agriculture suggest that food production could fallsignificantly below the quantities projected by GOLdue to soil deterioration, pest- or pathogen-controlproblems, water shortages caused by deforestation, andunstable supplies of energy-related inputs. U.S. and for-eign government policies will play a large role indeciding whether or not environmental problems seri-

14[t)l~, , \X>l. 2, pp. 77-W.

ously erode global agricultural production potential.These concerns will probably seem more important tothe industrialized nations than to the LDCs, which aremore likely to face the pressing problem of expandingproduction to meet rapidly expanding food needs, oftenregardless of long-term environmental costs. Model-comparison exercises similarly suggest that if GOL weremore integrated—i.e., if it gave greater attention to theinteractions between agriculture and other economicsectors or environmental conditions—its results wouldprobably be less optimistic:

The rising food prices and regional food shortages pro-jected in the agricultural model would be intensified bythe fact that agriculture is not the only sector wantingcapital to cope with increasing population demands anddiminishing returns. Land degradation caused by intensepressure on the land and by pollution would tend to

151bd. , voi. ~ , p, 89.


Table B-5.–Per Capita Grain and Total Food Production, Consumption and Trade (Alternatives 1, II, Ill)As projected by the GOL Model for Global 2000

1985 2000

Grain Food Grain Food(kilograms) (1989-71 = 100) (kilograms) (1989-71 = 100)

I II Ill I II

115.2115.2

Ill

105.0102.1

I

838.5- 769.8735.0- 692.4

+ 103.5- + 77.4

II Ill

847.5 716.9798.3 619.2

+ 49.2 + 97.7

1,719.1 1,479.51,383.3 917.7+ 355.8 + 581.8

I II Ill

Industrializated countries:Production . . . . . . . . . . . . . . .Consumption. . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . .

United States:Production . . . . . . . . . . . . . . .Consumption. . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . . .

Centrally planned countries:Production. . . . . . . . . . . . . . .Consumption. . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . . .

Less developed countries:Production . . . . . . . . . . . . . . .Consumption. . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . .

Latin America:Production . . . . . . . . . . . . . . .Consumption. . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . . .

North Africa/Middle East:Production . . . . . . . . . . . . . . .Consumption. . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . . .

OtherAfrican LDCs:Production . . . . . . . . . . . . . . .Consumption. . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . . .

South Asia:Production . . . . . . . . . . . . . . .Consumption . . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . .

SoutheastAsia:Production . . . . . . . . . . . . . . .Consumption . . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . . .

East Asia:Production . . . . . . . . . . . . . . .Consumption . . . . . . . . . . . . .Trade . . . . . . . . . . . . . . . . . . .

World:Production . . . . . . . . . . . . . . .Consumption. . . . . . . . . . . . .

716.9- 883.8613.7- 587.3

+105.1- +76.5

719.2 889.7858.9 569.4

+62.3 +100.3

112.9-104.5108.8-104.9

128.8-118.4127.7-121.2

131.8 108.8139.1 110.0

1,331.2- 1,301.0923.5- 874.9

+407.7-+428.1

1,324.6 1,322.11,021.9 629.9+302.7 +492.2

124.8-122.2108.5-103.4

124.2118.9

124.098.7

1,697.4- 1,840.31,183.3- 1,111.5+514.1-+528.8

158.0-151.1135.9-128.3

157.8 137.4154.8 107.9

411.5432.5

-21.0

452.5 369.6458.5 400.4- 6 . 0 -30.8

116.7122.4

127.6125.0

107.2115.9

451.1473.9

–22.8

489.2 375.3495.1 398.5- 5 . 9 -21.2

210.2 176.6219.4 189.5- 9 . 2 -12.9

129.6135.6

135.6 112.8138.4 119.0

182.0- 189.4203.0- 201.6

-21.0- -12.2

190.4 178.3207.8 191.8

-17.4 -13.5

101.7-106.5107.7-106.7

108.7110.6

99.1101.8

195.6- 197.1210.2- 205.5

- 1 4 . 6 - – 8 . 4

109.5-110.8111.0-108.6

119.5 99.1116.7 99.9

247.7- 247.4244.0- 240.8+3.7- +33.6

201.8- 203.9289.4- 285.8

–87.6- –81.9

264.3 259.6262.8 234.5+1.5 +25.1

209.0 188.4295.1 264.1

-88.1 -95.6

108.2-118.9110.9-109.6

114.9118.7

113.0108.9

305.9- 311.4282.8- 278.1

+23.1- +33.3

218.3- 222.5301.8- 292.8

-83.6- -70.3

348.6 288.0308.0 243.8

+40.6 +42.2

239.9 188.6318.6 283.7

-78.7 -95.0

131.5-133.7127.1-125.1

147.6 123.6138.7 110.8

87.2- 88.3101.8-100.3

91.0104.2

80.199.7

95.9- 98.2105.9-102.2

107.4 80.3112.9 98.4

130.7- 138.7144.0- 142.9

– 1 3 . 3 - - 4 . 2

138.2 125.5144.4 133.7-8.1 - 6 . 3

98.1-104.3105.0-104.2

102.3105.3

94.097.3

109.0- 113.2112.5- 112.0- 3 . 6 - + 1 . 2

123.8 108.0119.1 108.8+4.7 –0.8

81.2- 84.581.3- 80.9

92.7 80.588.3 78.5

170.0- 171.7184.3- 183.7

-14.3- –12.0

177.1 160.7188.4 167.7–9.3 - 7 . 0

104.6-105.6107.8-107.4

108.9109.0

98.898.0

174.0- 170.0186.7- 181.0

–12.7- –11.0

176.1 152.1192.4 184.9

-14.3 -12.8

322.7 282.5237.1 219.9

+85.6 +62.6

107.0-104.6109.2-105.8

109.6 93.5112.5 98.4

273.6- 295.8217.9- 217.9

+55.7- +77.9

282.0 276.1218.5 215.6

+63.6 +62.5

116.3-128.4108.9-108.9

120.1109.2

118.4107.6

301.9- 316.5233.2- 228.5

+68.7- +87.5

129.2-135.9117.1-114.6

138.7 120.4119.2 110.0

139 .9- 138 .9198 .8- 197 .8

- 5 8 . 9 - - 5 8 . 9

148.4 131.9203.3 187.1

–54.9 -55.2

104.6-104.9116.2-115.6

111.2118.9

98.5109.2

161.1- 163.5219.5- 217.3

–58.5- –53.8

188.7 140.4221.3 195.5

-52.6 -55.1

121.1-122.6128.7-127.3

126.9 105.0129.7 114.2

337 .7- 332 .6337 .7- 332 .6

354.4 315.4354.4 315.4

109.5- 108.5109.5-108.5

114.0114.0

103.0103.0

352.2- 343.2352.0- 343.2

373.0 302.0373.0 302.0

117.0- 114.5117.0 -114.5

126.0 104.0128.0 104.0

NOTE: In trade figures, + indicates export; minus sign Indicates import.

SOURCE: The Global 2000 Report to the President, VOl

make the projection of agriculturalgloomy. 16

Other Agricultural Projections

One of the most disaggregated and

2, pp. 93-94.

o u t p u t more global food situation in terms of its underlying causalfactors; and 2) to evaluate policy measures, particularlyinternational ones, that might redirect future develop-ments towards improvements in the world food situa-tion. The latter purpose focuses on the growth of fooddeficits in poor countries and the effect of agriculturaland trade policies in the rich countries on the develop-ment of food production and consumption in the ThirdWorld.

Structurally, MOIRA is similar to GOL in that bothof them address only the agricultural sector, leavingcritical factors such as population, GNP, and energy in-puts as exogenous variables. Both of them rely heavilyon traditional economic theories and techniques to rep-resent a world food system controlled by market supply,demand, and prices. They differ, however, in what theydisaggregate: whereas GOL disaggregates crops but con-siders only one class of consumers in each of its 28

mathematicallyelegant models of the world food system is the Model ofInternational Relations in Agriculture (MOIRA), com-missioned by the Club of Rome in 1973 and carried outby Dutch economists and agronomists with the supportof the Government of the Netherlands. 17 It was origi-nally constructed to investigate the consequences forthe food system of a doubling of the world’s population,and it is explicitly concerned with the problem of worldhunger. Its specific purposes are: 1) to describe the

lblbld., vol. 2, p. 672.17H. Linnemann et al., MO[~—MO&/ oj Internuttord Rekmons m Agncukure

(Amsterdam: North-Holland Pubhshlng Co., 1979).


Figure B.9.–World Grain Yields, Actual 1960-1976 and Projections to 2000 Under Alternatives 1, Ii, and Ill ofGOL for Global 2000

3.25

1960 1965 Actual 1970 1975

SOURCE: The Global 2000 Report to the President, vol. 2, p. 76

regions, MOIRA considers only one agricultural out-put—consumable protein—but disaggregates consumersinto 12 different income-and~ccupation classes in eachof 106 individual nations, which are linked through anequilibrium model of international food trade. Thisstructure allows the model to simulate conflicts of in-terest between producing and consuming nations orbetween agricultural and nonagricultural sectors withinnations, which in some ways makes it a better tool forfood policy analysis.

The model is solved by yearly increments over a50-year simulation (1960-2010), with the world marketassumed to clear each year and the world market price(along with factor costs and technical considerations)assumed to affect the next year’s production decisions.Each nation’s producers are assumed to operate in sucha way as to maximize expected sectoral—not individ-ual—income. National markets buffer themselves frominternational markets by tariffs or price subsidies thataffect the motivations of producers, but a “seepage” ef-fect tends to drive domestic prices toward world prices.

1985 Projected 2000

Time

In its normal mode of operation, MOIRA’s structureshows demand being steadily increased by populationgrowth and income growth, which leads to higherprices, which in turn stimulate additional supply. Dueto the costs of expanding production, however, supplyincreases more slowly than demand, and low-incomeconsumers who cannot purchase food at the going mar-ket price “demand” less food than they actually need toavoid malnutrition. In short, only high prices can in-crease production, but high food prices relative to con-sumers’ incomes will result in people going hungry. Thisoutcome reveals how, in one critic’s view, “this model

structure emphasizes the fact that, in today’s world—and in the foreseeable future-it is poverty, muchmore than supply constraints, that is the cause of worldhunger.” 18

18D, H, Mea&u,~, J. R1chard~on, and G, Bruckmann (eds. ) , G r o p i n g [n [h

Dark The FWS[ Decade o/ Global Nfodellng (?Qeu York: Wiley, forthcoming), pp.113-115.


Figure B-10.- Projected Regional Per Capita Food Consumption for South Asiaand Sub-Saharan Africa in 1985 and 2000 Under Alternatives i, ii, and iii of GOL

for Global 2000(Index 1969-71 = 100)

120

110

100

90

80

1970 1985 2000

Table B-6.–Projected Net Exporters of Wheat Under Alternative I-A (Medium Growth, Rising Energy Prices)of GOL for Global 2000a

(Million metric tons) Average annual1970 1985 2000 growth percent

Exports Percent Exports Percent Exports Percent 1970-85 1985-2000share share share

United States . . . . . . . . . . . . . . . 17,881 39 48,838 58 58,228 57Australia-New Zealand . . . . . . . 8,300 18 12,165 15 16,084 16 3 2Argentina, . . . . . . . . . . . . . . . . . . 1,640 4 6,410 8 13,974 14 10 5Canada . . . . . . . . . . . . . . . . . . . . 11,750 26 15,288 18 7,311 7 2 – 1South Africa . . . . . . . . . . . . . . . . 60 – 839 1 4,108 4 19 11U.S.S.R. . . . . . . . . . . . . . . . . . . . . 4,799 11 127 — 1,995 2 – 2 2 20India . . . . . . . . . . . . . . . . . . . . . . . — — — — 186 –Euro Six . . . . . . . . . . . . . . . . . . . .

