Achieving the goals of the Employment act of 1946 - thirtieth anniversary review


Material Information

Achieving the goals of the Employment act of 1946 - thirtieth anniversary review thirtieth anniversary review
Physical Description:
United States -- Congress. -- Joint Economic Committee. -- Subcommitte on Energy
U.S. Govt. Print. Off. ( Washington )
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 22347678
System ID:

Table of Contents
    Front Cover
        Page i
        Page ii
    Letter of transmittal
        Page iii
        Page iv
    Table of Contents
        Page v
        Page vi
    I. Introduction
        Page 1
        Page 2
        Page 3
    II. Model development
        Page 4
        Page 5
        Page 6
        Page 7
    III. Model simulations and policy implications
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
    Appendix A. Empirical equations of the model
        Page 20
        Page 21
        Page 22
    Appendix B. Technical notes
        Page 23
        Page 24
        Page 25
    Comments by George M. von Furstenberg
        Page 26
        Page 27
        Page 28
    Back Cover
        Page 29
        Page 30
Full Text
. ... . .

Eel1 : Vrl/

94th-----/w Conres
2d Session f
'_. ^ *_'

\^ '- ,..'




Volume 1-Employment









DECEMBER 3, 1976

Printed for the use of the Joint Economic Committee




For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 Price 45 Cents
There is a minimum charge of $1.00 for each mail order







1+-/LI. 1 /ppr it-



(Created pursuant to sec. 5(a) of Public Law 304, 79th Cong.)
HUBERT H. HUMPHREY, Minnesota, Chairman
IRICHARD BOLLING, Missouri, Vice Chairman

EDWARD M. KENNEDY, Massachusetts
WILLIAM V. ROTH, JR., Delaware

HENRY S. REUSS, Wisconsin
GILLIS W. LONG, Louisiana
OTIS G. PIKE, New York
MARGARET M. HECKLER, Massachusetts

JOHN R. STARK, Executive Director
RJCHARD F. KAUFMAN, General Counsel











NOVEMBER 29, 1976.
To the Members of the Joint Economic Committee:
Transmitted herewith is a study entitled "Estimating Potential
Output for the U.S. Economy in a Model Framework." I believe
Committee Members, other Members of Congress and other persons
concerned with economic policy will find this study a useful aid to
assessing the magnitude of the task of restoring full employment in
the U.S. economy and to evaluating various policy proposals.
I would like to express my thanks to the authors of the study,
Albert J. Eckstein and Dale M. Heien, and also to George M. von
Furstenberg of the American Enterprise Institute and Ronald Kutcher
of the Bureau of Labor Statistics, who reviewed the study at the
Committee's request. The views expressed in the study are those of
the authors and do not necessarily reflect the views of the Joint
Economic Committee, individual members thereof, or members of
the Committee staff.
Chairman, Joint Economic Committee.

NOVEMBER 24, 1976.
Chairman, Joint Economic Committee,
U.S. Congress, Washington, D.C.
DEAR MR. CHAIRMAN: Transmitted herewith is a study entitled
"Estimating Potential Output for the U.S. Economy in a Model
Framework" by Albert J. Eckstein and Dale M. Heien, together with
a comment on the study by George M. von Furstenberg. This study
utilizes an econometric model to measure the potential output which
the U.S. economy could produce based on the availability not only
of labor but of capital equipment and raw materials as well.
Accurate measurement of potential output is essential to an
informed judgement as to what the appropriate levels of employment
and unemployment are in our economy and how much economic
growth is required to reach and sustain those levels. Basing the
calculation of potential on supplies of capital and materials as well as
labor helps us to determine whether there are supply constraints on
production which might prohibit reaching full utilization of the labor
The study by Eckstein and Heien is an attempt to use a new and
more comprehensive approach to the calculation of our economic
potential. Its results indicate that potential output may be somewhat
higher than the official estimates prepared by the Bureau of Labor

Statistics and the Council of Economic Advisers. The authors conclude
that "the likelihood of supply restrictions and bottlenecks preventing
the attainment, of a 4.5 percent unemployment rate appear to be
relatively small." Certainly, this is an encouraging conclusion. I
hope this study, which is admittedly preliminary, will provoke further
exploration of this vital question.
In addition to the published comment by Mr. von Furstenberg,
this study has also been reviewed by Ronald Kutcher of the Bureau
of Labor Statistics, whose comments were of assistance to the authors
in making revisions in the study.
The views expressed in the study are those of the authors and do
not necessarily represent the views of the members of the Joint Eco-
nomic Committee or the committee staff.
Executive Director, Joint Economic Committee.


Letter of transmittals_ _-------------------------------------------- III
I. Introduction-- ----------------------------------------------- 1
II. Model development------------------------------------------- 4
III. Model simulations and policy implications------------------------ 8
Measurement of potential output---------------------------- 8
Simulation with the model---------------------------------- 10
Policy implications---------------------------------------- 16
Recommendations for further research ----------------------- 17
Appendix A. Empirical equations of the model ------------------------ 20
Appendix B. Technical notes---------------------------------------- 23
References-------------------------------------------------------- 24
Comments by George M. von Furstenberg-------------------------- --26

Digitized by the Internet Archive
in 2013

By Albert J. Eckstein and Dale M. Heien

This study presents the results of initial research in developing
an analytical framework for estimated potential output for the U.S.
economy. The vehicle for this analysis is an econometric model which
emphasizes the supply of labor and the production of output. The
basic purpose of the paper is to demonstrate the advantages inherent
in a formalistic model of potential output and to compare results
obtained from the model with those presently in use. In so doing we
hope to point out and disentangle some of the many separate, but
interrelated issues surrounding potential GNP.
Foremost among these issues is the question of the level of unem-
ployment which is chosen for the potential GNP computation. Em-
pirical measures of potential GNP have specified potential output in
terms of an arbitrarily defined level of utilization of the labor force,
usually that associated with a four percent unemployment rate.1 The
historical precedence for such a definition has relied on analysis and
judgments with respect to the adjustment processes within labor
markets. The conventional wisdom has been that certain structural
problems within labor markets must be solved before target rates
below four percent can be achieved. Recently, there has been concern
that changes in the age-sex-skill mix of the labor force have been such
as to warrant a redefining of "full employment" at an unemployment
rate higher than the traditional four percent. The evidence at hand
does not demonstrate conclusively that new entrants into the labor
force today have lower skill levels than in previous years. However,
the structure of demand may have shifted in such a way that their
average productivity is lower.2 The question of the appropriate un-
employment rate for potential output is probably, in the last analysis,
not answerable in a precise manner. However, more analysis could be
done on the level of frictional unemployment, the effects of legislated
wage standards and the changing skills of the labor force. The question
of the appropriate level of unemployment depends heavily on the
trade-off between inflation and unemployment. This paper has little
to say regarding the determination of the level of unemployment con-
sistent with potential GNP. However, a methodology is given wherein
parametric solutions (i.e., solutions dependent on various unemploy-
ment rates) for potential GNP can be compared. This comparison is a
I See, for example, Okun [24] and Denison [10] for discussion of the underlying concepts regarding potential
SM Perry [27] and Nordhaus [22]. It should be noted that in the approach taken by Denison [10], increased
education of the labor force is assumed to augment the skill and productivity of the labor input over time,
and thus add to potential output.

