U.S. economic growth from 1976 to 1986 : prospects, problems, and patterns


Material Information

U.S. economic growth from 1976 to 1986 : prospects, problems, and patterns
Physical Description:
United States -- Congress. -- Joint Economic Committee
U.S. Govt. Print. Off. ( Washington )
Publication Date:


serial   ( sobekcm )

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 22347714
System ID:

Table of Contents
    Front Cover
        Page i
        Page ii
    Letters of transmittal
        Page iii
        Page iv
        Page v
        Page vi
    Table of Contents
        Page vii
        Page viii
    Forecasts of long-run economic growth
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
    Understanding the changing basis for economic growth in the United States
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
    Long-term economic growth forecasts in the Federal government
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
Full Text

94th Congress
2d Session



Volume 6-Forecasts of Long-Run Economic Growth



DECEMBER 15, 1976

Printed for the use of the Joint Economic Committee

78-733 WASHINGTON : 1976

(Created pursuant to sec. 5(a) of Public Law 304, 79th Cong.)
HUBERT HI. HITUMPtREY, Minnesota, Chairman
RICHARD BOLLING, Missouri, Vice Chairman
EDWARD M. KENNEDY, Massachusetts OTIS G. PIKE, New York
WILLIAM V. ROTH, Ja., Delaware JOHN H. ROUSSELOT, California
JOHN R. STARK, Exaecutive Director RICHARD F. KAUFMAN, General Counsel




DECENIBER 10, 19,76.
To th4e Members of the Johdi Ecoiiornc Commrittee:
Transmitted herewith is the sixthll volume of the Joint Eooilnic,
Conrutte stuy series entitled "1.S. E-oiiomic (Gi-owt !i from I )(Y19861: Prospects, Problems. and Patterns."hssre o vrfr
studies forms an important part of thie Joint- 1Econoimic Commlni Thirtieth Anniversary st ud'V series. whvllch w\.)- u undertaken to- prmvide insght to the Members of Congress and to thp ul i !it ]- r ur oil tt important subject of full employment and economic growth. The Employment Act of 1946, which established the Joint Econoii Conmittee. requires that the committee iniake reports ajid recominein ations; to the Congress on the subject of maximizingc employ i tgl. pro-,0 duction and pu rchasino, power.
Volume 6, a 'Key one in the series. contains three papers -\ihiob )1C
sent the latest for-ecasts of long-run economic growth in the UitedC! States as w\ ell as analysis of the iiodels that vYenerate thesefoeat Private as well as Federal Government foreecass are examined. A lso. in addition to the stancardl ecmoometric models ulsedl for proje coti~ons, two of the papers discuss the newer system dynt-mi1cs model which has potential use in economic forecasting.
The authors of the studies 'in this volume gre Dr. Gairv F~~n
Professor -Nathaniel M-Nass and Professor Jay Forrester, anl Dr..Joszephi W. Duncan. The committee is grateful to these authors for their fine contributions, all of which are highly intformative and stimulativ0.
The views expressed are those of the authors anid (10 not necess~lrilv represent the views of the. Members of the committee 01' the comititee st aff.
Chiairman, Joint E"conomnic Cornmn;teeq.

DECEMBER 6. 19 76.
ion. HUB3ERT I-I. H~umPhirErY.
Cl mrrnan, Joint Econon ic Cor? n mttec, V .S. Con qreSS, lWa. hknton, D.C.
DEAR Mn[. CHII3ANMA: IranUsmitted herewvith are thi-.1e stlies e,]titled "KForecasts of Long-Run Economic (Groiwth," by 1 )r.(%
Fronm,"Understanding the Changing Basis for Econom11ic, (i'roxvi in the, United State,;,' by Prof. -Nathianiel Mass aInd P~rof. Jay l-"orrester. andl "Long-Term Economnic Growth Forecasts in thle 11 "1lera 1 Government." by IDr. Joseph WV. 1)unran.

TIwese three studies comIprise volume G of the Joint Economic CoimmIIittee's study series. .. Economic Growth from 1,Z G-lo98. Prostx,. P'rohlems arnd tPatterox This series forms a substantial part of le Joint Economic ('onmnitee's Tlirtieth Anniversary study series.
Each of these papers examines modelling efforts which have ken deised to yield projections of econOfliC growth over the 1on run. Not only are the forecasts presented. but the palwl's take an in-depth.
mI 1-te-scene.-ent illulinatuing d1cusso1ns which contrast the more ttlilna tOll)Ilfllt ic Io(1el aIp)roach with the newer system dynamics
alI mo cl 1.
The paper by )r. Frommi contains a comprehensive review of the n jr ong-term economic growth forecasting efforts in the private sector. Nearly half of the paper is devoted to a discussion of the alternative projection methods: Simple extrapolation. reduced form projection, system dynamics and feedback control models, and economIetric models. Pointing out the problems associated with relying on any one of these techniques at this time, he concludes that more structurally realistic and accurate models can be built in the future. lo'-t of the forecasts present a favorable picture for output, inflation and income over the next 10 years with growth rates projected to exceed those of most 5- and 10-year post recession intervals during t lhe years subsequent to World War II. The median forecast for the annual compound growth rate for real GNP is 4.8 percent for 1975-80 anid 3.5 percent for 1980-85.
The paper by Mass and Forrester is challenging in that it presents a hypothesis concerning long-run economic growth that is not widely shared. They contend that the greater instability in the economy appears to be caused by two principal modes of economic behavior whose existence is not widely recognized and whose causes are only poorly understood: The Kondratieff cycle, or long wave, and the life cycle of economic development. The significance of the long-wave phenomenon to public policies lies in the fact that, if the long wave in a real recurring element of the national economy, then the United States may be approaching another trough of the approximately fifty
--ar cycle. The problem, as the authors see it, is that economic sta,ilization policy today is predicated chiefly on prevailing theories of t Le Ishort-tenrm business cycle, whereas current economic developments probably arise from the interaction of both short- and long-term modes of economic behavior. Ten recommendations related to national policy regarding economic growth-including expanding the tihe frame of stabilization policy, abandoning the Phillips curve conceptt as an indicator for public policy and conducting increased research into the dynamics of the national economy-are presented which take into account these hypothesized longer term phenomena.
The paper by Dr. Duncan describes the major ongoing economic forecasting models developed by major Federal agencies. Each of the major models related to economic growth is described in terms of its haic structure, the most recent findings, and the users of the fore4ants. The paper examines the degree of coordination between forecastiiig groups of various agencies, concluding that there is a high degree


of informal coordination, but that there is a, gro \vini- need for- nioP( formal coordination efforts. Another of its major points is that the establishment of a central economic forecasting model would be counterproductive and too constricting in developing helpful decisionmaking tools.
The committee is indebted to these authors for their work in developing these highly informative papers for this study series. Dr. Fromm is an economist with the Stanford Research Institute and the National Bureau of Economic Research, Professors Mass and Forrester are at MIT, where they serve both as f aculty members and as members of the System Dynamics National Modeling Project, and Dr. Duncan is the Deputy Associate Director for Statistical Policy at the Office of Management and Budget.
Dr. Robert I-amrin of the committee staff is responsible f or the planning and compilation of this study series with suggestions f rom other members of the staff. The administrative assistance of Beverly Mitchell of the committee staff is also appreciated.
The views expressed are those of the authors and do not necessarily represent the views of the Members of the committee or the committee staff.
Executive Director,
Joint Economic Committee.

Digitized by the Internet Archive
in 2013

http://archive.org/details/usecon86u nit


Letters of transmittal ------------------------------------------------ III

By Gary Fromm
Introduction and summary ------------------------------------------- I
Alternative projection methods --------------------------------------- 3
Growth projections --------------------------------------------------- 15
Output, inflation, and income ---------------------------------------- 21
Employment, productivity, and finance --------------------------------- 25
Capital requirements ------------------------------------------------Growth strategies and policy alternatives ------------------------------ 33
Conclusion ---------------------------------------------------------- 36


By Nathaniel J. Mass and Jay W. Forrester
S u m ma ry ----------------------------------------------------------- 38
I. Time horizons f or analyzing economic growth -------------------- 41
IT. System. dynamics methodology as a tool for analyzing growth issues-- 40 III. Simultaneous modes of economic behavior ------------------------ 47
IV. Policy directions and implications -------------------------------- 69
Bibliography -------------------------------------------------------- 73


By Joseph W. Duncan
Summary ----------------------------------------------------------- 76
I. Introduction ---------------------------------------------------- 77
11. National Iong-run economic growth models ------------------------ 78
ITT. Regional long-run economic growth models ------------------------ 88
IV. Long-run environment and energy growth model ------------------ 90
V. Model users: Treasury, O.NIB, CEA, and FRB -------------------- 92
V I. Coordination of modeling within. the Federal Government ---------- 93
Appendix I. Other modeling efforts ---------------------------------- 9T
Appendix IT. Attendees of the interagency group on long-term growth
projections ------------------------------------------------------- 98
Bibliography -------------------------------------------------------- 98

By GARY FRo:L3t **

During the past decade there has been much public discussion about the desirability of reducing U.S. economic growth and that of other highly developed countries so as to conserve environmental and natural resource assets and improve the quality of life by lowering tensions associated with strong pursuit of material well-being. More recently, with a world-wide slowdown in growth rates and a severe recession in 1975, concerns about low growth have surfaced anew, especially with the realization that changing demographic patterns (an aging population) and lower productivity (for reasons to be explained) might limit possibilities for transfers from the young to the old or disabled via health and income security outlays and from the rich to the poor via redistribution of tax burdens. With high growth and rapidly increasing income and wealth per capita, more equal sharing of a rising prosperity is politically more feasible and more easily accomplished. Standards of living of those most disadvantaged by accidents of birth, age, health, past discrimination, or other social conditions, can significantly be raised while making everyone better off, or at least no worse off. In economists' terms, high growth accommodates Pareto efficiency. High growth is not necessary, nor even sufficient to guarantee ease of redistribution, but it surely helps. Apart from redistribution it also permits devoting substantially more resources to improvements in environment, public health and safety, arts and the humanities, education, national security. foreign aid, and other socially useful purposes. Still, the picture is not one-sided and there probably are many persons who would opt for a low-growth scenario. However, whatever one's preferences, the elements which influence growth and its prospects should be examined carefully so that suitable policies can be considered and appropriate choices made among alternative growth strategies.
Methods for projecting growth vary greatly in sophistication, structural detail, dynamic characteristics and costs. The simplest, quickest and cheapest, and perhaps least accurate, are extrapolations of past performance. These contain little or no structural or causal interactions and depend mainly on observed empirical regularities. Forecasts in this category include those based on combination of periodic shortand long-run swings of population growth and business activity-the
*Paper prepared for U.S. Joint Economic Committee study series, U.S. Economic Growth from 1976-1986: Prospects, Problems and Patterns. This research was in part supported by the National Science Foundation. I am Indebted to E. C. Hwa and Milton Kelenson for valuable assistance and to the individuals and organizations who kindly supplied forecasts.
**Member and Senior Research Staff. National Bureaiu of Economic Research. and Director. Center for Economic Policy Research, Washington, D.C., Stanford Research Institute.


7\ fc ,KIi'r 1( it ,hi ,l. 0Tn : r. 1Ktwc s. anl Komindratieff cycles--and those ld']t-kl \ TviI Ol ,I) e, ,:\' A"i" ,r td 110 6i14' ave. age statistical fits to, llizt'iiical 2_,t'o wlI ralte K'1aa.
I0idttc(1 foIT ::roWl m!' l s in Vl icl predictions depe(ITId on some raL-al linkages bIut aive drive hrgel y if nllot exclisivelv by exogenous forc''s are s)whVlVat more cot ly t an extra)oative teclminiques. However, th ,e sho(M) l,! e I moreie ieial)e 1)ecause normally tley incorporate ill fore)dItlonll oni hlb force par{ticip-ationi, unemployment. and produetivitv. all of which have strongr influences on growth. The principal deficiency of the reduced form approach is that it obscures dynamic interactions and endogenous causal or behavioral linkages which are determinants of growth.
In contrast. systems dynamics and feedback control models place primary emphasis on assumed, detailed patterns of behavior and their time sequences. Little, if any allowance is made for exogenous influences and predicted growth paths are determined almost solely by the mathematical characteristics of the model. The systems dynamics approach, at least as implemented until now, utilizes little data on the actual path of the economy but relies on extensive, intensive, and costly computer simulation to develop growth scenarios. The strength of the technique is its emphasis on dynamics and behavior; its weakness, as currently implemented, is the failure to allow adequately for external forces and shocks and sufficiently to undertake validation tests, especially those of comparing predicted and actual performance.
Econometric models share many of the characteristics of systems dynamics models in their emphasis on causal linkages and time varying paths. They differ in grounding their behavioral specifications in economic theory and in a more rigorous approach in delineating structural relationships. Parameters of equations largely are estimated by "fitting" historical data using a variety of statistical methods. In general, too, most econometric models are subjected to batteries of tests to ascertain their predictive accuracy and response patterns to shifts in policy variables and parameters. Data requirements and costs are the primary drawback of this forecasting technique, especially when the model is large in scale. Yet, given the value of accurate predictions in enabling improved formulation of government policies, the costs might be considered negligible.
Forecasts of U.S. economic growth over the 1975-85 interval were obtained from 22 organizations. While the sample is limited, the predictions are felt to be representative of the "best" and currently most widely used U.S. economic projections for the next decade. Most of the forecasts assume that fiscal and monetary policy will neither be highly stimulative nor restrictive and will steer a middle course which permits the private sector to grow at its own natural rate. But, some of the respondents anticipate a moderate or growth recession during 1977-79. largely due to a tight monetary policy in 1977-78. The median prediction for real GNP growth for 1975-80 is 4.8 percent per annum and for 1980-85 is 3.5 percent. Inflation during the first five years is projected at a 5.7 percent annual rate and in the first half of the 1980's at 5.0 percent. Unemployment rates for the remainder of this decade are predicted to average 6.5 percent and then fall to
5.0 percent for 1981-85.


This scenario is more favorable than that of the last five years. 11t, it is subject to a number of major uncertaiinties, including errors inl fiscal and monetary policy management, potential capacity shiortacres brought about by poor financial structure of capital intensive in(histries, and the effects of possible shocks from cartels, world food shor-tages, military or political upheavals, or other unanticipated sources. A successful growth outcome will depend upon pursuit of a galaxy of policies designed to affect both demand and supply and to maintain proper balance between them. Better analytical tools are needed to achieve that end and government-sponsored research for that purpose should be given high priority.
Naive Extrapolation
The variety of techniques for predicting growth is great, ranging from simplistic extrapolation of past performance to exhaust ive, highly detailed structural annalyses. At the bottom end of the s)ectrum, there is the naive method of using averages of output growth over long intervals or over the recent past. For example, real GNP arew at the following annual compoud rates over the stated intervals (measured in 1972 prices).
.Percen t
1930-75 ----------------------------------------------------------3.29
1940-75 ----------------------------------------------------------3.61
1950-75 --------------------------------------------------------3.27
1960-75 ------------------------------------------------------------- 3. 26
1965-75 ----------------------------------------------------------2. 56
1965-70 ----------------------------------------------------------3.04
1970-75 -- 08
Last year, 1975, of course was one of deep recession; using this as a terminal date would tend to lower growth rates depending on whether the initial year was at a cyclical peak or trough and the Ienith of the interval over which the growth rates are calculated. The longer the time span, the less the effect of initial and terminal year boomrecession conditions. Still, end-year conditions can make significant differences for growth measurement.
This is evident by inspection of real growth rates over succeeding five year intervals. (See table 1.) For example, starting in 19,50, the real growth rate drops from 4.2 percent for the 1950-55 period to 3.3 percent for the ten-year span 1950-60, and then rises to 3.7 percent for the fifteen year interval 1950-65. It then falls to 3.6 and 3.3 percent for the 1950-70 and 1950-75 periods, respectively. With a starting point of the boom year 1955, the pattern is similar but the 20-year growth rate (1955-75) is only 3.0 percent. Over shorter intervals, differences in growth rates tend to be even greater.



1 1930 1935 1940 11945 1950 1955 1 1960 1965 1970

1935 ............. -1.6 ..........................................................
1 -40 -........... 2.2 6.0 ..................................................
1945 t... ..... 4.5 7.7 9.4 ............................. ..
-95-............. 3.3 5.0 4.5 -0. 2 ...............................
195............. 3.5 4.8 4.4 2.0 4. 2 ....................
19O6u ........... ----------- 3.3 4.3 3.9 2.1 3.3 2.4 .............
1965............. 3.5 4.4 4.0 2.8 3.7 3.5 4.1....................
19701-......... 3.4 4.2 3.9 2.8 3.6 3.4 3.9 3.0 ..........
19751........... 3.3 3.9 3.6 2.7 3.3 3.0 3.3 2.6 2.1

I Recession ear.
2 Peak pior to 1961 occurred in April 1960.
Note.-Real GNP measured in 1972 dollars.

Therefore, projection of growth rates based on simple extra olations of past performance are subject to large errors. Compounddiffierences in growth rates can accumulate to large amounts over even short intervals. For instance, the difference between growth rates of 0.5 and 3.0 percent would result in a gap between the two projections of 6.7 percent of real GNP after ten years which, for the 1975--85 period, is nearly sufficient to allow for doubling the Federal budget without any increase in tax rates.
A complication with simple extrapolation of past growth is that increases in output, even when averaged, do not take place along a smooth exponential path but tend to exhibit waves or cycles. This was observed in the 19th century by Clement Juglar, who was the first to
isolate major industrial fluctuations of prices, production, employmenit, and so forth, over a period of nine to ten years." This is one of three cycles used by Joseph Schumpeter in his explanation of capitalist development innovations and the dynamics of economic growth.2 Of t he other two cycles, one is a regular forty month fluctuation named for Joseph Kitchin, who was the first to study it in detail, and the other is a lone wave of 50-60 years first identified in 1919-20 by a Russian economist. Nikolai Kondratieff.3 Taking the types together, there are three Kitchin vcycles of 40 months each to every Juglar cycle of 9-10 ,ears and five to six Juglar cycles in every Kondratieff cycle. Under1ying phenomena associated with those fluctuations are inventory
investment, business fixed investment, innovation and changes in capital accumulation, and opening and expansion of new markets.4

SThis summary description is from Douglas Greenwald and Associates, "Dictionsry of 1 2 Tbd.
sph Kitchin. "Cvcles and Trends in Economic Factors." Review of Economic Statistie. voL 5, No. 1, February 1923, pp. 10-16. Kondratieff's findings were first published in 5Movw in 1922. then in German "Die langen wellen du Koniunktor," Archvy flir Sozialwissenschaft and Sozialpolitik," vol. 56, No. 3, 1926, pp. 573-609, and then in 1:lisli. "The Long Waves in Economic Life". Review of Economic Statistics, vol. 17, No. ,. November 1 .935, pp. 105-15. The last is reprinted in "Readings in Business Cycle T, rv" (Richard D. Irwin, Homewood, Ill., 1951).
SI'-r a critical appraisal of Kondratieff cycles see George Garry "Kondratieff's Theory of Lon cycless" Review of Economic Statistics, vol. 25, No. 4. November 1943, pp. 203'.'n: Arthur F. Burns and Wesley C. Mitchell, Measuring Business Cycles (National B ofHiur" of Economic Research, New York. 1946), chapter 11. On Juglar movements see
1:. C. (. Matthews, The Business Cycle (Chicago, 1959), ch. 12. Many names are associ~ld with work on inventory cycles but most analyses are directly or indirectly related to 1,k'vI A. ietzler, "The Nature and Stability of Inventory Cycles," Review of Economic Stnr i;ji. vol. 21. No. 3. Auist 1941; reprinted in P. A. Gordon and L. R. Klein (eds.), AE\ Radings in Business Cycles (Richard D. Irwin, Homewood, Ill., 1975).

Another phienomnoni that has been identified, which has beeiial vl
ate(1 with the name of Simon Kilziets, is anl approximate 20-yeai' waNve in population grrowthi, which impacts onl the demland for r'dH u
housing, and related public utility and coiminit v facilities.
The analyses of K'uziiets. Kondratieff. Kitehm. N(Yilar. anid Tliitzu. Mitchell, and others date to the 1930's or (tlrIli(T tand were canledl oiit oi U.S. and European data for the 19th and early 90th centuries. Tlw, techniques they employed were, relatIx-el V unsophistioate(d and~ to a large degree depended onl datingr peaks ami trougzhs in n-ovin' aver ages of economic activities. A more riuyorous method is to calculate t li spectral density of annual gYrowth rates.G; Tbis was done fop .S. real GNP1 data, for the Period 19,30-1915 : the results are shown in ficguiv L, They accord rougly with the cycles or swings found by Kitclin.
J ti' ar and K uz iets. T he density fuu fiion p a sat annual r q el o e
of 0.5. 0.351 0.08. and 0.05, corresponding to 1)eriodicities (equal to reciprocals of the frequencies) of about 2-3. 12 .anid 20 yeaivs. The. two-year cycle may be attributed in part to a Nvane11tv of mi'easuremnent problems, including seasonality, observation errors, and a need for short-run smoothing. The periodicities of about four and 00 years also are consistent with findings from stochastic simuilationls with the( 'Wharton model using a post-World War II sample period."
Concentrations in the spectral density function m-ay be usedl to predict future activity levels. However, it should be recognizedl thlat there miay be signlificant variances about average lpel'iodicit ie:s and amplitudes of fluctuations at these periodicities. Moreover, there can be trencls and cycles in thoselimgnitudes caused by both deteriinmms--tic (caus--al) and stochastic (ranidom) factors. As a consequence, precise path predictions based only onl past growth rate experience are subject to considerable error.
This can be seen more readily using a nearly equivalent technique in the time domain to the spectral representation in the frequency (domain. Table 2 presents a uto regressive, integrated, moving average (AIRIMA) models of -real GNP growth rates over three sampie periods.9 The results reported were obtained by searching ov-er each stinple, space (period) for the parameters ( p and q1) which iimiized the standard err-or of estimate of the AIIIMA equation. It is eiiiidenlt that the mranitudes of the parameters shift signifi cantly with chang1fes in sample periods. The same holds true for extrapolations ( forecast-dz f rom these equations. (See table 3.)
5 See WV. Arthur Lewis and] Paul JT. O*Leary. "S'ecular Swing~s in Produetion :1ml Tn llo 1870-1913,'' The Mlanchester School of RLconomic and Social 'Sttidivcs, vol. 18. N" 2. Ml'v 1 955. pp. 11i3-52 and Moses Abramllowitz, "The Nature and Si iutii~uc Ku 'p.'rs1 Cycles.'' Economic Development aind C ultural Change. vol. 91, Apil 1 961. Bot. i rt icles are reprinted, in AEA ReadingIs in Business Cycles. op. Cit.
1Any continuou-s, Sta-tiouary ( trendless ) times series mnay he trans formed in to in equiv'llf'I1t linear cominfat ion of 'omplIonenlts in the frequency d44infl (8in anam b .y is A M -tild FMI radio signals). If all frequency components have e(111.1 weiglt tilh- slwetr'l I 4lil'i t function has a rectangular formi. If some comiponen ts of a ifiute setis's fr t r.' it removed, account for more of the total variation tin n .rlerz, ti s i,, revealed 1).\ 4' c-1 ceni ration of weights of its speelral density function at those frequeutes'.
,Daya used were G.NP in 1972 dollars. which kindl were provided by the Io re: i #i Economic A nalysis. D.S1epa rtment of Comnmierce. These were eonvA'ot ed to a iintii 1p\ thI rates. 1930-75. The analysis was performed with the sjpecti-AI suitittite of lio N I VIZ TfROLL sy-stem using a Irizing~ular smoothing window. it rang-e 4 Sf 10) \VPIr. prv\- Ii'e'a with it para;imeter of (IS. anrd removo,'il of the ltif in g-rowiV i ratte of 2.4 1 .oel'(ent p er mimm4 its standard deviation is 6.31 percentl.
8 See Ml. K. Evans. L. R. Klein, anid 'M- Saito .'Short Run Plredict i' i andit I~on 1 z'un imltion of the Whatrton 'Mode]."' in Bert Gy. 11i4*km it. ed.. Econiomietri 1' MIodilk of Cycial Behavior, vol. 1 ('National Bureau of LE( onontuc Reser rh New Yortik. I 9721, pp. 1-19-85 and Lawrence R. Klein, A TIextbook of Eceonomtetrics ( Pren tice-I10 I a ll.l::4 wood Cliffs. N.J., 1974), pp. 252-9.
9 For a description and examiples, of this techniques see G. E. P. Box : id (8 0. l'4 JWi', Ti1im- Series Analysis: Forecasting4 an(] Control t lulden D ay, San Franricico, 1 970)


-- -~:~~--- ------~--= =

2.01 :

a a
I *
a a a a a U
Nb, C C C S U
a U
a S U
a U
1.50 1 U

a C a a U
C p p U

.00 ~ (4/f S U U a 0 0 0 0 *U

P P a 'UUU

og J . . . . . . . * *
C C C S a U
a S a = U
Spectrum .
a P P U

a a P U
S a a

a S a a
C P p
P a U
0.00 U S S a


*~'i ~ P 5 ~ ~ S S S I

a I a a 8 P U

a C S a
a a

.5 5 a S I
a a U
S a P a U
5 U
a a
C a P a p p

____ a

0.100 0.200 0.300 0.4Oci 0.500 O.6c~1

Frequency (vp~r-~

FIGURE 1.-Spectral Density of U.S. Economic Growth Rates: 1930-75.



p q

Sample period c1 02 0, 01 02 C SER ((t) U(t) D.W.

1933-75 ---------- 0.806 --0.529 ----------0.58 -0.9S9 2. 808 4.036 4. 159 0.970 1.64
(23.5) (15.4) (39)
1950-75 -------- -0. 145 --0. 190 --0. 239 -0.255 5. 786 2.901 3.512 0. 826 1.79
(0.7) (0.8) (1.3) (3.2)
1954-75---.----- -0. 787 -0.482 -0. 454 -0.999 8. 317 2.436 3.036 0. 892 1.61
(3.7) (1.5) (2.2) (3.5)

1. Real GNP measured in 1972 dollars.
2. Symbols:
G(t)=annual rate of change of GNP in 1972 dollars (percent).
C=constant term.
e=Tandom error.
G(t)=sample mean of G(t).
SER=standard error of the equation.
DW=Durbin-Watson statistic.
3. Magnitudes in parentheses underneath respective parameters are t-statistics.

1933-75 1950-75 1954-75

forecast year:
1976 -------------------------------------------------------- 4.09 4.80 8.17
1977 ----------------------------------------- 4.20 5.93 3.61
1978_ -- ---------------------------------------------- 2.93 4.55 2.43
1979------------------------------------------------2.89 2.62 1.01
1980 ------------------------------------------------ 3.16 2.33 4.77
1981------------------------------------------------3.17 3.57 3.02
1982 -------------------------------------------------------- 3. 12 3.98 3.23
1983-3.11 3.72 2.21
1984 -------------------------------------------------------- 3.13 3.46 3.71
1985------------------------------------------------3.13 3.43 2.93
198 ------------------------------------------------ 3.12 3.56 3.28
1987 -------------------------------------------------------- 3.12 3.62 2.70
1998 -------------------------------------------------------- 3.12 3.60 3.34
1989 ----------------------------------------- 3.12 3.55 2.95
1990------------------------------------------------3. 12 3.54 3.2i
1991 -------------------------------------------------------- 3.12 3.56 2.90
1992 -------------------------------------------------------- 3. 12 3.57 3.20
1993 -------------------------------------------------------- 3.12 3.57 3.00
1994 -------------------------------------------------------- 3. 12 3.56 3.15
1995 -------------------------------------------------------- 3.12 3.56 2.99
1996 ------------------------------------------------ 3.12 3.56 3.14
1997 -------------------------------------------------------- 3.12 3.57 3.03
1998 -------------------------------------------------------- 3. 12 3.57 3.12
1999 --------------------------------------------------------- 3.12 3.5G 3.09
2000 -------------------------------------------------------- 3.12 3.56 3.11

Note: GNP measured in 1972 dollars. Actual growth rate in 1975 equals -1.83 percent.
Using the equation estimated over the period 1931-75, the extrapolation for 1975-2000 of annual real GNP growth rates (starting
from an, Initial condition of actiii 1 experience iII 19 of -I.8: per.cent) ii ps to 4.09 percent in-I 1,97C. peaks Cat 4.20 percent in 1977 and,
after twvo smail growth cycles of deereasinr amplitude, approa.bcls a
stea(lv state rate of 3.1-2 percent in the late 9S(Vs. M[ovino" to a post'Wo'ld 'War 1I salHp)e eiiod of 1, ) T0-,,i a similar pattern is evidenced, Nit the I1")iit lIte of cis Wnr'('r. hr f lll.
imlv daplled cycles, the steacl" state is not reached until the late
19)0 's, and the projected long-term c-rowth rate is 3.,56 percent. An
even more marked contrast in the extrapolated path results when just


fouri vez(trS ate d(lIjjwId flmii thle be~nj~ofth OwSample period."Usn the in1tervall 1.)93 -, h allplitlde of growth cylsthereafter be11s i~rznifican1t lv greIater thall for the longer sample periods and I here "Ire 11111 ore and -hortcr cycles. H owever. the steady state is about the sae s 1,nr thle l -75 sample period and( appears to be,
S-t.l1)ilizin.2 m t te year 2001) around a 0.10 poi-cent growth rate. A cornja ijj*_nj oft the( tlmi-(q gi'o()\Ntlj.t i s, m1a v be found in figure 2. Growth rate
(perc ent)
It i 9.00

8.00 ~80
I I t

7.00 If II
I 7.00

5.00II 1

it A I- N it

.00 If 2-- - - - -- - - - - .00
it -f I
.00 -e - - - -- - - - 9 1.00,

: Sapl peid 193-7
------ :~ Sapl perod 195075
II 0 a : S m l er o : .9 4 7
Futui-i-: -.A11.Frcfso elGP nulGot ae:17-00

The steady state projections of these equations should not be interpreted literally. Tie economy is expected to exhibit cont inuin" growtll yeveles. But, this is not reflected in the AINMA forecasts I)e volld a certain point because of the characteristics of tliis techlli(jqule and its dependence on stochastic terms. As the projection moves incereasi ngly beyond the sample period. no further stochastic information is entered and gradual damping takes place through the A IMA process. The
failure to include any structural information other tlian the historical path heightens this effect.
In general, extrapolation, whlethler of a naive (such as sinmpl)le projection of past performance) or sophisticated fornil (such as single variable ARIMA models) is a poor method of forecasting', because it
ignores causal factors which may radically alter fut ire character t Is and paths. The variance of prediction errors fromt these types of folecasts normally will be significantly greater than those fron models which account for structural linkages.'o Use of extrapolation as a forecasting tool can be justified if underlying striicturl alnd ca"isal processes are poorly understood or cannot be modeled for reasons of lack of data or expense, or with a l)predictioll accuracy that surpa'es that of simpler extrapolation alternatives. This is inilikely to le true in the case of models of long-term growth of the economy.
Reduced Form Prediction
In a reduced form, variables to be predicted are expressed only as functions of exogenous or predetermined information and do not depend simultaneously on each other. Reduced forms can be naive or complex, and can range from single to hundreds of equations. An example of the former is the orthodox monetarist doctrine that the rate of growth of GNP is dependent almost exclusively on the rate of growth of the money supply. While popular in some circles, this theory has not found widespread acceptance. Yet another relatively simple reduced form model has been utilized widely for growth prediction."
It may be characterized as follows. Assume that population is given exogenously. Assume further that labor force participation as a function of population is constant or can be extrapolated as a smooth trend. Next, take a given rate of unemployment as a full employment target and assume that it will be achieved. Then, extrapolate the rate of growth of labor productivity (output per employee) from p:ist data or by other methods. Combining these steps then yields a growth rate. Roughly, if labor force participation and unemployment rates are held constant. the output growth rate is the sum of the population and productivity growth rates. For instance, for the 1975-85 decade the U.S. Bureau of the Census has estimated that population will ,grow at a compound rate of approximately one percent. For the period 1968-73 (that is, excluding the 1974-5 recession), the annual rate of
"o In faet. eomparlon of extrapolation and structural model foreevstling errorsz i a 0-t of the reliability and potential validity of the models: residuaIls from the latter should! le smaller than those from extrapolation.
1 For the past 15 years until now, this technique has been the primary basis for U F Government estimation of potential output of the economy. see ,Report of the 4Con'nIil of Economic Advisers, 1962 and subsequent annual issues. It has been utilived in *ligh tl more elaborate form by the Organization -Tor Economic Cooperation aind Develoinment (OECD) in The Growth of Output 1960-80 : Retrospeet. Prospect and Problems of Policy. ID)cember 1970, and for several of the forecasts sunimmarized below.