— —1,170 3 — — — — — —

Total. ., . . . . . . . . . . . . . . . . . . 45,600 101b 83,887 100 101,864 100 4 1


The standard run of MOIRA, based on a continua-tion of present trends, assumes moderate populationgrowth, “rela t ively high” growth in nonagriculturalGNP, and no new policy interventions. The results ofthis scenario shows steady increases in agricultural pro-duction (fig. B-1 1) and per capita protein consumption(fig. B-12), along with higher but unstable world marketprices (fig. B-13). Nevertheless, there is also a large in-crease in the number of people below the minimumfood standard (fig. B-14). Nutritional gains are greaterin the developed regions than in the LDCs, and signifi-cantly lower in South Asia than in the LDCs as awhole. The LDCs show a decrease in food self-sufficien-cy, and North America becomes even more dominantas the leading food exporter. Sensitivity tests conductedto assess the effect of exogenous variables on these re-sults produced the following results:

●

●

lower growth rates in nonagricultural GNP (3.5rather than 7 percent in the LDCs) reduces de-mand, prices, and output, resulting in a 35-percentincrease in world hunger;lower population growth rates (about half the rateof the standard run) also results in lower prices and

output, but does produce a 30-percent reduction inworld hunger; andinternal income redistribution outside each coun-try’s agricultural sector (gradually reducing presentinequities by half their magnitude over the 1975-2010 period) leads to significantly higher effectivedemand and to price increases 50 percent greaterthan in the standard run, and, although it in-creases the food imports of the LDCs, it also re-duces world hunger by about 50 percent.

The authors conclude from this last test that “thehunger problem is, to a large extent, a problem of in-come distribution, ” but they are quick to point out thelimitations of their model. The general trends it projectsare more significant than its precise numerical results,but these sensitivity tests all point to a similar outcome:

. , . all simulation runs with alternative assumptions re-garding exogenous variables have one thing in common:if policies remain unchanged, the number of people whocannot obtain sufficient food will increase. 19

As a consequence, the authors have used MOIRA ex-tensivel y for policy testing, with the objective of discov-

Figure B-n .—Projections of Population and Food Production in MOIRA, Standard

300

260 -

220

180

140 -

100 ~ . .

1 I I I I i’ ! t.

I 1 1’ 1 i’ I’75 ’80 ’85 ’90 ’95 2000 ’05 ‘lo ’75 ’80 ’85 ’90 ’95 2000 ’05 ‘lo

’75 ’80 ’85 ’90 ’95 2000 ’05 ‘'10

— P o p u l a t i o n--- Food production

Run

NOTE: In all regions production increases faster than population, although in Southern Asia gains are quite modest,

SOURCE: Extracted from MOiRA


Figure B-12.– Projection of Annual Per CapitaProtein Consumption From MOIRA, Standard Run

’75 ’60 ’85 ’90 ’95 2000 ’05 ‘1O

Developedcountries

Centrallyplannedcountries

Developingcountries

SouthernAsia

NOTE: Greatest gains occur in the developed countries. There is a steady gainin developing regions, but Southern Asia shows very little gain.

SOURCE: Extracted from MOIRA, p. 29.

Figure B-13.— Food Prices on World Market inMOIRA, Standard Run

+ 600

+ 400

+ 200+ 100

0-100

’75 ’80 ' 8 5 ' 9 0 ' 9 5 2 0 0 0 2 0 0 5 2 0 1 0


ering what actions might be taken by the rich nationsto contribute effectively to the goal of reducing worldhunger. The four basic policy tests involve two meas-ures intended to achieve a redistribution of availablefood and two measures intended to stimulate food pro-duction in developing regions:

Reduction of foodconsumption in the richcountries, which might be achieved by shiftingconsumption patterns (i.e., fewer animal products),results in a low world market price that weakensthe incentives to production and thereby leads toincreases in total world hunger.Food aid, if it is purchased by the rich countries atthe prevailing world market price and distributedto those under the food norm in such a way that itdoes not disturb local markets, is capable of elimi-nating world hunger almost completely. This opti-mistic scenario, which requires the developed na-tions to devote about 0.5 percent of their GNP to

Figure B-14.– Total World Hunger (WHUNG) aHunger in the Agricultural Sector (AWHUNG)

MOIRA, Standard Run

cons10” Kg Million

umable protein people

ndin

’75 ’80 ’85 ’ 9 0 ’ 9 5 2 0 0 0 2 0 0 5 2 0 1 0

NOTE: Right axis shows number of people with a consumption level below theminimum food standard. Left shows total food deficit in 10 Kg con-sumable protein.


food aid, differs from the first in that it raises worldmarket prices, particularly between 1980 and 1990,and thereby stimulates additional production.Regulating international food trade, in order tostabilize world market prices at a relatively high lev-el, also proves to be an effective way to stimulateproduction and improve the food situation in theLDCs, particularly under the assumption of loweconomic growth outside the agricultural sector.Such a policy might be difficult to implement, how-ever, because it would require the rich nations tocreate buffer stock and to regulate their importsand exports in order to keep prices at desired levels;under some conditions this might require NorthAmerica to give up its leading position as a food ex-porter.Liberalizing international food trade, whichwould require rich countries to cease protectin g

their domestic markets from world food prices,causes lower food prices and (over the long term)considerably more hunger. Because this policylowers’ production in the LDCs, it also increasesthe developed regions’ share of food exports, withthe greatest increases coming from North America;in short, it benefits the rich at the expense of thoseless fortunate.

The polic y package the authors find most effectivecombines price stabilization and food aid. They con-clude that “[as] long as the rich nations are able andwilling to provide the funds” there is little potential forconflict between these apparentl y contradictor y objec-tives. However, they also found that deliberate changesin income distribution in the poorest countries would


“remarkably strengthen the positive effect of food aidand world market regulation. ”20 This “global food sup-ply policy” would nevertheless require a concerted ef-fort on the part of the rich nations to adapt their do-mestic production to the international demand andsupply situation.

Strengths and Weaknesses ofProjection Techniques

Anyone putting an agricultural sector into a globalmodel must arrive at formulations for the determinantsof agricultural production (supply) and consumption(demand). What he includes and how he includes it de-termines the results his model will produce, althoughhis biases may strongly influence the way he interpretsthese results. For example, all of the models exceptUNIOWM indicate there will be problems supplyingfood to at least some of the world’s people over the next20 years. The reason for this finding is fairly straightfor-ward: all of the models except UNIOWM show increas-ing costs for land development as the amount of unde-veloped land decreases, and all except UNIOWM showdiminishing returns to agricultural inputs. In short, themodels show agricultural problems because they in-clude agricultural limits.

However, their other major similarity–advocacy ofdrastic social and political change—appears to be some-thing modelers interject into their models rather thansomething they learn directly from the models. Theconclusions drawn from global models often suggestthat the only way to avert major catastrophe is to altersome difficult-to-change trend such as fertility, invest-ment rate, or income distribution; but the modelsthemselves are insufficiently detailed to tell what thesechanges would mean in practice.

Table B-7 compares how the six models treat some ofthe basic factors influencing agricultural systems, withfactors affecting demand on the left, factors affectingsupply on the right, and factors affecting both supplyand demand in the center. It shows that models aremore alike in what they leave out (make exogenous)than in what they include, but even when a variable isendogenous there are differences in how models treat it.All of these differences affect results in some way, al-though it is sometimes difficult to establish how. For ex-ample:

• Differences in aggregation cause large differences inthe form of model output, but they frequently haveno effect on model behavior other than making re-sults more or less detailed, or more or less difficultto interpret. Experimentation with different levels

201 bLd., p. 329.

of aggregation has led many modelers to prefermore aggregated structures, which give nearly iden-tical results at a much lower cost. On the otherhand, sometimes disaggregation does matter: if onewere to aggregate grains and livestock in GOL,UNIOWM, or any other model that differentiatesbetween agricultural products, one would probablyobserve a change in the way the model responds tohigh prices and/or poor harvests.Structural differences that have no effect under oneset of circumstances may be all-important underanother. For example, the soil degradation mecha-nism in World 3 has almost no effect on themodel’s standard run, but when nonrenewable re-source constraints are removed the resulting pollu-tion causes agricultural yields to plummet, and thesystem must begin devoting a large part of its in-dustrial output to rescuing the agricultural sector.Seemingly different structures can behave similarly,while seemingly similar structures behave very dif-ferently. For example, World 3, UNIOWM, andLAWM all lack price mechanisms; however, boththe LAWM and World 3 will cause investment toflow in the direction of a sector that shows signs ofsupply shortfalls, while UNIOWM contains nosuch mechanism.Important structural features may be buried in ac-counting matrices. For example, a zero entry in aninput/output matrix, or an assumption of non-sub-stitutability between two classes of agriculturalcommodities, could greatly affect intersectoralflows in any of the multicommodity models (GOL,UNIOWM, and WIM).

In short, the way a model’s structures affects itsresults can be complicated, so much so that modelersthemselves are often at a loss to explain system behav-ior. There are, however, some situations in which mod-el behavior is easily traced to model structure, and thereare many other situations where this influence is at leasta plausible explanation. LAWM’s pronounced tenden-cy to rapid urbanization, for instance, results from thefact that its optimization routines are based on statisti-cal research that showed a high correlation between liv-ing in cities and life expectancy; LAWM’s behaviorwith regard to housing and education can be explainedin the same way. Similarly, the relatively pessimisticprojections of GOL and World 3 with respect to foodper capita certainly result in part from the fact thatneither model includes labor in its agricultural produc-tion functions; the relatively optimistic findings ofLAWM and MOIRA, on the other hand, probably owesomething to their inclusion of labor.

MOIRA assumes that farmers produce the quantityof outputs that will maximize their profits at a givenprice level for agricultural inputs, which explains why


Table B-7.—Comparison of Structural Assumptions in Six Global Models

Demand Supply and demand Supply

Income IntersectoralModel Population growth interaction Prices Trade Constraints Technology Environment, J . , L. — ——.—. Agriculture, indus-

try, compete for in-vestment. If onesector fails itbrings othersdown.

Multisectoral,Input-output withdetail on energyand machinery,energy shortagereduces produc-tion, agriculturaldevelopment needsindustrial growth.

40 sector input-output, no intersec.toral competition,just accounting ofintersectoral flows.

Omitted, im- Omitted.plicit in linksbetween supply

Diminishing returnsfor land investmentand. agricultural in-puts, depletablemineral resources.

Diminishing returnson land investmentand agricultural in-puts, depletableenergy resources.

, worm

WIM

Births and endogenous,deaths driven byendogenous. supply of4 cohorts, 1 capital andregion. inputs.

embodied in capitalstock, properties may bechanged by policy, noautomatic progress indisembodied technology

soil delerlor-ation and pol-lution affectyields.and demand.

Births and Endogenous,deaths semi- driven byendogenous, supply of10+ capital andregions, 85 inputs.cohorts.

Explicit,endogenous,by sector.

Food and other im-ports constrained bybalance of payments.

Embodied in capitalstocks, investmentshifted by price and pro-fit criteria, no automaticprogress in disembodiedtechnology.

Omitted.

UNIOWM

MOIRA

Exogenous, Exogenous,15 regions. fast enough

to meet UNtargets.

Exogenous, bysector for allsectors, basedon input costprojections.

All sectors, Importsubstitution exogen-ously determined.

None, linear returnsto inputs.

Exogenous updates ofinput-output coefficientschanges production effi-ciencies, i.e., automaticdisembodied technologi-cal progress.

Emissions ofpollutantscalculated foreach sector,no feedbackto yields.

Exogenous, Exogenous.106 nations,6 rural and5 urbanincomegroups,urban-ruralmigrationendogenous.

Nonagriculturalgrowth affects in-come, henceagricultural de-mand. Relative in-come levels inagriculture andnonagriculture af-fect mlgratlon.