first step in assessing the cost of the inflation-output trade-off. The
question of the relation between output and the price level, or the
process of inflation itself, is still a matter of considerable professional
debate.3 Related questions, which also are not dealt with here, include
the problem of proper monetary and fiscal policy to achieve full
employment, policies for promoting maximum growth of output, and
evaluation of the efficiency with which current resources are used.
While the economy has recently been forced to cope with new
constraints in capital and material inputs, the goal of achieving full
employment has been made more difficult by changes in demographic
and institutional features of the labor force. Most persons born during
the "baby boom" have now entered the labor force. Moreover, because
of changes in education, attitudes toward women working, and real
income growth, participation rates of the female labor force have
increased considerably since 1960. Table 1 shows the changing dis-
tribution of the population by age group, along with participation
rates for selected periods. As more workers enter the 25-44 age group,
and as female participation continues to rise, achieving unemployment
rates in the 4-4.5 percent range presents new difficulties.4 By incor-
porating the behavior of the labor force, average weekly hours, wages
and prices into a model with a production function dependent upon
capital, labor and material inputs, it is possible to analyze the "cost"
of closing any gap between actual and potential output in terms of
wage and price effects. For reasons discussed below, in this initial
modeling effort, the price determination process in all of the factor
markets has not yet been developed. Hence, evaluation of target
inflation rates along a defined potential output path is somewhat
limited. However, as labor cost is the main component of overall cost
and price movements, main effects are captured. An important aspect
of the modeling effort is that it permits potential output at each point
in time to vary in response to underlying relationships between
endogenous and exogenous economic variables.5 As will be seen in
Chapter 3, this is particularly important with respect to the incorpora-
tion of higher rates of growth of the labor force during the 1970's.
[In percent]
1980 1990
Age group 1950 1960 1970 (projected) (projected)
Below 16---------------------------........ ........ 28.3 32.6 31.1 25.4 25.8
16 to 24 ----------------------------........................... 13.3 12.1 15.0 16.7 12.8
25to44 ---------------------------- 30.0 26.1 23.6 27.8 31.9
45to64 ----------------------------.......................... 20.3 20.0 20.5 19.4 18.2
65 and over......--------------------.................... -8.1 9.2 9.8 10.7 11.3

it See, for example, Okun [23].
4 Awareness of such policy problems is evidenced in the testimony of John Dunlop [111] before the Sub-
committee on Economic Growth of the Joint Economic Committee.
IA main criticism of the "official" (trend) method for computing potential GNP is that many factors
which affect long-term output growth in the economy do not change smoothly. Relationships between
factors of production in the current period, for example, may not correspond to those obtained in the base
period from which the trend-type calculation is made. For further discussion of this point, see Denison [10].


Age group 1960 1970 1980 1990

16 years and over.-------8-------- 82.4 79.2 78.0 78.4
16 to 19 years-------.....--.----------------------- 58.6 57.5 56.0 55.4
20 to 24 years-.-.-.......- ---- 88.9 85.1 83.0 82.1
25 to 34 years.---------------- ---------------- 96.4 95.0 94.6 94.4
35 to 44 years.------------------------------- 96.4 95.7 95.1 94.7
45 to 54 years.- ---- --- 94.3 92.9 91.9 91.5
55 to 64 years----------- --------------------- 85.2 81.5 79.1 77.5
65andover------------------------- ----- 32.2 25.8 21.2 19.3
16 years and over----------------------------- 37.1 42.8 45.0 45.9
16 to 19 years.------------------------------- 39.1 43.7 45.5 47.0
20to 24 years.----------- -------------------- 46.1 57.5 63.4 66.2
25to 34years.. ------------ 35.8 44.8 50.2 51.5
35 to 44 years.------------------------------- 43.1 50.9 53.2 55.2
45to54years------------------------------- 49.3 54.0 56.2 58.0
55to64years..-------------------------------- 36.7 35.6 37.5 38.8
Both sexes 16 and over. -------------------59.2 60.3 60.8 61.5

Source: "Manpower Report of the President," 1975.



The model developed here relates to the private nonfarm sector of
the U.S. economy. This coverage is appropriate since most of the
fluctuations in employment occur in that sector. Employment in the
Government and farm sectors does not fluctuate greatly, and the
forces determining the levels of output and employment in these
sectors are less market-determined.1 The core of the potential GNP
model developed here consists of two relationships: the aggregate
production function and the civilian labor force determination equa-
Traditional production function analysis has concentrated on the
role of capital and labor in the determination of net output (net of in-
termediate production) of goods and services. Nonetheless, as wit-
nessed recently, primary inputs (raw materials) do play a crucial role
in the determination of levels of output.2 Considering this factor, a
production function was specified with manhours, capital stock times
capacity utilization (to approximate the flow of services from the
capital input), and raw materials as inputs. Technological change was
specified as a constant rate of growth. The empirical form of the
function selected was Cobb-Douglas, restricted to constant returns to
scale. The estimated rate of technical change was 1.8 percent per year,
and the elasticities of output with respect to labor, capital and raw
material inputs were 64, 25 and 11 percent, respectively.3
Initially, we had planned to estimate the production function
along the lines followed by Berndt and Wood [2]. Their framework
considers gross output (value added plus material and energy inputs)
in relation to capital, labor, energy and material inputs, and provides a
consistent definition for proper estimation of the production function
for a subsector of the economy. At higher levels of aggregation of
the private economy, extending this framework becomes more difficult.
Time series data for value added and gross output are available from
BEA only for the manufacturing sector, which accounts for about 28
percent of total GNP. It is possible though not easy to construct a
gross output measure for perhaps 40-45 percent of GNP. An alterna-
tive which was considered involved the use of data developed by
Faucett [13], which was used by Berndt and Wood. These data relate
to a highly aggregated input-output coverage of the economy, em-
phasizing energy input sectors, with a few basic demand sectors.
Unfortunately, these data covered 1947-1971. Extending the series
through 1974 is a task beyond the resources of this study.'
I Government sector output is defined in terms of employment and constant dollars of compensation per
employee of Government workers. By definition, the Government sector always fully utilizes its supply
potential. Hence, variations in potential output in the overall economy do not depend upon Government
sector output, and potential GNP is only different from potential private output by a scalar factor. Most
variation in farm output is due to factors which affect supply, principally weather conditions. Farm policies
which affect land utilization also affect farm output levels, but these are of less importance in the short-run
than weather and decisions taken by farmers.
2 This became evident even before the energy crisis, when the effects of Government price freezes on raw
materials markets were observed in the summer of 1973.
s The empirical estimates are given in Appendix A.
4 See Appendix B, Note 1.

This study has concentrated on a more complete specification of
the input side of the production function. It is conventional practice
to estimate production functions in terms of labor and capital inputs,
but returns to land (as a primary factor of production) are usually
omitted. Christensen and Jorgensen [5] have treated returns to land as
a component of total capital in measuring real capital input in the
economy. We have not split the capital stock into land and other types
of capital, but have approximated this by the introduction of an input
variable for primary materials, in addition to the usual capital stock
variable representing structures and equipment.,
Items included in the raw materials input index used in the model
have been constructed from the following data series:
1. Domestic crude oil production (45.8)
2. Domestic natural gas production (14.9)
3. Domestic bituminous coal production (12.0)
4. Domestic copper ore production (1.3)
5. Domestic lead ore production (.5)
6. Domestic zinc ore production (1.8)
7. Domestic iron ore production (3.9)
8. Imported crude oil (5.4)
9. Imported natural gas (6.8)
10. Imported refined petroleum products (3.1)
11. Imported copper ore (1.2)
12. Imported iron ore (2.5)
13. Imported bauxite (.8)
The included items are obviously only a partial listing of possible
choices. The energy items give a reasonably comprehensive coverage in
that area, while the other items pertain to the most important metals
utilized by the industrial sector of the economy.7
A demand relationship for manhours was specified as a function of
output, the real wage, and the ratio of goods output to total output
(MIX). The latter variable was introduced to handle problems
associated with aggregating over goods sector and service sector
employment: Average weekly hours are determined by the rate of
change of output, the real wage,8 and the MIX variable. The per-
centage change in money wages is determined by the inverse of the
unemployment rate, the'change in that rate, and the percentage
change in current prices. Prices in turn are a function of unit labor
costs and raw materials and energy prices. The capital stock is taken
as exogenous-which for short-run (one year) computations of
potential GNP seems plausible. Capacity utilization was treated as a
function of the unemployment rate and a time trend. Raw material
inputs were made to depend on the level of output, and were thus
treated as a derived demand.9
SSee Appendix B, Note 2.
The figures in parentheses indicate theopercentage share of each item in the index based upon 1967 values.
1 The main source of data for the raw materials is the Bureau of Mines' Minerals Yearbook. In ,some cases
prices were given; in others, unit values are used, reflecting all the mix-shift anomalies inherent in that
measure. All of the important prices are represented by unit value series. The various items were added
together by construction of a constant dollar value series from the physical quantity series. A price deflator
for the overall series was developed by dividing the current and constant dollar estimates, and this price
series was used as a variable in the price function for total output. ,
The real wage variable can be interpreted to reflect both demand and supply factors. Other things equal,
workers will desifre'more leisure as real wages rise. Firms, by contrast, can be expected to increase' average
hours as real wbgfall. 'I'
' The own elativee price did not prove to have significant statistical explanatory power. Also, unlike for
labor input, factor market price determination in terms of a price-excess demand relationship or factor
demand-supply determination, was not successfully modeled. Since raw materials prices are highly de-
pendent upon world market factors, these prices were taken as exogenous.