7S 73.1. 76..


growth of real GNP per employee hour was 1.41 percent12 If this mne product ivity growth were extended to the next decade, then real (N P would be predicted to increase at about a two and one-half percllt. coO)oulIld rate.
Actually, labor force participation is expected to increase somewhat in thle years ahead due to chaning age-composition of the population and to a still rising proportion of women seeking paid employment. (See Figure 3.) Also, for various reasons to be discussed shortly, productivity could well grow faster than it has since 1968. Therefore, barring another severe cyclical downturn or other untoward circumstances, real GNP most likely will rise faster than the two and one-half percent rate just given.
FIGURE 3.-Population of the United States: 1870-2050
300oo Total
3 0! ....i . t l ...... ..... I .... . I , 1 .. .. I W ,
Percent .... Percent of Total, by Age Group
0-24 -'*~ **. .
40- -2 ".....
25-44, ]
3o-25-4 ..........***,,--,,,-- ,. .. ........... ..20
201- __ _,40*''
i o -1, 11 Illiiill 1 0i 0 0 4 0 1 0 Wl I $ I a n I .* I oio
10- 65 and over .
-mti liI' r I iII! mI
0 I i I I I I j I III I I I
1870 1890 1910 193G 1950 1970 1990 2010 2030 2050 2070
NOTE.-1980-2050 Census Estimates, Series II. Source: Figure provided by Office of the Secretary of the Treasury, Office of Debt Analysis.
A more sophisticated redicedl-form growth model underlies the heralded work of Edward F. Denison.'3 As I have observed elsewhere, his technique bn-ically is one in which rates of growth of various inputs are weighed by their earnings share in national income to obtain the contribution of each input to the rate of growth of national income."4 Denison disaggregates the inputs as well as the overall productivity of their use in great detail. An example of the output of his painstaking efforts may be found in table 4. As can be seen therein, necount is taken of capital as well as labor inputs, and allowances are made for age-sex composition, education, shifts in hours, economies
2 Estimate provided bv M. D. McCarthy, Council of Economic Advisers. See his Accouintine for United States Economic Growth 1929-1969 (Brooklngs Instit ion, Wahindton. D.C., 1974).
FGalry Frnniu, "Rviw of the Sources of Economic Growth andl the Alternatives Before Vs." by I:(lw: rd F. I)nison. JTournal of the American Statistical Association, vol. 58, No. .o04. December 19G, pp. 1168-71.


of scale, irregular infle es. arid OilI 1(r fOr St I P i1i 11a1 t

characteristics of Denison's approach is that these forces are exogenous and largely independent. In his methodology they are not determined simultaneously, and there are few if any feedback. When
structural linkages such as interaction between compel S "atiofn rates,
productivity advances, and increases in factor inputs are ignored. the
overall effect is likely to lead to biased and inconsistent esthnates of
the contributions of different elements to growth.

[Contributions to growth rate in percentage points]

1948-53 1953-64 1964-69 1969-80 1

National income .............................................------------------------------------------ 4.54 3.23 4.54 4. 13
Total factor input .... ----------------------------------------- 2.95 1.30 3.08 2.39
Labor .... ...... .............---------------------------------------------- 2.07 .60 2.15 1.37
General government, households, and institutions --------------- .96 27 .55 .23
Employment ............................------------------------------------. 91 31 .51 .24
Hours and shifting weights...... --------------------------- .05 -.04 .04 -. Cl
Nonresidential business...-------------------------------.................................. 1. 11 .33 1.60 1.14
Employment..........................................------------------------------------- .72 .20 1.75 1.12
Hours ----------------------------------------.C6 -.21 -.24 -.15
Average hours ................................... -------------------------------.29 -.26 -. 45 -.20
Efficiency offset ........------------------------------- 12 -.02 .10 .01
Intergroup shifts -------------------------------- .11 .07 .11 .04
Age-sex composition ....... ------------------------------- .07 -.09 -.31 -.05
Correction to hours and age-sex composition --------------- (2) (2) (2) -. G8
Education..... ..................--------------------------------------. 38 .43 .40 .30
Capital------------------------------------------.............................................---..........---. 88 .70 .93 1.02
Inventories ..............................................----------------------------------------. 18 .08 .18 .12
Dwellings ...........................................-----------------------------------------. .31 .27 .29 .39
Nonresidential structures and equipment---------------------....................... 38 .29 .45 .47
International assets ...............................-----------------------------------.. 01 .06 .01 .04
Land.. ...............--------------------------------------- ------------ 0 0 0 0
Output per unit of input ----...-. ....-... .....-. ........ 1.59 1.93 1.46 1.74
Advances in knowledge and not elsewhere classified.............. 1.34 1. 13 1. 15 .. 16
Inproved resource allocation................................ --------------------------------. 41 .24 .34 .10
Farm......................... ....................--------------------------------------------.. 33 .21 .19 .07
Nonfarm self-employment...... ------------------------------. 08 .03 .15 .03
Dwellings occupancy ratio ...............---------------------------------. 03 -. 01 .01 0
Economies of scale ...................................--------------------------------------. 48 .32 .56 .46
Irregular factors ..........----------------------------------------.61 .25 -.60 .16
Weather in farming -.-----------------------------------. 03 -. 02 .02 0
Labor disputes........................................-------------------------------------.. 0 0 -.01 0
Intensity of demand.................... .... ----------------------------------.58 .27 -.61 .16
Antipollution costs 0.-------------------------------------- 0 0 0 -. 14

I Change from actual national income in 1969 to potential national income in 1980.
2 Not relevant.
Source: Edward F. Denison, "Sources of Growth Accounting as the Basis for Long-Term Projections," in Tigran S. Khachaturov (ed.), "Methods of Long-Term Planning and Forecasting" (International Economic Association, MacMilan Press Ltd., 1976), pp. 241-59.

This is typical of the reduced form approach to preblictionl. ,While
structural information is incorpor-ated into such mod-ls, it is unidirectional, from a set of exogenous .forces or assumptions directly
to a set of outcomes. This pocC(,o too, obscures dynamic ilteractiolns
and does not coiristitute an adequate representation of dvyniiamic loirrun behavior.'~ In an economic growth context. it is mIost useful,
perhaps, for examinations of past history d for making ballpark
estimates and internal consistency checks of future perf ormance.

Systems Dynam ics and Feedback Conitrol 1o, is

The systems dynamics approach to modeling growth st:nd(ls in
marked contrast to those of naive extrapolation and reduced form

16 This also is the conclusion of PIhoebus .T. Dhrym,. "Econonttics : Statistical Foundations and Application" (Springer-Veria New York, 1974), i. 505.


metlods. In the Forrester-Mass formulation, the model is constructed to depict social and economic change following certain general 1urincilples.2'
Thle, e include:
1. D)ccisionmaking within sectors is modeled widely on observed business and government practices. Behavior in the models is not based on theories of optimal general equilibrium
nor oI profit or utility maximization or cost minimization.
2. Special attention is given to stocks of inventories, capital.
,order backlogs. finance, and so forth, and distinctions are made between desired and actual levels. Provision also is made for
grladuial adjustments toward desired states.
3. Highly nonlinear relationships are incorporated, especially
those representing limiting conditions such as capacity constraints
and maximum rates of change.
4. Quantitative computer simulation is used to derive qualitative behavior of the system. Policy alternatives are explored byv altering parameters of the model and observing consequences
for solution paths.
5. The model is wholly self-contained. parameters and structure are fixed. there are no exogenous influences, and stochastic
shocks are not admitted or damped very rapidly.
The application of systems dynamic techniques to analysis and prediction of economic performance has much to reconumend it. Emphasis on observed behavioral relations can lead to a model which more realistically depicts the actual economy than mlodels which rely heavily on exogenous elements and static equilibrium conditions.
Rarely does the world seem to be in equilibrium but rather it appears to be groping, from one disequilibrium state to another in the search for dynamic, shifting, equilibrium targets. To the extent that this process is mirrored in a systems dynamics model, it has advantages over models which are based on less realistic premises. On the other hland, the present assumptions of the Forrester-MAlass formulation al*e unduly restrictive and unrealistic in a number of respects, which tends to invalidate conclusions derived from simulations of their
iiodel. These assumptions are convenient in that they simplify the modeling process, but they are not necessary.
The most tenuous and least likely characteristic of the ForresterMass model is that the structure of the economy either has not changed or if it has, its essential dynamic properties as revealed in cyclical waves-Kondratieff. Kuznets (business), or (in their terms) life cycle-are unaltered. This seems incredible in the light of introduction of automatic stabilizers and the use of discretionary fiscal and monetary policy to abort or avert cyclical episodes. It is farfetched to believe that all economic and political behavior can be made endogenous and that the path of the systems is preordained. Even with a fixed
V For s nontechnical overview, see Nathaniel J. Mass. "Mo eline Cve'e in the National Economy." Technology Review. vol. 78, No. 5,. March/April 1976. pp. 42-52. Other deseriptions may be found in Jay W. Forrester. "Business Structure. Economic Cycles and National Polley." Systems Dynamics Group Working Paper D-2245-2. Alfred P. Sloan School of Mniagement, M.I.T.. Cambridge, 1975, Jay W. Forrester. Nathaniel J. Mass, iid Charles J. Ryan, "The Systems Dynamics National Model: Understanding SocloEconomic Change and Policy Alternatives." System Dynamics Group Working Paper D2245-2. Alfred P. Sloan School of Management. M.I.T.. Cambridge. 1975 and Nathaniel J. Ms. "Economlc (Cycles: An Analysis of Underlying Causes" (Wright-Allen Press, Cambridge, 1975).


structure of paramiieters and relationships, does it really make no difference whether a Democratic or Republican administration guiides fiscal policy? Few observers would believe it.
A related issue is the selection of specifications and parameters of relationships in systems dynamics models. The approach apparently followed for the most part has been to base these on judgments of individuals involved in the processes being modeled, on a priori reasoning, and on rough approximations to historical proportions, trends, or other magnitudes. Statistical estimation appears to have been utilized little if at all, and verification of the possible reliability of individual equations seems to have for the most part been ignored or carried out in minimal fashion. Validation of properties appears to have consisted mainly of comparing the cyclical characteristics of dynamic solutions with those of the economy and not of subjecting the models to a battery of predictive tests. Without results from such tests, how can policvmakers have any assurance that simulation studies with systems dynamics, national economy models provide reliable guidance for the possible course of future events or for the selection of policy options?
On another note, the treatment of stochastic shocks, can it really be true, for instance, that the simultaneous existence of high unemployment and inflation in the United States in 1974-5 can be blamed on an inevitable Kondratieff cycle? Or is a more reasonable explanation one that places heavy reliance on OPEC pricing, foreign harvest failures, and the conduct of wage-price controls policy in 1971-73? Stated more generally, is it plausible that operations of the economy can be modeled solely in deterministic fashion or must continuing stochastic influences be taken into account? Isn't this especially necessary if relationships are presumed to be non-linear and dynamic paths depend on gaps between desired and actual states? Notwithstanding that random shocks may be damped and gradually disappear, can't the transients he sufficiently large and last sufficiently long that they have significant impacts on the path of the economy ? Unequivocally, the answers are yes.
As presently constituted by Forrester, Mass, Meadows, and others, systems dynamics models of economic growth may have a role in providing broad perspectives and insights into economic development processes. They can be useful to Federal government policymakers primarily as a tool for exploring different scenarios in the use of policy tools and for examining the potential consequences of chances in the characteristics and operation of the economy. But, the techniones need refinement, especially in the area of specification and estimation of structure and parameters. Validation and predictive tests of the models, too, are crucial and without them. systems dynamics simltions should mainly be viewed as interesting and suggestive academic exercises.
Econometric Model
Econometric models structurally bear a strong resemblancme to present systems dynamics models but differ in a number of important respects. While both strive to mirror behavior of economic agents. formulations in moqt large-scale econometric models begin with sets of theorotical propositions and principles (such as profit and utility maximiization goals for producers and households, respectively) and derive


equatio 1 speciFiications tleirefrom. Allowance is made, as in systems .VIInamici liodels. for desired states or targets and for gradual adjust1ineilt to di-eiuilibritnn positions. Parameters of equations largely are estimlated by hitting historical data. but selected coefficients are chosen or constrained onil a priori grounds. Bayesian principles, using notions of distribution of variances and covariances of parameters are beginning to be employed in est nation of some large-scale econometric models, too. Such priors have been specified by model builders based both on their own intuitions and knowledge and on information obtained from participants in the processes being modeled.
Therefore, the primary difference in specification of the structure of models between present systems dynamics and econometric appron ehes is the greater reliance on deductive theory rather than inductive observation by the econometric technique. Which degree of reliance is superior depends on the accuracy with which the specifications accord with actual behavior, a question which can only be answered by empirical observation, predictive tests, and conformity with ineeting other validation criteria. Econometricians have begun to apply systems dynamic techniques in specification and estimation of selected sectors of their models, which is a partial indication that a combination of the two methods may be superior to either alternative.'7 Given the relative infancy of the systems dynamics approach, it is difficult to predict the precise nature of the combination.
However, the final amalgam probably will lie closer to an econometric formulation in at least several other features. First, the model would picture the world in a combination deterministic and stochastic f rinework. Second. provision would be made for exogenous influences. which impinge from outside the system and are not controllable from within. Third. the structure of the economy would not be presumed to be fixed but rather to be subject to rapid and evolving shifts. Such clhanges are attributable to a variety of forces, including the impacts of innovation and technology. revisions in tastes and preferences. relative availabilities of natural resources, and so forth.
Unfortunately. the most widely used econometric models of the U.S. economy are oriented toward short-run analyses and prediction and. for the purpose of long-run forecasts. make inadequate allowance for st ructural change. They take little account, for example. of shifting demographic patterns, evolving technology, or relative resource availabilities. Most of these models have quarterly time frames and their parameters are estimated usinfr data for approximately the 20 year period following the Korean War. In general, the shorter the sample nerio(l. the more tenuous are long-range predictions. Moreover, the fiirther the forecast from the sample period, the larger is the potential error band. For these reasons. whatever the prediction errors of the models in short-run. or near-term forecasting, they are likely to be larcrer for longr-run proiections.
It is possible to specify and construct econometric models that can serve the a(1l purpose of short- and long-run economic analysis and prediction. Efforts in this direction have begun, partially in response to growing scientific. policynimaker, and public recognition that policy decisions dealing with short-run issues (for example, an energy crisis)
2~ See. for instance JaTmes Tobin and IWalter Dolde. "Wealth. Liquidity and Consumption.'" in Consumer endingg and Monetary Policu": The Linkaqe, Federal Reserve Bank of Boston, Conference Series No. 5 (1971), pp. 99-147. These methods also are being used by William C. Brninard. Ga ry Fromnm, and James Tobin in the construction of a flow-of-funds model of the U.S. economy.


may have important and costly lol-run inpliention.s whi n should he taken into account.'" Also. there 1a Is been rising i terest ii i~cking. and predicting the (conllSe(luences of evolving. lanes in the size alnd (distribution of th e population, resotirce ava:i labti lite, tlie ,(e s-S'rvices COnsuimptiOll nuN. tfle integra iion of the wo(rl(l (cOHOIiyI. ad So forth. Skeptics of large-seale econolietric lo(tlels p)()ilt to stilet ihmes poor forecasting performance and other (ldeficiencies of such systeils. Occasionally there have been large prediction errors, but there also have been notable successes. The subject of how such models can be improved further (it might be noted that average prediction errors have decreased over the past decade) is too broad to be taken up here. But, given time and adequate resources, more structurally realistic and accurate models can be built incorporating the best elements of statistical, systems dynamics, and econometric te hiniques.


The Sample, Methodological Characteristics, ad Assumptions
Forecasts of U.S. economic performance over the 1975-85 interval were solicited from respondents to the American Statistical Association-National Bureau of Economic Research short-run forecasting survey and a variety of other sources. Twenty-two forecasts were obtained: many of those solicited indicated that they prepared only short-run projections or that their long-run predictions were fragmentary and incomplete. While the sample is small and includes only 22 forecasts, it is felt to be representative of the range and character of the "best" and currently most widely used U.S. economic projections for the next decade.
The forecasts are not strictly comparable to each other because they were prepared at different times (from mid 1975 through fall 1976). The projections made most recently have advantages of later initial conditions and the availability of officially reviewed national income and product account statistics. Given opportunities for modifications, this minight cause some differences in the earlier predictions but, judging g from the long-term projections made by some respondents in 1975 and then revised in 197, should not alter them greatly. Large changes in predictions could occur, of course, if a forecaster drastically mnodifies his views on the future course of the economy.
It might also be noted that a number of forecasts in the sample are not completely'independent, but are partially based directly or indirectly on projections of others, including the widely available services of Chase Economnetrics Associates. Data Resources. Inc.. and VWharton Econometric Forecasting Associates. Where known, such dependence has been indicated in the summary~ of selected characteristics of the forecasts. (See table 5) All the respondents utilize a variety and coinmbination of methods in preparing their projections, but rely to a greater degree on some chosen technique. Of the 22 forecasts, six place primary emphasis on econometric models and seven use such models as an important input to their analyses. The remaining nine forecasts rely more heavily on reduced form and judgment methods.
28 A new macro-econometric model for studying the medium-term growth path of the U.S. economy recently was completed by Bert G. Hickman and Robert M. Coen, An Annual Growth Model of the U.S. Economy (North-Holland, Amsterdam, 1976). Forecasts from this model are not yet available and it requires further testing.



financialDisaggregation of real
Re son1ent Primary method Time frame Scale I production interactionI

Chase ----------- Model ........... --....... Quarterly Large ---------- Limited- _ Medium
DRI ------------- do ........................-do ------- Very large------ Medium (recursive)... Strong.
Wharton-----..........-do-.......Annual_ --------do -------- High--------- Medium.
SSG ------------do .................do --------- do-- -do------------ Weak.
NPA ---------- Model assisted 4 ------------- do ------ Large-------- Limited ------------ Medium.
GE_ ........Model ------------------- Quarterly --- Medium ------Medium------------. Do.
14FORUM.------ Model ------------------- Annual ------ Very large ----- High --------------- Weak.
NYSE ----------- Judgment -------------1 10-year -----Very small ---------------------------BDC ----------- Analysis based on models. 5-yr -------- Medium ------Limited- ------------Do.
BF -------------- Model assisted--------- 5-yr -------- Small ------------- do --------------- Do.
A------------Analysis ---------------- Annual ---------------------- Limited -------------- Do.
B ------------- Model assisted 5 ---------------- do -------------- -High------------- Weak.
C -------------- Analysis -------------- 5-yr ------------------------ Limited -------------- Do.
D.---_--_-_--Judgment------------ 5-yr -------------------------- do ------------ None.
E ---------------.-do...---.2, 3-yr---------------- --- ----------- Do.
F_ --------- Model assisted ---------5-yr ------- Medium---------do -------------- Do.
G ..------------ Judgment -------------- Annual -------------------- do -------------- Do.
H------------Analysis-......-5-yr ........... do Weak.
I -------------- Model assisted 6 --------- Annual -----------------None ---------------- Do.
J -------------- Judgment -------------- -do------------Limited------------- Do.
K ------------------ do ---------do------------------- d-------------- Do.
L -------------- Model assisted --------------- do -----------------Medium -------------- Do.

I Based on number of equations: very small equals 9 or less; small equals 10 to 49; medium equals 50 to 119; large equals 120 to 199; very large equals 200 or more.
2 Based on sector detail: limited equals 2 to 5 sectors; medium equals 6 to 20 sectors; high equals 21 or more sectors. 3 Based on qualitative judgments on pervasiveness of financial variables in real sector equations or impacts and real variables in financial sector equations of impacts.
4 Chase.
C DM1, Wharton, Michigan.

Chase--Michael K. Evans, "Long-Term Macroeconomic Forecast" (Chase Econometric Associates, Inc. June 1976). DRI-The Data Resources U.S. Long-Term Bulletin. "The Economic Outlook 1975-90." Summer 1976. Wharton-Wharton Econometric Forecasting Associates, Inc. "Wharton Annual and Industry Forecasting Model," July 20, 1976.
SSG-Economic Policy Board Special Study Group, unpublished materials partially based on "The Structure of the U.S. Economy in 1980 and 1985," BLS Bulletin 1831 (U.S. Department of Labor, 1975). NPA-Henry Townsend, Timothy Sivia, Mark Kenda II, David Fay and Jessica Townsend, "The Next Ten Years," report No. 76-N-2 (National Planning Association, September 1976). GE- "Economic Prospects: 1975-85" (General Electric, March 1975) and Supplementary materials. INFORUM-Interindustry Forecasting Project of University of Maryland.
NYSE-"The Capital Needs and Savings Potential of the U.S. Economy; Projections Through 1985" (the New York Stock Exchange, September 1974).
BDC-Barry Bosworth, James S. Duesenberry, and Andrew S. Carron, "Capital Needs in the Seventies" (Brooking Institution 1975).
BF-Benjamin M. Friedman, "Financing the Next Years of Fixed Investment," Sloan Management Review, Vol. 16 No. 3, Spring 1975, pp. 51-74.
The remaining respondents include 7 large industrial enterprises, 3 financial organizations and 1 academic institution

Tho basic time unit of the forecasts varies from a ten-year span for
one of them to quarterly projections for three others. Eleven forecasts
are done year-by-year (annually) and the remaining seven are 3-5 year
predictions. The models utilized range from very large with hundreds
of equations to small with nine or fewer equations. As a rule, disagrrecation of production is "limited" to "medium" and interactions of
the financial and real (production) sectors is of weak to medium
strength. Overall, the variety of characteristics of the approaches to
1oner-term forecasting is more notable for its diversity than its uniformity. A priori, those that contain greater structural detail and informai on should be more useful. W hether these also are more accurate
l)redlitions of aggregates, such as overall real GNP growth rates cannot be determined at this time because of the extremely short history
of 1;-, of models for long-term forecasting.
]l[owever, sofmle general patterns emerge on the assumptions made
by dilferent respondents. (See table 6.) Assumptions about fiscal and
monetary policy are especially important because of their potentially


strong impact on growth paths. For those forecasters that provided l lie information, it was found that most anticipated only modest increases in federal spending and no real (constant dollar) tax reduction. Tlherefore, with economic growth, federal deficits are expected to decline from their recent high levels and, in some cases turn into surpluses. Monetary policy is seen as continuing in a tight to acconmmod:tin:f(r vein. Where inflation and unemployment are assumed exogenou- l.lv, both are predicted to fall from present rates and approximate or he below historical averages of the past decade during the first half of the 1980's.


Respondent Federal Government expenditures Tax policy Federal budget position Monetary policy Other assumptions

Chase ------------- No change in spending programs-- No change in tax laws ---------- Continuing high Federal deficits--- Tight monetary policy with double 1. Recession in 1978 -79 vithi UnemL IIdigit inflation in 1978. ployrneiitrate ris ng to more
thall 10 pefcenlt.
2. Negative niet exports ri noniinal
te,nlis in1 1980 8M.
DRI --------------- Tightening of spending; real pur- Personal tax cuts, holding tax to Moderately high deficit but do- If stop-go monetary policy is Damp~ening of State anti local
chases growing less than poten- about 11.V2 percent of income; declining in 1980-85. avoided, short-term interest spending because or hlighier costs
tial output; real transfer pay- effective corporate tax rates rates would stabilize at a rate of borrowing.
ments grow faster thain output. gradually decline with invest- near 6V percent.
ment tax credit assumed permanent at current 10 percent value.
Wharton (stimulative Increase in nondefense purchases Investment tax credit maintained Full employment and greater tax 1. More stimulative policy begin- 1. Corrected version of Census
scenario). of $15,000,000,000 by 1979 and at 10 percent. revenues from higher level of ning in 1977. series No. 2 used in estimates,
maintained at this level through- Social insurance tax rate pro- economic activity lead to turn- 2. Reduction of discount rate and of population and labor force out 1979-85 with peak increase gramed as law now stands for around from deficit position in increase in nonborrowed growth.
of 250,009 Federal employees 1975-85 with rise to 12.3 per- 1976-80 to surplus in 1981-85. reserves. 2. World trade activity assumed
by 1978, declining slowly for cent in combined employee- to grow at average annua I
remainder of period, employer rate in 1981; earn- rate of 5.2 percent during
ing base to rise at rate of 1978-85.
SSG----------1 No ew rogrms.$1,200O per year.
SSG------------- N nw pogrms.1.1975 personal tax cut ($8,000,- Declining deficit to $8,900,000,000, Accommodating, stable --------- 1. Full employment target of
2. Growth in transfer payments to 000,000) made permanent. 1985. 4-percent unemployment rate 00
reflect real income mainte- 2. $6,000,000,000 personal tax cut by 1985.
nance. In 1976. 2. Inflation down to 4 percerit per
3. Grants-in-aid increase less rap- 3. $6,000,000,000 per year in per- annum, 1980-85.
idly than recent past (3.5 per- sonal tax cuts 1977-82 to 3. Average 3-percent productivity
cent real growth). maintain real tax effect, gain through 0985.
4. Permanent 10- to 11-percent
investment tax credit.
5. Corporate profit tax rate lowered to 45 percent.
6. Depreciations allowances increased by 5 percent.
7. Gas tax increase to 7 cents per
gallon from 4 cents.
N PA -------------- 1. Real rise in defense expendi- No significant change in income or Deficit of $81,000,000,000 in 1979; Moderate to accommodating mon- Constant prices for imported petrotures of 2 percent per year; corporate taxes; small increase $21,000,030,000 in 1986. etary policies. leum; modest inflation in other
nominal outlays rise 8.6 per- in Social Security taxes. import prices.
cent per year.
2. Nondefense purchases rise at
2.2 percent real arid 9 percent nominal rates, 1975-86.
3. Transfers increase at 9.5 percent rate and grants-in-aid
at 8.6 percent rate, 1975-86.
GE---------------- 1. Expenditures increase at 10.2 1. Reduction in corporate income Deficits: 1976, $65,000,000,000; 1. Emphasis on contaiiiing infla- 1. Lower growth in labor force.
percent rate, 1976-85. tax rates from 48 to 43 per- 1977, $39,000,000,000; 1978, tion with "real" growth of 2. Lower productivity gains.

2. Defense outlays gain slightly cent, 1977. $27,000,000,000; 1979--85, $12- MI comparable to late 1950's 3. Continuing high unemployment
through 1985. 2. Permanent 12-percent invest- 000,000,000 to $22,000,000,000. to early 1960's: 1973-80, 0.7 rate.
3. Transfers rise to 63 percent of nient tax credit, 1977. percent; 1980-85, 1.3 percent.
outlays, 1977-80, and to 66 3. Special tax treatment for ailing 2. Nominal growth around 9 lperpercent by 1985.1 industries (e.g., railroads and cent per year.
public utilities).
INFORUM---------- 1. Constant real defense expendi ------------------------------------------------------------------------------------- 1. Continuing high inflation.
tures. 2. Steady reduction in unemploy2. Modest increases in nondefense ment rate to level of 5.6 perspending. cent in 1985.
NYSE ------------- Projects deficit ol- ----Assumes no chrange------------ $3,500,000,000; annual deficit No mention
(based on average deficit, 195463).
BO-------1. No net new Federal programs. No change; revenues rise 11.1 $82 000 000,000 initial surplus, Because of fiscal restraint (sur2. Expenditures grow 8.7 percent percent per year (higher infla- 180;' used to offset State and plus), easier monetary policy, per year. tion rate would increase revenue local financing gap of $25,000- lower interest rates than 1974.
3. Grants-in-aid grow 6.2 percent growth; tax elasticity equals 000,000 and increase Fedeial per year for continuation of 1.2). purchases $44,000,000,000. Net
existing programs. surplus equals $13,000,000,000.
4. Transfer payments increase (Note: Offsets not included in 1st
10.9 percent per year for column.)
funding existing laws.
BF---------------- 1. Only modest new spending in- Tax reductions to offset inflation Balance on average during Relatively tight; less rapid crea- 1. Real GNP growth of 3'2 to 4
itiatives. impact on revenues so that 1977 81. tion of bank reserves than in percent annually.
2. Constant expenditure share of budget is balanced, last 10 years. 2. Inflation a ta of 5 percent
GNP (excluding tra nsfers). yea;Ily._3. Transfers grow faster than GNP4. Expansion in real terms con- c
sisterit with real GNP fprowt'n.
A----------------- Expenditures to decline slightly as Coiit.acyclical tax measures------ Declining deficit -------------- Contiracyclical, achieving stabiliza- Unemployrment rate to drop to 5
percent of GNP. tion of interest rates. percent level in 1980's.
B----------------- 1. Goods and services purchases ------------------------------------------Assumes 2 mild growth cycles
in real terms will grow less with cyclical lows in 1978 and
than real GNP. 1982.
2. Transfer payments will increase
C----------------- Real outlays rise at 4 percent, Small net increase in total tax Balance at near full employment Mildly restrictive to neutral 1. Real GNP trend at 3! percent
1970 85; further shift goods and burden with a decline in Federal annually.
services to transfers, individual rates but increases in 2. Inflation rate of 3.1 percent in
corporate arid State and local 1980's.
D --------------------------------------------------------------------------------------------------------------------------------- 1. Annual population growth less
than 1 percent.
2. Labor force will increase at
higher 1 31e because of
changing age mix but will
slow down in 1980's.
E --------------------------------------------------------------------------------------------------------------------------------- Labor force growth between I and
2 percent; declining average
workweek; productivity growth
about 2 percent per year.
F ---------------------------------------------------------------------------------------- ----------------------------------------- Decreasing unemployment rate,
approaching 5 percent by 1985.
See footnotes at end of table.