5 sectors competefor capital andlabor; allocationbased on sectoralcontribution tobasic needs fulfill-ment. Agricultureneeds capital sec-tor inputs.

Sensitive to energyprices, little otherinteraction; withinagriculture, highlevel of interactionbetween grain andlivestock.

Domestic foodprices endo-genous, policycontrolled,world foodprice balancessupply anddemand.

Food only, detailedrepresentation oftrade policies, assure.ing policy tries tokeep agricultural in-comes in line withnonagriculturalincomes.

Diminishing returnson land Investmentand agriculturalinputs.

Embodied shifts throughproducers investing tomaximize profits. Disem-bodied progressautomatic.

Omitted (tobe includedin MOIRA2)

LAWM Births and Endogenous,deaths en- driven bydogenous, 4 supply ofregions, capital andmaximizes labor.life expec-tancy.

Omitted, Unimportant, mostlyomitted.

Diminishing returnson land investmentand yields.

Disembodied progressautomatic, in agriculture

Omitted.

1 percent per year gainin efficiency. I

G2000 Exogenous, Exogenous.28 regions.

Agriculture, Multicommodity trade Apparently linearreturns on inputs,no limits.

Disembodied progressautomatic, followingtrends observed in therecent past.

Omitted inGOL, ana-lyzed verbally.

endogenous by in agricultural prod-commodity; ucts, some policyenergy, representation par-exogenous. ticularly for United

States.


measures that drive up agricultural prices are successful Factors Affecting the Reliabil. . ity andas a means of reducing hunger. However, profit maxi-mization may be a better approximation of the behav-ior of rich farmers than of poor farmers (who may notbe able to borrow funds for expanding production, andwho may resist giving up their traditional agriculturalpractices), and for this reason MOIRA probably overes-timates the response to price incentives in the develop-ing countries, thereby underestimating hunger in itsrapid-economic-growth scenarios. Finally, as discussedbelow, the structural decision to make important partsof the system exogenous often results in a model whosereliabilit y and accuracy are heavil y dependent on theprojections used to drive the model.

Accuracy of Agricultural Projections

Model Structure and Assumptions

Most global modelers enjoin readers against takingtheir numerical forecasts as precise estimates, Nonethe-less, comparison of numerical results, in conjunctionwith comparison of structures and assumptions, is auseful means of determining why various global modelsmake the projections they do. Table B-8 compares theprojections made by six models for key agricultural vari-ables in 2000, including food production, prices, expan-sion of cultivated land, and yields. To eliminate differ-

App. B—Agricultural Projections Ž 101

Table B-8.—Comparison of Projected Values of Critical Agricultural Variables in Different Global Models

Food Price Globalproduction increase Cultivated land Yield food/capita Regional food/capital

World 3 280 Not relevant 130 120 109 Not included

UNIOWM 285 114 Region Arable Land 158 Calories Proteingrain Developed market: land productivity grain South Asia . . . . . . . . . . . . . 120 127274 North America . . . . . . . . . . . . 111 194 155 Africa (tropical) . . . . . . .

livestock127 140

Western Europe (high-income) . . . 100 162 animal Latin America (low income) ., 125 143Japan. . . . . . . . . . . . . . . . . ., . . 100 269 products Latin America (high income) 132 148Oceania. . . . . . . . . . . . . . . . 183 162

Centrally planned:Soviet Union ... . . . . . . . . . . . . 100 215Eastern Europe, . . . . . 100 186Asia (centrally planned). . . . . . . 120 278

Developing market:Latin America (medium-income) 166 311Latin America (low-income) . . . . . . 140 328Middle East. . . . . . . . . . . . . . . . . 126 487Asia (low-income), . . . . . . . . . 113 331Africa (arid) ... . . . . 131 282Africa (tropical). . . . . . . . . . . . . . 152 324

MOIRA 244 522 Not documented Not documented 140 High Lowhigh growth high growth high growth South Asia ., . . . . . . . . . . . 125 75

185 113 108 Africa (tropical) . . . . .low growth

164 100low growth low growth Latin America ... . . . . . . . . . 168 117

Global 191 to 201 130 104 160 126 High Low2000 grain optimistic to o p t i m i s t i c S o u t h A s i a 113 96

191 to 198 to 190 104 Other African LDCs. ... . . . . . 86 78food 215 pessimistic Latin America, . . . . . . . . . . 137 111

pessimistic

LAWM Not documented Africa . . . . . . . . . . . . . . . . . - 1 3 0Asia, . . . . . , . . . . . . . . . . . - 1 5 0Latin America . . . . . . . . . - 1 2 0

WIM Not documented South Asia . . . . . . . . . . . . . . 68

NOTE: Scaled such that 1970 = 100. (For illustrative purposes only, Some of these numbers have been read off of imprecise plotted output and may err by as much as 5or 10 percent).


ences in measurement, values for 2000 have been in-dexed to their values in 1970. Generally, world aggre-gates are used because the models use very different re-gional aggregations. It has not been possible to includevalues for all models, because not all models calculateall variables and model documentation often fails topresent needed data. Indeed, lack of information hasmade it impossible to include either the WIM or theLAWM in most calculations.

Projected food production in 2000 ranges from 185 to285 percent of 1970 figures, an increase equivalent toaverage annual compound growth rates varying from2.0 to 3.5 percent. UNIOWM and World 3 give thehighest figures; the MOIRA low-growth scenario andthe Global 2000/GOL projections give lowest. Priceprojections vary far more than supply projections: theMOIRA’s high-growth scenario, at one extreme, showsa price increase of over 400 percent (most of which, inci-dentally, occurs before 1985); on the low side, theMOIRA low-growth scenario and UNIOWM projectprice increases of below 15 percent. Columns 3 and 4show how much of the increase in output is attributableto expanded cultivation and how much to increasedyields; the figures show that World 3 anticipates landexpansion, GOL yield increases, and UNIOWM both.

Finally, columns 5 and 6 show global and selected re-gional figures for per capita food intake. From the dif-ferences between columns 5 and 1, one can infer thatglobal population growth is faster in World 3 than inother models and slower in Global 2000. The regionalfigures, however, show that there is greater variation inregional projections than global projections, with themost severe problems in South Asia and Tropical Afri-ca. These figures also hint that WIM may be consider-ably more pessimistic than other global models. Itshould also be noted, however, that model results arehighly dependent on time horizon. If one truncates theWorld 3 standard run at 2000 its projections look opti-mistic, and similar results arise from truncating theLAWM standard run for Asia at 2000.

Population, economic growth, and disembodied tech-nological progress are exogenous variables in at leastthree of the six models (see table B-7). All of these vari-ables have important influence on model results, as dotheir assumptions about the amount of potentially ara-ble land, the cost of developing that land, and the costof other inputs such as fertilizer, irrigation, and farmmachinery. Vital though these factors are to accuracy

and reliability, however, there is little agreement among

the models on what values they should be assigned. In


some cases this disagreement cannot be resolved be-cause of a lack of statistical evidence or theoretical un-derstanding.

Disembodied Technological Progress

It has become conventional for economists to identifythat fraction of income growth (or input cost reduction)that cannot be statist icall y ascribed to increases inlabor, capital, or other inputs as “disembodied techno-logical progress,” and to make projections of economicgrowth under the assumption that rates of technolog-ical progress observed in the recent past will continue inthe future. This amounts to an assumption that pro-ductivity will increase automatically. In most such for-mulations, the productive contribution of technologicalprogress is very significant: rates above 1 percent peryear are common, and it is not uncommon for im-proved technology to appear to contribute more to eco-

nomic growth than either labor or capital.Formulations of this type have been employed in the

agricultural sectors of LAWM, UNIOWM, GOL, andMOIRA, but not in World 3 or in some versions o fWIM. The specific rates of technological progress as-sumed in various global models are generall y not re-ported in model documentation, however: LAWM as-sumes a 1.O-percent annual growth in agricultural pro-ductivity, 1.5 percent for capital goods, and rates of ei-ther 0.5 or 1.0 percent for other sectors; but analogousfigures for UNIOWM, GOL, and MOIRA are notavailable, Nevertheless, testing shows that model re-sults are quite sensitive to both the presence and therates of disembodied technological progress. Whentechnological progress is omitted from LAWM, for ex-ample, Africa joins Asia in being forced against its landconstraints and facing economic collapse. On the otherhand, if sufficiently strong assumptions about techno-logical progress are introduced into World 3, the over-shoot-and-collapse mode can be eliminated from themodel altogether. z’ It should be added, however, thatnobody has a very good understanding of technologicalchange; its exclusion or inclusion in global modelstherefore stems ultimatel y from hunches, beliefs, andvalues—not from scientific understanding.

Population

Population growth is exogenous in UNIOWM, GOL,and MOIRA. UNIOWM assumes somewhat higherpopulation growth rates than the two other models, soits optimistic projections of food per capita cannot beascribed to low population growth rates. MOIRA, onthe other hand, assumes rapidly declining rates of popu-

lation growth, especially for countries whose 1970growth rates were high, and this may contribute toMOIRA’S tendency to become more stable in the lastdecades of simulation than it is in the first decades.GOL assumes significantly lower population growthrates in the industrialized countries than those pro-jected by the Census Bureau (see app. A). Better linkagein Global 2000 might thus have resulted in less optimis-tic food trade projections.

Economic Growth

The right side of table B-7 shows the rates of eco-nomic growth projected as exogenous drives for theUNIOWM, GOL, and MOIRA. In all three modelsthese projections are used (in combination with thedemographic projections) to project income per capita,which in turn is used in projecting demand for agricul-tural products. The standard runs of UNIOWM andMOIRA assume rapid economic growth in the develop-ing countries (7.2 percent and 7.6 percent per year,respectively). This leads to rapid increases in agricul-tural demand, due to the fact that demand for food inpoor countries is quite income elastic. In UNIOWM, in-creased demand causes increased production directly,while in MOIRA it causes price increases, which in turnstimulate increased production. When lower rates ofeconomic growth are assumed in MOIRA, prices andproduction stay low and the food situation improvesless rapidly (it even deteriorates, in South Asia). TheGlobal 2000 income growth projections are generally

lower, particularly for non-OPEC developing countries,which undoubtedly contributes to the fact that GOLprojections of food availability in the LDCs are muchmore pessimistic than analagous projections in theUNIOWM and MOIRA,

Potential Agricultural Land

MOIRA, World 3, and WIM all use formulations inwhich the costs of further increases in land develop-ment are dependent on the amount of land already un-der cultivation. A prodigious amount of work went intoestimating, these relationships for MOIRA, includingderivations of absolute maximum dry-matter produc-tion potential based on FAO soil maps and plant physi-ology models. 22 The figures used in World 3 derive fromthe President’s Science Advisory Committee (PSAC)report, The World Food Problem ( 1967), and are likewisebased on detailed studies of soil maps and climatologicaldata. The derivation of the figures used in WIM has notbeen documented.

21H. s. D. Cole, op. cit., p. ~. 22 LLnnemann, et al., 0p. cit., PP. 19”74

App. B—Agricultural Projections ● 103

A comparison of the values used in these three mod-els shows that differences between MOIRA and thePSAC/World 3 are relatively minor–PSAC givesSouth America somewhat more agricultural land,North America slightly less, and Europe quite a bit less,The WIM estimates, however, proved to be markedlylower for all regions except Western Europe, and WIM’sestimate of the world’s total potentially arable land areonly about 70 percent of the MOIRA figures and about76 percent of those used in World 3. WIM’s relatively

pessimistic assumptions about land availability, particu-larly in the developing regions, undoubtedly contrib-utes to its dire predictions of impending famine. How-ever, as shown in sensitivity tests of the World 3 agricul-ture sector, an increase in potentially available land of30 percent or more does not change the net outcome ofthe model—it merely postpones by a few years the dateat which scarcity becomes acute. 23

Other Factorsand Accuracy

Affecting Reliability

Global agricultural data are often poor, and all mod-els necessarily contain a lot of guesswork. For example,no one has very many or very good economic data onChina, nor are there reliable data on the soils (andhence the agricultural potential and costs of agriculturaldevelopment) in Amazonia. In addition, agriculture willundoubtedly be affected in the coming decades by sev-eral trends for which no global model accounts. Noneof these six models accounts for potential competitionbetween the agriculture and energy sectors for water orland. None takes the global monetary system into ac-count. None looks at the effects of climatic change orincreased levels of atmospheric carbon dioxide. Nordoes any of them examine the agricultural conse-quences of unusally bad weather or a serious destabiliza--tion of oil and gas supplies.