Other than the standard identities, the remaining relations of in-
terest in the model are those for the determination of the male and
female labor force participants. We found that female labor force par-
ticipation could be estimated in a time series analysis using employ-
ment, real wages and the mix of output between goods and total
output. The trend in real wages closely approximates the trend in
female participation rates over most of the postwar period. It is
doubtful if indexes reflecting other variables considered by Orcutt or
Cohen 10 would provide statistically significant additions to the time
series explanation of female participation rates.
Male participation rates have generally declined over most of the
period since 1960, reflecting such factors as earlier retirement among
those 55 years of age and over, and increased time spent in school by
those in the younger age groups. Interestingly, participation by Negro
and other races is lower than for whites, and has declined even more
rapidly .This may reflect a number of factors such as migration to
urban areas and the corresponding lack of match-up between skills
and jobs, plus outright discrimination. These conditions no doubt
explain the higher rate of "participation" of these groups in the
Armed Forces. After some analysis, we found that for the male group-
ings most of the change in the aggregate male labor force could be
explained by the differential movements of the various age groups in
the population. That is, with constant (1960) participation rates,
changes in population in the various age groups explained most of the
variation in overall male labor force growth. We have not yet extended
this hypothesis to the female labor force, but these results may sug-
gest that participation rates are not changing in response to economic
variables as much as some research suggests, but rather that much of
the variation in labor force growth merely reflects changes in the age
distribution of the population. In our judgment, research on this
subject has not yet resolved the question of which are the most im-
portant determinants of labor force growth: the age distribution of the
population, cyclical economic factors, or strictly demographic factors
such as marriage, number of children, extent of education, and so
In studying the behavior of average weekly hours, we found that
inclusion of a variable reflecting the share of goods production in total
production measurably improved the explanation of the cyclical be-
havior of average weekly hours. Perry [27] has also attempted to
handle mix problems in the aggregate function for average hours by
using an unemployment rate weighted by employment of different
groups in the labor force as the cyclical variable.
Since the determination of average hours and the labor force are
important to estimation of the endogenous unemployment rate, these
functions are also indirectly important to the determination of the
wage rate and price level within the model. We tried several simple
hypotheses about the basic Phillips equation, but none was particu-
10 Early work by Orcutt, et al, [251 shows that labor force participation is related to a number of demogra-
phic-economic variables such as real wages, property-type income, debt and liquid asset positions, the age
and number of children in a family, the level of education, and the structure of labor market demand in a
particular locality. Cohen and others [9] have taken a similar approach, but also have investigated the "dis-
couraged worker" hypothesis using cross-section data. and the general finding was that "discouragement"
was greater in areas (labor markets) where the growth of employment was modest or nonexistent for long
periods of time. This work supports the aggregate sort of time series analysis mostly attributed to Simler and
Tella [281, in which female participation is hypothesized to be highly correlated with changes ineffective
employment demand;

larly rewarding. One idea was to approximate that a wage earner's
perception of real wage changes due to price inflation is most influ-
enced by frequently purchased items (such as food, gasoline or regu-
larly paid utility bills), and that possibilities for money illusion are
greatest with respect to infrequently purchased goods such as durable
items. However, we found that a mere re-weighting of the CPI is not
adequate as a means for approximating the frequency of purchase
concept. Additional work would have to be done here to develop the
correct data concept. Other factors which are important to aggregate
wage behavior include an adequate incorporation of effects of collective
bargaining cycles, and differences in the response of wages in the
market-oriented segments of the labor market compared to Govern-
ment sectors. With respect to the latter, some recent work by Hall [16]
has shown that wages in the "nonentrepreneurial" sector (Govern-
ment, communications, etc.) are unresponsive to cyclical forces,
whereas wages in the "competitive" sector (trade and much of manu-
facturing) are responsive to changes in cyclical conditions. The
implication he draws from this is that average wages in the com-
petitive sector are higher than otherwise, since this sector has to com-
pete with the nonentrepreneurial sector to retain its talented workers.
He ascribes less of the non-cyclical wage push to the union-nonunion
dichotomy in the labor force. However, there is some evidence that
collective bargaining cycles do affect the change in wages on a period-
to-period basis. A year of heavy collective bargaining with large
first-year contract settlements may lead to changes in wage rates
which cannot be fully explained by concurrent changes in cyclical
conditions. The difficulty in introducing this sort of information into
an annual time series wage equation has to do with the availability of
data. BLS has published data on collective bargaining settlements
covering 1,000 workers or more only since 1964, but it is a considerable
task to construct a proxy from information in various issues of Current
Wage Developments for prior years. It is our opinion that a consider-
able specification improvement can be made by incorporating some of
these ideas into a more elaborate wage equation, and this would be one
objective of further research.
It should be pointed out that the model is based upon annual time
series, fitted from 1950 to 1974. There are several reasons for using
annual data. First, data on energy and material inputs are not pub-
lished by the Bureau of Mines on a quarterly basis. Similarly, capital
stock figures would have to be interpolated from the annual figures.
Since we wanted to focus upon structural relationships, and since the
potential capacities of the economy can only be changed over the
longer-run, we felt that an annual model would be appropriate. How,
ever, there are areas in which a quarterly model would provide some
advantages. For example, in the adjustment of manhours to output,
the lags may be less than one year. Similarly, average weekly hours
may respond to changes in output in less than a one-year period.

This chapter briefly discusses some of the main results from the
model. The measurement of potential output using the model is
considered first. The essential economic behavior embodied in the
model is discussed in terms of elasticities between principal endogenous
and exogenous variables. Events of recent years are reviewed in a
hypothetical way, and absorption of the labor force during the re-
mainder of the 1970's is considered in light of energy supplies and
growth and utilization of the capital stock.
The definition of potential output is arbitrarily dependent upon
the selection of utilization rates for the labor force, capital stock and
material inputs. It is conventional to assume that the capital stock is
the fixed input in the short-run, with labor and materials being
relatively more variable inputs. It is also assumed that a four percent
unemployment rate corresponds to "full employment." In order to
"solve" the model for four percent unemployment, the equation
defining unemployment was set equal to .04. Technically, the model
at this point had n variables and n+ 1 equations. Hence one equation
had to be deleted. The appropriate equation to drop was the labor
(or more exactly, manhours) demand relation. Manhours is thus
removed from the model and is required to be at a level consistent
with a four percent unemployment rate. Capital services and raw
material inputs are then determined at levels consistent with "four
percent unemployment" output, with no regard to supply constraints.
The remaining relations in the model, labor force, wages, prices, and
average weekly hours are also jointly determined.
Table 2 presents figures on actual and potential output in the
private nonfarm economy, 1951-1975, using four percent as the full
employment unemployment rate. In comparison with "official"
figures published in the Commerce Department's Business Conditions
Digest, the model produces larger estimates of the GNP-gap. For
example, in 1974 the model generated a gap between actual and
potential output of $77.7 billion while the BCD estimate is $64.5
million for a more inclusive definition of GNP. The corresponding per-
centage gaps for total GNP are 8.6 and 7.3 percent, respectively.
This difference is mainly attributable to the fact that labor force
growth in the model is considerably higher than that assumed in
official estimates, particularly for recent years. Also, average weekly
hours decline at a slightly slower rate in the model. For example,
official estimates for the period since the fourth quarter of 1969
embody an annual growth rate of potential output of four percent,
while the model produces a much higher average annual growth in
the private nonfarm sector of 4.6 percent. Official estimates assume
that the potential labor force will grow at an annual rate of 1.8

percent, while average hours will decline by .3 percent, and produc-
tivity growth will average 2.5 percent per annum. The model generates
about the same average annual productivity growth rate since 1969,
but average weekly hours decline by .28 percent. However, in the
model, the average annual growth rate of the potential labor force is
2.4 percent over the 1969-1974 period, about the same as the actual
growth rate of the labor force during this period, but well in excess
of the 1.8 percent assumed in the official estimate.