Respondent Federal Government expenditures Tax policy Federal budget position Monetary policy Other assumptions

G -------------------- Federal spending focused on con -------------------------------- Lower Federal deficits, but no -------------------------------- 1. Moderate recession for 1978sumption sector at expense of surplus. 79 per od.
production side of economy. 2. Continuing rise in inflation rate
ap reaching double-digit level in 978.
3. Slow decline in unemployment rate with floor of 612 percent. H ------------------------------------------------------------------------------------ Deficits to continue ---------------------------------------------- Unemployment rate will average
more than 5 percent but with long-term trend toward full employment
-------------------------------------------------------------------------------------------------------------------- Up to 1978 monetary policy is ac- 1. Moderate global recession in
commodating to rapid recovery. 1978-79. Federal Reserve sticks to money 2. Unemployment rate will decline growth targets and as inflation but persist at higher than accelerates liquidity is squeezed. socia I y acceptable levels.
3. Peak inflation rate at 8 percent in 1978; falling to 4.5 percent in 1981.
---------------------------------------------------------------------------------------------------------------------------------------------------- 1. Persisting high unemployment
during 1976-80, with rate of
7.5 percent in 1980.
2. Inflation rate declines to 5 percent by 1980.
K ---------------------------------------------------------------------------------------------------------------------------------------------------- Concurs with BLS projecUon of
3.6 to 3.7 percent growth during
L -------------------- 1. Fiscal policy geared toward 1. Personal tax cut of $8,000,000,- Deficits will continue at high level Will be aimed at avoiding excessive 1980-85 period.
gradually reducing unemploy- 000 in 1978. through 1982 and then decline expansion of money supply but
ment and increasing utiliza- 2. Other changes in personal and sharply. will accommodate economic
tion of productive capacity. corporate taxes assumed to growth.
2. Health insurance program in- provide offsetting gains and stituted by 1979* expendi- losses in Treasury receipts. tures starting at $4,000,000,. 3. Combined social security tax 000 and rising to $27 OOOp- rate will rise to 12.1 percent 000,000 bbyy 1T,95; partly fi- in 1978, 12.7 percent in 1979, nanced payroll taxes. and 13.3 percent in 1981.
3. Social security benefits will rise Taxable income base will rise about 5 percent annually over to $28,500 by 1985. next decade.
4. Federal Government will start to absorb local governmerd's welfare costs with expenditures starting at around $4,000,000,000 in 1980 and rising to $20,000,000,000 by 1985.

All Federal outlays other than purchases of goods and services (includes transfers to persons, interest payments, grants to State and local governments, and subsidies to Government enterprises).

Most of the forecasters see a generally favorable picture for output, inflation, and income over the next ten years. (See Table 7.) With a recovery from the 1974-75 recession, the median forecast for the ainiual compoun( growth rate for real GN"P is 4.8 I)erceiit for 0 71-80 and 3.5 percent for 19$0-85.19 These rates exceed those of most fiveand ten-year post recession intervals during the post-World-War I period (cf. Table 1 above for partial comparison). The range of real (,NP forecasts for 1975-80 is from 2.6 to 6.5 percent annual rates. Those at the low end, notably Chase, G, and I, anticipate a recession in 1977-78 or 1978-79, which lowers real GNP hiringg tlioso
years and 1980 and their respective 1975-80 growth rates. In all these cases the recession is attributed to reactions to a tight monetary policy which the Federal Reserve is expected to undertake during 1977 in an attempt to lower the rate of inflation.
9 These and other compound growth rates herein are computed between magnitudes of variables at the terminal years. They are not averages over the path.


[In percent]

Real capital formation Real net e~rpor ts
Real GNP Inflation (GNP deflator) (fixed investment) Real disposable income total real ttade'
Respondent 1975-85 1975-80 1980-85 1975-85 1975-80 1980-85 1975-85 1975-80 1980-85 1975-85 1975-80 1980-85 1976-85 1976-80 1980-85 3

Median forecast ... 4.1 4.8 3.5 5.3 5.7 5.0 5.6 7.4 4.2 3.7 4.2 3.3 6.4 7.3 5.1
Chase--------- -------3.9 3.6 4.1 5.6 5.8 5.5 6.1 6.1 6.0 3.5 3 4 3.7 8.7 8.8 8.6
R-------------4.0 4.9 3.2 4.5 5.0 4.0 5.2 7.1 3.2 4.0 4.7 3.2 8.3 9.8 7.0
Wharton -------- ---3.6 4.4 2.9 4.9 5.7 4.0 5.6 7.5 3.8 3.3 3.6 3.0 2.4 2.7 2.1
SSG -------------------------- 5.0 6.5 3.6 4.8 5 0 4.4 6.8 9.5 4.2 4.8 6.0 3.6 3.8 5.3 2.8
NPA 3___-__-_-__-_-__-_-____-___-__ 4.0 4.5 3.5 5.8 5.9 5.6 5.4 7.1 4.1 3.8 4.1 3.6 4.4 5.7 3.3
GE------------------------ 4.1 4.6 3.6 6.8 7.5 6.1 6.0 7.3 4.8 3.8 4.4 3.2
NFR--------- 3.2 4.0 2.4 9.5 8.7 10.3 5.2 8.2 2.3 3.1 3.7 2.5 --------------A --------------------------- 4.2 5.0 3.4 5.3 5.6 5.1 5.6 7.5 3.8 3.9 4.3 3.6
B4-- ---------- 3.9 4.9 3.2 5.7 5.9 5.5 5.0 6.9 2.9 3.7 4.2 3.1 16.8 5.4 6.3
C --------------------------- 4.2 5.0 3.4 4.0 4.4 3.5---------------------------- 4.3 5.2 3.3 -----------------------D------------------4.7 5.4 4.1---------------------- 6.2 8.3 4.3- ----E---------------4.5 5.5 3.5--------------------------4.2 4.5 3.8
F -------------------------- 4.2 4.9 3.6 5.1 5.7 4.5 -- - - -- - - -- - -- - - -- - -- -- - -- -- ---
G ------------------------------------ 2.6 72--------------------7.2--------------------------------------H------------3538----------------------------------------------35383.5-3.8
I ----------- --------- --------.--------------------5.7----------------------------------3.4
J6 ------4.0--------------------5.----------- 4057------------------------------ 69.3
K----------- 74.0 4.3 3.7 5.5 6.0 5.0 5.5 6.5 4. 3. 3.2 3.8
L --------------------------- 4.5 4.9 4.1 5.0 5.4 4.4 7.2 8.4 6-0 3.5 4.1 3.0 9,9 9.4 10.3

See footnotes on p. 23.

Government real expenditures
Real exports Real imports Federal Government State and local
Respondent 1975-85 1975-80 1980-85 1975-85 1975-80 1980-85 1975-85 1975-80 1980-85 1975-85 1975-80 1980-85

Median forecast ----------------------------------------------- $4.3 4.3 4.3 6.3 7.7 5.1 81.8 1.9 1.6 83.4 3.3 3.5
Chase -------------------------------------------------------- 2.7 1.8 3.7 3.7 4.6 2.8 .9 1.7 .1 ').9 2.8 4.9
DRI ---------------------------------------------------------- 5.3 6.4 4.2 7.4 9.0 'j. 8 2.0 1.8 2.2 3.2 3.0 3.4
Wharton 3_ 5.3 4.6 6.0 7.7 10.8 4.7 2.6 3.7 1.5 2. 6 2. 6 2. 7
SSG 6.4 7.6 5.1 8.0 10.3 5.6 2.3 2.1 2.5 4.0 4.2 3.9
NPA --------------------------------------------------------- 3.2 2.6 3.8 5.4 6.6 4.5 2.0 2.5 1.5 3.7 3.5 3.9
GE ----------------------------------------------------------------------------------------------------------------------- 1.4 1.2 1.6 3.7 3.9 3.6
1 N FORU M ----------------------------------------------------- 2.4 1.4 3.5 4.1 4.5 3.8 1.2 1.2 1.3 2.0 2.2 1.8
A -1.3 -2.5 0 .1 -.4 .7
D4 ----------------------------------------------------------- 4.2 4.1, 4.3 5.8 7.5 3.9 2.2 2.9 1.3 3.9 4.3 3.5

2.4 2.5 2.3 5.4 5.6 5.3
F -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------H -------------------------------------------- I ----------------------------------------------------------------------------------------------------------------------------------------- tND
K ------------------------------------------------------------------------------------------------------------------------ 2.1 2.0 2.2 3.0 2.7 3.3
6.1 6.0 6.1 6.8 7.9 5.7 2.1 1.8 2.4 4.0 3.5 4.4

Average ratios of levels. 5 197633-81.
2 StiMtjl-tiVe. 8 1976-80.
3 1975--86; 1975-80; 1980-86. 7 Nominal rate.
4 1975-86; 1975-81; 1981-86; except real GNP and inflation which are 197546, 1975-80, 1980-861 Average of 2 periods medians, and ratios of real net exports to total'trade which are for single years: 1975, 1981, 1986,


A moderate fall in the rate of increase of the overall GNP deflator is cxjpected by all forecasters, with performance improving gradually ,ver dhe next decade. Still, inflation rates remain above five percent; Monly four respondents predict that inflation may drop below five per,ent. IBy way of comilparison, during 1966-75 the GNP deflator inrvieased at a compound annual rate of 5.8 percent. The essential mes.-we in the predictions is that given fiscal policies of mild expenditure ("I %wtlh and tax cuts. accommodating money supply growth, and an ah1-.11nce of external shocks, inflation stays below double-digit levels aI I does not accelerate.
These are favorable conditions for real capital formation and is reflected in the anticipated scenario of the forecasts. During 1975-80, real fixed capital spending grows on average (across the orecasts) abotit 50 percent fater than real total product. This faster rise is attitutable to needs to compensate for low rates of capacity expansion during the past decade. continuing adjustment to higher energy prices, 1anl outlays for pollution control facilities. In some forecasts, much of Ihe ( at h-up) is completed by the early 1980's and thereafter, real invet ment. and real (GINP grow at nearly the same rates. In other forecasts, especially those with recessions in 1977-79, the process, and higher rates of investment than output, continues in the 1980's.
(Q)iiestions of capital shortage are discussed in a following section.)
The relative rise in investment is accompanied by a relative fall in (,0onsumption. The latter is brought about by a slower rate of increase Of: real disposable income than real GNP and is a consequence of a l)progressive income tax structure and tax cut policies. The elasticity of tax revenues with respect to nominal GNP is greater than unity (taxes ri -c more than proportionately to product or income), so that unless tax rates are re(lduced so as fully to offset inflation, rates of growth of rcal disposable income and consumption expenditures fall relative to r.al pre-tax income growth. This is not necessarily undesirable if incon e growth itself depends on pursuit of such policies.
(h)ie oIf the IIore difficult areas to forecast is the foreign trade sector, wlher, developments depend on the evolution of both the domestic and wn1d economies. U.S. experts are anticipated to rise in the years ahead but in ports, under stimulus of recovery from the 1974-75 recession (d growing demands for foreign oil, spurt even faster. Therefore, the hicihly favorable 1975 trade balance is eroded until 1980 under most forests. The proportion of real net exports to total trade averages nearly 61, percent for the ten-year period 1976-85, but faster rising inports reduce the ratio from 7.3 percent in 1976-80 to 5.1 percent fo," 1981-85. Only one respondent predicts a higher ratio of net exports to total trade in the latter portion of the ten-year interval.
The last six columns of table 7 list rates of growth of real Federal and state and local expenditures. Federal outlays grow far more slowly than real GNP under the assumption by all respondents that the governinent will pursue conservative fiscal policies. The rate of growth of stale and local outlays, too, is slowed from that of the last decade due i W voter resistance to increased taxation and bond issues, active reanit on ex)penllitures so as to preserve credit worthiness and potential b)ankliuptcy difficulties such as those confronting New York City and other municipalities, and lesser needs as total and school-age popuIlat ion growth slows.


Most of the forecasts utilize the Ile(1liant )projectiolls of thle 1.S. IM1reau of the Census to extrapolate population growth which., for all age groups combined, is anticipated to increase at one l)recent )per anniun over the next decade. The population above 15 years of ae is anlticipated to rise somewhat faster than the total due to falling birth and death rates. Therefore, the potential labor force grows more rapidly than population and also is swelled by the continuing rise in female labor force participation. During the 1975-80 period, the labor force is augmented, too, by a decrease in the number of discouraged workers who, with a decline in unemployment, again seek employment. These forces are reflected in the median predictions reported in table 8. which show labor force compound growth of 1.9 percent from 1975-80 and 1.3 percent from 1980-85. The lower rate in the latter five years is attributable to slower real growth and tapering of increases in female and discouraged worker participation rates.



Population (percent) Labor force (percent) Unemployment rate (percent) Employment (percent)
Respondent 1975-85 1975-80 1980-85. 1975-85 1975-80 1980-85 1976-85 197640 1981-85 1975-85 1975-80 1980-85

Median forecast ----------------------------------------------- 1.0 1.0 1.0 1.5 1.9 1.3 5.7 6.5 5.0 2.0 2.5 1.5
Chase 2 -------------------------------------------------------- 1.3 1.5 1.0 1.7 2.0 1.5 7.3 8.4 6.2 2.1 1.8 2.3
DRI ---------------------------------------------------------- 1.0 0.9 1.0 1.6 1.9 1.5 5.4 5.8 5.0 2.1 2.8 1.5
Wharton 3 ---------------------------------------------------- 1.3 1.5 1.0 1.6 1.9 1.4 5.3 5.6 4.6 1.9 2.7 1.2
SSG ---------------------------------------------------------- 1.0 0.9 1.0 1.5 1.8 1.2 5.3 6.4 4.3 2.2 2.8 1.5
NPA -------- ------------------------------------------------ 0.9 1.0 0.9 1.6 1.8 1.4 45.6 6.3 45.0 1.6 2.4 1.5
GE ----------------------------------------------------------- 0.8 o.8 0.8 1.5 1.8 1.2 56.5 7.1 65.9 1.7 2.o 1.4
INFORUM ---------------------------------------------------- 1.0 0.9 1.0 1.3 1.4 1. 1 6.1 6.6 5.7 1.7 1.9 1.5
A ------------------------------------------------------------ 1.0 1.0 1.0 1.5 1.9 1. 1 5.8 6.5 5.0 1.9 2.5 1.3
86 ---------------------------------------------------------- 0.9 0.9 1.0 1.5 2.0 0.9 5.9 6.7 75.0 1.9 2.6 0.9
C ------------------------------------------------------------ 1.0 1.0 1.0 1.5 --------------------------------------------------- 1.9 -------------------D ------------------------------------------------------------ 1.0 1.0 1.0 1.6 1.7 1.4 -----------------------------------------------------------E ------------------------------------------------------------------------------------------ 1.5 2.0 1.0 -----------------------------------------------------------Fs ---------------------------------------------------------- 1.0 1.0 1.0 1.7 2.0 1.4 6.5 7.4 5.6 2.1 2.5 1.6
G -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------H ------------------------------------------------------------------------------------------ 1.4 1.7 1.0 5.0+ 5.0+ 95.0 ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------K ---------------------------------------------------------------------- 0.9 ---------------------------------------------------------------------------------------------------L ------------------------------------------------------------ 1.0 0.9 1.0 1.7 1.9 1.5 5.7 6.5 4.9 2.1 2.6 1.7

Average government surplus or deficit 11 (in billions of dollars)
Productivity 10 (percent) Money supply (percent) Aaa bond rate I (percent) Federal(D) State and local (S)
Respondent 1975-85 1975-80 198045 1975-85 1975-80 198"5 1976-85 1976-80 1981-85 1976-85 1976-80 1981-85 1976-85 1976-80 1981-85

Median forecast----------------- 2.3 2.3 2.1 127.0 7.6 6.6 8.3 8.8 8.0 -31.9 -37.9 -25.0 1214.0 14.6 13.3
Chase 2 . . . . . . . . . . . . 1.8 1.8 1.8 6.7 6.4 7.1 8.6 8.8 8.4 -86.3 -88.0 -84.7 15.7 16.2 15.2
DRI ---------------------------- 1.9 2.2 1.7 6.6 7.2 6.0 8.1 8.5 7.7 -31.9 -37.9 -25.9 8.6 10.0 7.2
Wharton 3 ----------------------- 1.6 1.5 1.7 7.9 9.7 6.2 7.4 8.3 6.6 -12.5 -31.7 +6.7 17.5 18.3 16.7
SSG-- 2.9 3.6 2.1 ------------------------------------------------------------ -20.0 -23.5 -16.4 10.4 7.7 13.2
NPA ----------------------------- 42.2 2.2 42.1 7.0 7.2 6.9 9.0 9.4 8.7 -55.0 -70 4 -40.0 17.7 20.4 15.0
GE_____ 2.3 2.5 2.1 8.4 9.4 7.5 5 9.8 10.2 a 9.5 5 -27.3 -33.1 5 -21.5
1.4 1.6 1.2 ------------------------------------------------------------------------------------------------------------------------A ----------- 2.3 2.4 2.1 ------------------------------ 8.2 8.4 8.0 -24.2 -32.4 -16.0 -----------------------------B 2.3 2.4 2.1
2.4 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------2.1 2.3 2.0
2.3 4 -------- 2.2 -------- 7. 0_ 8_._O_ 6. 0 -------- 8.3- 8.8 7.8 -33.0 -40.0 -25.0 ---------------------------------------------- 2.3 2.3 2.4 ------------------------- ---------------------------------- -48.8 -62.0 -35.6 9.9 14.0 5.8

I Average of annual rates of all years in each period (not rate of change). 8 Unemployment rates are 1975-85, 1975-80, 1980-85.
2 Population 16 years and more. 9 Declining during this period.
3 Stimulative scenario; population 15 years and more. if Real GNP per employee (derived).
4 End year of forecast is 1986. 11 Averages of annual surplus or deficit: (D) Deficit unless otherwise specified; (S) Surplus.
6 Estimated from incomplete statistics. 12 Average of 2 periods medians.
6 Middle and end years of forecast are 1981 and 1986. is Surplus.
7 Drops to 5 percent by 1985. Note: Where figures not supplied, information unavailable.

A vca pl)roject lons for memlployinei rates over these intervals :re t;.1) and1 .( ).0 percent. respectively. Chase has the most pessimistic outlook and their n11 pie]ployment rate averages 8.4 percent, which in lalre meai re is dlu1e to prediction of a 1977-78 recession. Respondents F and I I also predict a recession at that time. but their unemiploynient rate forecasts are lower. All respondents forsee that the unemploynelit pi.tuire will improve significantly over the second half of the 197-') (Ie(ade. (despite decliness in rates of increase of real growth rate. Emplovymeillnt growth rates, of course, mirror those of labor force. llellp)loyvlellt. and output. They are more rapid in 1975-80 than 19)o-S5. When they are more rapid (less rapid) than labor force growt h. unemployment falls (rises).
The (declines in unemployment and growth in employment are accompl anied( in most of the projections by a recovery in productivity, as measured by real GNP per employee or per worker-hour. The slowdown in p)roductivitv growth in the early 1970's has been traced to very slow growthI in capital-labor ratios and to entrance into the work force of unusually large numbers of young and inexperienced workers. which tended to offset productivity enhancing factors such as increased education.20 As the demographic mix changes and the young work force gains experience, the effects should be to increase productivity growth rates.
Other factors which held productivity growth down to near-zero levels (setting aside cyclical influences, real GNP per hour of labor grew at only about one percent between 1970-75) were the energy crisis (which made a portion of the capital stock economically obsolete), environmental regulations which forced production cutbacks and modifications of procedures and equipment and, possibly, the tempora ry effects of the wage-price controls program. Adjustments to these factors. while still not complete, should become less important in the years ahead.
High rates of investment relative to employment growth should provide a strong stimulus to productivity in the last half of the 1970's. Changes in the composition of the workforce will help. too. The median forecast for the rate of growth of real GNP per employee (which is slightly greater but approximately the same as per worker hour) is 2.3 percent for 1975,-80 and 2.1 percent for 1980-85. While these are below the 2.7 percent productivity growth rates of 1965-75. they are far better than recent experience and contribute greatly to the revival of rapid real output growth.
The feasibility of achieving the high rates of investment necessary for this productivityv growth depends in part on questions of finance. D)eclininog Federal deficits and decreases in state and local borrowing will reduce delminds on monev markets and thereby ease private access. Most re,-pondlents expect the money supply to grow at a rate of S e ent 1es than the rate of increase of nominal GNP. so some fin.(ancial tilht ness is envisage(d. Long-term interest rates are anticilpated(l to remain around 8-9 percent. which is consistent with a forecast in flation rate of about 5-6 percent and a real rate of return of about 3 percent.
I-ii *lublislhd ma trials prepared for the Interngency Task Force on VI.S. Productivity GrC, "h. Sept1' :hor 1976).
T4 m11dian forecast for nominal GNP growth can he apuroximated by the sum of the rntes for r.al (GNP and inflation plus their croso-products. That is, e.g. from table 7, for 197~ [.1 5.3+ (4.1) (5.: /100=9.6 percent.



Returns to capital and investment have figr-ed promiiel tlv In th e debate of the last few years; oni the po.-siIilit-(.1 Oef -ft a(a1)it ii li4 e The controversy has raged between those who delny that a Amliol a could ever exist and those, who b)elileve serI0ous (Yaps between el i and actual cal-tacity vmnay occur. 'Ille former(,I grou)tp stles tt ii forces will leadl to adj u~tiient of produIlct. labor, an11( capita 1 good. prices so that(enaI and supplY for- calitl will be in eqiiilibriuiii. If capital goods prices rise i'elative to othcr l-lces. then1 rut ('s ol ret-1nih1
fall. Reus to -svin then alIso would fall. The latter group ad IiS
that this may b)e trule. bu"t that tIw te quilibiilm Ini('it occu I at, lev-1s of savings, investment. cap ital-labor r'atios, and ca'(pacities below those ( deemed socially desirable. If the target is a high rate of ,ri-owth ofi output, and unutilized capacity alnd labor force growth are low, thlits requires, a. sumlilig 1)r-(lctivity adx ance cannot diaiiiaitic-a li be spurred,, a hig-h rate, of (rr-owtL of investment.
Inlvestmenht requimeflit s over the next decade a i'e predcel t~d1o be; higrh not only1 for the reasons of a social ta1get' of lowverinl't) 1uneiployinent rates, but, also b)ecauvme of conitinuing1( ene11r reC0InVeI1lof allil augmentation needs, and polluion abatement and enviromneiitald goals. Most of the studies surveyed here show a siinficanltly igher proportion of ()''N1P devoted to Investmlent in 1975-S5 than In 1(,"6(-7,5. (See Table 9.). IDespite- substantial differences in these predicted lproportions, in other (+NP expenditure shares, and in nomlinal and reval WINP growth rates, there appears to be a consensus on a numbei- of points:
TsIsup s iii this section are examined in greater detail in Gary Fromm, "Investnent R~equiremnents and Financing: 1975-5,*" National Bureau of Economic Research, Oc'tober 1975.



History Chase DRI Wharton I
1966-75 1975-85 1975-80 1980-85 1975-85 1975-80 1980-85 1975-85 1975-80 1980-86

GNP growth rate---------------8.1 9.7 9.6 9.9 8.9 10.4 7.3 8.7 10.3 7.0
Inflation growth rt---- --------5.8 5.6 5.8 5.5 4.5 5.0 4.0 4.9 5.7 4.0
Real GNP growth rate---------------------- 2.2 3.9 3.6 4.1 4.0 4.9 3.2 3.6 4.4 2.9
Unemployment rate2------------5.0 7.3 8.4 6.2 5.4 5.8 5.0 5.1 5.6 4.6 A
High-g~ade (Aaa) corporate2 bond rate, new0 issues --------------------------------- 7.1 8.6 8.3 8.4 8.1 8.5 7.7 7.4 8.3 6.8

1966-75 1976-85 1976-0 1981-85 1976-85 1976-80 1981-85 1976-85 1976-80 1981-85

As percent of GNP:
Gross private domestic investment ----15.1 16.7 14.8 17.9 15.8 15.7 15.8 17.1 16.3 17.6
Nonresidential------------------------ 10.2 11.9 10.8 12.6 10.7 10.5 10.8 12.4 11.2 13.2
Inventory------------------------- .7 1. 1 .3 1.6 .9 1. 1 .8 .8 1.0 .7
Residential------------------------ 4.2 3.7 3.7 3.6 4.2 4.1 4.3 3.9 4.1 3.8
Total saig-----------15.1 16.7 14.8 17.9 15.8 15.7 15.8 17.1 16.3 17.6
Business--------------------- 11.0 13.3 12.6 13.7 11.4 11.2 11.5 13.7 13.2 14.0
Personal---------------------- 4.9 5.4 5.3 5.5 5.1 5.5 4.8 3.5 3.9 3.1
Government ------------------- -1.0 -2.7 -3.6 -2.2 -.9 -1.3 -.6 .2 -.6 .8
Federal------------------- -1.4 -3.3 -4.4 -2.6 -1.2 -1.8 -.8 -.5 -1.5 .2
State and lal.5 .6 .8 .5 .3 .5 .2 .7 .9 .5
Ohr--------------.2 .7 .5 .9 .2 .4 .1 -.3 -.3 -.3
Investm ent a i gs- - -- - -- - -- - -- - -- - -- - -- - ---less- -- - -- - -- - ---savings- -- - -- - -- - -- - -- -
GNP expenditures (percent distribution):
Gross private domestic investment--- 15.1 16.7 14.8 17.9 15.8 15.7 15.8 17.1 16.3 17.6
Personal consumption----------------- 62.6 61.9 63.7 60.7 62.7 62.8 62.7 61.9 62.6 61.4
Net exports------------------------- .5 0 .2 -.1 .5 .8 .3 -.2 -.1 -.2
Government purchases---------------- 21.9 1.4 21.3 21.5 21.0 20.7 21.2 21.2 21.2 21.2

1975-85 1975-80 1980-85 1975-86 1975-85 1974-85 1973-80 1977-81

GNP growth rate ------------------------------------------------------ 10.0 11.8 8.2 10.0 11. 1 8.5 9.2 8.9
Inflation growth rate --------------------------------------------------- 4.8 5.0 4.4 5.8 6.8 5.0 4.7 5.0
Real GNP growth rate --------------------------------------------------- 5.0 6.5 3.6 4.0 4.1 3.6 4.3 3.7
Unemployment rate 2 . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 6.4 4.3 45.6 46.5 -------------- 44.3 -------------High-grade (Aaa) corporate bond rate, new issues 2 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 49.8 -------------- 4 7.5 -------------1976-85 1976-80 1981-85 1974-85 1976-85 1974-85 1973-80 1977-81

As percent of GNP:
Gross private domestic investment ----------------------------------- 15.4 15.4 15.4 16.1 14.9 16.4 15.6 15.8
Nonresidential ------------------------------------------------ 11.2 11. 1 11.2 11.2 10.7 59.4 10.9 11.5
Inventory ----------------------------------------------------- .9 1. 1 .8 .7 .5 3. 1 .7 .8
Residential --------------------------------------------------- 3.3 3.2 3.4 4.2 3.7 4.0 4.0 3.5
Total savings ----------------------------------------------------- 15.4 15.4 15.4 16.1 14.9 15.0 15.6 15.8
Business ----------------------------------------------------- 11.2 11.3 11. 1 12.4 10.9 10.6 10.6 10.8
Personal ----------------------------------------------------- 4.7 5.0 4.4 4.7 5.8 4.0 4.6 4.9
Government -------------------------------------------------- -.4 -.8 _. 1 -.14 -1.4 .3 .2 1
Federal -------------------------- -.7 -1. 1 -.5 -1.9 -1. 1 -.2 .3 0
State and local -------------------------------------------- .4 .4 .4 5 -.3 .5 _. 1 _. 1 W
OthOr 3 ------------------------------------------------------- _. I _. 1 0 :4 -.4 0 .1 .2 k-A
Investment less savings ------------------------------------------------------------------------------------------------------------------ 1.4 ---------------------------GNP by expenditures: 7
G ross private domestic investment ----------------------------------- 15.4 15.4 15.4 16.1 14.9 16.4 15.6 15.8
Personal consumption ---------------------------------------------- 62.5 63.0 62.2 62.3 62.0 -------------- 60.6 61.8
Net exports ------------------------------------------------------- .3 .3 .3 _. 1 _. 1 .2 -.2
Government purchases --------------------------------------------- 21.8 21.3 22.2 21.7 23.2 -------------- 22.4 22.6

See footnotes on p. 32.


1975 85 1975 80 1980 85

GNP gRowth rate---. -- - - - - - - - - - - --- - --- - - - - - - - - 9.6 105 8.7

U e l oyNP geo th rate -- - - - - - - - - - - -- - -- - - -- - - - --.. 4.9 4.1
Un m p oy en at "- ---- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -5.7 6,5 4.9
High-grade (Aaa) corporate bond rate, new issues -- - - - - - -- - - - - - - - - - - -

1976 85 1976-80 1981 85

As percent of GNP*
Gross private domestic investment -----.....................----- --- -----------------------_---- 16.3 15.4 16.9
Nonresidential ----------- ------------------ ------------------------------ ---- ------- 11. 1 10.2 11.6

Personal ---- --- ----- ------ - - - - - - - - - - -- - ---- -- - - -- - - - 5.2 5.5 5.0
Government -- - - - - - -- -- ----------- - - - -- - - - --- - - -- -- -- -- - -- .- -1.5 -2.3 -.9
FeleraI.- -- -- - - -- - - -- - - -- - -- - - -- - - ~ - ------ 1.8 3.0 _1.1I
S tate and local - - - - - - - - - - -- - - - - - - - - - - - - .4 .7 .2
o th r . . . . .- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -3. ..1

GNP expenditures (percent distribution):
Gross private domestic investment --- -------- -------------------------------------------------------- 16.3 15.4 16. 9
Personal consumption ----------_--------- ------------------------------------------------------ 61.8 62.8 61.3
Not exports--- ------------------------------------------------------------------------- --------------- .4 .4 .4
Government purchases. ------------------- _--------------------------------------------------------------- ------------ 21.4 21.4 21.4

Stimulative scenario. 6 Includes inventories and business other nonresidential investment.
2 Periods are 1976-85, 1976-80, 1981-5 for Chase, DRI, Wharton and SSG; 1976-85 for GE. 7' BDC model excludes 1.2 percent unallocated resources.
3 Statistical discrepancy plus capital grants received by the United States less net foreign investment. Note.- Results estimated when figures in sources are incomplete or presented in other forms. De* Estimated fromi Incomplete statistics, tail may not add to totals due to rounding; equals not available or not applicable.
6 1 cluJesplan an eqipmet oly.Source: History, Survey of Current Business, January 1976, July 1976.