“J. Ran&rs and E. K. O. Zahn, “The Agricultural Sector,” In D>namlcs ofGrouth m a Flnitc W’O,M, D. Lfeadcrws, et al. (eds. ) (Camhndge, Mass.. Y1.I.T.Press, 19?4), pp. J39-N8.

APPENDIX C

Energy Projections

Introduction

The future availability and price of natural resourcesare crucial variables in long-range projections of worldeconomic development. Energy supply and demand inparticular affect not only industrial production but alsoagriculture, transportation, and general living condi-tions. As a result, the findings and conclusions of thesefive global modeling studies are significantly influencedb y their treatment of the global energy system and theirassumptions about future energy trends. Some of themodels do not address energy specifically, or in detail,and others merely assume that energy will not be a con-straint within their restricted time horizons.

In general, those models that include a finite resourcestock tend to show that depletion will raise prices, slowproduction, and dampen global economic growth; inshort, resource constraints make them susceptible toeconomic collapse. Even the most optimistic findings,however, suggest that the world faces a difficult transi-tion away from dependence on oil. Coal, nuclear, andsolar power are offered as the principal alternatives forthe future energy system. The accuracy and reliabilityof these projections are also influenced by their assump-tions about population and economic growth, potentialtechnological progress in extraction and conservation,the potential for substitution and alternative sources,and the future political and economic behavior of bothproducers and consumers.

Purposes, Structures, and Findings

World 3

Although The Limits to Growth is sometimes said tohave “predicted” the energy crisis of the 1970’s, theWorld 3 model itself does not specifically address futuretrends in energy supply and demand. The purpose ofthe model is to describe the world as a general systemrather than to predict its parts in detail, so energyresources—petroleum, natura l gas , and coal—arelumped together with 16 other raw materials (primarily

metals) in a single category called “nonrenewable re-sources. ” The authors cite U.S. Government estimatesshowing total reserves of individual resources rangingfrom 7 to 5,100 years, but they assume that,on average,there are about 250 years worth of nonrenewable re-

IThe followlng material IS based In part on the draft workin g paper , “TheRole of Energy in the Global System)” by Paul Werbos, Office of Energy Infor-mation Vahdarlon, Energy Information Admln[strat[on, U.S. Department ofEnergy.

sources at 1970 consumption levels. They also assume,however, that the quantity of resources consumed percapita is a fixed function of average income per capita,and that continued population and economic growthwill lead to a 4-percent growth rate in total worldresource consumption. Consequently, this 250-yearreserve would be completely used up by about 2040 ifgrowth continues unabated. In addition, the modelassumes that the cost of obtaining resources will risedramaticall y once 50 percent of the reserves have beendepleted, due to declining resource quality and increas-ing transportation costs.

The standard run of World 3 demonstrates the conse-quences of this combination of assumptions for the be-havior of the nonrenewable resource sector and for theentire world system (fig. C-1). Rising population, com-bined with rising industrial production per capita, re-sults in the rapid depletion of resources and an equally

rapid increase in the costs of obtaining the resources.As more and more capital is diverted from agriculturaland industrial production to the obtaining of raw mate-rials, per capita food and industrial output alike beginto decline. This leads eventually to mass starvation andthe collapse of the world economic system,

Sensitivity tests conducted by the authors of World 3indicate that the general behavior of the nonrenewableresource sector and of the integrated world system arenot particularly sensitive to their assumptions about re-source reserves, consumption rates, or extraction costs.

●

●

●

●

If reserves are set at twice their initial value, the col-lapse is delayed by only 15 years (fig. C-2).A - tenfold increase in initial reserves will eliminateresources as a constraint to growth—at least before2 100—but the general behavior of the world systemremains unchanged. In this case, persistent pollu-tion causes a decline in production followed bystarvation and a decline in population (fig. C-3).Improved extraction technologies, by reducing theshort-term cost of obtaining virgin materials,would eliminate the economic incentives for con-servation and substitution. These technologiescould delay the collapse a few years, but they wouldcause a faster depletion of resources and a sharpereventual decline in industrial output (fig. C-4).Improved conservation technologies, sufficient toreduce per capita consumption by a factor of four,allow much higher industrial output and postponesits decline for about 40 years, but they cannot pre-vent eventual collapse (fig. C-5).

The only model run that succeeds in postponing thecollapse beyond 2100 is based on the combination of

104

App. C—Energy Projections ● 105

Impact of Resource Depletion on the NonrenewableWorld 3 Standard Run

//

1950 2050 2100

Run 5-1: standard run for the nonrenewable resourcesector.

Resource Sector and Integrated Model,

II

0 0 00 0 0 1950 2000 2050 2100

Run 7-6A: World 3 reference run.

This is the World 3 reference run, to be compared with thesensitivity and policy tests that follow. Both populationPOP and industrial output per capita IOPC grow beyondsustainable levels and subsequently decline. The cause oftheir decline is traceable to the depletion of nonrenewableresources.

Figure C-2.— Behavior of the Nonrenewable Resource Sector and Integrated World 3 ModelWith Doubled Reserves

//

m

.

1900

Run 5-2:

1950 2000 2050

behavior of the sector with doublevalue of nonrenewable resources.

the initial

.1900 1950 2000 2050 21C

Run 7-7: sensitivity of the initial value of nonrenewableresources to a doubling of NRI.

To test the sensitivity of the reference run to an error in theestimate of initial nonrenewable resources, NRI is doubled.As a result, industrialization continues for an additional 15years until growth is again halted by the effects of resourcedepletion.


Figure C-3.—Behavior of the World 3 Model When Initial Resource Reserves

0 0 0 00 0 0 0

increased Tenfold

1900 1950 2000 2050 2100

Are

Run 7-8: sensitivity of the initial value of nonrenewable resources to a tenfold increase inNRI.

The initial value of nonrenewable resources NRI is increased by a factor of 10, to a value welloutside its most likely range. Under this optimistic assumption, the effects of nonrenewableresource depletion are no longer a constraint to growth. Note that there is no dynamic differ-ence in this run between setting resources at 10 times their reference value or assuming an in-finite value of resources. However, population and capital continue to grow until constrainedby the rising level of pollution.

SOURCE: Dynamics of Growth in a Finite World.

cost-reducing and resource-conservin g t echnologieswith zero population growth after 1975 (fig. C-6). Eventhis run, however, shows the beginning of the charac-teristic rise in the amount of capital that must be allo-cated to obtain resources—the collapse of the economicsystem, although delayed beyond the model’s time hori-zon, will undoubtedly still occur in the 22d century. Notest was performed to measure the model’s sensitivity tothe assumption that industrial output per capita willcontinue to increase exponentially as it has in the past.


WIM is both more detailed and more flexible in itstreatment of energy resources than World 3, and it hasbeen used extensively for testing alternative energy fu-tures. WIM specifically represents both the supply andthe demand for five different energy sources—oil, gas,coal, hydroelectric, and nuclear—within each of its 10

or more geographical regions.2 The model also repre-sents energy trade between regions, a significant advan-tage in modeling a world system where some 110 na-tions import over two-thirds of their energy needs and90 percent of all oil supplies move through the interna-tional trade network. J

WIM’s structural equations reflect the assumptionthat rising oil prices will lead to both conservation andthe development of alternative sources of energy. In-vestment in energy development is partially controlledby price, which in turn is determined by supply and de-mand. The model also assumes that oil will cover a de-clining portion of total world energy demand after

‘The fifth category, “nuclear energy, ” also includes solar energ y and othercapital-intensive alternatives. Source: Command and Control Technical Cen-ter (CCTC), Wodd Integrated Model Muhtkvel Hterarchumi Theoretic Concepts,

CCTC Technical Memorandum TM 197-79 (Washington, D. C.: U.S. DefenseCommunications Agency, June 15, 1979), pp. 2-17.

‘M. D. Mesarovic and E. Pestel, Mankind at the Tummg Potnt (New York: Dut-ton, 1974), p. 180.

App. C—Energy Projections . 107

Figure C-4.– lmpact of Cost-Reducing Extraction Technologies on theIntegrated World 3-Model

Run 5-3: the effects of cost. reducing technologies on thebehavior of the nonrenewable resource sector


Figure C-5.–impact of

0.

Oooa

II

II

0 0 0 00 0 0 0

.

\ 1.

1950 2000 2050 210(

Resource. Conserving Technologies on theIntegrated World 3 Model

Run 7-12: improved resource exploration and extractiontechnologies.

The implementation of improved resource exploration andextraction technologies in 1975 is modeled by lowering thecapital cost of obtaining resources for Industrial pro-duction. This policy allows industrial production to con-tinue growing for a few more years than in the referencerun, but it is ineffective in avoiding the effects of resourcedepletion.

1 .1950 2050 2100

the effects of resource-conserving technologieson the behavior of the nonrenewable resourcesector.

a!kax

x

n.

az

0 0 0 00 0 0 0

Resource Sector and

Resource Sector and

1950 2054 21OQ

Run 7-13: recycling technologies.

The advances in resource exploration and extractiontechnologies of Run 7-12 are supplemented by an improve-ment in recycling technologies that reduces per capitaresource usage by a factor of eight in 1975. That policyremoves the constraining effects of resource depletion andallows population and capital growth to continue untilchecked by persistent pollution



Figure C-6.—Combined Impacts of Zero Population Growth, Resource Conserva-tion, and Improved Extraction Technologies on the Behavior of the World 3

Nonrenewable Resource Sector

Kz

0 0 0 01900 1950 2000 2050 2100

Run 5-5: the effects of zero population growth and advanced technological policies on thebehavior of the nonrenewable resource sector.

SOURCE: Dynamics of Growth in a Finitefe World.

1990, and that substitution between energy sources willbe based largely on cross-price elasticities, Consequent-ly, although oil and oil substitutes are the most bindingresources in WIM, energy supply appears to be con-strained less by absolute scarcity than by the speed withwhich substitutes can be developed—energy is an eco-nomic and engineering problem, rather than a physicalone .4

For use as a policymaking tool, the model provides“scenario variables” or “policy levers” with which userscan test the consequences of nonmarket pricing behav-ior by producers or shifting patterns of substitution by

consumers. s In general, the model runs suggest that co-operation would benefit both producers and consumersand would also ease the transition away from oil, andthat continued long-term economic growth may bepossible under either of two alternatives: the rapid andwidespread deployment of breeder reactors; or the con-struction of vast “solar farms” and hydrogen plants inthe deserts of the Middle East.

4P. VanderWerf, “Energy, ” in The Global 2000 Report to the Pres/denr (Wash-ington, D. C.: Council on Enwronmental Qual[ty and Department of State,

1980), VOI, 2, p. 618.

5CCTC, Op. cit . , pp. 2-17.