[Dollar amounts in billions of 1958 dollars]

Actual output Potential Percentage
Year (A) output (P) Gap (P-A) gap (P-A)/P

1951 ...- -- ---------------------------------- $326.2 $323.9 -$2.3 -0.7
1952...---------------------------------------- 334.2 332.8 -1.4 -.4
1953.... --------------------------- 351.1 350.5 -.6 -.2
1954..----------------------------------------- 345.8 362.9 17.1 4.7
1955 ---------------------------------------- 376.3 377.5 1.2 .3
1956---------------------------------------- 384.0 390.3 6.3 1.6
1957 ----------------------------------------- 390.2 406.6 16.4 4.0
1958......- ------------------------------------- 384.4 421.5 37.1 8.8
1959----------.--------------------- 412.3 436.8 24.5 5.6
1960----------------------- 422.1 453.2 31.1 6.9
1961.--------------------- 430.1 470.2 40.1 8.5
1962 ----------------------------------------- 460.8 486.4 25.6 5.3
1963----------.--------------------- 480.4 506.2 25.8 5.1
1964 --------------------------------------- 509.7 527.2 17.5 3.3
1965------------------------ 543.3 551.0 7.7 1.4
1966...---------------------------------------- 581.1 574.3 -6.8 -1.2
1967---------.--------------------- 593.6 595.8 2.2 .4
1968 ---------------------------------------- 623.6 621.5 -2.1 -.3
1969--------------------- 640.8 651.1 10.3 1.6
1970 --------------------- 639.6 680.2 40.6 6.0
1971 ---------------------------------------- 658.7 710.8 52.1 7.3
1972---------------------------------------- 705.7 744.2 38.5 5.2
1973.---------------------------------------- 749.5 776.2 26.7 3.4
1974---------------------------------------- 729.5 807.2 77.7 9.6
1975----------------------- 713.6 840.1 126.5 15.1

As it is customary to discuss the gap between actual and potential
output in terms of total GNP, rather than private nonfarm GNP,
corresponding figures are presented in Table 3 for comparison. As
explained above, the path of potential, output is taken as independent
of the growth of farm and government sector outputs. Hence, com-
puted percentage GNP gaps vary only with the comprehensiveness
of the GNP concept employed


[Dollar amounts in billions of 1958 dollars]

Potential output Percentage gap measures
Total Private non- Total Private non- Unemployment
Year GNP2 farm GNP GapI GNP farm GNP rate (UR)

1951------------------- $381.1 $323.9 -$2.3 -0.6 -0.7 3.3
1952------------- 393.6 332.8 -1.4 -.4 4 3.0
1953------------- 412.2 350.5 -.6 -.1 -.2 2.9
1954 ------------------- 424.2 362.9 17.1 4.0 4.7 5.5
1955------- 439.1 377.5 1.2 .3 .3 4.4
1956- -------- 452.4 390.3 6.3 1.4 1.6 4.1
1957..-------- ..----------- 468.8 406.6 16.4 3.5 4.0 4.3
1958 ------------------ 484.4 421.5 37.1 7.7 8.8 6.8
1959 ----- -- 500.4 436.8 24.5 4.9 5.6 5.5
1960 ------------------- 518.8 453.2 31.1 6.0 6.9 5.5
1961 ------------------- 537.2 470.2 40.1 7.5 8.5 6.7
1962--------- 555.4 486.4 25.6 4.6 5.3 5.5
1963------------------- 576.8 506.2 25.8 4.5 5.1 5.7
1964---------- --------- 598.6 527.2 17.5 2.9 3.3 5.2
1965 ------------------- 625.5 551.0 7.7 1.2 1.4 4.5
1966 ------------------- 651.3 574.3 -6.8 -1.0 -1.2 3.8
1967.------- __ 677.3 595.8 2.2 .3 .4 3.8
1968 ------------------- 704.6 621.5 -2.1 -.3 -.3 3.6
1969 ------------------- 735.9 651.1 10.3 1.4 1.6 3.5
1970------------------- 765.7 680.2 40.6 5.3 6.0 4.9
1971 ------------------- 797.9 710.8 52.1 6.5 7.3 5.9
1972------------------- 830.9 744.2 38.5 4.6 5.2 5.6
1973------------------- 865.9 776.2 26.7 3.1 3.4 4.9
1974.-------- ----------- 898.8 807.2 77.7 8.6 9.6 5.6
1975 ------------------ 945.7 840.1 126.5 13.4 15.1 8.5

1 From table 2, col. 3.
2 Private nonfarm potential output (table 2) plus farm and Government sector output.
I Preliminary estimate.


Table 4 presents a set of elasticities derived from the model. For
percentage changes in exogenous variables changes in key endogenous
variables can be calculated. Most of the figures in the table are self-
explanatory. It may be noted that changes in the capital stock, mate-
rial inputs or the mix of goods output relative to total output have no
impact upon the male labor force. This follows from the fact that the
male labor force responds only to changes in the male population and
a time trend. The female labor force, on the other hand, is responsive
to a number of variables generated within the model. Similarly, be-
cause a price-quantity relationship has not been established for mate-
rial inputs, the level of material inputs does not affect prices. Shifts
in the output mix deserve explanation. Changes in the ratio of goods
production to total production (thus, production of goods versus
services) have a powerful effect through determination of the demand
for inputs. Relatively more goods production reduces employment
(presumably reflecting capital for labor substitution, and/or higher
capital-labor coefficients associated with goods production), and the
growth of the female labor force. Negative effects here are, para-
doxically, large enough to generate a reduction in total unemployment
and the unemployment rate. This suggests that a larger share of
service sector production will induce more female participation in the
labor force and raise total employment, but will simultaneously raise
the unemployment rate.



Exogenous variablesI
Materials input Male population by age groups
Endogenous variables' Capital stock Quantity Price Output mix 16 to 17 18 to 19 20 to 24 25 to 34 35 to 44 45 to 54 55 to 64

Average weekly hours........ 0.02 0.01 0.01 0. 19 0 0 0.01 0.02 0.01 0.01 0.01
Real output.---------------- .37 .16 .02 .09 0 0 .01 .02 .01 .01 .01
Total employment---- .------- .14 .06 .01 -.04 0 0 0 .01 .01 .01 0
Total man-hours ------------ .19 .08 .02 .14 0 .01 .02 .03 .02 .02 .02
Wage rate---.---..------------ .22 .08 .14 .13 -.04 -.05 -.15 -.26 -.21 21 -.16
Civilian labor force .....----- .03 .02 0 -.08 .02 .02 .07 .12 .09 .10 .07
Male.-.------------------ 0 0 0 0 .03 .04 .11 .20 .15 .16 .11
Female-..---------------- .09 .04 .01 -.19 0 0 0 .01 0 0 0
Total unemployment--.--- -2.49 -1.10 -.48 -.93 .38 .54 1.57 2.92 2.25 2.33 1.68
Unemployment rate---------- -2.52 -1.11 -.49 -.86 .36 .51 1.50 2.77 2.14 2.21 1.60
Price level-...---------------.03 0 .18 .13J -.03 -.04 -.10 -.17 -.14 -.14 -.11

I Measured as the percent change in the engdoenous variable given a 1-percent change in the
exogenous variable.
2 Variables are defined in app. A.