1. The economy has the ability to neiat, ~licieint savites to
meet investment needs of the next decade, including increased outlays for energy COnversioI, 1pollutionl abatement, andli( capacity
2. To make this possible, Federal expenditures should be restrained so that current high deficits are reduced and government
saving is raised.
3. Individual income tax cuts will be needed to offset a progressive tax rate schedule and limit reductions in real consunier
incomes arising from inflation.
4. Monetary policy should be accommodating and should not
foster but seek to prevent episodes of highly restrictive credit
5. The principal problem is financing increased investment in an
inflationary setting when nonfinancial corporate business exposure to working capital needs are swollen, depreciation falls short of replacement costs, growth in retained earnings is insufficient to fund much higher capital outlays, and relative rates of return are too low and risks too high to attract much greater equity funding.
Because of this financial situation, a mlajoritv of the analyses provide or recommend reduction in corporate taxes so as to raise rates
of return and augment financial cash flow.
The last conclusion holds notwithstanding a stock market recovery and improvement in conditions of equity financing during the past year, recent surges in corporate margins and profits, and the 1976 tax reform act which extended the 10 percent investment credit. These 1)rovide greater internal cash flow and equity capital but. given high investment demands, greater resort to borrowing will be required and debt/equity ratios are predicted to continue to rise. For some companies and sectors these already are at high levels, and both borrowers and investors are exposed to substantial risks of default. If investment can be accomplished only by further weakening of financial structure. many companies may decide to forego capacity expansion even in the face of strong demands for their outputs. Obviously, the situation does not apply equally to all industries. It is most severe for capital intensive sectors whose capital structure already is highly leveraged. whose rates of return are below average, and whose prices or returns (profit rates) are subject to a high degree of government regulation. Transportation. electric utilities, steel, paper, and a few other industries may be particularly hard pressed by finances. demand, and environmental and safety requirements.

Whether policy actions should be undertaken to allevi:lte investment problems confronted by capital intensive industries is -0 clestion that has aroused strong emotions amono those opposed( to lowering ta x burdens of corporate business, others who are as concerned hout thle adequacy of profit returns and the continuing viability of the free enterprise system, and yet others on)osed to high rowth rates and further eOPvironmionta! encraclment. Tbere is no easy nc'wer because any : ,t(itel'A (t solltion depel'l)ds Izreatlv on b"lancing the preferences and


nel of diverse intere ts. Somehow it is necessary to achieve a national consensus on the relative importance of different goals, or at least a weak preference ordering among them, so that policy choices can be made which come closer to maximizing social welfare.
It is true that some apparent conflicts between achievement of different goals can be eased. if not eliminated, by use of a wide range of vaolicv instruments which offset or deter undesired effects. For instance, while high growth may entail greater natural resource use. marginal resource requirements can be reduced by greater emphasis on conservation and recycling, shifts toward relatively less scarce resources, and channeling The first step. of course, is to determine the social goals and targets. This set most likely will include some goals that are complementary and others which are competitive or substitutes. For example, lowering the proportion of the population with incomes below poverty levels is compatible with raising economic growth rates, but may be accompanied. unless otherwise offset, by stimulation of inflation. Other things being equal, the greater the concern about inflation, the less the emphasis that can be given to growth and income redistribution. There are situations in which higher growth and redistribution lead to lower inflation (for example, when there is a large pool of foreign migrant workers that augment the labor supply on an as-needed basis), but these are far less common than the reverse.
The establishment of national priorities and tradeoff rates between them is critical because this conditions the set of policies which can best be used for their achievement. Relative preferences for different goals lead to relative emphases on alternative policy strategies and tools. For any set of objectives, there is a socially most efficient set of policies which has the highest expectation of achieving the goals. Social efficiency is defined in terms of maximization of a social welfare function wherein weight is given to the utility or disutility of policy instruments as well as to the resource and other consequences of policy actions. While it may not be simple to place values on such elements as the degree of intervention or loss of freedom of choice associated with given policies (such as wage and price controls), this should not be viewed as impossible. A variety of techniques, including referenda. surveys, and experiments can be employed to help ascertain social preferences in these and other areas which involve tradeoffs between psychic and material returns.
Unfortunately, little research has been conducted on defining and measuring social welfare functions. In the economic growth field, economists often have simply used an objective of maximization of the discounted value of real (constant dollar) personal consumption outlays. This might be sufficient if all other net positive effects were proportional to this criterion, but this is unlikely to be true in reality.


Growth prescriptions w-81lhich enployv tlis o)biective as the )ai for choice between policyv alternatives are ilikelv to be hiase,| iII the Nevertheless, some useful infornlition can be obtained when such simplistic objective functions are utilized together with models of the e('()noI()Lyv to explore policies to attain feasible productioll-po) ibility growth paths in the neighborhood of this or a similar criterion. Simulations of this nature with econometric models of the Inited States were run for examinations of cyclical stabilization policies about a decade ago and during the past few years, but they have not vet been done in an economic growth context. 2 These stabilization studies and mnultip)lier results from long-run sinmulations lay. however, have a few significant implications for growth policies.24
First, after allowing for a delay of 6-9 months, stimulative nonetary policy can have sizable imnI)acts on real aggregate output. In solme models there is a deeay in these effects after three or four years, while in others they contin le to grow. Tax and expenditure (fiscal) policy is more effective in the initial year but, again decavs in impacts are found after a few years. For all stimulative policies of reasonable magnitude, very little impetus is liven to inflation 'i te first v, tr or two. But, the long-run cumulative impact on price levels is high because period-to-period changes in prices remain positive throughout the 10-year spans over which the simiulations were conducted.
The principal message from these simulations may well be that
policies which are oriented primarily toward stimulation of demand, which may be necessary, may at the same time be insufficient to generate adequate supply. That is, there is a counterpart to Say's law. Just as supply may not beget its own demand, so may demand not beget its own supply. Growth policies that are to be effective probably must be designed to work on both sides of the demand-supply equation.
DI)emand stimulus will, to some extent, create investment incentives to augment capacity. Still, these may not be sufficiently strong to sustain contmuing capacity growth, especially in capital intensive sectors. Similarly, greater job opportunities may induce workers to seek additional training and education but. again, of inad(lequate miagnitude or types to match growth in labor requirements. Research and development, too. may not he carried out with sufficient intensity on a broad scale and in high risk areas so as to promote strong prod110tivity advance. Government regulations designed to counter recession maladies and monopoly abuses may at the same time restrict competition. thereby fueling inflation and inhibiting growth. Other inefficiencies in the U.S. economic system abound, and whatever rate of growth
1'These simulations utilized more coinplex objective functions which admitted Additional arguments (variables) such as investment and government expenditures. inflaion, unemployment rates, capacity utilization, and foreign trade balances. See G. Frommn and P. Taubman. "Policy Simulations With an Econometric Model" (Brookings. Washington. D.C.. 19CS), J. II. Kalchhrenner ind P. A. Tinsley. "On the Use of Fodha(,ek Control in the Design of Aggregate Monetary Policy." American Economic Review. vol. No. 2. May 1976, pp. 349-53, and Albert Ando and earl Pali-h. "Some Stabilization Probileis of 1971-75. With an Application of Optimal Control Algorithms." In Michael Intriligator (ed.), Frontiers of Quantititive Economics.- ol. III (North-Holland, Anmsterdam. 1976).
24'Multiplier paths for various models, including DRI and Wharton, may be found in G. Fromm and L. R. Klein. "The NBER/NSF Model Comparison Seminar: An Analysis of Results." "Annals of Economic and Social Measurement," vol. 5, No. 1. Winter 1976, pp. 1-28.


is to be sought. government actions to reduce them could aid productivity and cut resource waste. (desirable goals in their own right.
In formulating growth h strategies and examining policy alternatives it also is important to pay heed to international considerations. The worbl .z eCOnomyv is bleconmin increasingly integrated, and U.S. dependence on foreign demIand and supply, as indicated by the ratios of exports and ill)orts to ( NP, has doubled over the past twenty years. These trends are likely to continue as foreign nations become larger puirlars of I.S. foodstuffs and manufactures and I'.S. import requirewints for raw materials and petroleum rise. Therefore, it is nece-s rvy to consider the effect of U.S. growth policies on the econonlies of other countries and the feedback reactions this may generate.
Evidence from simulations with the LINK model suggests that an additional one percentage point of growth of the U.S. economy in 1976 might produce increases in world trade of 0.2 percent in 1976 and 0.. percent in 1971.- While the impacts of this higher world trade on gross domestic product (GNP) of various countries would individually be small (the largest being that for Canada which would have a GNP 0.4 percent higher in 1977), the total income effect on other nations could easily exceed the amount of initial growth stimulus from the U.S. This, in turn, would lead to rises in demand for U.S. exports. thereby raising U.S. incomes further.
Oni the other side of the coin, economic and political developments may have either beneficial or harmful long-term effects on the U.S. economy. For instance, formation and operation of effective cartelsOPEC is an obvious example--can dramatically raise crude material costs and drastically alter the set of efficient production and growth possibilities. Customs unions, free trade areas, and other multinational forms of economic coopertaion may either stimulate or retard demand for U.S. exports in the directly affected countries or in other world markets. Instability in international finance payments mechanisms could inhibit both capital and trade flows, thereby limiting foreign income growth and impinging on domestic finance and product. To the degree that U.S. policy aids or abets such developments, this should be taken into account in the formulation of growth strategies.

About a decade ago there were great expectations that the 1970's would be years of strong economic expansion and of rapid advance in U.S. living standards. The record to date surely has been disappointing. The next five years should prove far more satisfying as most forecasters predict a moderately strong recovery from the 1974-75 recession and downward tapering of unemployment and inflation rate. There fter. growth is ex)etrted to be more modest and declines in ineiIplovment and inflation less rapid.
Thelwre are a number of major uncertainties in this scenario, including possible errors in management of fiscal and monetary policy, a potential capital s hortage brought about by poor financial strucure of capital intensive industries, and the effects of shocks from cartels,
SThe'e calculations were performed under the direction of Lawrence R. Klein at the University of Pennsylvania and ae summunarized in the 1976 Economic Report of the President, p. 135.


world food shortages, military or political upheavals, or other unanticipated sources. A successful growth outcome will depend upon pursuit of a galaxy of policies designed to affect both demald and supply and to maintain proper balance between them.
The tools available to formulate such policies and to accurately project growth under different strategies still are highly imperfect andl( in need of refinement. Comnibinations of econometric models and stems (lVnamlic approaches would appear to provide t1ie greatest p)otential for fulfilling growth forecasting and analysis needs. (' nen the great benefits which a stronger analytical capability of growtlth strategies would provide, government Sipport for such research s iould receive high priority. Research on social welfare functions should also be accorded st rong emphasis in order that better glides to national preferences for tradeotfs between goals such as growth, income distribution. environmental quality and so forth, would be available for the formulation and choice of public policies.


JII 1974 1111d 197.) d ie V... (WoDioJiI en-comitered a deeper recession thn hi occurred sinlce World War 11. With the recession came a IFiji rate, of inflation. falteringr 4rowt~h in real output. and unusually severe iin iloymenit. These diflicul I ues all measure ain increase in econotimc iistabiliy as sugg ested vividly by the title of a recent report issued by the Conference Board, The TVide'ninq Cycle.
The cre-ater inistabilitv in the economy appears to be caused by two principal mi-odes of" economic behavior whose existence is not widely recognized and whose cause:-- are onlV poorly understood: the I'- old [:11 c* 1 ()I 41 a i W '. a" I the 1*te C el of (1(I) O1 ic develop ]Yleflt. -I odatief Icycle is a long-term, approximately fit-year, fluctuation in prices, interest rates. employment, and production of capital goods. The signiificance of the long-wave phenomenon to public policies; lies in the. fact th-at, if the-long wave is a real recurring element of tlie national economy, then the Great Depression of the 1930's prol ,-,Yy represented a typical low point, of such a cycle. In the 1970's we are now about fifty years past the Great Depression, and s;Cnls are that the grrowth'l process is again faltering. The question arise's, liz the long wave of underlying structural origin, and is a. severe economic downturn arising from the long wave likely or
The second mode of economic behavior that, is receiving insufficient id telitioll i's the life cycle of economic development. Tfhe life cycle spaiis the lerio(l. of approximately two hundcred years duration, duirnlt' whI'ich Iponul ation anid industrialization grrow,, and are eventually r(estir2imed by a. range of physical, environmental, and social limits. Sivwh limits *nclude risinrr enierpw anid food pirices anid growing social coMpl7exity. Evidence. suggrests th-,t thie United S'-tates ma ,Y be entering a1 trillsition- 1)erio(1 during which 'Yrowth beis to slow. evenutallv
le:Iinato one form of future equilibiii.
Rrhreis economic stailiza tion policy is today predlicated chiefly on i~oaiincr theories of the short-term buins cyIcurn eco
niomnic (levolopmients p)i'o1ably arise from the initeraction of both shortand loii-fermn jinodes of econo i''( bellavior. This paper therefore attimiiips 't describe soime of the principal lonszf-term forces that will iifhue~ie the niat jonal economic over the. next 10 to 30 years. and suggests Some of their impllications for public attitudes and national
OP-wf" o. r M8 i, nn assistant nrrof0qor of management and director. System DYnamicsR Nation4 Md'Iiii~r Pro'ovt : nld Prtife;-sor Forrester Is a Czernieghawen professor of the Al]f:-("! P. Sn iSchool of Mann grenient. Nras ;achusetts Institute of Technoiogy.

policy regarding economic growth. The preliminary recommendations deriving from the paper are summarized below:
(1) Iheo/d/r/o f t/ "t//)l Imod.s of economic bu/aor.-Present economic difficulties may arise from a confluence of three modes of economic behavior: the business cycle, the long wave.
and the life vcycle of economic development. Misattribution of all symptoms of difficulty to the short-term business cycle can lead to adoption of ineffective government policies. Proposed Congressional actions and legislation should be evaluated from the standpoint of which modes of behavior they are designed to address.
and how any proposed policy will affect causes of the several
very different modes.
(2) / pderstandi( th' na/tIhrc of the recenit e (eNio.-The
greater severity of the recent recession may be an indication of the top of a long-term capital cycle, leading to an excess of capital plant. Evidence for suchli an emerging capital excess is seen in the decline of capacity utilization rates. high unemployment rates of college-educated persons. growing debt burden on corporations.
faltering growth in housing construction and machine tools, appearance of excess office space in major urban areas, and other
(3) EL'1dg the tirn of ito ol/ly.- The
long wave of some 45 to 60 years duration may exert a far greater disturbing force on the economy than the short-term business cycle. In order to address this long-term mode. the scope and time horizon of economic stabilization policy must, be broadened
from the present outlook of a few months to several decades.

term business cycle appears to be caused primarily by interactions between inventories and employment. Capital investment does not seem to be a fundamental contributing cause of the business cycle.
Seeing the business cycle in terms of inventories and employment implies that present monetary policies designed to stabilize the business cycle by influencing the incentives for capital invetment may exert only low leverage for control. Alternative bases for stabilization policy consistent with causes of the b,,siness
cycle must be sought.
(5) Ioli'ie (lor d'lln/! ,I f/h f' lonyg wa/.(.---Appro)rite p-l icies for controlling the long-wave mode may differ quite subtantially from presently-advocated economic policies, or even frolii policies to stabilize the short-term business cycle. To the extent that the long-wave downturn is a consequence of overinvestmnt in physical capital. investment tax credits or rapid expansion of money supply by the Federal Reserve may provide little stimulus for investment. Alternatively, the opposite policy of restrainin-gr monetary growth may lower infstion with little contractionarv effect on output and employment. There may also be a possibilitof cushioning the downturn by encouraging the diver.-ion of rsources from capital-excess areas arising from the long wave to 'rea s of capital slinrtam, t sucl as nc(u1'r, pro(lu 't io. 'reared lV the life cycle of economic development. Such policies as dostribe I above are tentative proposals. but they need to be considered if we are to manage the economy thro, uh the canes emlerging from
the long wave.


(6) (Coer/slenerw of high inflation and une ployment.-Unem plovmelt and inflation may coexist if inflation is caused by increase in the maney supply. unem )loyment is caused by reaching a peak in the lo1ng wave. and money supply has little effect on the long-wave fluctuation. simultaneouss high inflation and unemployment may also stemin from a downswing of the long wave superimposed on life-cycle pressures limiting growth, especially if money supply is expanded in an atteImpt to offset the recessionary forces.
(7) Limitations of the Ph/illip s CuOre as a Uide to policy.Much of economic policy is predicated on the Phillips-Curve concept which implies the existence of a stable tradeoff between unemployment and inflation. However, the IPhillips-Curve relationshi) has probably been misinterp)reted as a general relationship between all sources of inflation and all sources of unemployment. Our work thus far indicates that the balance of inflation and unemployment in the economy depends in a complex way on the many modes of behavior in the economy as well as on the governmental policies being followed. This implies that the PhillipsCurve concept is not a reliable indicator for public policy. The Phillips Curve needs to be broadened into a deeper understanding of the relationships between inflation and unemployment deriving from the interactions between the various modes of economic behavior.
(8) Policies for th/w life cycle.-Policies for adapting to the transition region of the life cycle will need to be developed in concert with stabilization policies for the business cycle and long wave. As suggested earlier, use of conventional monetary and fiscal policy may be more inflationary than stimulative in an era of growth restraints set by physical and social forces. New policies will be needed in areas such as energy, materials, agriculture, and water resources.
(9) creased emphasis on policy rather than decision-making.-A policy is an enduring rationale or decision rule that describes how a decision will be made under any possible set of circumstances. A policy can endure until the economic system is better understood and an improved policy has been determined. By contrast, decisions continuously change and are the action of the moment occurring as a consequence of applying the established policy to the changing conditions of the economic system. Too much attention is given to current government decisions and not enough to the background policies that govern the decisions. A clear statement of economic policy, or general decision-making criteria, is lacking in most governmental agencies. More emphasis should be given to underlying policies guiding government decision-making, and less emphasis to isolated decisions.
(10) I creased research into the dynamic of the national economy.-The policy directions summarized here represent only a tarting point in developing improved management of economic behavior. Much additional work is needed to refine the analysis of the business cycle. long wave, and life cycle: to organize additional evidence for the resulting theories of economic behavior; to test the effects of alternative policies: and to disseminate resutilts to a broad public audience. Economic problems confronting


the nation are of sufficient magnitude to mnerit a largre-scale coitimitinent, of government R( & 1) futnding~ to till t~d ~~mbt i'
Such work should be. conducted b~y jtiultiple groups usingdilcliii met hodologies ill order to foster eoi ii pet t I(,)fl b(,Let een groups
and inocase, the likelihood of ueccss. It is im1por tn, h1oNee
that the meth.odologi~es used by capable of (Walingr wi-th loiig-terinl behavior; multiple niodes of chiangre and feed-backi inter-acti1A
between economic, social, psychological, and demiog-raphic ore.
With new efforts we can hope to acieve greater ptiblic ndr
standing of prospects foi- ecolionic gro t aid m~ore effective government, policies for responding to social and econom ic

In 1974 anid 1973 the U.S. economy encountered a deeper recession than had occurred since World W~ar 11. With thie recession came a high rate of inflation, faltering growth in real output, and unusually severe unemnployiellit. Tiese dilficulties all measure an increase in economiic instability. its suggested vividly by the title of a recent report issued
bthe Cionference Board, The 'Widening Cycle.
In response to increased economic instability government, and the public have0 emphasized the sam-Le nionetary and fiscal policies emnjployed in preceding recessions-expalision of money supply and gov e rnment deficit. However, despite apparent effectiveness in the past, 1-4uelm policies today sem unable to restore economic health and achieve a favorable balance between inflation and tineinl) loyin ent. At the sam~e time. opinion is rowingr that the recent recession is signilficantly diffe rent, in cliarmeter friomi prev-ious recessions..
Our revenlt investigations suggest two principal modes of behavior in the national economy as contributing to present difficulties-the Kondratieff cycle. or- long wave, and the life cycle of economic developiient. The existence of these two modes is not, widely recognized and their causes are only poorly ulnderstood. The, Konclratiefi cycle is a longr-term., approximately fiftyv-year, fluctuation ini prices," interest rates, ellployi ient. amid 1)roditiOl of cap~it al g-oods. This longi(-wave phenomenon was originally identified in statistical studies con-ducted in the 1920!S by the Ruissian economist _Nikiolai K*8onciratieti.'- It hias recently beruin to receive renewed attention in YneWspalper articles, financial p)ublicatons. (roNvernihnent agencies, academic literature, and even the popular press.2 PresenIt interest, in the long wave derives f romn a recocrimi tion that. if thle 1loniy wave is a real1 recurringr elem-ient of the liatioiial econioiliv, flten thie G'reat I)pression of the 1 930s 1)rolbably relieseiI ((1 a tylpical low point of such a cyl.Since thle I 930's the coim it as expe(rienceed relativelyv viirorous and sustamnedl ci-owth of economic activity. Butt we are pi'esently about fiftv years past the Great DI emesioln. an1d :itnis are that thie gmiowth Ii rocess; is againn
faltrin. The question arises. is the ]on, wave of undierliN 11( strucetural origin, qind is a severe economic downturn arisingc from the longr
-wave likely or avoidable?
I S ee Nonra tieff (19)85) for I-li Engilsh tr'insiatlon of Konhdr:~ieff"s w,)rk. 8ewo. fo r eN8l 1111do Rostow (1 975a) :p. 12 :Bank Crodit Analyst Mayi 1972 ml 1 T) ol bor 1974) : Levv- 2a 'iAscj 97--) 1 -tow (19751)) Forrester (1975a) : InI(l S11~"1 'I 8! R~'oseam (1972,.


[Ille Zeonld mTodle of VC01oiwic beh.,avior that is receiving insufficient :ittcilt iol is the life cyNcle of economilic development. The life cycle ~pas'he period. of ppxituatci twNo hundred years duration, duilring- which I)opmlit ion and industrialization grow and are. eventuially r1v 1,; 1 171 bN a rans.re phys cal nvironmnental, and social limis Sucih liuuur it1 ,1clude rIiing( ene11rY an1d food prices and grrowingr social (Thlmllexit v. EvidenIce, ugst tat the United states may be entering l!1111 I(. ranit on erid dur11ingz which l(.rOWtll begins to slow, eventually Wa n~tosoefo0rm1 of fulture-( equ Ilibriium.-I
Ibi aper attcnlI)tS. to clarify the causes of, and interaction amongY. Iii rue pt'iiicilpal m1odes of econoic behavior: the short-term business cylthl~ W'ave0 and the life Cycle Of economic development.
XX]~c e~seconom~ ic stail lizat ioni polic v is to~lay predicated chliefly on prtx-u~ngthe(ories Of theP Sho01t-0e 1t business cycle, current economic
Ieveop'ciP robblyarise from the interaction of both short- and
lufL~ei'~1mod'5of ecoromie behaia or. If the identity of the separate
1101 *snot rcog 1i ved. symptoms arising from one part of the system,
niY](,n*-' imste(rpveted andI applied to poli-v contopinsIIole
cvi el A- 11brent part of thie sy stemt. Policy is then ineffective because it is Ni ~sed onl an ,111pe diagnios is, of symptoms that led to adoption cf !lie policy, Iiese,( oblserations indicate an urgent need to expand
1 li 2 ~ 111dyi of w,(onomic behavior and stabilization policies beyond cir. tfoculs oil Ie shor01t-termi busiiess cycle. This paper describes S n )1'pinlcipal lmog-terni forces that wilinfluence the national econoimi over the next 10 to 30 years, anld suggests implications for niat ioual policy regrardling economic. growth.

The viwon~dvlpdin this pepei-- ref-ar(Iincreconomnic behavior and economiic polic,% stem from the systemn dynamics" approach for analyzing~ social systems. This s ectioni give-,s a brief description of sys-tem (lnainics, and outlines; the ma.-in features of the Anethodolog0 ry that suit i1t for analvzingYlong-run economic growth i~e.

IL.A fL'.vrdpf~oim of System IDynaiudr8
System dynamics is a way of combining~ all available information, nclilkiii ur written description. numerical data. and personal experi(bll..Q. Wirthl coinlpltet- sitiiilationl to yieldl a better' understanding of socemf Tlvsisl~' sx-stein (Ivuil es approach starts by constructing
1. 1,110l10 of lloxv Ole r~irnoiis actors in a social system go about making decis~ons. Suchi aI modeI consists of a set of cause-and-effect statements W-(lribn hlow differenT-t circumstances, p~resslures, and motivations influeiie he eciio-maerand lead to action. The apIproach draws hieaiv II*y on descr-iptive information and observation about the infor'la-tin ,_ources available to the (lecision-niaker, andl the way information, is converted into action. A system. dynamics model is designed to
:, Fr m~irptait (1iS(i-Slofl of tlipse i1--ws, ;ep Forreqter (1971) : Meadows Pt al. (19721 Forr -tir (1(975bt~ : ;nd Mladden (i)7-5i. 4'Reade(rs familiar with q systeni dynamics or primarily interested In the behavioral and p)oil,\- n~iect-, of the, pnper rmy wish to skip directly to section L11.


capture the intangible ()I- 11011-111OZIS1111(ld Influellol-es oil real-life decisionmaking, such as attitudes, valtw-,. md exi)ectatimis. in additimi to Incorlmratiliff 11MIS111104 N-ttriablesand mimeri%-al data.
A well-constrilcted system dvilamics model ;As a cal)tive replica of
*i. to I)ortray. 'I'llo, "llo(Oel cIIII be S11111,
the actual system d(,,s*()-11(,(
lated oil a computer to sve If lie 1)('11avior ower time that- al-ises from tile interaction of policies ill dilrevent J)"Ifts of' Ow :- vsfem. III otlier Avol-ds.
SI ys t1w r()1e-, of.
he Computer 111111ates, ol. l0a., t1w IIIaIIN I III
tile system to I've lmw t1ley ]n! J.-itil ope ,jjIo)tljer to M(r patterns of' bel avior. Iji ol)eratimi, a system dviikinlics model should repro(Iluct, intornally the same modes of bellavlor SCIOD il) t1w actual social svSteill and sliould ex1libit tile same, kinds, mf observe(l prohlems, such, as 111flation, Insl'.1bilitv. aildt III().h 1111e)III)IOV111011t. As flie model beh,).vlov is explored iii (](, all. new insiglits emel-ge (,lbout
-ior. Moreover, onee fit(, model exhibitiiicr
(III-111ses of behaN ]I e
havior and difficulties Seen ill t1w. real system. it- call bo, "Iltered ret)(111v to incorporate new or prol)osed policies for (_ovvrnillo, the sv.stelll. Bv shitulating the model -with the new policies, and coml-)arMcr tile re0 havior to the performance of the system wifil Ow
salting b old policies, the effectiveness of alternative policies can be assessed.