The world oil crisis is a major focus of the many sce-narios reported in Mankind at the Turning Point. In thefirst pair of computer runs, the model was used to dem-onstrate the long-term economic benefits of an “opti-mal” oil pricing policy (fig. C-7). Continuation of lowoil prices would encourage overexploitation and rapiddepletion, discourage the development of substitutes,and lead to major dislocations in the developed regionswhen reserves are exhausted. Both exporters and im-porters would fare better under an “optimal price sce-nario, “ in which the real price of oil rises 3 percent an-nually until it reaches an optimal level (about 50 per-cent above the initial price of $13.50/barrel, as deter-mined in a separate analysis), at which level it stabilizesthereafter. Optimal oil pricing is assumed in a secondset of computer runs that indicates that both exportersand the developed regions benefit under scenarios inwhich the flow of oil from the Middle East is unim-peded and international energy trade is “governed sole-ly by the economic forces without undue interferencefrom the political level” (fig. C-8).6 All of these runs,however, project a transient world oil deficit between

‘Mesarowc and Pestel, op. cit., p. 112.


Figure C-7.—Comparison of Long-Term World Development for World integrated Model “Fixed Low Oil Price”Scenario and World integrated Model “Optimal Oil Price” Scenario

(A)

I Developed world(Regions 1-4) I

1 9 7 5 1 9 8 0 1 9 8 5 1 9 9 0 1995 2000 2005 2010 2015 20202025

Year

Cheap energy in the form of oil has been a prime fuel for theunprecedented growth of the world economy in the 1950’sand 1960’s. The dramatic increase in oil prices in 1973 wasviewed as a catastrophe. However, computer analysis ofour world system model indicates that the continuation ofwhat amounts to overexploitation of oil, spurred by anunreasonably low price, would lead to major dislocationsbecause of the exhaustion of reserves and the lack ofmotivation to develop substitutes in time. Pursuance ofshort term objectives would lead to major dislocations in


1997 and 2002 and a severe, persistent deficit beginningaround 2020; substitutes or alternative sources wouldbe necessary after that date.

One possible energy future, based on nuclear power,has been proposed by some technological optimists.Tests using the WIM “fast-nuclear scenario” raise ques-tions about the short-term feasibility and long-termconsequences of this alternative. After testing short-term scenarios based on Herman Kahn’s The Next TW O

Hundred Years, the authors conclude that:It is impossible to design any energy program in Wes-

tern Europe or Japan which could, over a ten-year period,reduce energy demand and increase production of energyfrom non-petroleum sources sufficiently to compensatefor the loss of the Persian Gulf by 1987. The Hudsonreport statement regarding the ease of adjustment to aquick disappearance of oil reserves is therefore errone-ous. ’Other WIM tests indicate that the longer term feasi-

bility of the nuclear option is just as questionable. TheU.S. Atomic Energy Commission estimated in the early1970’s that nuclear energy would provide 30 percent of

TB. B. Hughes and M.D. Mesarowc, “Testing the Hudson Institute Scenarios”(Washington, D. C.: US Assoclatlon for the Club of Rome, September 1979),mimeograph, p. 22.

(B)

1975 1980 1985 1990 1995 2000 2005 2010 2015 20202025

Year

the long run (see A). A much more beneficial developmentfor all concerned results from the “optimal price scenario”in which the price is gradually increased up to the “op-timum” level. Such a policy would bring in the substitutesin a more regular fashion while prolonging the reserves.Both exporting and importing regions would fare better(see B). it is only by taking a global and long term view thatsuch a course of development, most beneficial to all con-cerned, can be identified.

the developed world’s energy needs by 2000, based onhigher demand growth than is now expected. By ex-trapolating from this figure, the authors found that by2025 nuclear power would have to provide 60 percentand by 2075 almost 100 percent of all energy needs (seefig. C-9). The social, economic, and security impacts ofsuch a course would be enormous: sole reliance on fis-sion nuclear power would require 24,000 fast-breederreactors worldwide by 2075, which in turn would re-quire the construction of 4 reactors per week for a cen-tury and the eventual construction of about 2 reactorsper day just to replace reactors that have reached theend of their 30-year lifespans, at a cost of $2 trillion peryear for replacements alone (see fig. C-lO).S This sce-nario would also require the energy sector to processand transport 33 million pounds of plutonium eachyear; only 10 pounds of this element are needed to con-struct a nuclear bomb.

Mankind at the Turning Point concludes that an energy

future based on nuclear power would be a “Faustianbargain,” but this conclusion involves several impor-tant assumptions. For one thing, fusion as well as fis-sion could provide the growing nuclear share of energy

sMesarovic and Pestel, op. cit., pp. 132-1 M


Figure C-8.—Comparison of Middle East Oii Pro-duction, World Oil Deficits, Developed=Worid Gross

Regional Product. and Middle East Accumulated‘Wealth Under Two World integrated Model

Scenarios

.25

,20

,15

,10

.05

Solid line: Middle East withholds oil, Imposes ultimate ceilino of 14 billionbbllyr.

Dotted line: Production and trade governed solely by economic factors.

SOURCE: Mankind at tha Turning Point,

supply, although fusion power on such a scale would in-volve similar financial and engineering problems and(at present) even greater technical problems. Anothersignificant factor is the assumed rate of growth for ener-gy demand, which seems unrealistically high in view ofsubsequent events; in updated projections, higherenergy prices and slower demand growth would makethe nuclear share—whether fission or fusion-substan-tially smaller. The authors prefer an energy future thatplaces mid-term reliance on coal and coal-derived syn-fuels, combined with long-term development of huge“solar energy farms” in the present oil-producingregions. This alternative has not been tested with themodel, but it too would involve massive capital andengineering requirements, even with slower energy de-mand growth. Because of its technical uncertainties (see

below), and because it is not based on explicit analysiswith WIM, this solar alternative remains speculative.


LAWM assumes the nuclear-energy future that WIMrejects, but unlike the other global models it containsno representation of energy or any other resource. Thestructure of LAWM is based on the assumption that,“for the foreseeable future, the environment and its nat-ural resources will not impose barriers of absolute physi-cal limits” on the satisfaction of basic human needs.9

The authors base this assumption on two studies con-ducted independentl y of the model itself: a survey o fcurrently known reserves of fossil fuels, which foundenough oil and gas to last 100 years, and enough coal tolast 400 years, at present consumption levels; to and ananalysis of future production costs for energy, whichfound that:

. . . the so-called energy crisis . . . is of a conjuncturalcharacter, such as others of similar importance that oc-curred in the past. And it may be perceived that the mainreactions of the system will be to establish a newequilibrium, which, generally speaking, in the long termwill not differ from the previously observed trends. ’ 11

The latter conclusion appears inconsistent with theauthors’ own reserve estimates, which represent slightly

more oil and gas but considerably less coal than thoseused in World 3 and WIM, However, the authors assertthat “the most important fuels for the future are nuclearfuels.” The y cite predictions that nuclear power willgenerate 50 percent of the world’s electricity by 2000,and, although their own estimate of uranium oxide re-serves reflects only 33 years supply even at 1970 levels,they suggest that “a small increase in the price or an ad-vance in technology” will make it economical to extractvast amounts of uranium from granite or seawater. 12

The model itself, however, does not reflect the capitalcosts or potential technical bottlenecks involved in thisnuclear scenario.

United Nations Input-Output World Model(UN IOWM)

The central concern of UNIOWM is to reduce the in-come gap between the rich and poor nations before the

year 2000. In order to determine whether U.N. develop-ment targets are consistent with the availability of non-renewable resources, the model projects the levels ofproduction and world trade in six metals and three en-

‘A. 0. Herrera, et al., Catastrophe or NeuI Soctety? A .Latm American WoddModel (Ottawa: [nternatlonal Development Research Centre, 1976), p. 8.

IOIbid., pp. 32-33.“lbId., p. 34.121 bid., p. 33.


Figure C:9.- Projected U.S. Energy Supply Distribution for World IntegratedModel “Fast-Nuclear” scenario

8“Imports and Oil Shortages”

#(fission and

/ / ’ fusion)& B “Demand” / / /

laska oil

omestic gas

t Hydro

Geothermal

1

Lower 48 States oil —I I 1 1 I I


ergy sources (oil, gas, and coal) that would be requiredto support its relatively high rates of population andeconomic growth. Its major conclusion:

The problem of the supply of mineral resources for ac-celerated development is not a problem of absolute scarci-ty in the present century but, at worst, a problem of ex-ploiting less productive and more costly deposits . . . andof intensive exploration for new deposits . . . .13

Reserves and prices are determined independently ofthe model itself, and the amount of each resource re-quired for expansion in each economic sector is a d -justed for assumptions about increased efficiency due to

●

1 ~w. ~ontlef, A. Caner, and P, Petrt, Tk Future of [k WOdd Ecmmv: AUmted F+’atIons Study (New York: Oxford, 1977), p. 11; emphasis added.

technology. Most of these technical assumptions arenot reported, although the model does assume 55-percent recycling of all materials, worldwide, by 2000.14

The model also assumes that all nations will rapidly de-velop domestic reserves, but that extraction costs willrise as high-grade deposits are exhausted.

The authors project that 77 percent of the world’sknown petroleum reserves will be depleted by 2000, buttheir confidence about future energy supplies rests onthe world’s plentiful reserves of coal, which they “con-servatively” estimate at 9 trillion metric tons (roughlythe same as World 3). The model assumes “autonomous

,—“lbld., pp. 5, 45.


Figure C-10.– Long-Term Consequences of U.S.Energy Self-Sufficiency Under World integrated

Model “Fast-Nuclear’S Scenario

----- Cars of coal produced, United States— Days between starts of new

nuclear plants, United States— — World price of oil~ GNP of United States— World oil production

1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025

Year

SOURCE: Systems Research Center, Case Western HeSe~e university.

substitution” between energy sources, but it does notrely on cross-price elasticities nor does it compute pricesendogenously. For example, it assumes that shale oiland gasified coal will replace petroleum and natural gasin North America before 2000, but it also assumes thatthe price of coal will decline despite a sevenfold increasein the price of natural gas.

15 In addition, UNIOWM as-sumes a growing substitution of nuclear for conven-tional fuels in the utility sector, although this assump-tion is not reported in the documentation.l6

In all of its many scenarios, UNIOWM projects a tre-mendous increase in world consumption of mineralsand energy between 1970 and 2000. The global demandfor petroleum is projected to increase 5.2 times, naturalgas 4.5 times, and coal 5.0 times 1970 levels. Rapid in-dustrialization in the developing regions causes them tomore than double their share of world energy consump-tion. The Middle East remains the major net exporterof petroleum, with output projected to rise almost eight-fold—a projection that now appears unrealistic in viewof subsequent OPEC production decisions, Non-OPECdeveloping regions, along with Western Europe and

‘Ybld., p. 65.16A. Carter, private commumcatlon, April 1981.

Japan, become increasingly dependent on imported pe-troleum; but the U.S. oil deficit disappears after 1990,apparently due to the development of shale oil and asudden doubling of domestic coal consumption in the1990’s. Despite anticipated advances in mining technol-ogy and industrial efficiency, the percentage of capitalstock used in resource extraction increases steadily in allregions, just as it did in World 3.

In the standard scenario, based on the U.N.’S Inter-national Development Strategy, the real price of natu-ral gas increases by 656 percent and that of petroleumby 225 percent over the 1970-2000 period; but the priceof coal—the most plentiful energy resource—actually

declines by 14 percent in constant dollars.17 In a secondscenario based on “more generous” resource endow-ments, production and consumption levels change verylittle and extraction costs, although they rise later, arejust as high by 2000. The most significant impact ofhigher reserve estimates is on trade deficits: this sce-nario reduces the balance-of-payments deficit of the de-veloped regions by 50 percent, but the non-OPEC de-veloping regions, after being better off in 1990, have ac-cumulated greater debts by 2000 than they do in thestandard, “conservative” scenario.