The gap between actual and potential output has been substantial
since 1970, as shown in Table 2. While a good part of this may be
attributable to demand management policies, there have been im-
portant changes in the supply capabilities of the economy which
have become a concern only recently. Labor force participation and
growth have both increased. Domestic crude oil production peaked by
1972, and natural gas production slowed, partly in response to reg-
ulation of prices in that sector. Energy imports rose dramatically
after 1970, until the extraordinary price increases and export policies
imposed in 1974 dramatically altered both demand and supply.
These factors underlie the movements of total energy supply and its
components as shown in Table 5. During this period, an increasing
proportion of capital investment was allocated to pollution abatement
activities, as opposed to meeting requirements for capacity expansion
or replacement. Investment to meet pollution requirements may have
introduced some anomalies in the measurement of effective capital
stock utilization. Simple addition of investment for pollution abate-
ment to the overall capital stock may overstate the growth of the
productive capital stock, and thereby understate actual capacity
utilization. For example, in 1973 the capacity utilization rate was
83 percent, while in the peak year of 1966, it was 91.9 percent. In
both years, however, the ratio of goods production to total output
was approximately the same, there is no reason to think that capital-
output ratios have fallen significantly over so short a period.

Crude petroleum Refined Natural gas
petroleum Bituminous Total
Year Domestic Imported imports Domestic Imported coal energy

Indexes 1967 equals 100:
1967--------------- 100.0 100.0 100.0 100.0 100.0 100.0 100.0
1968--------------- 104.7 114.6 110.3 105.8 115.5 93.3 104.4
1969 ------ 105.1 124.8 124.7 113.6 128.8 95.6 108.1
1970._-- ---- 110.2 117.2 148.8 121.8 145.5 104.2 114.9
1971----- 107.4 148.8 159.3 124.5 165.6 94.5 115.4
1972--------------- 107.6 196.8 179.8 123.6 180.7 108.4 121.9
1973------ 104.9 315.0 198.4 125.9 180.0 100.5 128.5
1974-.-------------- 100.7 330.6 166.3 119.8 171.4 102.3 123.9
Percentage changes:
1968-------------- 4.7 14.6 10.3 5.8 15.5 -6.7 4.4
1969-----------.... ---- .4 8.9 13.0 7.4 11.5 2.5 3.5
1970--------------- 5.8 -6.0 19.4 7.2 12.9 9.0 6.3
1971.--------------- -2.5 26.9 7.1 2.2 13.9 -9.3 .5
1972------.1 32.3 12.8 -.7 9.1 14.7 5.6
1973-------------- -2.5 60.0 10.4 1.9 -.4 -7.3 5.4
1974 ----------------4.0 4.9 -16.2 -4.9 -4.8 1.8 -3.6
I Represent those inputs embodied in the model. Indexes are based on constant dollar figures. Basic data are from the
U.S. Bureau of Mines,

We have attempted to use the model to evaluate the 1970-1974
period in terms of changes in supply variables. It is of some interest
to determine how much more of the growing labor force could have
been absorbed during these years if "normal" supply conditions had
prevailed, and demand management had been reasonably astute. The
assumptions made below might have looked reasonable if planning
ahead, say, from 1970:
1. Energy inputs grow at 5.5 percent per annum, reflecting deregu-
lation of natural gas prices and demand characteristics incorporating


higher fuel consumption by4automobiles imposed by pollution stand-
ards and heavier weights.
2. Energy prices rise by 7.5 percent in each year, which is approx-
imately the weighted average rate of increase for the 1970-1973 period.
In the model, this converts into a 6.5 percent increase in the material
input price variable.
3. To offset requirements for pollution abatement, the capital stock
expands by an additional 5 percent in each year. This is a modest
change which disregards any changes in investment demand stem-
ming from the devaluation of the dollar.
4. Capacity utilization is mainained at the 1973 level of 83 percent
in all years. With the growth in the capital stock, this produces a total
growth in services from the capital stock of 6 percent for each year.
5. In conjunction with a higher rate of energy and capital stock
utilization, the share of goods output in total production is higher, on
average, by 1.4 percent (that is, all years are set at the 1973 value for
this variable).
Using the appropriate elasticities from Table 4, the following results
are obtained:
[In percent
Impacts due to changes in i-
K QRM PRM MIX impact
Average weekly hours-..-----.---.-------- 0.12 0.05 0.07 0.27 0.51
Real output------.-------------------- 2.22 .80 .13 13 3.28
Employment-.---.----..----------------- .84 .30 .07 -.06 1.15
Total man-hours -.--------------------- 1.14 .40 .13 .20 1.87
Wage rate-.-. ------- 1.32 .40 .91 .18 2.81
Total labor force-- .------ .18 .10 .02 -.11 .19
Unemployment rate----------..------- -15.12 -5.55 -3.18 -1.20 -25.05
Price level--------------------------- .18 0 1.17 .18 1.53
SAbbreviations: K equals capital, QRM equals materials, PRM equals price of materials, and MIX equals ratio, goods
output to total output

These results indicate that the unemployment rate would have
been 1.2 percentage points lower-i.e., instead of an annual average
rate of 5.4 percent during these years, the unemployment rate would
have been 4.2 percent. Actual output would have been 3.3 percent
higher in each year, and the average gap between actual and potential
output would have been reduced from $45.1 billion per year to $34.2
billion. The gap as a percent of potential output would have averaged
4.5 percent instead of 6.6 percent. On average, there would have been
1.0 million more jobs in the economy each year between 1970 and 1974.
Under the above assumptions, the overall price level rises by 1.53
percent per annum. Because assumption (2) approximates the actual
energy price change which occurred, the net increment to the price
level due to differences between assumed and actual values for other
variables is only .36 percent. This net price increase is consistent with
the implied increment to the annual growth of productivity and real
wages of 1.4 percent. Private nonfarm real wages actually grew at an
annual rate of only 1.1 percent, on average, during 1970-74. Con-
sequently, achievement of a somewhat higher rate of growth and
utilization of the economy's resources, as under these assumptions,
implies a rather substantial increment to the growth in real wages
without significant supply-side price inflation.

The model can also be used to trace out the direct impact of the
"energy crisis" of 1974. Energy prices increased 111 percent from 1973
to 1974, while the quantity of inputs declined by 3.6 percent. The
decline in energy inputs increased the unemployment rate by two-
tenths, and the price increase contributed 15.6 percent to the overall
price level, and through the wage-price feedback, 8.6 percent to the
wage level. While the model does not incorporate indirect effects of
these changes upon utilization of the capital stock and the composition
of total output, it is fair to assume that a portion of reductions here
(those not attributable to tightened monetary policy in response to
the price effects) are due to changes in the price and quantity of
energy available. Goods production as a share of total output fell
by 1.1 percent, and capacity utilization fell from 83 to 79 percent.
Even though the real capital stock grew by an estimated four percent
in 1974, the decline in capacity utilization reduced total capital
services by one percent from 1973 levels. These changes would add
an additional one to two-tenths to the unemployment rate. Thus,
the model would suggest that about four of the seven-tenths increase
in the unemployment rate between 1973 and 1974 was attributable
to the decline in energy inputs and their higher price.
In order to consider the impact of future trends in energy inputs
during the remainder of the 1970's, we have used the FEA's Project
Independence study. Growth patterns in crude oil production are based
upon the maintenance of an $11 per barrel price in constant (1974)
dollars through 1985. Natural gas prices are decontrolled, and are
assumed to rise to an average of $1 per MCF in 1974 prices. Coal is
assumed to have a price of $15 per ton in constant prices. Nominal
prices for oil are assumed to rise at an annual rate of 3.5 percent.
Exact assumptions are shown below.
[in percent]
Real Price
growth increase
Domestic crude oil--------------------------------------- 2.8 3.5
Imported crude oil-------------- ---------------------------------------2.0 3.5
Imported refined products...---.----------------------------------- 0.0 3.5
Domestic natural gas------.. .-----------.------- 2.0 8.5
Imported natural gas..- -------------------------------------------- 0.0 8.5
Domestic coal-.-------------------. -------------------------.-----.. --- 4.0 8.5