TI.B. SYsten? Dynaip;cs twl;es of *' oc;al
System dynamics has been under development at MIT and ekewhe're since 19.:56. The approach has been applied to studyincr keliavior of a range of social systems, spanning from mrporate* to urbaii to (rlol-.al behavior.-5 These earlier st-tidies provide the basis for discussion in Section TIT of the life cycle of economic development. Most recontly, the ATIT System DynamiCs Group. under direction of t1w aiAbors. has bevii workin(y oil developim-Y a comprelmisive model of social and economic chan(re in the United Sfates. The System Dynamics National Model is coniposed or' seven basic secto-111s: prool lict 1011. financial, household. demographic, labor. government. and forel"ll
1 1_ e7l ?I,
trade. These sectors describe the major determinants of productlOll. cons-umptioi), investment, employinent, prices. government policy. balance of pavement and other activities and Didices ()f economic performance. ectors of the Atodel are intervomiected by flows of information, people, Money, "Oods. services, and orders.
The National Model i's far more comprehensive than earlier systell, dviiamics models. The Model is highly detailedd. and c0iltaills ..I o internal spanning from short-ter,
price-setting policies to ("apital invest-1-wilt poliICK's, 1111d lomr-term d-mographic and environmental forces. By encompassiii(r a diven-41of s1hort-term and lomr-term f orces. tile 'Natiollal Model will de(.Ij "viidl
lomr-ran(re, Issile'l-I Of (111CO110111 je (rl-0117th. re, Zojjr(.e 4. % popul'olon.
and capital investineiit. as well "Is xvitll Sliorter-terill oil, Ow
busiiiess wele and economy o, stal)ilization I)ojj(,iVS. Til e det"Ifled stl llcture and ability to integrate IOTI(),-t(,I.lll '111d short-1 e I I I I bellavior z1ve 11("CeSSIMIN71 foi- comprellellsive Iml and f or escrib-my -dternative f UtUTTIS.r)
-,ev Form ter (1961 Forrester (19flf9) Ma-s (1974) Schroeder, SWeeTiev. .1ii(I A.Ifeld (1975) : Forre ,ter (1971) 'Mendows et ,,I. (1972) : MeIdOws 111d 'Mpadows (1!17:.)
d 'Mefldowl, et aL (1974).
",ev Forrester and Nlass W )7 for a 1*1111 r (141SCriPtion of the National lfodel.


i\~t idiitljalsetois f the( i-tiNalMode(l have beenl forinu1i1.I d iLI I -(,(It[ are I10 I( \ )Ai( testedl a 11(1 1-improed. Assern>2j~a at e (10't mto ],(, li full -Natiwnal Mode4 i underwa.
A--cii J Ivof: flr-st VQ$lOf the Ntional -Model NvilI take abouttw
h) U 'M II(O L11(ft e l ll urino(
a-lII) n~aii behav-Ior wvill be iniade availale to a ran,,ec of mdi~ik Idil Il adeIti 11-t it II Bls. Ih pr Ivate sector, and( ge' flm,1,11 flt
r ciii ISII liaf caii he a b~asis for iiilproveinenlt. But (een at the
1111(l~t m. Pa11tial ass-eliil les of the ationl1 MNodel are exhibitnvlr ki 'laiaV( I"* 1 hat r 1ies 1h))rtant (lii-;tsti:4i about 1)111 onal 1)olicy ai('l II (J-1) C I fr eCOnoIiic .i"romth. The res-ults reportedly inl Sectin III of 1.Iis i~peV regMAriif: the causes of the business cycle and the long(w,,ave uijniiaizf tle-e lpreliImiln'y fnd'ingsl. S

ILC. (tv/tr.ic~of ASt, cwf bi~~.
As a prelude to Section III on grTowth issues, this section outlines thie main characteristics of the systemn dynamics inethodlolgvA that
ikeit especially useful for analyzin'r economic growthi issues andI iiitoiimr short-iiin and lono-run behavior. The (liscullssion dr1aws il eitiple frmteKainl o(lel to illust rate major l)')mts:
(1) The sy-stemi dymainics approach focuses onl repress enting
adlaltive behavior of decision-makers, andl requires no ass umpt ions about equilibrium or market clearing. Economic growth and econ1101liC fluctuation are inlherently V (i-sequililbriilni modes of lw~lhavior. I1reliniinary cotn puter simuiilationis of the prodluctioni .,ectoi' of the National Model show, for example, that both) the business cycle andl the long wave may be caused by typ Ni)cal eorp)orate-iaaC111C it policies regarding labor acquisitioni and capital investment as they are influenced l)y inventonries,- order
backlogs, and growth expectations within p)ro(1ucing' sectors.
()A system dynamics model distinguishes sharply between,
and interrelates, stock and flow variables. The model conserves all1 flows of money, br-ders, goods, peoplIe, aii(l liiancial ass--ets nI thle sense that the effects of rates of flow ini changing system levels (stock variables) are all represented explicitly. For example, in the 'Nation-al M.-odel. shipment of capital goods depletes the output inventory of the capital -producingr sect or andl adds to the factor inventories of capital equipment inl reCeiVill~r sector-s. Ill an economic system. sto ck variab~les such as invenitoriesz ind lbackbo!~ Y(l.zecoulple rates of flow such as production andl sh~ipenits.
IRe1)r(-s(it at ion of i nterveiiing stocks varialbles. and their Impact onl (lecision-mnaking.n is essential for portraying changes in economic activity that occur when rates of flow are out of eqjuilib'r~uin. In turn. such lportrayal of dlisequilibritii behavior isi necessa ryT for uinderst anding~ the processes of economic (.rowthI and

(3 *) A sytmdynamics model initegxrates supply and dlemand
co~iideiatio 4 o'r example. the National Model portrays bo'th
})flce ad(1jus:tmients a risino- from imblalanices bet weeni supply and(
V ;1~ .! ; (1 fPfiifor fi rt hor (I n'-'n


demand as well as quantity adjlstments-for example, piroduction-rate changes induced by inadequate inventories or high order backlogs. Stock and flow measures of supply and demand-for example, resource stocks, inventories, and order backlogs on the one hand, and production and consumption rates on the otherare both incorporated comprehensively. Market clearing occurs through both price and availability. Thus, for example, demand for a particular comiunoditv or imterial 1mav be (liscoua(t bLv high price or by low availability, as manifested by shortages and a long delivery delay. In contrast to the integration of supply and demand in the National Model, many economic models focus primarily on supply variables s or on demand variables?. This deficiency of economic models is noted by Carl Madden who states, ". . we must concern ourselves with supply as well as demand.
We must look at long-term relationships of population, food, raw
materials, energy, invention and innovation." o10
(4) The structure of a system dynamics model is derived from
observation of real-life decision-making processes, not inferred from numerical data. For example, the production sector of the National Model contains a rich description of how capital investment decisions in a typical corporation are made on the basis of influences such as average shipment rate, actual and desired output inventory, actual and desired backlog of orders for output, delivery delay for capital equipment, price of capital equipment, capital equipment on order, price of output, long-term and short-term growth expectations, money balances, interest rates, and return on investment. In addition to containing a rich policy structure, a system dynamics model can also include relatioinship)ils which lie outside the range of observed data (such as effects of resource depletion on energy costs), but which may become influential in future modes of behavior.
(5) A system dynamics model can interrelate economic, social,
and psychological variables. Any statement or hypothesis which can be expressed clearly in English about the relationship between variables in a system can be incorporated in a system dynamics model. Various hypotheses about social and attitudinal variables can thus be included as part of the policy structure of a system dynamics model and examined in terms of their behavioral imiplications. For example, the National Model will contain such social variables as the effect of attitudes toward public support of individuals on the evolution of welfare and social -ecurity payments, influences on birth rates and desired family size, pressures for environmental control, and attitudes toward retirement age.
Such issues are important due to the tight coupling that exists between social, demographic. and economic variables. Such interrelationship is seen, for example, in thle growing icon'erns about the future of the Social Security system and the effects of large Social Security obligations on economic growth rates as
population growth slows and average retirement age declines.
SSee, for example. Solow (1970).
,oe. for example, Samoelson (1939), which fostered the modern theories of the 'multiplier" and acceleratorr."
SMadden (1975), p. 11.


(C) A sv-e idyniamics model can incorporate any nonlinear
1,la ilidq)-- tlitexist in reality. Nonlinearities imps physical
re- mitson licion and are influential in causing shifts from one
II ioMI( of1 cr0110111 IQ behavior to another. Stich shifts, manifested
by isluzeiiei,(r prices or increasing' economic instability, can
OcCIIr raIpidy a vnid muist b~e understood if economic policy is to be cetf(ct VIII innticipating future behavior of the economy and forml atiu IIa1ppropriate responses.
7) A svNt cii dynamics model can be used to understand the
(lulalitat ive' behavior' of a system, that is, to discern the various poss4 ible ItIodes of behav1ior01 and how they can be influenced by
clm i inlicies .20so, a system dynamics model can be used
to anticipate sifts iI the doiuiiiit miode of behavior of the socio-ecoflotiic, systen 1. Uniiderstand(1in g how the different modes (If beliavior ari!5(' is critical in diagnosing symptoms of difficulty 1111d developi ng ar-ppropri ate policies. For example, if the economy iS; in recess,-ion due: to long-termn forces. the policies appropriate to a normal business-cycle downturn may be ineffective or even
counterproductive (see Sections III and IV).
(8) A system dynainics model can be used to formulate policy
guidelines" or rules for improving system behavior. Very often, (lecision-niakers. in Washing'ton will rely on a forecast of future 1 ehat~vior as a basis for mn1akin-Iga decision at the present time. Such a process h;as two major difficuilties. First,, in a complex social Sy1stem such as the economy, the presence of random disturbances tends to preclude forecasting the state of the system far enough
ahad to allow time for effective action.' However, while forecasting( tihe, precise condition of the economy at some point in the future may he infeasible, it will preners lv be possible to forecast and understand the general mode of economic behavior. Second.
the, emphatsis onl isolated decision-making at the present time detracts from a broader understanding of hiow policy should be madle in ;general. How should we be reacting to high inflation Or to uinemployvment? A broader focus in government on policiakin.ty is essential as the complexity and interconnectedness of tho na1tional system grow. and as the fulndamentsl mode of behavior of
the system undergoes change (see Section TV for elaboration).
Fi-'mj the above (Incip1 F V sys tem va ncdifferences, in qtriic
ture, and lise can be identifrnd between system dynamics models and the econometric models used widely in eomi lvi.Sever-.ql of
these differePnces are summarized 'briefly below. In particular, comnpared with a system dynamics model.
(1) The structure of an econometric model is derived largely from interpretin f economic theory in the liaht of available data, rather thaqn from observation of rpeel deocion processes:
(2) An econometric model tends to focus- more on equilibrium concentc; nnd bebh, v7ior', as (1istinmuisbod from ada,)ptive disequilibriuim belinuvior:
ST; nf1-fr WWI (1mnhn'47-s th'n mnortnrnep of~ dPed!!Trinc lmr~Yovd T pq" for -frp(linfl (v eiqi~z 'iipr flt' sttPm fl r nrpeliotion. Tnr AppnA'rv F the' prc'-0T1-Oo f ralom (1jstiir1)aflf' is: qliowri to nrehide neeivrate forpcist~ntr of the Nurp
~~(4 ;'. -'s'fj p~ i,- f-.~ .'anl On VhP otber bnnd. if two p~ollee qre enrn,rt,,1 .1, 1..4 i O' j-1,)PrT81t fo r1 'ii'''n Iz nlwnvq less vilnorable r e*'' r,1o, of tl(4 p r i- rp rlot "p 'P PP a~"t un ')jnP "'P on tim qvqtoin.


(3) An econometric model does not adhere as strict ly to thie prinlciple of conservation of flows;
(4) Dynamic behavior inl ail econometric iiiod(,l a ~rises mo,1re I101n. exogenous, inputs and distr-ibtited-lag foritilations t han from exlic it processes of integration embedded in interacting stocks and flows;
(5) The structure of an econometric iiiodel is limited to iiairll
concepts for which numerical data exist rather than allIowinig flill incorporation of descriptive information ;
(6) An econometric model whichi is derived fromt time-series dJata can only incorporate relationships between variables thiat. lie xvitilii the range of observed data and whlichi reflect the past mnode of siystein behavior;
(7) An econometric model tends more, to inlcorp~orate linear approximations of relationships between ivai bkes ratlIier t I an vig general nonlinear formuilations;
(8) An econometric model is used primarily for shlort-termn point foecast in-r raflie tlhai fo f( d( vlpifli of vePW(1ra1 p0' icy crit-0il1 a01r for understanding possible future mnodes of behavior;
(9) The process of constricting an econometric model tendls more to emphasize numerical precision in parameters estimates for shor-t-t erm forecasting purposes rather than identifying struictur-es that underlie long-term change.

This section describes three primary modles of economies behi."'or: the business cycle, the long wave. and thie life cycle of economic (levelopment. In each case, a, mode of behavior is described, an explaination of behavior is given, and broad implications of behavior mode are (uS cussed. Section TV subsequently integrates these viewpoints into a set of policy directions for dealing with interacting modes of behatvior.

TITA. Bu'- ijie s ('yucle

TIT.A.1. BCURUTb~e business cycle is the predominant focuis of (ilissions of ec (onomic instability in the business, press and] economic literature. The business cycle, as noted by Arthuir Burns. is a reemrring fluctuations, of averau'e. four--year periodicity. in emlloymenIt. 011t1)11t, inventories. prices, and capital investment.2 Ninmerous theories of thle bus-,iniess cycle have, been advINfneel. M.Nost of these f-he(ones,'q ineclidhifl. for exaiiipie9 the theories of Pauli Saninulson, Johln Hicks. andl James 1 musenberry, attribute bulsiness<-cycle fliietiiations to fluct-iations in fixed callital inves'tmient. Such thieiories hiave been accel)ted widely, and thiey under] ie ecOtlOlule stabilization policies, incnd(ing moneta rvlolK' wh-Iich focus onl re"'ulatinlr imi estillent oppotii it_ (e .7S uliel o ontrollingr thie ov-erall cycle. Front t he PjrsI )ct i ye of tispper the focus o1 ecolinmic, policy Onl the llisine -s cycle, UIIll the 11nderlvin" capita-investmient thecory of biisiness fliuctuat ions. is two prir'a defectss. First of all, attrihiitinir all oconontic behl.vior to the bs I e* cycle i oinores thle contrilt ion to hellav ior of tilme I oi1Ler-te'rmnii iodes
12 Burnis (19039), p. 14.


aI< th 11gwave d life *cycle of econlomlic dIevelopment. Such
P1'ITrVCrt It t iIiltio ()l f behimvior to unolerlyingy causes may lead to in1 jf)'O~)rate~ O1 cyact 10115;. Sec-ond, sinmulation stuldies with the SysI e I vnawsNational 'Modlel suggest, that capital investment policies aIreot finWeta~ involved III gerating,161 the business Cycle,
kit Mi t a thIat cap~itai investment principally mnderhies longer-term (,oOrIMC NC can, ges,. It appear's that the shiort-termi bus,-iness cycle arises fnoiii the interact ions of inventories, order backlogs, production rate, .IIId luring( and termiination pice.3This alternativ-e theory of the cllises Of the bineiiss cycle in turn calls for a differentt basis for ecoIMIoniic Stabiliza'tion, pohicy.

14i'.iire 1 shows a, compiller simulation of the production sector of the 'National Model, here representing a typical consumner-gcood s pro(h1iig sector of time economyv. For this simulation, capital equipment w within thie sector' is held constant andl production rate is changed by varia-tilons in labor only. .2 monthly, five-percent, random variation is stiporiniposed on an otherwvise-constant incomingr order rate to the sector Suich a test input is used to exhibit the inherent dynamic pertOdic'ities of the sector.

:'BacklooOf : :
orders for output:
110La bor : .. .... .. .. .. . . . . .. . . .. . . .
~Inventor of : :

*.*.. . . . . . . . .. . . . .

90............ ....... ... ... .................................'0 I 2 1 1...

1Li~uREF 1.-Buisiness-Cycle fhwctiatioiis appearing in labor, inventory, and backlog.
II lFigiire 1, the p)roductionl sector generates a sequience of fluctliations typical of thie nornal biiSineSs cycle. Interv-als between. peaks var ivaromnd five years. Relative timing Of backlog, production rate as sliown 1)'y lab~or. aninvnorlr typical of i noltist rial lbehavxior.
Time SIgr~ifiaiice of Figur ie 1 lies in its generation of the lu'IIW~eSS c-ycle witliout variations in consumer income or capital inv-etment. Prices are, no0f (.halVlIlg", leifaitot is conistanit on the average, itrs rates, are ( oiistant. and1( capital investment is niot involved. The cyclic, fluctuat ion in 1igurre 1 has the major characteristics of the busineCss,
'-Se as 19T ; 'Mass (19761); and Firrester (1975) for more detailed treatment.


evele and 111-ises from t1w interaetioll of Lacklog. illvent(wY, pl-odw.tdoii, and eiiiploviliviit. TIii,-; is ilot, to tImt Ow b1i'silless cple
operates Nv thout, i I) fl I Ie 11(-* 114r ()t Ile I ,I ct-i I I t 11(, (-(-()1 j()Ij I N,. Figure 1 does raise flie, questioii. of wfietliev cow uiuei- 111COMO. I I I veA inetit, and moiieltarv clutiiffe are celiti-til 1() t1w gelwn16011 al)(1, cwit 1.()I of the bushiess cycle oi- Nviietlier t1ley 'Iiv.i1wIvI.v 111duced by val.1.116011 arising fl-Oin eillployineilt, ('111d III vellkww: .
The fluctuatiii.r, busine s-,_-vcl(-Iike IK lrivioj- In F1,)-lire I Ill.P41S from policies that coi-itrol eniplo viiwiit III to 111v('11tot-ic ; '.111d
backloas. Such policies teii(I to 411111plify disturl)"111c('s ('111d to c()I1vel.t short-teriii random disturbailces 111to ail 111110,0111al. \v:1ve 111,1t I.Acct,i the, iiatui-al oscillatoi-y cli-iracter ()f the
The, reason for aiiiplificatiou CMd ovvi-.-Jio ot of owl)1o.villent .111d production caii be seoii by traciiig- t1w effects w" aii *11CI'1,1W (1 ill dvillaild (hieoniiiig orders) oil the behavior of a typlicId Iii-iii. Assiiiiie tlult a constant demand has existed foi- coiisi-iiiier gowk -flid t'Pat, fi-oin t!)is equilibrium condition deniand suddeii1v iiici-eas(ls 1,411,glitly (Figuiv The first coIlsequeiiec is aii mci-eflse iii orders. hicivase iii the backlo,(r for output, hicrease, of sliipiiie.iits. aiid reduction (4 Mveittol'y ()f output.. The. increase iii Nickloo- aiid (lepletloii ol" illvellto;.V Collill,110 until inaiiacreiiient. Iias confidence that, the iiew hiirfilei- A g ]('N of bw illv S
is not an aberratioi-I aild mitil additioilal factors of product-ioii (filbor iii this exaiiiple) can be acquired to ificl-etise prodlictMil. Betweell the time, deiliaild hicre-(ises and flit, tliiie that pi-oduct loli rl, e.s to eqilal the. ii(,..w demand, three. tbiii(-Ys occur. First, t,,acklo(r for output iiicretnses to aii midesirably I-ii(rIA level: secoiid, inventory of Olitput is wq)leted below its illitial desired levoI; and t1iii-d. because deiu,(ui,,I is i)ow hicrlier than before, niot-e iiiveiJoiw tliaii ,,it the 1wo-Milill", is iteeded to service flie, higher deiiiaiid., aild ilwi-efon, desh-edt im-em01W rises hiorber tliaii at the beginning. ,As, (I coi-iseqiIeii(.1,e of tfiese chaiigros. when production has risen to equal deil-itilid, the, systein is mit of equilibriuiii. Bticklog for output, is too lilgli. (aiid iiivelltorv I,-, too low. Productioii must. be pushed higher tliaii the iiew deiiiaJ'd to ivdiw( the backlog for ()utput, and to Ilivelitorv 11()t wily to it-,
old value but, up to the iiew higlier desired le*\-el. Wlieii hiveiitol-Y ftlid backlog rea-ch the desired levels, production is apt to be; too 16(rh so that iiiveiitory continues to rise. III titi-ii. as 111ve.1fol-v ijlcIv(1 :es above its desired value, production rate inust, fall Iehm"' 111CO111111(y 011ders to reduce ijivejitorv. Througli such ineclumj* -"Ijls, :1 1,02*111111, fluctuation in production 'rate ii N7vtitoi- aiid cinj)lovineiit, caii 6, (!ejwrafed.
It is from many su(-Ii depletioiis of stocks aiid flie iwed for exccs responses to recover fr()iii tliO, iiiibalailces that fluctitatilig illo(le of the economic systeiii arise. D i stu rbaii (-V's I')1*0pagate through a systeill 1 y changing a stoek fr()iii, a de ,sired It' Nvl, setting-- up (,I disc repalicy between actual and desli-ed coiiditioils, activatilig a policy to st.,iil a corrective sequelice (111d pi-oo-i-essively woi-kilw tlll.()Il(rll 11 C.Isclatle ofsta,es. Time lags iii t1w svA-eni delay tictioji ai-id evelitu.-Illy iliduce Coirectioiis greatere r thaii, tIie*iii1tiatijlg disturbtuice.


"- Production rate

".. Incoming orders
Ol (demand)

Order backlog

Desired inventory of output

Inventory of output

time -0.
FIGURE 2.-Overshoo t in production rate produced by interaction of inventory, order backlog, and production rate.
This preliminary examination of business-cycle behavior suggests that the business cycle primarily involves inventories and employment. Capital investment, although it will show fluctuation induced by the business cycle, need not be a necessary participant in creating the short-term business cycle. As shown later, capital investment policies appear to affect principally much longer-term modes of behavior than the business cycle." Furthermore, the business cycle can exist without inputs from money supply, interest rates, or changes in consumer income. Therefore, monetary policies aimed at diminishing the business cycle through affecting investment may be coupled only very loosely to the primary causes of business-cycle fluctuation, and probably provide little leverage for influence. To summarize, development of enhanced policies for economic stabilization must proceed from
(1) Recognition of the multiple modes of economic change that
influence economic activity; and
"See Mass (1975), Forrester (1975), and Mass (1976b) for more detailed descripti(n ,,f economic behf.ior generated by capital investment. In particular, these works ,-1.gest that caital investment contributes primarily to the Kuznets cycle of some fifteen to twenty-five years duration and to the Kondratleff cycle or long wave.


(2) Recognition of the business cycle, as arisingr principally
from labor acqjisit ion alhi iii Xei'titVI-Iai1i ienit policies
rather than from capital investment.
Section 1L' of the paper expands upon these implications.

1I.13. Loiin7 1J(U'


The long wave is a fltictuation in the economyv of some fifty years between peaks. It is characterized by sharp peaks in economic activity separated by, long valleys of dlepression.
The long wave has been most exteiisively examined 1.) the Ruissian economist iNikobii Kondratieli. Kondratiefi showed that mlany economnic time series in the IVestern econom-ies had fluctuated with lpeaks around forty-five to sixty years apart. These time series. wh-lich include prices. wages. and interest, rates:. exhibited troughrls around 11790. 18451 and 189.5, and peaks around 181.5, 1870, and 1920. Althiough K,,oni(lratieff's time series included (data Only through 192.5. the massive depression of the 1931' s would appear now as a trough of the longr wave lolwncrtepak around 1920.
Kondratieff did not himself propose a complete or sharply-defined set of mnechaniisms that coudI 8ecomnt, for the generation of ai fftv- oar' cycle. However, lie did ar~yiie fltat Hie cycle, arose from the internal structural dynamics of the ecoiiomic sys-temn. He didl so by showing hIoN events sich as maior teclinolofrir-al inni-ovations tiid flucwtua,,tionls in pop~ulat ion crovtb, which were conmionl reirarded as exoz, 'noi is or external caus es of the cycle, could in fact be seen as interiial l)I-o(, c, -ses that interact over time to produce a loncr-wave fluctuation.
Today. man- idiviiduials and research )roiips are shO Vs-n11 (rene1Wed interest in., and concern about, the lono-wa ye phnmnnas severe conTOToi difficulties are beirni encountered nearly fifty year iaftei' the Great Depression. (learly. the long-wave phenomenon woulld be of cyreat relevance to econoinic policy-makers if it is of underlvingrstructural origin. Opinion is strongly divided mn this point in the( business and academic press. Many Peconomists- believe that the occurec of the Great Depression wa-s either aclcidenital or a conlseflulence of mlismanagWeient lbv the F~ederal Reserve, and that the Depression coud have been avoided through more judiciouis use of mnonetary policy.
Our research with the System Dynamics "National ode suggests
that long-period cy click ehvr ("ti arve-- 4 fr-oli jnc to f4 invrileconomviic p rocess;es. and that the cycle i v r' it b e ew)it volla~ le rca ilily through convenitionial moiietfirv ald liscad 1 ol icy. Inprtcl.8b !wave behavior canl arise froml the physical stnilctiire coliect-ilr (.(c]nslilner-Croodis sect ors and capital prodingm sectors. A sihcienit caullse forl a .50-year fluctuation lies in the mnovemlent of people lbetweenisectors, the loncr time to change produictioni capacity of capital sectors. the w\ay capit-,il sectors provide, th eir owni ni ca-pitl :is I factor c { produlction. the need1 to develop excess catpaci t mo cat ili 111 p onIc 4' redlo mand. and the psych~ologrica i-s c itiy Ip fres of' ex pc I t 'ons
that can cause overexpansion in the- cAlpita 1 sectois lb isw theory, of~ the long wave is detailed below. and some evidence is provided to dlemonstrlate the pertinuence of the theory to the curlrenit ecoiuoilnic sitita t ion and as an explanation of the Great Depression.


11. 111. V 11"S () F T I 11: 11WN G. NVA N, E, I III (11*4,(Ill liecte(I production sectors. Eich produetl()Il -ccItil, I ()t, t1w tall(hll-d pr()d1lct loll seckw ()f tile
1),11-allicter4 foi- illvciltol-ies 111(1 the lillic Iv(ptlIVII, to Change pl.,)(IlIctloll tNpical of 11 collstillier (111ralfles
ill(, ()ther typii-,d ()I' a capital equipillelit sector. The cOndunables ector orders capital equipment from the capifid cquipIII(.III (,ctor 'Illd bas labor fl.(\ely av(,6la1)1e ( ill the sense that the (1(,I:i ill ti11,I1_(r N"W'111cies ft w the sector. i z to be cwi,4ant ). 'I'll(, clapit,.Il e(IIIII)IIIelit -:cctor ab o h&, lalwr fre(Ay available kit orders its caplfal k yllpjlwllt as la filctm. of product loll froill its own outpilt. RestrMlti(,'Il (M ;LV,I1ltbIlitN' 01' labor Nvill be include(l iii tLa, future when dw


Orders for

Delivery of Capital equipment
capital equipment

Orders for
Delivery of capital capital
equipment equipment



Avail 1 orders


FIGURE 3.-Two-sector structure of consumer durables and capital eqnipintnt.

15 4,e 'Section 11T.11 for fiirther dowriptlon of the standard production sector.


labor-mobility networks. are addled to Olie Nat lona 1 Mode I1. T il r]eODtrant strutcturie, in whiich the capital sevtor- pirox. de:s its owN-i cai pi ta equipment, implies that ani incr-ease In deniand for noisiner diiivabhes Niotrld catise the conisumer sector to try to increase bothl of its factoris of produictioni. It c-an obtain labor. but NwhJen It \\-,(lits mlore capital equipineit, the capital sector nmlui4 expaimh B11t If tile cllta :m 1 .ctol to expand in balanced manner, it nlee(1s bothl Labor (and ctipitd (1I.-;iMpuits. A "bootstrap" operation is in\volv-ed III NNhice i t li capita 1 sect ol mulst withhold oultpult IToi its cuistom11er ( the Coiisiumler sector ) s o it can expand first- in order to later. ieet thle need's of tile (( )1511 ner >etoi'. Stich an interrelationship of sectors czan create a mtode of 1be l avlor nlot seen in either sector separately.
In Figutre 4 the two-sector industrial. structure shio\\s a l11Li,, fluctuiation. in the capital sector of some fifty years duratioii. The :>iape hlas siinilarities to the classical description of the KondratiefV, wvave ini which steep peaks in economic activity are separately by lirozil \-alleys of depression. The model and its behavior in Figutre 4I coniisttute.- a theory of how the Ksondratieff cycle can be catised.

/.in capitol Output of
150.* sector capital ~ ..4,;.,t..:,,~..4. .

100 ... .......

50. . . ....... . . .......... . . . t . . **.......
-:Output r
in capital sector
0 ~ .... .......
46' ''0' 0 so 90 K~ go 20 130.i i~

F IGURE 4.-Kondratieff cycle appearing in the capital sector.

Although the behavior behind Figure 4 is not yet, completely understood, it seems reasonably certain that the pr-ocessels of prodlc .tin and capital eqlilipmnent proci'ement. and the relat ionshii) bet ween consumer and capital sectors have the potential for p)rodu~cing a ixo-Arat ieffcvcle-hike behavior. Thie mode of fluctuiation- in 14(rure 4 is stronlyI determined iiinternially amid is unstable for s iiai v-a atiois andi bound('ed by nonliniearities for large amplitudes. Such a itiode (rrows qulcl fr'om any tris-onr ditrac and tnsosuanisef Iies
igeic dis tib- e4n ed ositi tef ti s
Ipecially persistent and not easy to influence unless its nature is well enouioh UndOlrStoool to discover any available Polilits of lev'N*a1,!e1. I f sticl amodle exists iii real life, it Is pi-obabie that chlanges- over thle fi ftyMea interval in psychol ogjcal attitudes, propensity to takle riskss aiid efforts to sustain the. upwardl growth phase by ii oneta-v expanlsionl will all tend to accentuate thle fluctuation.


The m ,st basic cause of the fifty-year fluctuation in Figure 4 is si)ilarI to the mode inll Figlure 1 that involved depletion of inventory and then an amp)lified production rate to reestablish internal balance. To illustrate the counterpart in Figure 4, consider the UIT.S. economy at the end of World(l War II. After the Depression and the War, the capital plant of the country was depleted at both the manufacturing and consumer levels. Automobiles were worn out, housing was inadequate. conunercial buildings were old, and production equipment was obsolete. T'lie physic al capital stock of the country was at low ebb. But to refill the depleted pool of physical capital in a reasonable time, like twenty years. required a production rate greater than would be necessarV to sustain the capital stock once the pool was filled. In other words. thIe production rate required to replenish the depleted physical capital in an acceptable period of time was higher than could be sustaied. Moreover, as the order rate for capital in the consumer and goods-producing sectors of the economy began to decline, desired produIction in the capital sector declined, thereby lowering the need for capital equipment on the part of capital producers and accentuating the falling, demand for capital equipment. The capital sectors would consequently overexpand and. then he forced to retrench. : In more detail, the sequences in the long-wave mode, starting from tie depression years at the bottom of the cycle, seem to be:
(1) Slow growth: of their capital sector of the economy;
(2) Graduil decay of the eiitire capital plant of the economy while the capital sector is unable to supply even replacement
: () Initial recirculation of output of the capital sector to its own iniput whereby the capital sector initially competes with its
customers for capital equipment:
(4) Progressive increase in wages and development of labor
shortage in the consumer:sectors that encourage capital-intensive
prod idion and still higher demaxids for capital equipment;
(5) Overexpansion of the capital sector to a capacity greater
than required for replacement* rate in order to catch up on deferred demands;
(6) Excess accumulation of physical capital by consumers
(housing and durables) and by durable manufacturers (plants
and equipment)
(7) Developing failure of capital equipment users to absorb
the outn it of the overexpanded capital sectors:
(8) Sudden appearance of unemployment in the capital sectors
(9) Relative reduction of labor cost compared to capital to
favor a shift back to more labor-intensive production that further
diminishes the desire for new plant and equipment;
(10) Rapid collapse of the capital sector in the face of demand
below the long-term average needed by the economy to maintain
the existing capital stock in the face of depreciation; and
(11) Spreading discouragement and slow decline of the excess
capital stock through physical depreciation.
Such a chain of developments can underlie, as shown in Figure 4, large fluctuations in production and employment emanating from the capital-producing sectors of the economy.



Investigation of this lonr-wave mode is inwaiplete. Yet is is of sufficient. potential importance that eveii preliminary hvi) otlieses are Nxoif 1 sieriols Consideration. Present s l)tolli's il t1ie ecolloilly seeil consistent with the top of a long wave when thle top is viewed as a time of excess capital expansion. Olne exanlle of such n)pital overexpansion is provided by observing that, for the first time since the late 1920's, manv cities have today an excess of office space. A recent article in Fortune gives a good description of the psychological, financial, and speculative forces that lead to successive booms and depression of officeconstruction activity:
The vacancy rate in Manhattan is now up to 18 percent, the highest since 1939 (though still below the peak of 1934). Worst of all1, wvhite-co()llar employment in the city has actually declined since 1969. The basic problem, surplus space, is not going to disappear for ten years or more, even if New York's office jobs start growing strongly again-and the prospects for that are dim. . .
But whereas fluctuations in interest rates and capI)italiza tion can be temIporary, the key in the long run to the capital losses in Manhattan's office market today is the vast aerenge of unrented space and the (nsuing collapse of rental values. Indeed, if the financing of developers had proceeded in traditional fashion during the recent rampaging building boon, the matter of rates and capitalizations would be a minor issue ....
Following the overbuilding that began in the Ite 1920's and ended with the Depression bust, the New York financial community set up in effect a tight "production control," which for nearly forty years prevented a repetition of the fiasco. I)evelopers traditionally borrowed from the banks to finance projects, paying off construction loans with the proceeds from a permanent mortgage obt:'ined from long-term lenders. But traditionally the developer could not get a full mortgage commitment from institutions without pretty firm leases in hand Prom solid, reliable tenants for about 75 percent of the planned space-and he couldn't get all the bank money hlie needed to erect a building without such a mortgage commitment.
Over the years, however, this control weakened, and in the late 160's it all but collapsed. Because of an extraordinary boom in white-collar employment from 1965 to 1969, both interim and permanent lenders thought the demand for new space was a sure thing, even without tenants on the dotted line. They began competing furiously for the right to finance new building. "There was a watering down of requirements in order to get the business," says Robert Schlageter Equitable's senior vice president for mortgages.
Banks for their part became eager to lend without the backup of permanent "takeouts" commitments. In that rich rental market of the late 1960's, moreover, most calculations assumed virtually 100 percent occupancy, top rental rates, and minimum expenses for alterations or concessions ....
The result of this blithe approach was that flood of nearly 80 million square feet of space, 55 million of which reached the market after white-collar emphloyment peaked in 1969."'
The above quotations describe several processes that are probably contributing strongly to present recessionary pressures in the economy arising from long-wave behavior: Changing attitudes towar(l risk; declining fears of major recession and consequent relaxation of debt standards in the face of nearly three decades of vigorous economic growth; speculative construction activity; and retirement of business and financial executives who lived through the Great Depression.
Relaxation of lending standards of banks and other financial inst itutions in the years following World War II has stimIulated massive

N Carruth (1975).

caIital investment and ac iulation of debt on the part of nonfinancial MnrporInt ios. 1inlskY has lecribed these develo)l onents as follows:
:1 the ilid (f World War II the financial system was very robust : banks were n et miiaiingemnft institutions largely holding government debt. private 11s1ie 1s 301arge libldings of government debt and money and few debts,. A similar pilitre ri d for households. Over the post war period financial changes :miulat ed: banks heam liability management institutions when they ran
out )f debt for position making, corporations began to be heavy borrowers as well :s short teri leaders. and household debts grew enormously relative to ijmie. Thlle fifancial system evolved from being robust to being fragile."
Figure .5 presents a variety of measures showing tile increasing
fragilit v of the commercial hnking system in the U.S. since 1950.
suh fira.ilitv probably underlies, to a large extent, the growing incidence of bank failures and mounting concerns for the safety and
protjtability of many surviving banks.