In a later policy test conducted for the U.S. Depart-ment of Commerce, UNIOWM indicates that aggres-sive fossil-fuel conservation in developing regions couldreduce LDC trade deficits and thereby remove the maineconomic constraint to Third World development .18Such a course would also increase LDC capital require-ments and (implicitly) would require an even greaternuclear share to provide adequate energy for continuedindustrial expansion. In general, the authors findenergy to be an economic rather than a physical prob-lem, at least until 2000: internal reform in the LDCSand the creation of a “new international economicorder” are the necessary conditions for accelerateddevelopment in this century, although the model’s re-stricted time horizon prevents it from examining thesustainability of such growth in the next century.

Global 2000

The energy projections in the Global 2000 Report in-clude a variety of short-term and midterm forecasts,based on different methodologies and assumptions,whose purpose is “to define a range of credible futuresagains t which a l te rna t ive pol icy opt ions can betested. ”19 The report’s estimates of total world reserves

IT~ontlcf, Carter, and petri, op. cit., p. 65, table 61; these projections, likethose for production, appear unrealistic in view of subsequent events.

ISA. p. Cafier and A. K. Sire, “An EnerW Conservation Scenario for theWorld Model,” prepared for the Bureau of International and Economic Policy,U.S. Department of Commerce, November 1977, p. 2.

1~J. Pearson, et al., “Energy Projections,” In ThE Global 2(XM Report to the Presi-dent (Washington, D. C.: U.S. Council on Environmental Quality and Depart-ment of State, 1980), vol. 2, p. 173.


of fuel minerals come from figures prepared by theWorld Energy Conference, the Congressional ResearchService, and the U.S. Geological Survey; but thesereserve estimates are not used as inputs to the produc-tion or consumption forecasts, The short-term projec-tions (1975-90) were prepared by the Department ofEnergy’s Energy Information Administration (EIA)soon after its creation in late 1977, using two similarcomputer models: the Project Independence EvaluationSystem (PIES) for U.S. figures, and the InternationalEnergy Evaluation System (IEES) for global figures. Theforecasts made for the Global 2000 study use the study’slow, medium, and high growth-rate assumptions forpopulation and GNP, but the energy projections werenot used as inputs for other sectors; in essence, “GNP isimpl ic i t ly t rea ted as independent of the energymarket .“2°

Both PIES and IEES are equilibrium market simula-tions, in which producers compete to satisfy short-termworld demand and the entire global energy system is as-sumed to act in such a way as to minimize total costswithout regard for the future value of the resources. A sa result, both models assume that unlimited world oilsupplies will be available at prices (determined outsidethe models) that rise from $13/barrel in 1978 to$23/barrel in 1990. The models contain detailed repre-sentations of the OECD countries and the major fuels,but they contain no representation of resource deple-tion or political factors and only simplified demandequations for the growing LDC market. The modelsassume that coal and nuclear will substitute for oil inresponse to price elasticity, but they impose no limit onthe creation of new generating capacity in the utilitysector. (The number of new generating plants is deter-mined by expert judgment rather than price, a proce-dure that may be reasonable for forecasts through 1990:the lead time for developing new productive capacity isso long that current investment plans are a good guideto what will happen in the next 10 years.)

EIA was unwilling to extend its IEES forecasts be-yond 1990, and Global 2000’s midrange energy projec-tions (1985-2000) are based instead on four differentstudies representing diverse philosophical and method-ological approaches, as well as different assumptionsabout future demand growth and fuel substitution (seefig. C-1 1):

● the Workshop on Alternative Energy Strategiesstudy, based on estimates from independent na-tional experts, predicts global supply-demand“gaps” of 15 million to 20 million barrels of oil perday by 2000 and examines the consequences ofchoosing either nuclear or coal as the major re-placement;

lov~nder~’erf, op. cit . , VOI. 2, p. 571.

●

●

●

✎

the World Energy Conference study, after examin-ing different assumptions about world oil reservesand recovery rates, predicts that global productionwill peak at a ceiling of 82 million to 104 millionbarrels per day around 1990, and emphasizes coaland hydropower as replacements when oil supplyfalls short of demand;the Stanford Research Institute (SRI) model of theU.S. energy market through 2020 (see below),which foresees a relatively plentiful energy supp ly

based on coal but assumes perfect consumer fore-sight about future shortages and prices; anda Brookhaven National Laboratory/Dale Jorgen-son Associates study of the U.S. energy market,which emphasizes conservation and coal as well asthe development of nuclear and renewable sources.

in general, to the degree that these diverse projectionscan be compared, Global 2000 suggests that a rapid in-crease in the supply of energy will be needed throughthe end of the century, even with a declining rate ofeconomic growth. Demand growth will be moderatedonly b y price increases, and significantly higher oilprices will be needed to encourage substitution. Becausepetroleum production capacity is increasing more slow-ly than demand, a supply-constrained market is likelybefore 1990. Furthermore, because the rate of petro-leum reserve additions is falling, world production islikely to peak between 1990 and 2010 and gradually de-cline thereafter. As a result, “a world transition awayfrom petroleum dependence must take place, but thereis still much uncertainty as to how this transition willoccur. ”21

The findings suggest a considerable potential for coaland natural gas beyond 2000, but the short-term projec-tions indicate that nuclear power will expand far fasterthan any other source, particularly if oil prices continueto increase. The potential contribution of solar andother renewable sources is rather limited at best. How-ever, there does appear to be a substantial long-termpotential for “aggressive, conservation-induced reduc-tions in energy consumption. ”22 The alternative energysystems examined in the different studies indicate thatoptions do exist, however limited, and that currentdecisions about the future fuel mix will have increasing-ly significant impacts after 2000.

Other Energy Projections

Concurrently with its PIES/IEES short-term globalprojections, EIA also produced a set of long-term (1978-2020) energy projections for the United States using theLong-term Energy Analysis Program (LEAP), an up-

zlThe Global 2000 Report to the Pres;dent, vol. 1, P . 27.22pear~On, et a l . , Op. cit. I p. 161.


Figure C-Il.–Comparisons of Global 2000 Projections of World and U.S. Energy Consumption and SupplyMix, 1975-2000

6 0 0

500

400

a

; 300F

200

100

0

150

135

120

105

902

? 75b

60

45

30

15

0

(A) World projections Nuclear

Coal

Gas

Oii

tar

Low Medium High Low Medium High WEC case H-5 WAES cases D-7, C-21975 1965 1990 2000 2000

(B) U.S. projections

Low Medium High Low Medium High WAES WAES BNL/ SRI

Other

Nuclear

Coal

Gas

oil

case case DJA baseD-7 C-1

1975 1965 1990 2000

SOURCE: The Global 2000 Report to the President.

App. C—Energy Projections 115

dated version of the SRI -energy model (see above) .23This forecast, which was not reported in the G l o b a l2000 Report, describes the energy requirements for con-tinued economic growth in the United States during atransition from oil to coal and nuclear power. Itassumes the rapid development of shale oil and synfuelsto replace oil imports, as well as an ambitious level ofenergy conservation: a 48-percent reduction in the en-ergy-consumption-to-GNP ratio, based on improve-ments considered possible on the basis of known tech-nologies, with almost no increase in residential con-sumption and most of the demand growth coming inthe industrial sector. It also assumes that the rate of realGNP growth will decline to 2.4 percent by 1995. Despitethese assumptions, the SRI model’s projections wouldrequire U.S. coal production to triple by 2000 andreach 5 times present levels by 2020; nuclear powerwould increase eightfold by 2000 and reach 16 timespresent levels by 2020. This vast expansion is requiredpartly to satisfy growing industrial demand, but primar-ily because coal and nuclear must grow from 23 percentof primary energy to 72 percent in order to replacedepleted U.S. oil and gas. (This energy supply mix cor-responds roughly to the “fast-nuclear scenario” thatwas tested with WIM; see figs. C-9 and C-10, above).

EIA’s Office of Energy Information Validation hasconducted an evaluation of this 1978 forecast, and theirreport cites a number of factors that might modify theseprojections :24

. statistical studies of U.S. oil reserves suggest thatthere may be only half as much domestic oilavailable as previously estimated by the U.S.Geological Survey and assumed in the forecast;

. recent studies show that the actual costs for shaleoil and synfuels may be two to four times greaterthan earlier engineering estimates, and that thereare severe physical constraints on the developmentof a massive synfuels industry;

the forecast assumes that extraction costs for coaland uranium will not increase significantly overtime due to depletion or scale, and that environ-mental consideration will not impede the develop-ment of these industries;

the forecast does not address the capabilities of therelevant industries and therefore fails to considerpotential bottlenecks and constraints, includingthe need to build up the U.S. railroad system, a po-tential shortage of engineers and deep-seam miners,the risk aversion and capital constraints of poten-tial consumers, and the potential efforts of State

2~Energy Inforrnatlon Admlnlstratlon, Anntud Re/mt [O Confless 1978 (Wash-

ington, D. C.: Department of Energy, 1979), vol. 3.Z+ C) fflce of Energy Information Valldatlon, Amd?so @l[t> RePorr: 1978 Long-

Ter-rrr Forecasts and Methodolov (Washlrrgton, D. C.: Energy Information Ad-mlmstratlon, 1981).

●

governments to prevent rapid expansion in theWest; andmore recent data, and more realistic technical as-sumptions, suggest that electric cars can capture 50percent of the U.S. personal-transportation marketby 2000—despite optimistic oil prices and moderateconsumer prejudice—if automobile companies canacquire enough capital to keep up with the poten-tial market.

These and other problems are discussed in more re-cent forecasts by EIA, which continues to refine, vali-date, and expand its energy modeling capability; but de-spite a number of structural changes in the models, the1980 forecasts are based on many of the same assump-tions as the Global 2000 projections.25 Midterm globalenergy projections (1975-95) now come from an im-proved version of the IEES model that incorporates theannual oil production capacity forecasts provided byDOE’s Office of International Affairs and the oil priceforecasts generated by the Oil Market Simulation(OMS) Model. OMS assumes that OPEC will raise oilprices only when world demand requires them to use al-most all of their annual production capacity, and it alsoreflects lower rates of economic growth. Oil prices areprojected to reach $50/barrel in constant mid-1979 dol-lars by 1995, with almost no increase in oil consump-tion between now and then. This projection assumes,however, that all OECD nations will reach their officialconservation targets and that higher prices will lead tofurther substitution away from oil. Coal is projected toprovide a slightly larger share of total energy demand;nuclear projections are lower due to lower estimates ofthe speed with which new reactors will be built.