The weighted impact of the above assumptions yield a 4.3 percent
increase in material input prices, and a 2.1 percent increase in the
quantity. It may be noted that the latter value is considerably lower
than the 4.4 percent increase in energy inputs which prevailed during
1963-1973. These assumptions about energy inputs are assumed to
apply through 1980, and can thus be interfaced with population
projections for 1980. With assumptions about the composition of
output and growth in the capital stock, the extent of unemployment
imposed by plausible supply conditions can be evaluated.
The availability and price of imported energy is an important un-
known with respect to future supplies. Not only is imported energy
subject to policy changes of foreign governments, but under a regime of
flexible exchange rates, it is subject to fluctuations in the exchange
value of the dollar. On the basis of recent history, fluctuations in the

weighted exchange rate between the dollar and currencies of those
countries supply the U.S. economy with raw material inputs easily
could be in the range of plus or minus 20 percent. To stimulate the
impact of a change in the exchange rate, the following is assumed:
1. The short-run price elasticity of import demand for raw ma-
terial products is low-on the order of 0 to -.5-reflecting the fact
that substitution possibilities are limited in the short-run.
2. In 1974, the share of imports in the total materials input in the
model was about 28 percent. Although energy imports may decline
in terms of growth rates in future years, it is assumed that a rising
dependence upon other imported materials will continue. Hence,
the share of imports in total materials has been kept at 1974 values,
rather than reduced to pre-1970 levels.
If the price elasticity of demand for imports of raw materials is on
the order of -0.5, a 20 percent change in the exchange rate (plus or
minus) implies a 10 percent change (plus or minus) in the quantity of
imports, and a 2.8 percent change in the materials input variable
within the model. This produces a four-tenths of one percent decline
in output, a two-tenths of one percent decline in employment, and a
three percent rise in the unemployment rate, assuming there are no
domestic substitutes for these imported inputs. The overall price
level would be higher by about one percent.
Using the estimated relationship for the male labor force from the
model and the population projection figures from Table 6, we estimate
the male labor force in 1980 to be 63.126 million persons. This figure is
14.4 percent above the 1974 value of 55.190 million, and slightly
greater than the 1980 projection of 62.590 million given in the 1975
Manpower Report. Using a similar methodology for the female labor
force gives a 1980 figure of 40.970 million persons, as opposed to the
Manpower Report figure of 39.219 million. The Manpower Report
projection relies upon declines in the participation rate in certain
important age groups of the male population, while our projection is
based upon the model equation. Our figure implies a growth in the
female labor force of 14.4 percent. Hence, the growth in the civilian
labor force from 1974 to 1980 will be 14.4 percent, resulting in a labor
force of 104.1 million. Using our estimated production function and
allowing capital and technological change to grow at their historical
rates of 4.05 percent and 1.8 percent, and allowing materials to grow
at 2.1 percent,1 it will be necessary for private non-farm output to
grow at an annual rate of 4.62 percent in order to achieve a reduction
in the unemployment rate from 5.6 percent in 1974 to 4.0 percent in
1980.2 Historically, the average annual growth rate of output has
been 3.5 percent. In order to simply absorb the projected labor force
growth and maintain the unemployment rate at the 1974 figure of
5.6 percent, output will have to grow at an annual rate of 4.44 percent,
still far above the historical pattern.
SAlso assumed is a constancy in the average weekly hours.
I Since output declined by 2.2 percent in 1975 from its 1974 level, computing from 1975 indicates that output
would have to grow at a compound annual rate of 6 percent through 1980. If the economy were to grow by 7
percent in both 1976 and 1977, it would have to achieve an annual growth rate of 5.4 percent for the years
1977-1980 to reduce the unemployment rate to four percent by 1980.


Percent change in population
Age group (years) 1974-80 1974-90

16 to 17....-- ------------------------------------------------------------3.0 -24.3
18 to 19....------------------------------------------------------------- 4.4 -11.5
20 to 24.......----.----------------------------------------------------- 11.5 -4.1
25 to 34 ----------------------------------------------------------- 21.6 38.0
35 to 44....-- ----------------------------------------------------------- 12.5 60.8
45 to 54...---------------------------------- --------------------- -4.4 6.5
55 to 64 -------------------------------------------- ---------------- 7.7 5.5
65 and over......------- ------------------------------------------------- 10.6 28.5
All groups--- --------------------------------------------------- 9.4 19.5

I Based upon series II data of the "Current Population Reports," U.S. Bureau of Census.


The foregoing analysis has led us to the following conclusions con-
cerning potential GNP in the U.S. economy. First, present levels of
GNP are substantially below potential levels ($200 billion, measured
in current dollars). In the present context, the likelihood of supply
restrictions and bottlenecks preventing the attainment of a 4.5 percent
unemployment rate appear to be relatively small.3 This statement is
conditioned upon the assumption that there will be no arbitrary
disruption of energy supplies or radical changes in their prices. If this
can be assumed, much of the present gap between actual and potential
output can be viewed as a demand management problem.
Views differ as to the causes of the current slump. There is evidence,
however, that the restrictive monetary actions taken in 1974 in an
effort to stem inflationary forces in the economy were based upon a
misreading of the causes of that inflation and the appropriateness of
monetary policy for dealing with the problem. Most of the inflation
which occurred is attributable to radically higher fuel and food prices,
induced largely by changes in supply conditions. Higher energy prices
had a particularly large impact upon automobile demand. Consumers'
confidence was badly shaken by the oil embargo which made the
availability of fuel highly uncertain. The rapid increase in gasoline
prices coupled with lower fuel mileage due to Government-mandated
pollution equipment, made new automobiles even less attractive.
Radically higher prices of new cars reflecting added costs associated
with safety and pollution equipment and higher production costs
(due in part to higher energy costs) also served to severely dampen
consumer demand. The construction sector was badly victimized by
the higher interest rates associated with restrictive monetary policy.
Also, a good part of the over-accumulation of inventories in 1974 was
in response to supply disruptions in basic materials.
The following reduction in inventories may, however, have been
more in response to monetary factors affecting both demand for
output and the cost of holding inventories. In addition to these events,
the many disruptions and uncertainties for both consumers and pro-
3 For example, the unemployment rate for experienced wage and salary workers climbed from 4.5 percent
in 1973 to 8.2 percent in 1975. but has declined to 7.4 percent in early 1976. Unemployment rates for blacks,
teenagers and part-time workers have risen more sharply and may be more difficult to reduce through aggre-
gate demand policies than previously. However, experienced workers presently unemployed comprise over
three-fourths of total unemployment, and the employment for these workers is generally responsive to
aggregate demand conditions.


ducers which accompanied the period of price controls significantly
eroded confidence and upset expectations.
The above statements do not imply that demand management
policies needed to return the economy to full employment are easily
designed.4 However, the downward trend in the unemployment rate
achieved during 1973, which produced a 4.5 percent rate in October
of that year, would appear to be a feasible goal and attainable through
usual monetary and fiscal policy measures. Should supply-induced
inflationary forces reappear as in 1973-1974, restrictive monetary
policy will not serve to solve those problems-reduced output and
employment will work the other way. Lost opportunities for capital
accumulation imply a lower growth rate in trend productivity, and
other things equal, higher rates of inflation over the longer run.
Although in our judgment most of the present weakness in the
economy reflects deficient aggregate demand, we are cognizant that
supply side conditions are more relevant than in the past. For example,
in 1973-1974, there were important capacity constraints in some sectors
of the economy (steel, paper, petrochemicals, and gasoline refining).
In some cases these deficiencies reflected a history of inadequate
domestic investment due to prior over-valuation of the dollar. In
other cases, safety and pollution investment comprised a significant
share of total investment. At the present time, however, there is little
reason to expect that capacity problems in these sectors would lead to
sectoral bottlenecks or outsized price increases under conditions of
substantially greater stimulus to aggregate demand than contem-
plated by the Administration for 1976.5