Financial net No default Bought funds 2
worth divided by assess I divided divided by
total liabilities by total liabilities total liabilities

1951 ------------------------------------------------- 7.4 58.3 3.7
1952-------------------------------------------------7.4 56.2 3.9
1953-------------.-------------------.-----------.-.-7.5 55.9 3.8
1954 ------------------------------------------------------ 7.7 55.6 3.9
1955 ------------------------------------------------- 7.8 49.6 4.3
1956. ------------------------------------------------ 8.0 46.6 4.8
1957------ & ...------.-----...-..-.---.--.8.2 45.3 4.8
1958 ------------------------------------------------- 8.1 46.1 4.9
1959 ------------------------------------------------- 8. 3 40.7 5.1
1960-..- ----------------------------.---.----.---.-.8.6 39.8 5.1
1961--.------------------.------------------.-------.8.5 39.9 6.9
1962--. ---------------------------------------------8.3 37.6 7.1
1963 ------------------------------------------------- 7.7 33.6 9.0
1964------- ___ __- . _.._-- __..- __- __-..-.- -- - 7.6 31.3 10.1
1965----- .--.------------..--------------.-- 7.3 27.9 11.6
1966-.---- --.-.--------------------.--------------.7.3 26.0 11 2
1967_____ -----...-.... .. ... ..7.0 26.3 13.1
1968 .------------------------------------------------- 6.7 24.7 14.7
1969 -------------------------------------------------- 6.7 21.4 16.9
1970 ------------------------------------------------- 6.4 22.0 17.1
1971 ------- ----------------------------------------6.2 21.9 17.2
1972------------------------------------------- .92L5
19 2. .. .... .. .. .. .. .: .. .. .. .. .. .. .. .5. 9 20. 1 19. 5
1973------------------.--------.----.-.-----------..-5.7 17.7 24. 3

1 U.S. Government securities, vault cash, and member bank reserves.
2 Large negotiable L.D.'s, other inter-bank claims, credit market debt, liabilities to foreign affiliates, borrowing at Federal Reserve banks, and other miscellaneous liabilities.
Source: Board of Governors of the Federal Reserve System, Flow of Funds Accounts: 1945-72 (August 1973), and 1974 Supplement: 1965-73 (September 1974).

Fiiures 6 and 7 show the increasing burden of private debt. Figure
6 shows, for example, a more than doubling of the ratio of corporate
debt to profits before taxes between 1953 and 1975. As suggested by
Albert Sommers and Lucie 3lau, such expansion of corporate debt
reflects numerous factors, including "increased assurance with respect
to business-cycle risk, particularly in the years since the successes of
the 'new economics.

I MnT k (1974). p. 4.
14 S4mminirs and Blau (197."), p. 30.


IfGURE 6.-Burden of private debt.


Ratio, Corporate Debt
to Profits Before Taxes



2 Ratio, Individual Debt

to Personal income

0 11 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1953 55 57 59 61 63 6567 697 1 73 75

Source: Sommers and lau (1975), p. 23.

[in percent

Fixed Internal Demand Protected
investment funds deposits assets'1
divided by divided by divided by divided by
internal funds debts debts debts

1951------------------------------------- 106.9 14.4 18.8 33.0
1952--------------------------------- 95.8 14.7 18.5 31.5
1953------------------------------------- 112.2 14.2 17.9 31.5
1954------------------------------------- 100.4 15. 3 18. 5 30.7
1955------------------------------------- 90.5 16.7 16.8 29.7
196-----------------------166.6 15.4 15.4 25.1
1957------------------------------------- 110.9 15. 6 14. 7 23.6
1958------------------------------------- 100.3 14.2 14.5 23.5
1959------------------------------------- 93.0 15.5 12.9 23.7
1960------------------------------------- 103.8 14.5 11.5 20.6
191------------------------97.9 14.1 10.9 20.2
192-----------------------93. 1 15.5 10.4 19.7
1963 ------------------------------------- 93. 1 15.0 9.7 19.1
1964------------------------------------- 90. 6 16.1 9. 1 17.4
195-----------------------96.1 16.2 8.3 15.7
1966------------------------------------- 101.6 15.9 7.6 13.7
19G7------------------------------------- 104.3 14.8 7.4 12.9
1968------------------------------------- 111. 1 13.2 6.9 12.0
199-----------------------126.7 11.7 6.7 10.3
190-----------------------131.9 10.6 6. 4 10. 1
1971------------------------------------- 120.8 11. 3 6.0 10. 5
192-----------------------117.8 12.0 5.5 9.7
1973------------------------------------- 128.4 11.4 4.8 8.5

1 Demand deposits, time deposits plus Government securitift.
Source: Board of Governors of the Fedeal Reserve System ,Flow of Funds Accounts: 1945-72 (August 1973), and 1974 Supplement: 1965-73 (Saptember 1974).

78-733-76 -5

Increased debt is today threatening profitability and stability of many businesses. Such funds have largely been channeled into capital investment, as manifested in a very rapid expansion of nonresidential fixed investment since 1953, and particularly since 1965. For example, nnmiiers and Blau note that "the real growth in plant and e'quipmlent outlay mhas been faster since 1965 than it was in the preceding twelve vears.~' At the same time that capital plant has been growing rapidly, however, capacity utilization rates in manufacturing have declined significantly, from 90% in 1965 to around 65' in 1975.20
To suinnnarize thle preceding description, empirical evidence shows:
(1) A growing fragility of commercial banks;
(2) A rising burden of corporate debt;
(3) Rapid expansion of capital investment; and
(4) Declining rates of capital utilization.
Thek long-run economic developments suggest a variety of strong recessionary pressures that need to be understood better in the broad context of economic stabilization policy and the long-wave behavior.
In addition to the evidence cited above, the national economy today exhibits a number of other indications of overinvestment in capital plant: New tanke rs are leaving the shipyards and going directly to anchorage. Aircraft are going into storage. The interstate highway system is nearly complete and another is not needed soon. The condition of the auto industry only partly results from the oil shortage and is partly due to the consumer stock of automobiles having been filled. The financial plight of the real estate investment trusts and the decline in home construction suggest that we already have more housing than the economy can support. Most municipalities have built sufficient schools and hospitals. Finally, college-educated persons, a form of human capital with dynamic characteristics similar to those of physical capital, are in excess as evidence by high unemployment rates.
It is illuminating from the standpoint of the present economic situation to apply the capital-investment theory of long-wave behavior developed in this paper to the circumstances surrounding the Great depression. Although the Great Depression is frequently attributed to had luck or monetary mismanagement, a careful reading of the economic history of the 1920's portrays a strong investment boom leading to overexpansion and a drving-up of new investment opportunities. This history reembles closely the theory of long-wave behavior surgested by our research. In addition, several historical aspects of the Deprowsion, snch as the rapid building of corporate and household debt. increase in speculative construction, and high rate of capital investment. are similar to prevent condition 'n the TT.S. The following
(1)thf1neq derrihirng the 1)0's are from Robert A. Gordon's book, ERoomc Instailty and iroth: The Amrcan Record: 21
The ouitstanding fact about the movement of total capital formation in this decade [the 1920's] is the high level reached by 1923 and the maintenance of this level for seven years. We have here a prolonged period of high-level investment in producer durable goods and construction. . It i. sionif ioant flthiat both producer and consumer durables formed a larger fraction of the GNP during the 1920's than during any period before World War I. We thus have a picture of a prolonged in rextmen t boom, which supported a steady expansion in incomes and consumer demand, and at the same time provided the enlarged capacity necessary to mwet the risingg dunand for goods and services [emphasis added] ..
Ti ro d., p. 34.
I id., p) 15.
21 Gordon (1974), Chapters 2 and 3.


The investment boom and the rise in consumption during the 1920s were accomnpanied by a steady expansion in bank credit, the flotation of an enormous volume of new security issues, and a mounting tide of speculative fever reflected paricularly in the promotion of new enterprises, a boom in real estate, the development of a variety of unsound financial practices, and an upsurge in stock prices that culminated in the stock market crash of 1929 . .
We have already mentioned the importance of construction in maintaining investment in the 1920s . The most important single comIp)oent of new coi struction was residential building, which comprised 40 percent or more of the total through 1920, when a decline set in that lasted until 1933. . .
The decline in residential building after 1926 reflected a number of factors. The high level of construction in the early and mid-twenties eventually permiitted the supply of housing to catch up with demand; . The satisfaction of pent-up demand would by itself have called for some decline in building, and overbuilding made the situation worse. . Thus 1920 marked the peak of another long building cycle. . It is doubtful whether these monetary developments were of primary importance in creating the boom of the 1920s. In the terms used by some business cycle theorists, the "natural rate" of interest was higher than the market rate, and part of the large volume of investment was financed by credit expansion. But the chief reasons for this lay in the nonmonetary sphere-in the developments discussed earlier that made the marginal efficiency of capital high and in the wave of speculative optimism that raised it still higher. In addition, the upwise lending practices of the commercial banks encouraged speculation and unsound promotions, and weakened the banking system's ability to withstand the strains that were to come after 1021. . .
. Houses, apartment houses, office buildings, and hotels were built with almost reckless abandon under the spur of promoters' profits and the ease with which securities could be sold to finance the cost of construction. Banks loaned heavily on bonds and mortgages, without adequate safeguards as to amortization, and later found themselves with "frozen assets" whose values had to be scaled down drastically.
The consequences of these financial developments need no great elaboration. One result was a good deal of real investment that was not justified in terms of long-term profit possibilities. Capital goods were created that were to "hanz over the market" and discourage further investment for a decade after 1929. The banking system was seriously weakened. Many weak business ventures were saddled with a load of fixed charges that could eventually lead only to the bankruptcy court .
It is impossible to give a complete and precise sta.temrnent of the immediate causes of the downturn. Certainly the full explanation of the extent and severity of the Great Depression is not to be found merely in the sequence of events during 1928-1929; we must look at boom [sic] of the 1920s as a whole and at the course of developments during 1930-1933. . .
Although we cannot complete our explanation of the causes of the Great Depression until we look at developments during the 1930s, we can dispose of a number of possible hypotheses as to the major cause of the downturn in 1929. It was clearly not due to an encroachment of costs on profits .
Nor can the downturn be explained by monetary developments. The rise in interest rates was not great enough h to discourage business borrowing: the Federal Reserve authorities were careful not to restrict credit for legitimate business purposes. We have already seen that business was becoming increasingly independent of the banks, and commercial loans did not begin to decline until after the stock market crash . .
. What we have said in earlier sections snges'ts that the following factors were chiefly responsible for the magnitude of the catastrophe that occurred.
1. The exhaustion of investment opportunities resulting from (a) the working of the acceleration principle in industries app)roaehing maturity and (h) the creation of considerable excess capacity, particularly in residential and commercial building.
2. The financial excesses of the 1920s. which at the, same time led to too rq1i a rate of real investment in some industries and created a snnerstrnature in inflated caTnital values whose collapse nfokened the 1)banking system :nd eased both borrowers and lenders to take a pessimistic view of the feasiblility of further investment.

There remains the question: Why was the recovery of the 1930s so slow nd halting in the United States, and why did it stop so far short of full em;iynent? We have seen that the trouble lay primarily in the lack of inducement to invest. Even with abnormally low interest rates, the economy was unable to generate a volume of investment high enough, given the propensity to consume, to raise aggregate demand to the full employment level.
T() sunuarize the preceding quotations from Gordon, it appears that Ihe 1920's represented a periiod of rapid capital accumulation that led to an overexpansion of the capital-producing sectors of the econ11V. 11llch overexpaision caused a severe (diminution of investment Op)Ort iuities a11(1 consequylent stroii recessionary pressures. Gordon arnes that the Great Depression was thus in large measure a result of a mllapse of new capital formation., and that the Depression was neither adventitious nor a simple result of poorly-conceived governinent policy. Gordon's description lends support to the proposition that the long-wave mode is generated from the interplay of internal economic and social forces: in particular, it lends support to the capital investment theory of the long wave developed in this paper.

If we are indeed moving from a condition of rapid capital accumulation to one of excess capital stock both at the industrial and consuoner levels, the implications for business and economic policy are substantial. A downturn of the lon, wave arising from capital overinve-t went would yield retarded growth and long-term unemployment radiating from the capital sectors of the economy. Appropriate governmi:i,,ent policy for dealing with such changes will need to be developed.
Countercyclic monetary policy, discretionary fiscal policy, and other "fine tuning" mechanisms have often been given credit for reduHeinr business downturns between 1945 and 1970. However, one alternative explanation for the decreased severity of the business cycle in tlhe postwar years stems from considering the interaction of the business cycle and the long wave. If the UIT.S. economy is characterized by a short-term business cycle superimposed on a much longer-term c-c1e, then business-cycle expansions would naturally tend to be relatively long, and contractions short, during an upswing of the long capital cycle. Thus, it is possible that the apparent lessened vulnerability of the economy to recession following World War II is attributable to the rising phase of the long wave, with little or no contribution from government policy. This argument is illustrated further below.
Fie-nre 8 shows two sinusoids as stylized representations of the busineZ cycle and lone-term capital cycle.
Fi nre 9 and Fieure 10 are on an expanded time scale and show the simpr)le sinuvoids added together. The numbers give the time in vears for economic expansions and contractions. Figure 9 covers the rising segment of the long wave and Figure 10 the falling segment. Note that the upward thrust of the long wave before the peak in Figre 9 eives business cycles the appearance of having strong and long oxranrions with weak and short recessions. By contrast, in Figure 10. which shows the falling phase after the peak of the long wave, the long-term decline weakens and shortens the expansion phase of the


50yo pro

ye5 period perio


0 0
5 yqo perio



C 49 . .

9 I 9 OP
S I 9 9 1.5 1

.9 I Years

oFIGUR .A o of siuod duin riin par of th lon wave... ..................................


business cycle and deepens and lengthens the recession phase. With no other influences, the superposition of business. cycle on a longterm fluctuation would produce the milder recession since World War II, without relying on post-war monetary policy as an explanation.



I I l


FIU~O-Adto ofg siud duigfiigpr oIh og ae

I i ec i
c, t 3.
I S I I I'
t Years
FIGURE 10.-Addition of sinusoids during falling part of the leng- wave.
Much concern has been expressed about the failure of monetary policy to cope with faltering economic activity during the current recession. The assumption that monetary policy accented for milder recessins in the preceding two decades underlies disappoiatrent in the lack of present effectiveness. But, the explanation may be simply that monetary policy has at all times had little leverage over employmentr and the level of economic activity. I the economy is now at the top of a Kondratieff cycle the more severe recent recession is adequately explained by weakening of the long-wave upthrust that had given preceding business cycles their apparent buoyancy. These comments, and their mplcatins rewardig the possible ineffectiveness of government stabilization policies, are consistent wth Carl Madden's observation that "The era of the 'New Economaics' . is receding. The optimism for 'fine tuning' the New Economics generated now seems in the face of the economic history of the 1960's to be all of a piece with the intellectual hubris of the best and the brightest." 22
But even if monetary policy were effective in dealing with recessionary pressures emanating from the short-term business cycle, there is reason to suggest that expansionary monetary policy may be ineffective, or even counterproductive, for dealing with a long-wave downturn. A long-wave decline is a period of depressed capital investment
2Madden (1975), p. 2.


as a consequence of prior overinvestment. The interval of reduced investment allows time for the capital plant of the nation to wear down toward an appropriate balance with other aspects of the economy. Under conditions of excess capital plant, increasing the money supply will provide little incentive for purchase of physical capital and instead may only feed speculative and inflationary forces. Alternatively, to the extent that monetary policy has any influence on the long-wave behavior, the principal effect may be to prolong the peak of the long wave by stimulating investment in the short run, but causing an eventual steepr decline as even more excess capital plant is accumulated.
As discussed above, existence of a long wave in the economy poses numerous dilemmas for government policy. The long-wave theory provides an explanation for the precipitous depression of the 1930's, makes it unnecessary to invoke government policy to explain either the Great Depression or the milder recessions in the 1950's and 1960's, and accounts for the worse recession recently encountered. Section IV expands upons the implications of the long-wave mode for government policy.

III.C. Life-Cycle of Economic Development
Over most of the past 200 years, the United States has enjoyed a relatively sparsely-occupied geographic land area with abundant natural resources. Under such favorable circumstances, population and industrial activity have tended to reproduce themselves in such a way as to promote cumulative growth. For example, a higher population produces higher births, thereby contributing to further population expansion. Analogously, more capital plant (industrialization) yields a higher output stream and greater output allows more investment in capital plant through increased production of buildings, machinery, and other forms of physical capital. Exponentially-growing population and production have. in turn, generated rising demands for food, energy, and other renewable and nonrenewable resources.
After two centuries of relatively sustained and vigorous growth, the carrying capacity of the Unitel States is becoming stressed. Past growth in population, consumption, energy usage, and pollution generation is beginning to trigger strong counterpressures.
The life cycle of economic development refers to the S-shaped, or logristic, curve wherein growth gives -way to maturity that in turn is followed by either equilibrium or decline. It is from the dynamics of the life cycle that the phrase "limits to growth" emeres. If there were no limit to resources, energy, water, pollution-dissipation capacity, or land area, then growth of population and industrialization could continue indefinitely. iBut at some point growth becomes bounded by environmental capacity.
The life cycle of growth can bo subdividedl into four time phases (Figure 11). First is the period of exponential growth during which growth is uninhibited by proximity to ultimate limits. Second is the transition region lying about halfway up the growth curve, at which point enough resistance to growth is being encountered to move the system out of its growth mode alid into an equilibrium-seeking mode.


Third is an equilibrium region of no growth with population and enolnic activity restrained by environmental capacity. Fourth may be a decline phase such as has usually been encountered by nations and institutions.


16Growth Equilibrium-*

8Production rnito

2 e
I I I I 1 1
1890 19 10 '30 '50 1970 '90 2010 2030
F1GuRE 11.-Life cycle of economic development.

There are signs emerging that the American socio-economic sse is in the second phase-the transition region separating exponential growth from equilibrium. The transition region is at the point of inflection where upward curvature in growth gives way to downward, curvatuire as growth slows and moves toward its peak. The transition region marks the division between accelerating growth and deceleratingf growth.
It is in the transition region that the greatest social and economic stresses occur. In the transition region, the forces opposing growth become great enough to suppress the powerful mechanisms o owth and to change the behavior mode of the system from accelerating exponential growth to a slowing and stopping of growth. When eqi1ibrium has been reached, society will have come to terms with the new condition. It is in the transition region that the pressures are greatest and the maximum rate of change occurs in expectations, attitudes, and values. The transition region is consistent with the

social. environmental. and iiiflationary forces that are developl)in in the ['.S. Tie transition stage is a tile of turbulent change as the system loves out of the growth mode.
Thel litniting forces that arise in thle transition reoi to re- it
growth of population and production caln be divided broadly into two categories: physical and social limits. Each of these categories is discussed briefly below.
Physical limits
(1) Food Shortage.-Food shortage represents one physical lilit which may confront the United States before long. The nloat productive arable land in the United States is already under cultivation. Thus food production in the nation can be expanded in the future either by cultivating relatively infertile and unproductive land or through increases in yield per acre. Expanding available arable land, however, will entail high costs for irrigation, land development, and fertilization. Much higher per-acre yields than at present also appear unlikely. To expand yields significantly, increasing amounts of fertilizer and irrigation would be required at the same time as energy prices are rising rapidly and water tables are being drawn down as irrigation and other claims on underground water supplies begin to exceed the relcharge rate. Agricultural output. and food processing in the United States are heavily energy-intensive operations, requiring nearly nine calories of energy to produce one calorie of food output23 Such energy intensive agriculture will be less profitable or feasible in an era of high energy prices.
The agricultural problem is compounded by the fact that approximately one half million hectares of cropland are being converted annually to housing and other urban-industrial activity. This figure implies that over the past twenty years a land area larger than Ohio has been converted from farmland to cities and highways.24
Rising food needs have also caused agrimltural land to be used more intensively by reducing fallowed acreage. As Lester Brown and others have indicated, reduced fallowing tends to cause lower moisture content in soil and, consequently, lower agricultural yields and potential soil losses to wind and water. The potential seriousness of thle arimcultural output situation is suggested by rising food prices and the fact that the U.S. government had to impose unilateral export restrictions on soybeans and feedstuffs in June 1973, to fight domestic food inflation.
(2) Resource Shortagqe-Natural resources comprise a second important array of physical l limits. The Inited States is presently a net importer of 26 of the 36 "critical" raw material.5 For example, the recent ERDA report. "A National Plan for Eneray Research. Development. and Demonstration." notes that the TUnited States is almost entirely dependent on foreign supplies of aluminum, chromium. : nI
23s Stinhart and Stelnhnrt (1974). p. 112. The enlarie-eonversion fltnre (loted here Is for energy eC:laries input per energy e:1 nlari ontput t of food. In nutrition, the term "food calorie" actually mon.i ures one t1losand Tier! y e(nlri 24 Pinmentel. Dritschlo. Krummel. and Kutzman (1975), p. 760. 25Ewell (1970), p. 43.

alloying elements for steel. Supplies of many materials and natural resources are presently critically low, having been drawn down and depleted through two centuries of exponentially-growing population, consumption, and production. Moreover, the average grades of many resources such as copper are declining as progressively less efficient resources nMu-t be extracted and utilized in production processes.26 The combination of decreased resource availability, rising prices, and diminish1ied resource efliciency threatens to curtail continued expansion in national standard of living.
(3) Energwiy hAw ortage.-The issues surrounding energy shortage in many ways parallel natural resource questions. In 1973, petroleum con-umiption accounted for nearly 50 percent of total domestic energy usage. Natural gas accounted for an additional 31 percent.27 Thus, n arl y 80 percent of total U.S. energy requirements in 1973 were met through depletable energy sources that are in short domestic supply. Imported oil, for example. exceeds 50 percent of domestic oil production -o that domestic oil consumption exceeds production by over onethird.- The natural gas situation is, in some respects, still more severe. In 1969. domestic production of natural gas exceeded additions to reserves by a factor of approximately 2.5.29
The dynamics of energy shortage and resource shortage both develop out of growing production and consumption in the face of finite, nonrenewable supplies. As labor, capital, and other factors of production are used to generate output, nonrenewable resources, including fossil fuels, are consumed in the process. As resources are consumed, the total amount of resources extracted (measured, for example, in tons of resources) rises. Increased extraction, in turn, raises the "real costs" of resource extraction, processing, and distribution. "Real costs" here mean the amount of labor, capital equipment, energy. and planning that must be devoted to extracting and utilizing an additional unit of resources. As total resources extracted rise, mines must be dug deeper, and drilling must be attempted in less accessible locations, in order to obtain resources. At the same time, the probability of discovery declines. Moreover, to the extent that the most accessible and highest quality resources are extracted first, the average grade and quality of remaining resources becomes progressively lower. High-cost recycling techniques may also have to be adopted to sustain the flow of resources. All of these effects imply that more effort and more physical and financial resources must be diverted to the resource and energy sectors even to maintain a constant output. In other words, as a consequence of resource depletion, the production function of the energy and resource sectors shifts downward so that given amounts of labor, capital equipment, energy, and other factor inputs now yield less output of energy and resources. Growth in resource and energy usaire thus becomes progressively harder to sustain and the economy undergoes a "real" inflation caused by falling productivity and declininf extraction efficiency.
The sQirnifieancee of rising real energy costs can be seen in the fact that. for several ve rs now, petroleum discoveries in the Tnited States have been declining due to depletion of accessible and low-cost
'q rontDr*p on Pesonroes and fMan of the National Academy of Sciences (1969), p. 124. I r-, ,rv 1, .HI Prnirt n the vrt ln aon (11)74-). p 7.. Tn<.",-r Pr ,v Proect of tb Pord Foundation (1974b), p. 28.
SAmerican CGa Asc location (1970).


pet 'oiiiii eerves: at the same time thatt demland for and u. ageo of petI'(111 role hve been grOWing.301 rlie combination of falling, I)r-O(IIClio an riingdeiiad has yielded a grap that,inurasldt
(k1l~ in lroveli reserves of petroleum and to high import dlemand.
Vitia 1 1v dicra ep Q eece oil ifl}ottel oil has led to ri:-hig oil
prics. educd dmesic sandrd f fiving, and increase e internal ionial tens ions.
4) !'rodco iffs Rctv'e'nP Physical Lil77ts.-X range of (ldlitioflal physical l imlits could be dcib -o exampiile, oNV(TColillutmnt of pollution generation ca 'pacity; exces'-Ive thermal beat enrtin;alid sllorta ges of water, wood. and othi r reiiewvablle retiiree. purpose, however, is not so much to (leccri)C ill detail the1 functioning of euch. potential physical limit as to point ouit the existence of multiple phys*ical limits of which ener'(v sltorta~re is anl example. The Ipliysical li1, iits all tend to be triggered by rising p)opulationl and industrialization.
Strong tradeoffs can exist between the various physical lirniV to growth. Thus, alleviating the energy short age promotes Turl her growth that causes more pollution, higher demands for food andra materials, and conversion of agricultural land to u rb: in list riAl use. A transfer of pressures can thus occur between limitingr fa('tors so that an energTy shortatie shifts to a mnaterials shortage, a food shortage, or other manifestations of phylc'-dal stress. National policy must therefore be predicated on a broadi~ undlerstandingy of the tr-ldeoffs between physical limits engendered b y alternative policy actions.

Social limits
Growth is simultaneously encountering many limits. So far, world attention has focused on the physical limits of food, pollution, resources, and energy. But other limits lie just beyond the physical limits.
Social limits are already exertingr powerful forces. Thle social forces are related to rising population dlensitv. growingr ii(111strial i7'ltion and capital-intensive production, and.'expandig technology. As crowdingr becomes greater, life becomes more complex. Friction hetween individuals and groups increases. Political conflict intensifies. Inter-group rivalries increase. Overcrowdinlz of physical space leads to public conflict with business, as manifested by disputes over power plant siting, strip mining. and pollution emnissions. Ps vchiolo~rica~l stress, community breakdown, rising crime rates, drug~ addiction, grenocicle. revolutions, and war can all be seen as, manifestations of social stress enroenlered by rising crow\vlinvr andl popular ion denczitv.
Tihe O~~c ~5' .f f-rowt inll~ tel alt -~i .n1t~
nolo2jv :11iso foster social si ves. )re(r-a lnir1n-epnent pro0(1lotiofl todayl 1(eImfl)lia-izes, HLU 11n1,1In (011t11iMit iol to 11ork. 1lda(I to hl "Jelaioln 111rd peonal 5aifcion a1!c-CulP'IUtio
s. ni 1AI-s rod(ticll. (,il ii7d o l ivod h mi
vidiial tron it s-iise o-F pen~onal Contiil ~ii1 ion or ae e.1 iun. el i..~ of dist ru1st afl(1 q1l~!(,4olfl u _YOf fi~le~rtmc of inst litut ionsc tenld to deep 1. 'A_ cow leox teiooaaIsociet v is at the same111 time harder to understands(. mlore difficult to accept. aiid eas ier to disrup11t.
30Energy Poix Project of therFordlFoundlation (i974a), p. 75i.


Thle above discussion has tried to convey the general nature of social limits. Such influences are relatively subtle and intangible, stemming from the interplay of psychological motivations at the individual an group levels, and from societal attitudes and values. Nonetheless it apppears that social limits are exerting increasingly powerful influeW(,es on U.S. society. These limits are simultaneously making consensus on directions for the nation progressively harder to achieve, and making management of the society progressively more intractable.