Long-term U.S. energy projections (1975-2030) nowcome from LEAP, a descendent of the SRI model thatEIA used in 1978. LEAP still assumes an unlimited sup-ply of oil imports, at the OMS/IEES world price, but itpredicts that the United States will require far fewer im-ports over the next 40 years. This is due in part to con-servation and in part to massive deployment of shale oiland synfuels: U.S. oil consumption is cut in half be-tween 1978 and 2020; and of the remaining demand forliquid fuels, coal-based synthetics provide 50 percent,shale oil 27 percent, and conventional oil only 23 per-cent. This assumes that the goals of the Synthetic FuelsCorp. (1.5 million barrels per day by 1990 and 3.0 mil-lion by 1995) will be met or exceeded, and that, therewill be virtually no constraints on the expansion of thesynfuels industry after 1995. This projection also ac-cepts with little modification the current engineering es-timates of synfuel costs; sensitivity tests, using capitalcosts twice as high as these estimates, lead to an in-

IJEnergy Information Acfmlnlstratlon, 1980 Annwd Re@rt [O congress (wash-mgton, D, C.: Department of Energy, 1981 ), 1,01, 3.


crease in U.S. oil imports through 2020. LEAP also as-sumes an upper limit to the net contribution of renew-able energy sources (hydro, wind, biomass, small-scalesolar, etc.) of about 6 percent of U.S. demand in 2020.26

Strengths and Weakness of EnergyProjection Techniques

All of these models are generalized analytic tools thatcan accommodate a variety of assumptions and serve avariety of applications. Some of them have features thatlimit their usefulness in examining the future of the glo-bal energy system, but these features were usually ap-propriate to the models’ original purposes. World 3, forinstance, was intended to give a generalized descriptionof the long-term behavior of the entire global system ex-cept agricultural land; as a result, it treats energy only aspart of a highly aggregated “nonrenewable resources”sector and makes no specific projections of energy sup-plies or prices. LAWM explicitly excludes all resourceconstraints except agricultural land; it assumes that en-ergy and other resources will be available and concen-trates instead on how they should be allocated in orderto satisfy basic human needs. UNIOWM and Global2000, because of their shorter time horizons and theabsence of resource depletion in their structures, are ill-suited for examining the long-term effects of resourcedepletion. In addition, UNIOWM concerns itself pri-mar i ly wi th development ta rgets and t rade ba l -ances—although it treats the energy sector in detail, itdoes little more than tabulate resources as they are con-sumed in meeting those goals. Global 2000’s IEES pro-jections represents major producers and consumers ofoil in detail, but they contain considerably less detail forother fuels or for the rapidly growing Third World mar-ket. Furthermore, Global 2000’s energy projections donot interact with other sectors and therefore do not re-flect competing demands for energy resources. WIMstrikes a balance between detail and generality, and itspolicy levers provide more flexibility than the othermodels for testing alternative energy futures and dif-ferent producer and consumer behavior. Because of itscomplexity and lack of documentation, however, the

~sFor further assessment of the technologies and assumptions involved inthese projections, see the following OTA reports and technical memoranda:Nuclear Pmhferarim mrd Safeguards, OTA-E-48 (June 1977); Gas Po[entud From

Devoruan ShaIes o~rhe APp&chum Basin, OTA-E-57 (November 1977); Enhanced

Od Recotq Potenttal m [he United States, OTA-E-59 (January 1978); A Technol-o~ Assessment o] Coal Slu~ Ptpelmes, OTA-E-60 (March 1978); APp/imtion of%km Technolo~ m Today’s Ener~ Requirements, OTA-E-66 (June 1978); TheDwect Use of (20u1: Prospects and Probkrns OJ Production and Combu.men, OTA-E-86 (April 1979); Gasafui A Techmczd Memorandum, OTA-TM-E-1 (September1979); The Future o/ Lu@ied Natural Cm Imports, OTA-E-110 (July 1980);

Energy From Biologmd Processes, OTA-E-124 (July 1980); Wodd Petroleum Atai/-ahhty 1984) -2(3(XI: A Techmczd Memorandum, OTA-TM-E-5 (October 1980); Nu-clear Pouerpkmt .Randard[zation: Light Water Reactors, OTA-E-1 34 (April 1981).

value of WIM’s conclusions may not have been fullytested or understood.

Factors Affecting the Accuracy andReliability of Energy Projections

These differences in purpose and technique have aninfluence on the projections generated by the differentmodels, but the accuracy and reliability of these projec-tions are also affected by a number of factors and uncer-tainties on which there is presently little general agree-ment or understanding. Among these factors are thefollowing:

total resource reserves and future prices;● population and GNP growth rates;● conservation;● Third World energy choices; development bottlenecks; and● future energy breakthroughs.

Total Resource Reserves and Future Prices

Table C-1 shows the estimates of total world reservesof conventional energy resources on which the differentmodels are based. There is little agreement among themodels on the size of these reserves or, how they shouldbe measured, and even less agreement on the costs ofextracting them. In general, however, those models thatconsider prices show that lower grades and unconven-tional sources will become available as prices rise. Ex-traction cost will be higher for these low-grade deposits,however, and recovery rates will be significantly lower.Consequently, a steadily larger percentage of capitalwill have to be allocated to obtaining resources, leavingless capital for investment in other sectors. Improve-ments in extraction and processing technologies mightmodify this trend, but recent studies have shown thatcapital costs for shale oil and coal synfuels are likely tobe higher rather than lower than previously estimated.

Several of the models predict an energy future basedon abundant reserves of coal, but estimated total re-serves of coal have increased little since 1913, whenthey were estimated at 8,000 billion metric tons.27 O fthe 9,000 billion metric tons now generally agreedupon, 50 percent or less is recoverable with currenttechniques, and two-thirds are reserves claimed by theU.S.S.R. that some experts treat with skepticism. Untilrecently, however, there has been little incentive forfurther exploration; new deposits may soon be discov-ered due to renewed interest in coal.

Z?H. S. D. Co]e (eCl. ), Mo&/s of Doom: A Crmque oj the Limus to Grwth (NeW

York: Universe Books, 1973), p. 98; on the other hand, these reserves haveseemed so large relative to expected demand that there has been Ilttle incentivefor increased exploration.


Table C-l .—Energy Resource Reserves as Described by Five GLobaL Modeling Studies

Petroleum Natural gas Coal Uranium oxide(billions (trillions of ft) (billions (millions

Global model Year of barrels) of metric tons) of metric tons)

World 3 . . . . . . . . . . . . . . 1972 630 identa 1,000 ident 8,600 ident N/A1,200 hyp/specb 10,000 hyp/spec 6,600 hyp/spec

WIM . . . . . . . . . . . . . . . . . 1974 667 proven 285 proven 4,200 recoverable Implicitly2,300 ultimatec 8,400 totald unlimited

LAWM . . . . . . . . . . . . . . . 1976 1,800 total 103 total 9,640 total 0.76 @ $10/lb; prac-tically unlimited @$20/lb

UNIOWM . . . . . . . . . . . . . 1977 1,555 total N/A 9,080 total NIA

Global 2000 ......, . . . . 1978 646 proven 2,520 proven 786 proven 1.661 @ $10-$35/lb2,100 total 5,984 hyp/spec 12,682 total 2.794 additional at

higher prices

N/A = not available.aldentified resemes include both proven and inferred reserves, including deposits that are currentlY su~conomic.bHypothet~cal resources include undiscovered deposits in known districts; speculative resources include undiscovered deposits in districts nOt PreSOntlY known to

contain deposits.cljltimately recoverable reserves; total reserves greater.dAssuming that so percent of total is recoverable.


The rate at which new petroleum reserves are beingdiscovered, on the other hand, appears to be falling. Inaddition, many geologists are pessimistic about theprospects for discovering vast new reserves of naturalgas at greater depths than have been explored thus far.

There is as yet no indication of diminishing returns inuranium exploration. Current reserves (including spec-ulative reserves at much higher prices) represent lessthan 50 years of total world energy supplies at currentconsumption levels (assuming the use of conventionalnuclear reactors), although with breeder reactors thesereserves would last far longer. Technologies for extract-ing huge amounts of uranium from granite or seawaterremain speculative, and DOE has at times taken the po-sition that uranium is scarce enough to justify the de-ployment of the breeder reactor.28

Finally, none of the models deals with potentialreserves of lithium. Despite the LAWM team’s con-fidence that fusion power will solve the world’s long-term energy problems, successful development andwidespread deployment of this technology remainsspeculative (see below).

Population and GNP Growth Rates

Appendix A shows that there is general agreementamong the projections of world population in 2000,however much the projections differ in the longer term.There is far less agreement among the models on futurerates of economic growth. Both factors influence the

ZW, S, Atomic Energy Commlsslon (USAEC), Proposed FiruI[ Emwonmend

Statement Llqwd Metal Fast Breeder Program (Washlngton,D.C.: National Tech-mcal Information Service, 1974).

energy projections, althoughtions is somewhat greater.

The standard run of World

the effect of GNP projec-

3 assumes that populationwill continue to grow exponentially. It also assumesthat the recent rate of world economic growth, 1.7 per-cent per year measured in real GNP per capita, will alsocontinue through 2000. These two factors quickly forcethe model against its resource limits, and the global eco-nomic system collapses sometime after 2010 due to ris-ing extraction costs. Sensitivity tests indicate that poli-cies designed to slow population growth or limit indus-trial expansion could delay (but not prevent) thiscollapse. Critics have claimed, however, that World 3underestimates the ability of the free-market system toanticipate and thereby prevent a possible catastrophe,although they have not explained how it would be pos-sible to prevent such a catastrophe in the absence ofspecific new energy sources .29

UNIOWM assumes a slightly faster populationgrowth and a much higher growth rate for gross prod-uct per capita—3.O percent per year even in its most pes-simistic “business as usual” scenario, and as high as 6.0percent per year for the LDCs in other scenarios. Themodel does not examine the long-term sustainability ofthese growth rates, however, and the principal reasonwhy UNIOWM does not predict a catastrophe is that itstops at 2000.

LAWM’s population projections are higher still, par-ticularly in the longer term, and its optimization pro-cedures impose high investment rates that lead to eco-nomic growth rates of 4.0 percent per year for the devel-

2QCole, op. cit., p. 66

118 ● Global Models, World Futures, and public policy

oped regions and up to 6.0 percent for the LDCsthrough 2000, gradually declining thereafter to 2.0 and3.0 percent, respectively. These growth rates do notlead to an energy-related catastrophe because LAWMcontains no representation of resource availability.

WIM assumes relatively ambitious population controlin many of its scenarios. Even so, it shows that popula-tion will not stabilize before 2050 under the best of con-ditions. Economic growth varies considerably amongscenarios, but in at least one policy test the economicgrowth rate (supported by investment aid from devel-oped nations) remains at 7.0 percent in Latin Americaand 8.2 percent in South Asia through 2025.

Global 2000’s short-term projections test three dif-ferent sets of population and GNP growth assumptions,in order to illustrate a range of possible futures. Eco-nomic growth rates are highest for OPEC and medium-income LDCs, somewhat lower for the developed na-tions, and lowest for the low-income LDCs and Com-munist bloc. In all cases, economic growth slowssignificantly after 1985.

Conservation

Total demand for energy can be represented as theproduct of three variables: population and GNP percapita (see above), and the ratio of energy consumptionto GNP. Conservation can reduce this latter ratio (andthus the total demand for oil and energy at any givenlevel of population and GNP per capita) in either of twoways: 1) by improving the efficiency with which energyis used (e.g., through residential insulation or improvedindustrial machinery); or 2) by improving the way inwhich the overall energy system matches energy sourceswith particular end uses (e.g., a large-scale coal or nu-clear generator is more appropriate to the energy needsof a major industrial city than to those of a rural village,which might well be better served by a small-scale wind,hydro, or solar source). Biomass, dispersed solar, andother small-scale alternatives can contribute to the sec-ond form of conservation, but conservation of conven-tional fuels will depend primarily on the response oflarge-scale industrial and utility consumers.

Many economists believe that higher prices are an ef-ficient mechanism for inducing this kind of conserva-tion, and economic models generally assume that de-mand will fall in response to future price increases tothe same degree that it has in the past. For example,pre-embargo studies of energy demand in the UnitedStates generally showed little responsiveness to price:when prices were low, other variables (such as the priceof automobiles or new capital equipment) are more im-portant to consumers than energy costs. As prices firstbegan to rise rapidly, there was a large initial demandresponse due to “housekeeping” conservation and

other simple measures, but the long-term response wasexpected to be slower due to the slower conversion tomore efficient automobiles and capital equipment.

However, the response to the 1979 oil price hike sug-gests that long-term elasticity will be greater than manypeople had expected. Conservation has been greater—and demand growth slower—than was previously fore-seen, even when the effects of economic slowdown areeliminated. In addition, some economists claim thateconomic models based on the United States do not re-flect the full global potential for long-term conservationthrough more efficient capital equipment. JO Otherstudies, however, have shown that technological prog-ress in some critical industrial sectors requires an in-creasing use of energy per unit of output .31

The World 3 standard run assumes that resource con-sumption per capita will continue to be a fixed functionof GNP per capita. This assumption implicitly rejectsany significant potential for conservation and—although reasonable when the results were published in1972–it gives the standard run a pessimistic bias thathas been contradicted by subsequent events. However,World 3’s “recycling” run (fig. C-5) does show that con-servation, in combination with improved explorationand extraction technologies, can have a significant im-pact on resource depletion.