A conclusion we draw which pertains more directly to the results
of this study concerns the growth of potential output over time. Even
considering all of the disruptions in supply which occurred in the
1973-1974 period, most of the gap between actual and potential out-
put is due to the deficient demand, rather than a significant alteration
of the potential path itself. The above analysis indicated the substan-
tial growth which can be expected in the labor force over the next
few years, and on into the 1990's. What is needed at this point for
analyzing the employment issue, the energy question, and (thus) the
"capital shortage" issue, is a framework for analyzing the growth
potential of the U.S. economy. Input-output growth models (such as
the BLS model) or the approach of Denison do not properly address
the dynamics of the growth process. And, in terms of recent experi-
ences, existing models do not adequately incorporate price behavior-
hence, changes in relative prices of both inputs and outputs, which
4 For example, there appear to be increasingly difficult structural problems of aligning labor demands
with age-skill characteristics of the labor force in central city areas of large metropolitan areas.
5 With low worldwide stocks of grains, poor harvests could again set off a rapid inflation in food prices.
If combined with large fuel price increases, inflationary forces as existed during the 1972-74 period could
again appear. If the monetary authority does not want to "validate" supply-induced inflation, lower levels
of output and employment are implied. With supply inflation, it may be necessary to raise target inflation
rates In order to keep other resources in the economy fully utilized.


are most important, are largely ignored.6 Present macro-models used
for short-run forecasting are largely demand-oriented, and deficient in
the portrayal of price behavior. The model developed in this study is
also deficient in terms of portraying price behavior since our model
ignores the determination of output from the demand side.7
To successfully analyze the supply side of the economy, a model is
required which fully treats all inputs. Hence, labor supply by sex, age
and race would be an important component to incorporate in the
model. Equally important would be development of labor demand
relationships for the above categories, so that long-term structural
problems such as black and teenage unemployment, and effects of
minimum wage policies could be studied. Likewise, a sector would be
developed to explain the demand and supply of capital. This would
require expansion of the model to include income determination in
order to link saving and investment behavior to the growth of the
capital stock. Effects of changes in policy variables such as tax rates
and tax credits, depreciation policies, etc., upon output and employ-
ment demand could then be studied.8 To adequately trade the effects
of energy inputs and other primary inputs through the economy re-
quires a disaggregation of the economy into subsectors. An initial
model might include durable and nondurable manufacturing, whole-
sale and retail trade, agriculture, construction, transportation, serv-
ices and government. Only through some degree of disaggregation
accompanied by the specification of production functions with capital,
labor, energy and intermediate inputs can relative price changes be
adequately modeled, Such disaggregation would also permit examina-
tion of various hypotheses with respect to sector differences in wage
adjustment (such as Hall's hypothesis about wage adjustment in the
nonentrepreneurial versus competitive market sectors).
It is clear that a main reason why aggregate models have not
produced meaningful projections in recent years is the presumption of
constant relative price relationships between subsectors. The "washing
out" of these effects in the aggregate did not hold under conditions of
currency devaluation and radical changes in energy prices. Changes in
relative prices have important consequences for the composition of
output and, hence, employment demand. A fully detailed production
function in a multi-sector model including sectoral wage and price
relationships could be used to examine how much the growth potential
of the economy responds to changes in relative prices and distortions
in structure of relative prices (both between sectors and between
I Some recent work by Berndt and Wood [2], covering the manufacturing sector of the economy, considers
the role of energy and material inputs in production in terms of their relative prices. Their results show that
capital and energy are complements, while capital and labor, energy and labor, and energy and other mate-
rial inputs are substitutes. For example, a decontrol of energy prices will reduce capital and energy-inten-
siveness in production, and raise labor-intensiveness, which suggests that freeing energy prices may promote
greater employment in the longer run. On the other hand, expanding the investment tax credit will raise
the demand for capital, and raise the demand for energy (since they are complements). The final employ-
ment effect depends on whether the substitution or expansion impact of the tax credit dominates
7 In addition, no attempt was made to explicitly link the behavior of farm prices to prices in the nonfarm
sector. The impact of farm prices on nonfarm prices since 1972 has been particularly important and may
remain so in years to come.
8 These effects are not obvious. For example, some recent work by Coen [8] shows that the effects of tax
policies on investment and employment are crucially dependent upon the price determination process in
the economy. The effect of the credit on output and employment expansion as opposed to capital for labor
substitution depends upon the extent to which changes in the cost of capital are passed on to product prices
or retained as profits.

factor inputs for a particular sector). Distortions in the wage and price
structure could be evaluated in terms of Government policies (mini-
mum wage laws, wage-price controls, tax credits, depreciation rulings,
etc.) and other institutional arrangements which have contributed to
the existing market structure.
The magnitude of work involved in developing a multi-sectoral
model in which input and output price determination is more ad-
equately portrayed is large. As usual, much of the work is in the
development of appropriate data series to match the economic con-
cepts to be modeled. Consequently, such model development ought
to be carried out in cooperation with appropriate Government
agencies having responsibility for the construction of basic economic


(1) Labor Force Females
(LFF/FP160) =. 0658+. 388(TE/P20-65) +. 053(PNFWR/PNFPI)_
(1. 17) (3. 58) (9. 05)
.70 .32
-. 131 MIX
W2. 992 S.E.E.=. 0033 D.W.=1.00
(2) Labor Force Males
LFM=. 1008E+08+. 7935.SUM--98952. In TIME
(57.73) (1.33)
.80 .01
72=. 998 S.E.E.=. 1656E+06 D.W.=. 53
(3) Quantity of Raw Materials
QRM=. 3501E+10+. 0285.QGPNP
(34. 57)
72=. 981 S.E.E.=. 5639E+09 D.W.=1. 20
(4) Quantity of Gross Private Nonfarm Product
ln(QGPNP/PNFMH)=1. 2315+. 2508 ln(KS.CU/PNFMH)
+. 1109 ln(QRM/PNFMIH)-+. 0176. TIME
(1.4) (13.2)
R2=. 995 S.E.E.=. 013 D.W.=1. 28
(5) Private Nonfarm Manhours
PNFMH=. 6575E+11+. 1042 QGPNP-(. 7769E+10)
(7. 12) (1. 81)
.47 .10
(PNFWR/PNFPI)+(. 1965E+11) MIX
(1. 03)
7Z2=. 994 S.E.E.=. 87E+09 D.W.=1. 91
(6) Private Nonfarm Wage Rate
PNkWR=. 0043+. 00165 UR-1-. 8116 (UR-UR-1)
(3.64) (3.64)
.63 .001
+. 5638.PNFPI
W2=. 77 S.E.E.=. 0109 D.W,=1.81
(7) Private Nonfarm Price Index
PNFPI=. 219+ 1. 27 ULC+. 0466-PR.M+. 0054. TIME
(19.7) (5.5) (6.2)
.69 .045 .064
=.998 S.E.E.=. 0087 D.W.=1.47
I The figures below the regression coefficients in parenthesis are the t-values. The numbers below the
t-values are the elasticities (at the means). A dot above the symbol indicates percent change.