The United States may now be entering into a period of slower growth arising from the long-term growth restraints that emerge within the transition region of the life cycle of economic development. The life-cycle mode thus contributes to retarded growth, expan(led unemployment, and inflationary pressures focused in the areas of energy, food, land, and materials. In many respects, the growth restraints associated with the life cycle are more social, demographic, and environmental than they are economic in nature. In such a national environment, stimulative fiscal and monetary policy may exert more of an inflationary impact than it contributes to expanding real output. As Carl Madden notes:
. we may be in the early stages of a profound confrontation between the appetites of industrial civilization and the physical limits to growth on the planet. Conventional neo-Keynesian economics alone hardly suffices to cope with any such confrontation. . The United States, even though it is alfiuent, is not the limitless cornucopia of recent popular thought. . Caution should behoove us to concede the possibility that old ideas in economics, such as the Keynesian ideas of forty years ago, may not be safe to be fully trusted. Simple faith in faster money creation and still larger governmental deficits that is generated by the econometric models, say, of the new Congressional budget Office, may only bring uncomfortably queer results of stagflation, of recoveries stopped in mid-passage by renewed inflation and rising, not falling, unemployment.'
The severity and typeg of pressures encountered by the IT.S. in the transition region will certainly depend on future attitudes toward immigration and family size and on technological developments. Regarding population growth, present U.S. fertility rates are below the population replacement rate. But such fertility rates, even if sustained, would lead to a more-than-fifty-percent increase in population before population equilibriumll is attained. Such increase in population would result from the large number of adults of child-bearing age presently in the U.S. as a consequence of past rapid growth in population, especially around World War II.2 In the future, a slower rate of IT.S. population growth would tend to diminish physical and social stress. However, if lower stress and consequent greater well-being lessen concern over population and trigger a resurgence of population growth, then over the long term, pressures wil not be significantly red uced.
Teh hnological changes can also be expected to exert a future impact on e'onomlc growth and well-being. But such changes may not be unilaterally favorable or desirable. If technological developments push
'T: 1n (1T)7 ), Pn. 3. 4.
Fre~k' (10s). This the number of persons born per year depends both on fertility rates and on the age structure of the population.


backiei igr()wti limits lea(Iingo to (011t intled griow0 li th)at additionai 1 rwt 11 n iay qlitiikl swallow uit tI li argin of available sjice 81(1lM11eQ- while elol rl Cr~ 1,1(ae )op)llat ion. Teoll11lo)(6ial
le\'CO~uI~nt iiaY also I elld, as (1isoeussed(: earlier, mnerely t o shlift pr-esSiies [,()IIIon Ot rZKVoli-lilif111.2~ll( lact-oi to ()1? ot her: shell ttIIoltlcies illL Ie SCWith Ii ''o \vI Il" 1)o1)llaIt 101 (li1'I iio with IinilcrasiflO social
-'olnpleXitv. iNat joll I gr)\vIi, policy iumi, iot, li-efore, be pi-IdIcat('( 0il (11 Ahi~le ('N)et tati101 or hope t I at teciniologicai I rowth or tempoi-airy slower lpopIflationi growth INill relieve the growth1-fiilitingc 1)essirs con froit'llug thle nationl. but shwld( Ibe based1 ol a deep;ei' understandingr of interact ions between populatioi-I technological change, Nvell-uein'(r anol the veI arious physical and social limits.
The life Cycle may also interact with the short-terr business cycle and longf wave, to accentuate economic instab-ility. Throughi its inflationary ifluence in the transition region, thle life cycle contributes to rising pricess of goods and various raw materials. Such rising prices. in turn, lead to speculative inventory purchases in anticipation of still further price increases. These speculative purchases tend to buoy demnanol, inducing higher prices and expanded production in the short
P11.But, eventually as production capacity expands, price rises beg(ini to slow, and advance inventory purchases consequently (decline, leadme'r to o erexpansion and ovel:buildinog of inventories. Such excess inventories can be liquidated only through declining production and employment.33 Thie influence of the processes explained above in fostering, economic instab~ility issuggested by a quote taken from the Wall Street Journal in 1974:
A stream of revised statistics from the Commerce Department has led to increased worry among economists that the nation may, after all, be in for a classical case of what's known as an inventory recession. Coming on top of the current slump, this could mean that the economic downturn would be deeper and long, r than has been exlpected.... Most of the accumulation so far this year has come in manufacturing, and most of that i.-; in purchased materials and supplies. Some of the building undoubtedly stems from Luying to beat p~ric,,e increases. Manly firms are hoarding goods that already are scarce or that they fear may bev'1me SO. 11
To sumimarize, then. the life-cycle mode cofltribultes to slackened frrowthi. hiL nemlplovnient, rising prices of eiierowv, agrriciiltural collnnO(it'--. anld mlaterials,, and heighced inshlil IDevelopmlleit
ofgvrnetpolicies to (lea] withl suchb ('!1!!flWilties- mst stei fr-om a deeper 11 ll(t I'st "Ill1in ( of' the( divey-se limlits to !oovt encountered in the trans~it~on re,(ion.i their individual characteOristics, and the ways ;i which they itei'act, anld trade off against one another,

l~x~uiie of 0.)nlici(eS to) alleV;iatet I le oia,1i161 (1.j( I)hI A >1 '- 0 'e()i i~it :~dite todhtV iiiii -t I)I11l)ifl all ilI1A(tlIz 1Iid'IIii () I, sliort-teil~ (Ilv"iiaiiis cl Ole l)i511- cwi. h t ii uis ht )1 u a fifty-yea,_" )~ wave. anId (hle Ii I(t c Vt (1-eo tWO )i 0111 h C'\ 1 ,j a iie utjt-. V I' thle first timle Mi the niat oifs hi.-1l- rvwe mu ta face t le triple coicib'mi v.. of a t iW >wtuia olwa elpe.anl~ Ole prc,--rro> (r Ht
transit ion uc~~~~~~~i~~n1 le t 11 ree ('01111. con 11) 1C le c I~ ) 1'8 jV
~ ~ '~Ii-<'ctiol for 11 Inore 111p1' tt, ''e'(Iiptioll of l;~ ~ j r~~
34 NI791 StrI e J tinA (Anist 11 197-1). Iit.p 24.


and national standard of living unless public unde standing of these behavior modes is increased andl suitable national policies are adopted.
Thie purpose of this paper has been to expand awareness of the Ionger-term economic and social forces that will become increasingly influential over the next ten to thirty years. Failure to recognize these long-term forces can lead to governmental actions that are ineffective or which make matters worse. The result would be an intensified sense of public futility and increased dissatisfaction with government.
The issues raised here are based on work that is still in progress. We believe that much additional light can be shed on the issues as continued research into the dynamics of the national economy yields a deeper examination of the sources of unemployment and inflation, and the channels connecting government policy and private action. Even at the present time, however, the preliminary insights surrounding the business cycle, the long wave, and the life cycle of economic development, seem sufficiently important and persuasive to merit serious discussion and debate in this Committee and in the public forum. The time that will be required to achieve public understanding of longterm issues and to form a subsequent legislative consensus is sufficiently long to merit exposure of the issues to a broad audience even at a preliminary stage. Therefore, the following points summarize the principal implications and policy directions that can be derived at the present time from our analyses of long-terin economic growth.
(1) Recognizing the multiple mode,? of economic behavior.-Present
economic difficulties may arise from a confluence of three modes of economic behavior: the business cycle, the long wave, and the life cycle of economic development. Erroneously attributing all symptoms of difficult to the short-term. business cycle can lead to incorrect diagnosis of causes and to consequent adoption of ineffective government policies. Congressional debate on economic policy should recognize these three modes of behavior. In particular, proposed actions and legislation should be evaluated from the standpoint of which modes of behavior they are designed to address, and how the proposed policy will affect each mode.
(2) Understanding the nature of the recent recession.-The greater severity of the recent recession may be an indication of the top of a long-term capital cycle, leading to an excess of capital plant. The recent recession may thus be of a substantially different nature than other recessions since World War II. Evidence for an emerging capital excess is seen in the decline of capacity utilization rates. high unemployment rates of college-educated persons, growing debt burden on corporations, faltering growth in housing construction and machine tools, appearance of excess office space in major urban areas, and other symptoms. In the falling phase after the peak of a long wave, business-cycle downturns would tend to be relatively long and expansions brief, whereas the opposite relation held during the long-wave upsurge. The causes of the recent recession and possible future recessions must be understood in order to formulate correct policy responses.
(3) Expanding the time frame of stabilization policy.-The long wave, of some 45 to 60 years duration, may exert a far greater disturbing force on the economy than the short-term business cycle. Such a long-wave mode can arise from typical capital investment policies within the economy. Thus government policies affecting interest rates

an(d illvest tIent incentives may exert their primary effects oIn tIle (l)Igwaie mode rather than Oil the business cyele. The scope and tiie lorizoi (t' economIc stabilization policy 111tst be broadlened froill t 1 i enIt o0lt look of a few imontlhs to several decades.
(4) I'umstanding M] h u.ws o thc bUcics (e/.----le shlorttern business cycle appears to be caused primarilyy by ilteractions between invento ries and elmploylienlit. ( pital ilnvestiieint does not seem to be a fundamental conitribut1ing (tltse of the business cycle. Seeing the business cycle in terms of inventories and employment ihiplies that present monetary policies designed to stabilize the business cycle by iIntitiencing the incentives for capital ill vestileit IIa"IV exert only low leverage for control. Alternative bases for stabilization policy consistent with causes of the business cycle must be sought. New policies should attempt to dampen the amplifying eirffects of inventory-management policies on production and employimelt (see Seetion III.A). To this end, new policies such as rules requiring businesses and households to maintain money balances in proportion to the rates of flow through those balances should be considered.
(5) Policies for dealing with the long wave.-Appropriate policies for controlling the long-wave mode may differ quite substantially from presently-advocated economic )policies, or even from policies to stabilize the short-term business cycle. To the extent that the longwave downturn is a consequence of overinvestment in physical capital, investment tax credits or rapid expansion of money supply by the Federal Reserve may provide little stimulus for investment. Alternatively, the primary impact of expansive monetary policy might be to defer, and eventually accentuate, the long-wave downturn. It is possible that the opposite policy of restraining monetary growth may lower inflation with little contractionary effect on output and employment. There may also be a possibility of cushioning the downturn by encouraging the diversion of resources from capital-excess areas arising from the long wave to areas of capital shortage such as energy production created by the life cycle of economic development. in other words, resources may need to be diverted to the energy sector in the future, even to maintain a constant production rate in. the face of depleting fossil-fuel supplies. Such diversion can be accomplished successfully if we anticipate the impending capital excess and apply deliberate pressures to reallocate manpower and other resources. Alternatively, if unemployment radiating from the capital sectors is dealt with through increased welfare and uifleimploynlient coilnpensation, reallocation of manpower will be effectively slowed. Such policies as described above are tentative proposals that need additional study, but they need to be considered if we are to nmaunage the economy through t Hie clhanges emerging from the long wave.
(6) Coexistence of high inflaItion and u nlu plo Ioinent.--An explanationl for the existence of siiultanieous lhi~h inflation anId unemplovment in the U.S. lies in the three supernipoeIed economic modes. First, unemployment and i inflation may coexist. if inflation is caueI by increase in the money supply, unemployment is caused by reaching a peak in the long wave, and tioney slpply has little effect on the longwave fluctuation. Simultaneous high inflation and unemployment may also stem from a downswing of the long wave superimposed on life-


cycle pressures limiting growth, especially if money supply is expanded in an attempt to offset the recessionary forces. Such issues need further analysis and elaboration.
(7) Lmita'tion of the Phillips Cu're as a guide to polly.-Much
of economic policy is predicated on the Phillips-Curve concept which implies the existence of a stable tradeoff between unemployment and inflation. Yet such a stable relationship between inflation and unemplovment rate has seldom been observed in the real economy over long periods of time. The simple tradeoff concept is contradicted, for example, by the incidence of simultaneous high inflation and unemploymnt (see point #6 above). Perhaps the most serious shortcoming of t he Phillips-Curve relationship is its interpretation as a general relationship between all sources of inflation and all sources of unemployment. Our work thus far indicates that the balance of inflation and unemployment in the economy depends in a complex way on the many modes of behavior in the economy as well as on tlhe governmental policies being followed. For example, a simple Phillips-Curve relationship probably applies to wage changes, cost variations, and employment fluctuations that go on within the dynamics of the shortterm business cycle. However, changes in money supply or changes in the position of the economy relative to the long-wave fluctuation will tend to cause shifts in inflation and unemployment that cannot be described in terms of simple movements along a fixed tradeoff curve. This implies that the Phillips-Curve concept is not a reliable indicator for public policy. The Phillips Curve needs to be broadened into a deeper understanding of the relationships between inflation and unemployment deriving from the interaction between the various modes of economic behavior.
(8) Policies for the life cycle.-Policies for adapting to the transition region of the life cycle will need to be developed in concert with stabilization policies for the business cycle and long wave. As suggested earlier, use of conventional monetary and fiscal policy may be more inflationary than stimulative in an era of growth restraints set by physical and social limits. New policies will be needed in areas such as energy, materials. agriculture. and water resources. For example, such new energy policies may include: encouraging lower energy use; reduced emphasis on expanding energy production as a means for balancing energy supply and demand; imposition of a high tax on consumption of nonrenewable energy; and increased efforts to okerstand the interrel-ationships between ener gy sihortage and population and industrial growth.35 Comparable policies will be needed in the other major areas of social and economic activity related to the life cycle.
(9) Increased em phasis on policy rather than decision-makn.Government actions in the next decades will need to be predicated in a strong policy basis that comes out of increased understanding of national economic dynamics. Such a policy focus is lacking in government at present. Many government actions are frequently termed "policy." For example, at a particular time Federal Reserve "policy" is to expand the money supply at a certain annual rate. But such a use of the term policyc" is a misnomer. A policy is an enduring rationale or decision rule that describes how a decision will be made
rSe Forrester and Mass (1976) for detailed discussion of Qnergy policy options.


l~u~d' UIy ps~i~eset of circumstances. A pol icy can l i-er tinil i to ecoiiolinic ssti is better ujiderstood and -(u imsproved1 policy 11tasbei detvcu ried. Byv coni ast, dccisidons costntios ly cli 111e: tt e 1s
-tcL ion of the momiienit occuVViIJ a CO(~U.1C It Lpp \' iv 1 lie tabhlied policy to the chaningiIon~ditionls 01f th le COjiiliir>z According to this viewpoint, ,in annomiicciiieit 10y I lie lede 'ral I~rv to) expand the money supply tt say, 6%" p UWol tid [ )e 1(Th>I( )11 rather than a policy. The unii d,,rlyig poIicyV wo Uld Kae Io I )v le fliley supply should be iiianaged in general in response to iniLation 1.11, Uemploymient rate, and other cond~itionls. Buit a CICIar1 St tC11mt CI c nonuc policy is lacking in most go)vernmenCt a~rellcb>. Mlorke ('11ph"--IS needs to be given to the underlying policy guiding gvrn~i deci1sion-mnaking, and less emphasis to isolated decisions. In other-i WOPrds. how should we be responding to inflation, i i-inploinent, and W1IWI!.
dificltisnot just at present, but in general? Are gcneraiized policy responses to inflation, such as monetary policy, effective in dealing~ with inflation arising from any cause? Will the same kinds of policies for dealing with normal business-cycle recessions be effective i31 corn1batting a downturn of the long wave? Such questions can be add ues-'.-ed through a broad focus on economic policy. A more sound and clearlystated set of government policies will be needed in the future to gjiide government action and to prevent the misinterpretation of symptoms that can occur from confusion between the diverse modes of zo(-i oeconomic behavior.
(10) Increased research into the d!/nalr~cs of the 'a at m ,'i cr
omy.-The policy directions summarized here represent only a, start ing point in developing improved management of economic beli,-vior. Much additional work is needed to refine the analysis of thoe bilsilless cycle, long wave, and life cycle: to organize additional evidencev for the resulting theories of economic behavior; to test the effect-s of alt erna)tive policies; and to disseminate results to a broad public audience. Economic problems confronting the nation are of sufficient. m'.mcilit 11(1e to merit a large-scale commitment of government R, &, T) funn ii)Q t o understand them better. Such work should he conducted by mm-iipiile groups using different methodologies in order to foster conetif ionl between groups and increase, the likelihood of success. It is inmTOrta-nt. however, that the methodologies used be capable of dealing with 1on2rterm behavior; multiple modes of chance; and feed-back inter,-c% joli between economic, social, psycholocrical, and demographic forcos.Tes emphasis should be on short-term, foreca sting and st.9tic analvsijs. With new efforts we can hope to achieve greater public underst'Indill-r of prospects for economic growth and more effectivegornme-RInt, plicies for responding to social and economic difflcultiesz.

Anierionn Gas Association 1970-American Gas Associatioln. (ia.q (m~. Now York: 1970).
Baink Credit Anqi st-Rank Credit A nahilst. (Hamilton. Bernmula T m Research. Ltd.) May 19T3 and Deemnber 1974. Burns 19(19- -Burns. Arthur 'F. Thec RhI.91f(rq.9 Ci/C7( in a inqiq TVI'rld ('Nw York : National Buireaui of Ec'onomiv Resenrch. 19619).
Carruth 1975-Ca rrui Eleanore. "The Sk ysera ping 1,osse's ill M11nh ,iit an Offlce Buildings." Fortutne (February 1975)).


Committee on Resources and Man of the National Academy of Sciences 1969Committee on Resources and Man of the National Academy of Sciences. Resources and Man (San Francisco: W. H. Freeman and Company, 1909).
Energy Policy Project of the Ford Foundation 1974a-Energy Policy Project of the Ford Foundation. Ex.rploring Energy Choices (Preliminary Report, 1974).
Energy Policy Project of the Ford Foundation 1974b-Energy Policy Project of the Ford Foundation. A Time to Choose (Cambridge, Mass.: Ballinger Publishing Co.. 1974).
Ewell 1970-Ewell, Raymond. "U.S. will Lag U.S.S.R. in Raw Materials," Chemical and Engineering News (August 24, 1970).
Forrester 1961-Forrester, Jay W. Industrial Dynamics (Cambridge, Mass.: MIT Press. 1961).
Forrester 1969-Forrester, Jay W. Urban Dynamics (Cambridge, Mass.: MIT Press. 1969).
Forrester 1971-Forrester, Jay W. World Dynamics (Cambridge, Mass.: Wright-Allen Press, 1971).
Forrester 1975a-Forrester, Jay W. "Business Structure, Economic Cycles, and National Policy," System Dynamics Group Working Paper D-2245-2, Alfred P. Sloan School of Management (Cambridge, Mass.: MIT, 1975).
Forrester 1975b-Forrester, Jay W. "New Perspectives for Growth over the Next Thirty Years," System Dynamics Group Working Paper D-2251-2, Alfred P. Sloan School of Management (Cambridge, Mass.: MIT, 1975).
Forrester. Mass and Ryan 1975-Forrester, Jay W.. Nathaniel J. Mass, and Charles J. Ryan "The System Dynamics National Model: Understanding SocioEconomic Behavior and Policy Alternatives, System Dynamics Group Working Paper D-2248-3 (Revised May 1976), Alfred P. Sloan School of Management (Cambridge, Mass.: MIT, 1975).
Forrester and Mass 1976-Forrester, .Tay W., and Nathaniel J. Mass. "U.S. Long-Term Energy Policy in a Changing National Environment," Testimony for the U.S. House of Representatives, Subcommittee on Energy and Power, in Medium- and Long-Term Energy Policy (Washington: U.S. Government Printing Office. 1976).
Frejka 1968-Frejka, T. "'Reflections on the Demographic Conditions Needed to EIst:-iblish a United States Stationary Population Growth," Population Studies, vol. 22 (November 1968).
Gordon 1974--Gordon, Robert A. Economic Instability and Growth: The American Record (New York: Harper and Row, 1974).
Kondratieff 1935-Kondratieff, Nikolai D. "The Long Waves in Economic Life," Reriew of Economic Statistics, vol. 17 (November 1935), pp. 105-115.
Levy-Pascal 1975--Levy-Pascal, Ehud. "An Analysis of the Cyclical Dynamics of Industrialized Countries," Central Intelligence Agency, Office of Political Research, Working Draft (August 1, 1975).
Madden 1975-Madden, Carl H. "It's Time to Rethink Basic Economy Policy," Presidential Address to the National Association of Business Economists, Boca Raton, Florida (October 1975).
Mass 1974--Mass, Nathaniel J., ed. Readings in Urban Dynamios: Volume 1 (Cambridge. Mass.: Wright-Allen Press. 1974).
M:ass 1975---Mass, Nathaniel J. Economic Cycles: An Analysis of Underlying Cau.ses (Cambridge, Mass.: Wright-Allen Press, 1975).
MAlass 1976a-Mass, Nathaniel J. "Stock and Flow Variables and the Dynamics of Supply and Demand," Proceedings on the Fifth International System Dynam, ic (Conference (Bergen, Norway: 1976), System Dynamics Group Working Paper D-23151. Alfred P. Sloan School of Management (Cambridge, Mass.: MIT, 1976).
Mass 197s -Mass, Nathaniel .1. "Modeling Cycles in the National Economy," Tchnology H'ciew, vol. 78 (March-April 1976).
Meadows et al. 1972-Meadows, Donella IH., Dennis L. Meadows, Jorgen Randers, and William W. Behrens III. The Limits to Growth (New York: Universe Books, 1972).
Medows, and Meadows 1973-Meadows, Dennis L.. and I)onella I. Meadows, eds. Toward Global Equilibrium; Collected Papers (Cambridge, Mass.: WrightAllen Press. 1973).
Meadows et al. 1974-Meadows, Dennis L., William W. Behrens Ill, Roger F. Naill, Jorgen Randers, and Erich K. O. Zahn. Dynamics of Growth in a Finite World (Cambridge, Mass.: Wright-Allen Press, 1974).
Minsky 1974-Minsky Hyman P. "Comment on 'A Critical Look at the Keynesian Model'," paper delivered at American Economics Association meeting, San Francisco (ID)ecemnber 1974), p. 4 (unpublished).

'Qii~ .rlI[~zI* t ildi .1. Kut,,iiiuu. "EFiwr.-' anid Dtiil ( 'oisraiits i 1;w
ill~~ I!IItjII f'jric~tc' vol. 19) to~eIIV 21, 1975), ppJ. 751-761,

1 -st ~w I 9756- Wtost ow, Walter W. "The Dev-elopinlg World ill the Fifth Nn
hx~ iffUl):swing,'' A nuial.8 of the Anieriuuni Academny of Pol il d iq
>C~cn ~J ol. 420 (July 1975).
S, i iuel son 1939-Samuelson, Paul A. 1n1 eract jons Betkveen the Miultili er AiAlysis and the 1Principle of AeCeleration," lfrici of Econonow- and X' I ati tics, vol. 21 (MNay 19&()), pp. 75-4-8.
Schroeder, Swetney, and Alfeld 195Sbodr Walter W\., 111, Rubert 1]. Sweeney, and Louis EdwardI Alfeld. INeudigs in Urban Dynamiic': VoliUUic2 (Cambhridge, Mass. : Writght-Allen LPress, 1975).
Sluiuan and Roseanu 197-2-Shuian, Jaines B., and I-thvid loen.The Kwtdraticif Wave (New York: World Publishing, 1972).
Solow 1970--Solow, Robert M. Growth Theory (New York: Oxford Univer.sity Press, 1970).
Sommers and1 Blau 1975-Sommers, Albert T., and Lucie R. Blau. The Widening Cycle (New York: The Conference Board, 1975).
Steinhart and Steinbart, 1974-Steiinhart, John S., and Carol E. Stoinhiart. "Energy 1Jse in the U.S. Food Systems," Scientce, vol. 184 (April 19, 1974), pp. 307-316.
Wall Strect Journtal 1974-"Sharp Rise in Inventories Stirs Fears of Deep Slump-but May Slow Inflation," W~all Street Journal (August 16, 1974), 1). 24i.


By JOSEPh W. D-u-CAN* **

This paper describes the major ongoing economic forecasting models developed by major Federal agencies. The sections describe the purposes of the models, their general approach, recent findings and relations to other modeling efforts. Three general points emerge from the review of ongoing efforts:
1. The major forecasts of the national economy are made by the
Bureau of Economic Analysis of the Department of Commerce and the Bureau of Labor Statistics of the Department of Labor.
Many other Federal agencies (including the Federal Energy Administration, the Environmental Protection Agency, the Economic Research Service of the Department of Agriculture, the Energy Research and Development Administration, the U.S.
Department of the Interior,- and the Federal Preparedness Agency) prepare long-term forecasts for nore narrow'ly defined
sectors of the economy.
2. There is a high degree of informal coordination between forecasting groups of various agencies, but there is a growing need for a more formal "Interagency Committee on Growth Models,"
which would meet on a regularly scheduled basis for the timely exchange of planning and implementation information and advice. The Statistical Policy Division of the Office of Management
and Budget plans to establish this committee.
3. There is a strong feeling that the establishment of a central
economic forecasting model would be counterproductive and too binding in developing helpful decisionmaking tools. One reason for this is that the purposes of each model are so diverse that no one model could serve them all. Another reason is that the tech*Deputy Associate Director for Statistical Policy, Office of Management and Budget with assistance from the Ad Hoc Committee on Long-Term Growth Models.
The conclusions and findings are not official views of the Office of Management and 131udet.
The ad hoe committee members assisted in the review of sectionq concerning description of their agency models. The overall conclusions and findings are the views of the author and were not subjected to agency concurrence since this has been prepared as a background document for future policymaking.
**This paper was prepared with editorial help from Judith Strenlo (Harvard University) while she was serving as a research intern In the Statistical Policy Division (SPD) of the Office of Management and Budget (OMB). David Hulett, Chief of the Economic 5tatitics Branch, SPD. assisted throughout the meetings of the Ad Hoe Committee on LTnL-Term Growth Models. (See attendees listed in appendix IT.) Agency contributors to this report are: T. M. Albert, Dnergy Research and Development Administration, Alvin A. Cook, Jr., Federal Energy Administration, Daniel H. Garnick, George R. Green, and A. Ray Grin'es, U.S. Department of Commerce. Ronald E. Kutscber. U.S. Department of Labor, CpTlvin 0. Lawrence, Environmental Protection Agency, and Lenore Sek, U.S. Department of the Interior.
The author is deeply appreciative of the time, effort, and counsel provided by all of the above.
A spelal note of thanks to Gale Berghoefer for cheerfully typing drafts, revisions, and the final report.

niiill (lificuities of interface between separate models are such
I h'; t :t central model would not be efficient.
IThe (evelopilent of long-term economic projections is lot a liew tutilit : thIere have been several long-term econoinic foreca-ti ng etffor,-. including the Palev Commission in 152 and the Interaenriv ILong-ecrm (irowth Project in the early 1960's. Recently. however, the interest in long-terni growth projections has accelerated. The Statistical P(olicy Division of the Office of Management and lBuiget. with its responsibilities for establishing statistical policy for all (oveminent agencies, has recognized the appetite of long-term iemodelers for improved statistical series. As a result, an Ad IHoc Intera(ency Committee on Long-Term Growth Projections was eAlablished inl
1975 to review existing efforts in Federal agencies, to identify areas of common interest, and to examine policy options for imI)roved coordination and integration of some of the various models. Much of this report is a result of this Committee's work.
At the outset, it is appropriate to recognize the limitations and difficulties of long-term economic forecasting. In a recent report. Data Resources, Inc. (DRI) discussed several sources of instability in the economy at present including the disequilibrium of the international relations system, the world commodity situation, the legacy of doul)ledigit inflation, the rapid changes of relative prices, and the overall financial condition of the economy. DRI concluded that:
Under these circumstances, it is very difficult to develop serious long-run-e plans for government and business. Economic planning is offered as one of the solutions to our difficulties. There are long-range matters which deserve better attention from our government. But increasing frequency of shocks and the continued uncertainties make it totally inappropriate to draw up elaborate plans which assume that the future can be known. The rational strategy for businesses and governments in an environment such as this one is quite different: to develop quick responsive capabilities to new shocks as they may come along, and to devise policies which at least partially insulate institutions and systems from the many sources of instability.'
Thus, in discussing and evaluating long-term economic growth
projections, one must always keep in mind the fact that, under the present economic instability, the best forecasting efforts may not lie accurate enou'li in retrospect. However, it is important to try to forecast the impact of current governmentt actions and outside events on the level of economic activity as a whole, on individual sectors and regions of the economy, and on the Federal budget in particular. Some guide to the potential effect of proposed programs derived from long-term forecasting can be an iimortant policy tool when used in combination with other factors in comi)aring the impact of several possible alternative programs. Thus, the present use of long-terin economic projections lies more in contribute ing an additional analv ic dimension to the decisionmaking realin than in the area of acttual knowledge of the future.
A list: of several of the ageeies l)resently involved in lon-tem m projections illustrates the p)resenlt Scope o)f such activities wiflhin the Federal Governmient. The Econoimic Growth l'ranch and t ie Regional Economic Analysis D)ivision of the Bureau of Economnic Analysis of 'Quoted frn Daita Resources, V.S. Long-Term Ileview Sumner 1975 In "National vnergy oitlool:: February 197, Federal Energy Administration, F:VEA-N-75 713, pp. B-1 to W-:


the Department of Co The following modeling efforts are covered in some detail:
1. BEA Growth Model.
2. BLS Economic and Employment Projections Model.
3. The OBERS Program (BEA/ERS).
4. The ERS Economic Projections Program.
5. Federal Energy Administration Forecasts.
6. Energy Research and Development Administration Projections.
7. Environmental Protection Agency/SEAS.

BEA Growth Model
The Bureau of Economic Analysis (BEA) of the Department of Commerce is engaged in continuing development of, and projections with, a moderate sized annual growth model of the U.S. economy. The BEA Model provides a medium-term 2 projection of GNP and its components, productivity, inflation rates, income items, and other aspects of the national economy. The BEA group maintains communications with other governmental units interested or involved in related work, especially the Bureau of Labor Statistics (BLS), in arriving at assumptions to be used for the projections work.
Metlhodology of BEA Growth Mode.-A simplified flow chart of the model which identifies major sectors, but abstracts from the simultaneity in the model, is shown as Figure 1. The original model specification appeared in the June 1969 Survey of Current Business, although the model currently in use has been improved considerably from the beginnings described there.
The current model contains four principal sectors-supply, prices, income, and demand. The supply equations estimate the GNP that could be produced with different quantities of capital and labor. This is the GNP necessary to achieve an unemployment target, given labor force assumptions derived from population estimates. The price equations provide estimates of the overall GNP deflator as well as component deflators. These are used to inflate the supply GNP estimate for the income sector and to deflate the income estimates which enter into the determinations of the demand sector. The income equations estimate the various components of aggregate income. They are used to derive the distribution of income that is consistent with the supply estimate of GNP. Given the income flows and the price values, the
2 Recent uses of the model have involved projections to 1985, although the model has been used for projections as far out as 2000.


Simpriffed Flow Di2gra in of ',Odel Supply E uaCcns Su;Vy Ch? Estimate (in wastant 1) Price Equati ons--,*] SUPP4 GNP uimea
(in CLOMM kr=)e EquaCcas (to Current Income Deflato-rs

DemaH C NP Erimla (in cmstvt 1V)

Source U.S. Department of Commorce, Burean of Economic Analysl ;.

demand equations estimate the endogenous final demand components (Federal purchases, which represent fiscal policy decisions, and exports are exogenous; that is, supplied to the model from outside judgments or data).
'Summinc: the elements of demand provides the demand side estimate of GNP. w ich need not equal the supply estimate. If the two are not (c(Ilial, the unemployment target is inconsistent with the specified fiscal policy package. To eliminate this "gap," the fiscal policy input can be reformulated and the specified unemployment target maintained or the unemployment rate can vary to equate supply and demand with aiven !sQB policies.
kR -c ( Yovc1,s~ons of BIA modeIl.-The BEA econometric growth
model was designed to provide medium-term projections of the U.S. economy and to aid in the formulation and analysis of fiscal policies.
A model run is currently in progress, but the results will not be available until late November or early December 1976. As of November 1975 the most recent conclusions from this effort were:
1. After recovery from the recent recession, the U.S. economy
'ivl experience slower rates of growth in real GNP and productivity than Jn the full employment period 1964-1969. This is a consequence of a number of factors including: a slowing in the rate of growth in the labor force due to a shift in the age composition of the population; the growing amount of nonresidential, fixed investment for pollution abatement which can hamper productivity growth to the extent it substitutes for investment in productive capital: and some shift in the industrial mix of GNP
towards industries with lower levels of labor productivity.
2. The U.S. economy will continue to experience relatively high
rates of inflation reflecting both the increase in unit labor costs associated with lower productivity growth and the high wage increases which result from attempts to catch up with past price
3. Given the assmumiption that the 1975 tax cuts for persons and
corporations will be extended and that no additional tax cuts will occur, the most significant change in the distribution of income from the historical pattern is the decline in the ratio of disposable personal income to total personal income. As a result of the progiessive Federal tax system and the projected inflation rates, the effective average Federal personal tax rate increases significantly
o-er the period.
4. The composition of final demand shifts somewhat over the
period. The personal consumption share of GNP declines reflecting the reduction in the real disposable personal income share.
The nonresidential, fixed investment share increases as a result of the additional investment expenditures necessary to comply with pollution abatement requirements and to reduce dependence on foreign sources of petroleum. Though the total Government sh P re o f GNP continues to decline as it has for the past several years,3 the State and local share continues to move higher, reflecting BEA's assumption of continued growth in Federal grantsin-aid.