WIM, published after the 1973 embargo and pricehikes, does assume some conservation in response tohigher prices. In the case of oil, a l. O-percent increase inprice is converted into a direct decrease of 0.225 percentin consumption, plus additional decreases due to substi-tution. The authors conclude, however, that even opti-mistic assumptions about conservation will not preventa substantial increase in total energy demand—theworld will need to develop alternative sources, eithernuclear or solar.32

LAWM assumes that the global energy system willmake more efficient use of different energy sources inthe future, and that technological progress will increasethe productivity of the capital goods sector by 1.5 per-cent per year (a rate that would double the output-to-input ratio in 47 years). For the most part, LAWM’s op-timism about the availability of energy is based not onconservation but rather on unlimited supplies of nucle-ar power, including the deployment of fusion technol-ogy within 50 years.

UNIOWM also assumes that technological progresswill change the energy requirements of every sector.The documentation does not reveal, however, the pre-

~OR. S. Pindyck, “The Characteristics of Demand for Energy,” in J. c. Sawhill

(cd.), Ener~ Cmsemmon and Pub//c Pohcy (Englewood Cliffs, N.J.: Prentice-Hall,1979), pp. 38, 39.

IID,W. J o r g e n s o n a n d B.M. Fraumeni, “Relative Prices and Technical

Change,” prepared for AEA Annual Meeting in Denver, Colo., Sept. 5, 1980.IZMe~arC,vic and Pestel, OP. Cit., P. 136.


cise value of these assumptions or how sensitive the re-sults are to them.

Global 2000’s short-term projections assume energyconservation in response to price ($23/barrel in 1978dollars by 1990), above and beyond the official conser-vation targets of various national governments. Themidterm projections extrapolate from engineering andeconomic forecasts for 1995 and result in a final level ofconservation—48 percent reduction in the energy-to-GNP ratio by 2020–that indicates the technologicallimits of what can be achieved. More recent DOE fore-casts of conservation are based on higher prices($50/barrel in 1979 dollars by 1995) and on detailedanalyses of specific technological possibilities in everyend-use sector, many of them derived from engineeringstudies conducted by the Oak Ridge National Labora-tory. These and other studies, combined with the de-mand response to the 1979 oil price hikes, suggest thatlong-term conservation in the 50-percent range is in facta reasonable expectation.

Third World Energy Choices

None of the models except Global 2000 deal with thefirewood crisis, which may have severe social and envi--ronmental repercussions in many Third World coun-tries. Industrialization and economic development inthese countries will also require electricity, however,and all of the models assume, at least implicitly, thatnuclear power will be widely deployed in the future.This common assumption is of political as well as eco-nomic interest, particularly in the case of LAWM andUNIOWM. Both of these models represent ThirdWorld attempts to chart a future in which the gap be-tween the rich and poor nations is narrowed throughlocal and international efforts to accelerate develop-ment and increase industrial output. Nuclear power of-fers developing nations an alternative to their currentdependence on fossil fuels, whether to preserve their do-mestic resources or to reduce their energy-related tradedeficits. There is evidence that, whether or not theUnited States and other OECD nations finally acceptthe hazards of nuclear power, many Third World na-tions are likely to accept them on a very large scale, andvery soon, for lack of a credible large-scale alternative.

Business commentators already speak of “the nuclearpower boom getting under way in the Third World,” in-cluding such nations as Mexico, Egypt, Korea, Taiwan,and the People’s Republic of China.33 Convent iona lenriched-uranium reactors in such countries represent apotentially lucrative market for the U.S. and Europeannuclear industries. Many Third World governments arealso investigating natural-uranium technologies that

‘]H. Rowen, “Nuclear Reactors: Fear of Losing Export Race,” Wa.slungtonPost, May 17, 1981, p. K1.

would allow them to exploit domestic uranium ratherthan depending on Europe or the United States for ex-pensive enriched fuel. Argentina, Pakistan, India, andSouth Korea already have operational reactors basedon a Canadian natural-uranium, heavy-water system.Mexico’s recently published National Energy Programcalls for the construction of as many as 16 such reactorsto meet a tripling of demand for electricity by 2000, de-spite that nation’s abundant oil and gas reserves .34

The WIM “fast-nuclear” scenario, however, foreseesan energy future based not on conventional fission butrather on the more efficient “breeder” reactor, which ef-fectively produces more nuclear fuel than it consumes.As a result, breeder reactors might produce 60 to 100times as much energy from a pound of uranium as doconventional reactors. Both past U.S. Governmentstudies and current expert opinion suggest that, giventhe limited size of world uranium reserves (even in-cluding speculative reserves), nuclear fission may not beable to provide a large-scale, long-term contribution tothe world’s energy supply without the widespread use ofbreeder reactors.

35 Because breeders would also produce

large quantities of weapons-grade plutonium, however,their deployment throughout the Third World couldalso pose a serious threat to domestic and internationalsecurity.

Development Bottlenecks

Some economists would argue that the world’s majorenergy problem is not finding adequate resources, butrather overcoming the bottlenecks in getting those re-sources to market. Three of the models considered here—World 3, LAWM, and UNIOWM—implicitly assumethat there will be no major bottlenecks in the exploita-tion of energy resources. Although WIM’s authorspoint out the numbers and speed with which reactorswould have to be built for the “fast-nuclear” scenario,they do not report any further analysis of developmentbottlenecks. DOE’s models for Global 2000 involve aserious attempt to address problems of timing wherethey involve liquid fuels and the transition from oil tocoal and nuclear power. However, DOE’s own evalua-tion of the 1978 long-term forecasts points out that themodels describe the requiremenrs for U.S. energy in-dependence rather than the capabilities of the releventindustries, as well as making questionable assumptionsabout investor foresight, extraction costs, and environ-ment constraints (see above). Further consideration ofthese and other factors suggest both that considerablecooperation between Government and industry may beneeded to reduce the impact of these potential bot-

‘+C. D]ckey, “Scenic Mexican Reactor Site Entangles Indians, Umons, Na-tlonalwts,” W’a.shlngton Post, May 18, 1981, p. A12.

J5see, for ~xample, USAEC, oP cit.


tlenecks, and that other energy sources and mixesshould be examined more carefully.

Future Energy Breakthroughs

Finally, the accuracy and reliability of energy projec-tions will also be affected by the assumptions they makeabout the successful development and deployment ofentirely new energy sources. Some such systems are im-plicitly assumed by one or more of the models, butothers are not foreseen by any of them.

The widespread deployment of fusion technology, forinstance, could possibly invalidate the pessimistic as-sumptions of Limits to Growth if the supply of lithiumwere large enough or if deuterium-deuterium fusionwere developed. The LAWM study bases its exclusionof resource problems in part on the expectation that fu-sion power will in fact be deployed within 20 to 50years. Commercial-scale fusion power remains purelyspeculative at present, however. Although fusion re-searchers hope that a commercial fusion reactor designmight be possible by 1990, until then all claims aboutengineering problems, capital costs, and net energy pro-duction will also remain speculative. In addition, noneof the models discusses the potential world reserves oflithium. Other sources, however, have estimated thatminable U.S. lithium reserves alone are worth over160,000 quadrillion Btu prior to conversion losses, orabout 640 years total world energy supply at 1976 con-sumption levels.3b

The WIM solar scenario envisions a long-term energyfuture based on centralized solar power in the form ofhuge “solar farms” in the deserts of the Middle East, tobe financed by OPEC oil money .37 With higher energyprices due to scarcity, the capital costs of such facili-ties—which the authors estimate at $20 trillion to $50trillion in 1974 dollars—might become bearable; solarmight even prove cheaper than nuclear. Theoretically itwould be possible to supply all of the present U.S. de-mand for electricity from a “farm” of solar cells the sizeof Massachusetts, but DOE studies claim that there arenot enough economically feasible sites in the UnitedStates for centralized solar to make more than a mar-ginal contribution to U.S. energy supply in the foresee-able future.38 On the vast scale foreseen by Mankind atthe Turning Point— 1 percent of the world’s land sur-face–such facilities would involve scientific, engineer-ing, and planning problems that are beyond the currentstate of the art. Furthermore, some critics question

MS. s. penner (e&), ~uc~~r Errergy and Ener~ Politics (Reading, Mass.: Addi-

son-Wesley, 1976), p. 562.3TMesarovic and Pestel, op. cit., pp. 139-141.J8DoW~Clc po~im Re~lew, O/ SOIUT Ener~ (Wash ing ton , D . C . : Depa r tmen t o f

Energy, February 1979), table 8; DOE’s long-term forecasts therefore assume ana priori hmit of less than 1.0 Quad central solar by 2020.

whether the energy produced by these solar farmswould be greater than the total amount of energy in-volved in building them, maintaining them, and dis-tributing the hydrogen and electricity they would pro-duce, Since this scenario was not subjected to rigoroustesting with the model, it remains speculative.

A number of other groups have suggested another al-ternative: solar-power satellites orbiting in space andbeaming power down to earth in the form of micro-waves or lasers, The National Aeronautics and SpaceAdministration (NASA) has produced detailed designsfor such a system, for which it envisions commercial op-eration by 2000, or about 20 years before fusion is pre-dicted to become commercial. NASA studies have indi-cated that marginal costs and net energy productionwill be comparable to present energy sources, althoughthese findings are highly controversial. DOE claimsthat potential receiving sites are probably more thanadequate to supply several times the present level ofU.S. electricity demand, since more energy can be re-ceived per acre than with central solar and without nec-essarily prohibiting agricultural uses of the sites, al-though the effects of low-level microwave radiation areproblematic. Research and development costs for theNASA design, however, would be over $100 billion,and the cost per power station would be somewhathigher than that of conventional fission reactors re-gardless of the scale of production. An alternative de-sign developed at Princeton University involves lessR&D and employs small modular processing unitswhose cost and performance could be tested prior toany commitment to large-scale deployment. The Prince-ton design is somewhat riskier than NASA’s, but itmight be possible to deploy it sooner and to reduce en-ergy costs substantiall y as the scale increases. Thismight in turn make it possible to sell energy to theLDCs at prices much lower than would be possible withbreeder reactors.39

J~Further discussion and analysis of this technolog y can & found m the fol-lowing sources: .!iatehte Power System Concept Development and Evahatlon Pro-gram Reference System Report, DOE/ER-0323 (Washington, D. C.: U.S. Depart-

ment of Energy and National Aeronautics and Space Administration, October1978); The .Schr Power Satelhte Concept: The Past Decade and the Next 13ecde,NASA/JCS-14898 (Washington, D. C.: National Aeronautics and Space Ad-ministration, July 1979); R.A. Herendeen, T. Kary, and J. Rebltzer, “Energy

Analysis of the Solar Power Satellite,” hence, vol. 205 No. 4405, Aug. 3, 1979,pp. 451-454; Massachusetts Institute of Technology, Department of Aeronau-tics and Astronautics, Space Systems Laboratory, Extraterresticd Processing andManuJcsctunng of Large S@ce Systems, NASA contract CR-161293 (Washington,

D. C.: National Aeronautics and Space Administration, September 1979), vol.3, pp. 47-49; Of%ce of Energy Research, Solar Power Satellite Project Division,Program Assessment Report, DOE/ER-0085 (Washington, D. C.: U.S. Depart-ment of Energy, November 1980); and Office of Technology Assessment, U.S.Congress, solar Pou,er Satellites, OTA-E-144, (Washington, D. C.: U.S. Govern-ment Printing Office, August 1981).

Global Models, World Futures, and Public Policy: A Critique

Documents

Transcript of Global Models, World Futures, and Public Policy: A Critique