Federal purchases of goods and services declined from 11.4 percent of GNP in 1967 to 7.8 percent in 1073. A slight increase occurred in 1974 and 1975.

-(Ill w ith lllod(, -,t 4:Y1-oNN-(lI III I"c(lel-:11 I 14 0' 1 f
10 f \V I I IVS I I t I I r( ) Ll(,-l mnlv s!()NOY froill 11)(I recclit rec(-, Mfl.
VIw, of t1w 111.'A Model.-J."lie niod(,] 's clil rcl J] lwil)()- llse(l to Illake IllediIIIIIAC1,111 proje(I iolts of P "tll(,t its ('ollip"W kill 111 "1(1(11holl, t I I kI 11 f )( !e I I I-LIS SC I' VOCI selvera I I Ise III I i a I) I) I I ca(",4 )I, I I I mcludliio- :,,!i amalvs*, of tho sO11sItiN-JfV of t/he ecolioilly to clltil, I I I V zi r i Oil,,- 1 11 policy ill -" I- 1"11011ts m d -1-11c aiialys* i of capit'd r(-.11
meiit.s 1"or fuil -inpioyinciit piv(.1tictioii. -- Vlso, Ow 1,loat'i 1011s
111V to assi.-'-. othk-t ton k s witlilhi the of
aIIJ other Gr()VC 11*11!11elit Ifrencles ill their ali,, ; 'ses of futilre ( 'f-0110111W* coll(litiolls, iliclu(,Lilll().:
1. An analysis of the sources and uses of gross saviiicr ill t llo
1970's, for the Assistant Secret-ary for Edconomic
2. The implication,-,- of the existilip* tax stri-Icti-ire for sai-il"IY. illvestinent aiid economic growth through 1985, prepared for t I I e
1974 Economic Sunimit 21il-eetings; aild
3. Projections for tike, period 191,0 to to fla', e(-ojioiiiie, effects of -Ni-idely differilig assulliptloiis tabotit Hit, rate of population Growth and the American Future aiid the Department of State (for submission to the U.N. 1975 World Population ("Onference).

BLS L'conom lc and Employment P),oI)eet;ov..sJ1ode1
The program of economic growth studies in BLS d(n7-elops to
15-year economic and employment projections of the U.S. ecoilorily in industry detail. Recent projections have inx-olived 12.) The
projections involve a detailed study of the growth of the U.S. econonly under alternative scenarios, einbodvin(T assulliptions Iboiit Federal economic policy and other factors which shape the future c(, .!oilon-iic environment.
Methodology of BLS Olodel.-The BLS projections are developed using the following sequence:
1. Labor force projections are de. -eloped ushig population proectiolis for various a(re, and sex groups deN-elol.Ied by t-he I'Mroall
of the CenSUS.4
2. Potential GNP is projected as the product of: (.q) eiiiploiillelit, based on the projected labor force aiid the assilillplloll of pere-elit ul I(II-11 ployll 1("Ilt; (b) projected :11)IIIt(al jo(II-I, J(d): .111(i
(c) pi.-ojectedoiitputpei.-iiiai i-l ioiii-.
3. Distributioii oft-Potelitial G'NJ) into ui:ior cateo-orics of (leniajid is projectLed lroiigli flie it'4, of Ca IiIZI(I 1'000011011; 111
which starts with poteial-,i] GNP aiid develol),- of GoN-erilmelit revellue, Persoll.al illcome, 1 1 171(t bllsi7lv s illcoTlle. Tll(
ilicoli'le estil'ililtes are fliell Ilse(k to cd,
l1lelit plti-cha,,-,es of (roo(Is -m(l services :110
invesfiIleiit expeii(litures. Assuiiiptiow aboiii
illelit fiscal policy .."It-ed so that ;I I 1 (.1 ()1, 1 N.
of demand and stipply GMP are "Ille l11,I(.I*m,(,o11olIIiC
4 The latest labor force projvoti(In- Nvere cont ,iint it in Johrt,;ton. 1) e I -. ;.
Labor Force ProJe('00111, to 199(),- Monthly Labor Rov:vw, July Thestatwill be updated by a release planned for August 1976.


11odel wsed by BLS is the same model described earlier in the
I EA Growl h Model section.
4. Conversion of projected demand into detailed industry emplo.vment estimates is done in three substages:
(a) Major final demand components (i.e.. consumption,
investment, Government demand) are distributed into detailed projections of demand by industry.
(b) The potential demand for all final goods and services
is converted into industry output requirements through the use of inter-industry (input-output) relationships projected
to 1980 and 1985: and
(c) Projected industry output is derived and subsequently
converted into employment requirements based on projections of annual hours per job and output per man-hour.5
in the late 1960's the Tnterafency Growth Project (consisting of BLS. BEA and OMB, and chaired by the President's Council of Economic Advisers) miided and funded the development of a basic projection model by )r. Lester Thurow, then at Harvard. Both BLS and BEA have enlarged and modified this basic model to reflect their separate needs for detail and focus in economic projections. They maintain close communications to ensure comparability of results from the two models in the sense that, given the same fiscal policy assumptions, the models will project the same growth rates of GNP and the same unemployment levels.
Frequently, their uses of the models differ in that BLS sets an unemployment assumption and modifies the fiscal policy assumption to achieve the assumed level of unemployment. BEA's model can be used this way, but BEA generally assumes various proposed fiscal policy packages and observes what the resulting unemployment rate would )e for each.
Rcevit Conclusions of BiLS ,ode.-The latest BLS projections are in the March 1976 issue bf the Monthly Labor Review.6 The most recent rejectionss were revisions of projections published in a 1975 rlport entitled "The Structure of the Economy in 1980 and 1985." Some of the conclusions reached in these projections are as follows:
1. In the period 1973-1985, overall real GNP growth will be
slightlv lower than the 1955-1968 rate. This 1973-1985 rate reflects a somewhat more rapid rate of growth for 1980-1985 than for 19.73-1980. However, because of the 1 73-1975 decline in GNP tle projections show a growth in real GNP during 1975-1980 of 5.8 percent a year to reach a 5.5 percent unemployment level by
2. Employment projections show a moderation of the growth
of government employment at all levels but particularly for State and local governments, with a corresponding increase in the rate
SSource: U.S. Department of Labor, Bureau of Labor Statistics, Bulletin 1831 (1975), "The Structure of the U.S. Economy in 1980 and 1985," pp. 2-3. 6KOt-oher. Ronald E. "Revised BLS Projectons to 1980 and 1985: An Overview," Monthly Labor Review, March 1976, pp. 3-8. Bowman, Charles T. and Morlan, Terry H., "Rovised ~rojections of the U.S. Economy to 1980 and 1985," Monthly Labor Review, March 1976. pp. 9-21. The Bureau fo Labor Statistics has recently revised its labor force proJeetion, to 1990 These revisions are not reflected in these economic and employment projections de scribed here In general the new labor force projections would call for a modesty higher rate of economic growth in the 1980's but amounting to less than .1% a year.

4,t _- (,,vtlh of, private sectw nl~ lvi'l
o\tki will slow o(wn ini 1so--1!)s5.
:. Tie slhift in employment towvard(l n icsml awniy1 con,
1aolis-)produlciiig sectors is expected to continijo. .\2,,' hs.; ti uh pr)'ojected to decline severely iln tlhe 197: pwedlic .bil. LiNw been )projected to decline miucl less t hani I j b 1e c:i a i brought about by increased demand as a result of t1he new worl( food outlook. Mining (previously projected to decline) in the latest projections shows modest growth due to a change in tl(:, energy outlook, requiring a higher rate of coal production and oil
4. Federal receipts are projected to grow more rapidly than
ex enditures, but under the fiscal policy assumption used a small deficit is still projected for 1985. The rate of growthf of State and local expenditures is projected to increase at slower rates tlan
experienced in 1955-1973.
Two more articles describing revision of the 1973 projections wNill appear by the end of 1976 in the Monthly Labor Review. An article in the August 1976 issue of the Monthly Labor Review will review in detail the accuracy of the 1970 projections prepared by BLS.7
Uses of the BLS Model.-Attention is given to labor fore and productivity growth, capital and material requirements, and changes in technology and the patterns of demand from individuals, goveriments, business and foreigners. Projections of output levels as well alabor and material requirements are currently made using a 125-sector disaggregation of the U.S. economy. In addition, staff capabilities, data bases and models developed for the projection effort are regularly employed to estimate the impacts of various Government programs. legislative proposals, and other current or anticipated develol)lnieilts which may affect distribution of demand, rate of economic growth. )or level and distribution of employment.
The major use of the projections within the Departinent of Labl)or is to supply an economic and manpower framework upon whii,,h estimates of future occupational requirements are made. The Bureau of Labor Statistics regularly publishes detailed information on the outlook for employment is a large number of occupational categories.
The projections have also been used within thlie Labor )eI)artielnt and other parts of the Federal Government as a framework for a ssing a number of diverse economic problems such as capital requ(irements, manpower utilization and energy policy. In addition, Several State and regional agencies, private research groups and business organizations have used the projections as a "national" fraiiewvoirk within which to develop their own, generally more disaggregated. projections. In order to make the projections as generally available a possible, a large amount of detail is published and, in addition, historical and projected data bases are made available on computer (margnetic) tape.
SPersonick, Valerie A. and Sylvester, Robert A., "Evaluation of BLS 1970 Project! us,* Aug:ust 1976. Monthly Labor Review forthcoming. The BILS projections were in i'rin.Jctiions 1970, BLS Bulletin 15: 6.


Federal Energy Adminlstration Forecasts
Long-term Federal Energy Administration (FEA) projections are made through the Project Independence Evaluation System (PIES). This system generates planning estimates depicting possible states of the energy system.
The model is used in two ways: (1) to help the Administrator of FEA in his policy role by analyzing the impact of various energy policies and (2) in developing a set of projections of what the energy picture will be in the future. The principle result of PIES i- the determination of equilibrium prices and quantities of ener" by type and region at specified future time points, based on specified alternative energy policies.
Meth odology of PIES Projections.-A flow chart representation of PIES (shown as Figure 2) is found in the Nationald Encrgy Outlook.* February 1976,8 FEA. The central portion of the system estimates the energy supply and demand which are integrated to make the forecasts in an iterative way.
FIGURE 2.-Basic PIES flowchart.

Conserva!on Options E"Supply Options and
Economic Forecast Assumpions

Demand Model Supply Models
Sectors/Fuels E traction


D Reports
p S


Source: "National Energy Outlook: February 1976," FEA, p. A-4.
Input into these forecasts is of three types: (1) general indicators of economic activity such as GNP, inflation, and unemployment from an external macroeconomic model; (2) constraints to be imposed upon the system such as limits on material and equipment, capital, labor and transportation; and (3) foreign energy prices and domestic energy policies.
Op. cit., Figure A-3, p. A-4.

Tj;,I )!l"wr( ?CcOll 0I I I i( i I I w I, s are Ilot, fonw': .-t I )v F E A 1 It t t IXTI froni forecasts (venerated I)v flie
ecou(wi ric viodel developed by Data 1-tesom-ces, lilt'. inplits are a matter of judt-rillelit coillbillk"d AN'll 11 11 1111-11 1wr'l III i Wnation as call Jw ratliered. bitt no Illat1will"tt W"d To
crenerate thein. "I'lle iise of a macroecolloillic 11 1(,Oel III coll all One r(_rVsi)OCif 1 e niodet stich as llell:- -:III"
tween the ellm, -v sector and the rest of the emnoi) iv.
0vtput, froln the eciltral computation ilio(lels coiIie- 1ji tlj,- k)iIii o C
00 1 I I -1
some -eports in t1w areas of ecollo.-ii(. iillp'l(J. illterllw
irollinenta (IS 411 j,-j ,;!I
inent, and (IM, iilipact as iv, 11
The oricrinal p'Urpose in developino- 11TI-t"S, V-,, S f 0 ryp 1, (1 1, S () I I (I quantitative analvF.i.saboutenilrcry to 11 e 'It file flepo)l 9 of 197 4. But the purposes for wlik,1I *It is, aiid cai-i go far loeyond t11v,, one-1--inie, report. This is possible Cio
flexibility built into the systeni including: (1) the ability -,,) 1'!I(*0IT()rate constraints; (2) the ability to v._,z:v.me new ( apabillities: aiid (,3)) the ability to incorporate pid(rinents into Hie knkll--llvsl,, a n(I to analyze the sensitivity of the results to the accuracy of tliv- (, jiid nnents. The model also incorporates responsiveness to
11 C Ill
relative prices and analysis of externalities.
Rerc.-tt Con.cT*4.om o PIE'S.-The most recent pre(Ii.ctiojis ated by the PIES system are contained in -q1;0,(?(j7 J jje; qy
Feb iwary IfI'i'O',111 published by the FEA.
The. inajor findiiicro and conclusions of fliat, report are:
1. Over t1w n, Ixt 10 ve(ars, the Nation c(an Q-reatly n l it
domestic elier-yv production and ctit the rate of
7' Z!(10111'and and still ille(A its e(-.011011lic Obje (4ives.
2. The post-19.S5 prospects for maintfUnin" 11 1
less certain unless technological and e(1ojj()jjjj(1 I J- I- jl-T

3. SpecificallA. by 1198), as a re i,.It, of Iii-_1
demand will be im-1ch lo-wer thcm I-IiA o6e (4-vowth ,AjlJ all
active cni)-ervavion pro,(rrain could furtlier redti(T (I'll("",,
4. As for ne-,vv supplies, Ahlslm will be the 1!D11 111 Of
new oil pro(imetion. _Niltionally, projection is lli;1(1(, I)y
lack of knowled,<-re ,lT)oiJ t1w amwvllfs of
Uses of PIF.S.-Tlv PIES ,vstem is used exteiisivel-v by tlie A(]ininistrator of FEA_ in the O< of dcfin*n(r and t liaplmr pi-oi,,)<(
enerow policies. The quantitative amflvs*17 ANas perv.Llsive ill 114, ")I-cj):,lration of the oricrimal T'j"Oj(-'c' IwICI), 1,1.1, /!( ( flcpol-4 1 "U'(1 I,-Z ll ; d 1)v I- ss ,As the primary tool fot- evziblt the Administration and the, Coii-res inor ener ,ry initi-atives. Foil ex(Illiple, it i ii ed to flie effect of propose4-1eaislation. (e.cr., (ras derecriflation) ill the ,tiv:l.
Essentially, any policy (Juestioll that call be of
chancres In SlippIv ,Ind (1.ewaml curves. modicic'ItWit )f elwnv collversion and distribiition te -1111()]()Z7ri( S, or cw,A i-aml wl, the supply system call be examiiie(I-iii rlv ,, dellil !111-oll4,11 tll(.t
OProject IndepeDdence, Report, FF.A, GPO NO. -111 -f,"(12". No\ em1wr 1974.
10 01). Cit.


system.1' The system is focused upon the energy sector alone, without intent to interact with other sectors. However, for analysis outside the chert"gfield, the system is used to develop parameters for use in other models. For example, a proposed energy policy might imply a large investment in energy supplying sectors. One could then go outside PIES to other models to evaluate the impact of this form of investnient upon the economy.

Energy Research and Development Administration (ERDA) Projectios

Energy-related projections have been published as a part of "A National Plan for Energy Research, Development and Demonstration: Creating Energy Choices for the Future."2 These projections are the product of a system created for ERDA by the Brookhaven National Laboratory (BNL), which uses as its macroeconomic framework the Data Resources, Inc. projections. The projections, which are for the years 1985 and 2000, include total energy demand, imports required, electricity used, and other factors of the national energy system, projected under a variety of scenarios.
Methodology of ERDA Projectionw.-BNL's system deals with
future energy demand in terms of key end-use categories (for exampl% space heating and transportation) that are constant for a given test of scenarios. The demands are specified not in terms of fuels, but in terms of services required of energy end-use devices (passenger miles to be driven, square feet of floor space to be heated and cooled, tons of steel to be made). The services can be met for each scenario with

FIGURE 3.-Reference Energy System 1972.

06 (31) 0.2
fi2 (.)OPOWER (Electric) IRON & STEEL
14.1(34 27 63 _,/


23,3 (Liquid 29,6
Source: ERDA 48, Vol. I: The Plan, p. IV-3.
Hogan, William W., "Energy Policy Models for Project Independence," Discussion Paper, June 18, 1975, FEA, to appear in the Journal of Computers and Operations Research.
12 ERDA-48, Vol. I: The Plan, GPO No. (1975) 0-579-905. More recent projections are contained in ERDA-76-1; A National Plan for Energy Research, Demonstration, and Development: Creating Energy Choices for the Future: 1976, Vol. I: The Plan, GPO No. (1976) 052-010-00478-6.


: This new analysis links detailed mathematical process engineering and economic models to more aggregate econometric models. The four models which are used are:
1. The BNL engineering model, used to estimate physical flows
within the energy sector;
2. The BNL/University of Illinois input-output econometric
model, used to link general economic transactions to the engineering model;
3. The DRI interindustry model, used to analyze general economic structure and the interaction between energy and the rest
of the economy;
4. The DRI macroeconomic grrowth model used to specify the
long-run trends of U.S. economic growth.
The use of an expanded interaction between econometric and engineering models is the new feature for the ERDA analysis, and provides a much more detailed capacity to analyze the economic impact. in a broad sense, of various alternatives in energy research and development. The linkage of the four models is not completely automated yet, but work is going forward to do that before the next set of ERIDA projections is made.
BNL has developed several other capabilities for their energy and economy models. These include:
1. Technology assessment--assuming a given set of available
energy technologies, which one would be used?
2. Time tracing-if a certain technology were introduced, what
time path would its usage rate follow?
3. Timing assessment-if solar energy were economically viable
in 1990, would nuclear power ever be needed?
Recent Conclusions of ERDA Model.-In the Plan, submitted in June of 1975, projections are shown for six scenarios: (1) no new initiatives; (2) improved efficiencies in end-use; (3) synthetics from coal and shale; (4) intensive electrification; (5) limit on nuclear power;
(6) combination of all new technologies.
In the most pessimistic scenario (developed only as a reference point), current use and technology are assumed to continue unchanged; in this case, total annual consumption reaches 164 Quads (1 Quad= 1015 BTUs) in the year 2000, and projected oil imports are 13 million harIs Ibid., pp. IV-3.
The Relationship of Energy Growth to Economic Growth under Alternative Energy Policies, by David J. Behllng, Jr., Robert Dullien, and Edward Hudson, available from National Technical Information Service/U.S. Department of Commerce/5285 Port Royal Road/Springfield, Virginta 22161.


rels a day in 1983 and 28.5 million barrels a day in 2000. These projections contrast with the current level of about 6.5 million barrels a day. This scenario is clearly unacceptable, and overly pessimistic since it is highly unlikely that no new initiatives will be taken.
At the other end of the scale, assuming all new technologies are developed leads to the most optimistic projections. This scenario shows total energy consumption at about 130 Quads in the year 2000, with a surplus of two million barrels of oil per day. From this it can be concluded that if all technologies were pursued, successful, and fully implemented, it would be possible to meet our energy requirements with domestic supplies. Unfortunately, this scenario is an ideal which is also very unlikely.
These projections and those of the more likely intermediate scenarios are not predictions of future energy use, but rather they represent possible patterns which may emerge from the research and development and other energy policy decisions made now.
Uses of ERDA 1Iodel.-Projections of the energy system in 1985 and 2000 based on various scenarios have been used extensively by ERDA in developing the substance of their Plan for Energy Research, Development, and Demonstration. Implications of various alternatives, such as (1) conserving energy by developing greater efficiencies at end-use or (2) extracting more coal and oil from current locations by developing more effective recovery technology, are examined in terms of projected imports, demand, and other facets of the energy system. The results suggest which approaches are best for long-term and intermediate-term periods.

The OBERS Program (BEA/ERS)
The Regional Economic staff of the Bureau of Economic Analysis has a separate projection effort in cooperation with the Economic Research Service (ERS) of the Department of Agriculture to produce area economic projections of population, employment, personal income, and earnings for 37 industry groups. BEA produces the major economoic projections, while ERS produces only the agricultural parts of the projections.
This subnational projection program was begun at the request of. an with financing by. the U.S. Water Resources Council which uses the projections to assess water resources requirements and to evaluate programs. The projections involve a combination of econometric model ing and judgment.
Methodology of OBElRS.-In order to make area economic projections, projections of population, employment, and income are made first for the Nation and then disaggregated geographically. The projections of area population are derived from the projections of area income and employment on the assumption that persons in the labor force migrate to areas of economic opportunity and away from areas of economic decline or stagnation. Crucial assumptions used in making the forecasts include an unemployment rate of four percent during the projection years by decade to 2020. and the assumption that there will be no policy or program changes of an unusual nature or magnitude.5
7 "State Projections of Income, Employment, and Population to 1990," Survey of Current Business, April 1974, p. 26.

-Y oj 01) 7 RS.-Tia, followill(r cow-111s*
YNTIA pro sections (which owit
-apid1v risiny prwe Of fWZS*
orta -,III([
ges alld
During, the 19740"s and 19801s the Souflw& I-Wky Mountaiii States are projected to ().rov at abox-e-av" I J'y(*
the Fai- West and 'New England 'State: (11V 1j)
(Yrow at near-aN-ertage rates; and the Alideast. Gire(,4 Plains States are projected to (rrow at. below-averaov
-A- lite rapid growth proi(,Cted for 4-he S0jjtj)o,,,st tyl(I
is a continuation of past t-,I-ends (for the Rockv it is a
reveit-sal). Alanulactml-Incr will contill1w to boost sout 1101-11 o-rowtll.
but much of the growthh will occur in cheniic,-ils. machinery, falwicated inetals., paper, printing, and other inalillfactI111111i)* 11"ItIler than in tl,e current1v doiiiinaiit textile and apparel industries. In addition, a continue influx of retirees and vIwationelrs, maijily to
Floridat and Arizona, will spur inconie growt1i iii these State's.
3. The near-average orrowth proje(--ted foi- tli(.-- Far West States
contrasts with the rapi+J growth of the past. This chan(ro, in trezi(l becran in the late 1960's. with the cutback in the space pr9(ri-all-I 1-1714
leclining mil*tar-\- am-1 ek-ilian aircraft prc)ductioil. In dditi
the crowded conditions and the resultincr environnionital (JlwnaLre tlia.t develm_)ed in parts of California Itave dis('Ollracred
4. The below-averao-e cri-owt1i projected for the Af;&,:i 4, : lwl
Great Lakes Stat.-s refle(,ts the likely coiifilim, t-ion of f1w t,-irlel,'.
101, U.S. nianufacturin(r to becorne more dispersed (-reo0-r',1plIWa11v.
The gro-w-th performance of the Plains States will continue to be
shaped by th-_ slow crrowth of ao-riculture, find food proves sin,X.
Usr. of OBERS.-Other Federal aoxIncles OnO4 of -wliich liai-e fun(Il-d exte-11sions Of tile projections pr(,)() *rani) "their I],--,(,s of Ow projections a-re:
1. Council o'n Qvviity-for prepariiitr eiwiroiimental impact stateiiients,
transportation Department of Tjwt poi tation-f or
3. U.S. Postal projecting mail volume, and evaluating Post Office capacities,
4. Federal Emrcl? AdnI;)Iis raliov,-for aSscssincr the area iinpacts of energy policies,
5. Winter Ya ;qafiol), B 11'(1 (all interaIrr iwv autboritv)-for
studying the impact on industrial location patteriis of extending
the Great Lakes-St. Lawrence Seaway sliippiii(r season,
C. U.S. 2tr;Iny Corp, of FaRcy AuOor;ty-for addressing specialized problems in loc.fl tirea, plannincr
and the
estimatim-Ir waste
water facility iwe(ls by are:iIn addition to tfiese -by Federal :icreiwie -,. flw ze lonrr-teri)i area
economic -,r) roi ect ions are iii lIeavy demand I)v Sf:tko -Incl local crovoriiments. universities, and private (inebiolii1cr
roj eet J 0 rj, z i, i T);i I T' n i i (- A c t i, P v f t 11 (A I Serf,
tj()n), ,even volumes Nf-4,-,rco- C-mcil, 1,)T;) :111d Ar( .[ Fo"no'llic titans : 1990, a 1 474 slipplemont to the Soirvey of Ctirrow, 11111,inos,.

major construction- related, transportation, and marketing concerns, banks, and private economic antengineering consulting fimns). At least 38 State governments have contacted the Federal project team to discus the projections and their uses. Many of the Government units use BEA projections as a starting point in preparing intermediateterm economic outlooks and in evaluating the economic impact of proposed Government projects or private investment.
The ERS Economic Projections Program,
The agriculture portion of the OBERS projections are furnished by a larger program within ERS. When ERS was reorganized in 1973, the National Economic Analysis Division (NEAD) was given responsibility for developing an additive, ERS-wide Economic Projections Program with a quick-response capability. They have developed the core of the National-Interregional Agricultural Projections (NIRAP) system which provides OBERS data as one of its functions.
Methodology of NIRAP.-The NIRAP system is a computerized simulation of the food and fiber system, with a 10-year horizon for most projections. It can simulate alternative futures based on scenarios differing with respect to major uncertainties which have an impact on food and fiber, and with respect to policy decisions and programs designed to alleviate specific problems. By systematic scenario development and comparative analysis of alternative futures. the range of possible adjustment paths for food and fiber can be bracketed, an early warning of potential difficulties provided, and possible solutions to potential problems and trade-offs between policy goals evaluated.
The NIRAP system is still being enlarged and improved. Extensive intra-ERS coordination, with all program areas of ERS having input into what assumptions are made and how the model works, characterized the development and use of the mode].
Uses of NIRAP.-Each year, the NIRAP system is expanded to encompass a broader spectrum of the food and fiber system. A core set of scenarios is revised and resulting projections of alternative futures are analyzed to provide ., continual check on major issues and to provide current projections for analytical extensions of the core program. For example, the core program is used by ERS to support periodic national assessments of water and related land use resources needs and individual commodity or input subsector studies. Also. studies are conducted for special purposes such as ampraisin. the U.S. production capacity for food and fiber and providing the food and fiber projections for broader general economy studies conducted by other agencies or research groups.

IV. Long-Run Environment and Energy Growth Model
Environm ental Protection Agqeney!SEA iq
The Environmental Protection A qencv (EPA) established the Strate.oic Environmental Assessment Svstem (SEEAS) "I Tt i! a collection of interdependent models used to forecast the state of the environment which would result from alternative environmental policies ar1d :ocioeconomic trends. Forecasts are presented annually through 19fl,5.
'7 it should be noted thnt the President's Budget for fiscal year 1977 contains no funds or personnel for the SEAS project.


liThe socioeconomic trends are predicted outside the SEAS system, and ii environment "al policy alternatives are, of course, generated bv decisionmakers in EPA.
Methodology of SEAS.-The SEAS structure is modular in nature, consisting of 28 computational and input-output computer programs. Any program may be executed independently of the others. All programs are autonomous in the sense that they do not reiu'lre iiajndatolv user-supplied information. That is, the mandatory input files all exit within the system with default values that a user can override if lie wishes. The modules include: national economic modules, including macroeconomic forecasts, subsector growth and technological change; abatement cost; and national residuals. Other available modules include stocks of critical resources of materials, solid waste types, disposal methods and costs, transportation, energy used by fuel type and user category, and a set of models which permit regionalization of national economic and pollutant residual data. A flow chart of the system is shown as Figure 4.


Abatement SAS/INFRU
Cost /
Macroeconomic Energy

Input-Output Module

Stocks of Solid
non-renewablet\ Waste/ feiu
resources Recycle




LL c ,1t Co;Wnlusions of SEAS.-The most recent major SEAS forecasting effort is not yet published, but the review process is nearing coziiplet ion and it will be published as a Report to Congress on the Cost of Air and 'Water Pollution Control: 1976-1985. Six separate scenarios vere:malyzetl and the results will be available in the Report.
U.es of EAS.-SEAS is also used on smaller-scale projects, both l)v other divisions of the EPA and by other agencies. Within the EPA, tie ()ffices of Energy. Minerals, and Industry is making use of SEAS. The office e of Solid Waste Management Programs has used SEAS to forecast impacts of auto recvcling alternatives and its Industrial Sludge Task Force adapted SEAS to provide comprehensive input for much of its analytic effort. The annual report of 1974 of the Coimcil on Environmental Quality contains forecasts generated by SEAS. The National Commission on Water quality was another user agency. The Bureau of Land Management (Department of the InteriOr) offshore drilling evaluation relied on SEAS to support deter11ination of the onshore environmental impact of such drilling along til New York/New Jersey coast.

The discussion in this paper has focused primarily on long-term economic growth models built and utilized by the Federal Government from a model builder's perspective. Many governmental groups use long-term forecasting as input to their decisions, as a basis for policy advice to the President. or even to make projections of their own, without actually developing a large econometric model within their own agencv. Examples include the U.S. Department of the Treasury. the Federal Reserve Board (FRB). the 0rice of Management and Budget (0MB) and the Council of Economic Advisers (CEA). Each for reasons that vary finds it impractical to develop an internal long-term model. but uses the results of other modeling efforts to shape views of the future.
Each agency has different needs for long-term projections and different ways of dealing with those needs. The CEA, for instance, is an advisory group for the President. It is asked for expert opinions on complicated economic questions, usually with a very short time to develop these opinions. Because the questions it investigates are so diverse, a model suitable for all of them would be unfeasible. Instead of attempting to build an in-house model, they rely on many existing efforts, both public (BEA. BLS) and private (DRI, Wlarton, Chase). This works well because different models are well suited to answering different questions, and CEA is able to draw on the strengths of each model.
Each year the CEA,. Treasury and OMB develop five-year projections of major economic variables (such as national output, rate of inflation and unemployment rate) ; however, the longer range economic assumptions are "mechanical" projections or extrapolations as noted in the budget documents.
CEA and OMB's use of models offers an illustration also of how long-term models and analyses affect policy decisions. In CEA's case, a question is posed whose answer requires long-term projections. The