(8) Average Weekly Hours
AWH=-36. 29+4. 66 QGPNP-1. 86 (PNFWR/PNFPI)+ 11. 53-MIX
(3.5) (18.7) (2.6)
.005 .2 .17
W2==.948 S.E.E.=. 266 D.W.=. 89

(9) Civilian Labor Force
(10) Unemployment=CLFTE
(11) Total Employment
(12) Unemployment Rate
*AWH: Average weekly hours (BLS)
*CLF: Civilian labor force (BLS)
CU: Capacity utilization (FRB)
FE: Federal government employment (BLS)
FP160: Female population 16 and over (CENSUS)
FRE: Farm employment (BLS)
KS: Capital stock (BEA)
*LFF: Labor force female (BLS)
*LFM: Labor force male (BLS)
MIX: Ratio of goods output to total output (BEA)
MP16-17: Male population 16-17 (CENSUS)
MP18-19: Male population 18-19 (CENSUS)
MP20-24: Male population 20-24 (CENSUS)
MP25-34: Male population 25-34 (CENSUS)
MP35-44: Male population 35-44 (CENSUS)
MP45-54: Male population 45-54 (CENSUS)
MP55-64: Male population 55-64 (CENSUS)
MP65+ : Male population 65+ (CENSUS)
PC: Price of bituminous coal (BOM)
PCO: Price of crude oil (BOM)
PCOP: Price of copper (BOM)
PIO: Price of iron ore (BOM)
PL: Price of lead (BOM)
PMB: Price of imported bauxite (BOM)
PMCO: Price of imported crude oil (BOM)
PMCU: Price of imported copper ore (BOM)
PMIO: Price of imported iron ore (BOM)
PMNG: Price of imported natural gas (BOM)
PMRP: Price of imported refined petroleum (BOM)
*PNFPI: Private nonfarm price index (BEA)
*PNFWR: Private nonfarm wage rate (BEA)
PNG: Price of natural gas (BOM)
PNFMH: Private nonfarm manhours (BLS)
PRM: Price of raw materials-=VRM/QRM
P20-65: Female population 16 and over (CENSUS)
PZ: Price of zinc (BOM)
QC: Quantity of bituminous coal produced domestically (BOM)
QCO: Quantity of crude oil produced domestically (BOM)
QCOP: Quantity of copper produced domestically (BOM)
*QGPNP: Quantity of gross private nonfarm product (BEA)
QIO: Quantity of iron ore produced domestically (BOM)
QL: Quantity of lead produced domestically (BOM)
QMB: Quantity of bauxite imported (BOM)
* Denotes endogenous variable.


QMCO: Quantity of crude oil imported (BOM)
QMCU: Quantity of copper ore imported (BOM)
QMIO: Quantity of iron ore imported (BOM)
QMNG: Quantity of natural gas imported (BOM)
QMRP: Quantity of imported refined petroleum (BOM)
QNG: Quantity of natural gas produced domestically (BOM)
Quantity of raw materials=PZs7*QZ-PL67*QL-}-PC67*QC
QZ: Quantity of zincs produced domestically (BOM)
+.976*. MP35-44+.56*MP45-54+.862*MP55-69 +.303*MP65+
*TE: Total employment (BLS)
TIME: 1950=1.0, 1951=2.0, etc.
U: Unemployment (BLS)
ULC: PNFWR*PNF\IH/QGPNP-Unit labor costs
UR: Unemployment rate (BLS)
Value of raw materials= PZ*QZ+ PL*QL+ PC*QC+ PCO+ QCO
Denotes endogenous variable.


1. The Faucett data include an input-output flow table for each year. However,
for many sectors of the economy, time series data on gross output are not statistic-
ally very "rich," since the figures are derived by moving a base period ratio of
gross output to value added by value added figures for particular years. True
measures of output for non-industrial sectors of the economy are also often not
available, and deflated income measures are not always reliable proxies. It is also
the case, for highly aggregated subsectors of the economy, as treated by Berndt and
Wood, that intra-sectoral shipments of materials and energy inputs introduce
problems in double-counting; if not treated explicitly, true relationships between
exogenous inputs and total supply can be obscured.
2. Imports have also been treated somewhat differently in the definition of input
used here. In the national income accounts, net exports equal gross exports less
gross imports, a definition compatible with net income and product concepts. This
treatment does not differentiate between imports which enter (and affect) inter-
mediate production from those which enter final demand categories. Excepting
years in which input-output tables have been developed, time series data have not
been developed in which imports are distinguished as between inputs into produc-
tion and elements of final demand. To somewhat account for imports of primary
materials which do bear importantly upon the supply potential of the econGmy,
we have added to our materials input proxy variable key imported materials which
enter the producer goods sector. These are mainly energy and metallic ores. We
think this somewhat compensates for a deficiency in GNP accounting which can be
illustrated in the case of crude oil. Theoretically, changes in the costs and returns
of domestic crude production are captured in GNP and the GNP deflator, since
the value added in the mining sector is measured. Also, if there are no differences in
the measurement of prices of imports and prices of final demands, and no lags
between time of import and time of final consumption, then a change in price of
imported crude oil will not affect the GNP deflator, by definition. In reality,
domestic oil processors mix imported and domestic crude oil to achieve desire
technical characteristics; prices of petroleum products are a function of the input
price mix, and the domestic price cannot be determined independently of the cost
of foreign oil. The value added concept would indicate that it can be. Hence by
adding key imported raw materials and their prices into total input, supply price is
represented more realistically.


1. Berndt, Ernst R., and David 0. Wood, "Economic Interpretation of the
Energy-GNP Ratio," in Michael S. Macrakis, ed., Energy: Demand, Conserva-
tion and Institutional Problems (Cambridge: The M.I.T. Press, 1974), ch. 3.
2. "Technology, Prices and the Derived Demand for Energy," Review
of Economics and Statistics, vol. LVII, No. 3 (August 1975), pp. 259-268.
3. and Laurits R. Christensen, "Testing for the Existence of a Con-
sistent Aggregate Index of Labor Inputs," American Economic Review, vol.
LXIV, No. 3 (June 1974), pp. 301-403.
4. Black, Stanley W., and R. Robert Russell, "An Alternative Estimate of
Potential GNP," Review of Economics and Statistics, vol. LI, No. 1 (February
1969), pp. 70-76.
5. Christensen, Laurits R., and Dale W. Jorgenson, "The Measurement of
U.S. Real Capital Input, 1929-1967," Review of Income and Wealth, Series 16
(March 1970), pp. 19-50.
6. "U.S. Real Product and Real Factor Input, 1929-1967," Review of
Income and Wealth, series 15 (December 1969), pp. 293-320.
7. Coen, Robert M., "Investment Behavior and the Measurement of Deprecia-
tion," prepared for the Office of Industrial Economics IRS, U.S. Treasury De-
partment, Dec. 31, 1972, mimeo.
8. "Efficacy of the Investment Credit for Fiscal Purposes," mimeo of
paper presented at the annual meeting of the National Tax Association-Tax
Institute of America, Houston (Nov. 4, 1975).
9. Cohen, Malcolm S., Samuel A. Rea, Jr. and Robert L. Lerman, A Micro
Model of Labor Supply, BLS staff paper 4 (Washington: G.P.O., 1970).
10. Denison, Edward F., Accounting for United States Economic Growth,
1929-1969 (Washington: The Brookings Institution, 1974), ch. 7.
11. Dunlop, John T., "Labor Market Policies to Restore Full Employment,"
statement before the Subcommittee on Economic Growth of the Joint Economic
Committee, May 13, 1975.
12. Eisner, Robert, and M. I. Nadiri, "Investment Behavior and Neo-Classical
Theory," Review of Economics and Statistics (August 1968), pp. 369-382.
13. Jack Faucett Associates, Data Development for the 1-0 Energy Model,
submitted to the Energy Policy Project, Washington, D.C., May 1973.
14. Friedman, Benjamin M., and Michael L. Wachter, "Unemployment:
Okun's Law, Labor Force and Productivity," Review of Economics and Statistics,
vol. LVI, No. 2 (May 1974), pp. 167-176.
15. Gordon, Robert J., "The Impact of Aggregate Demand on Prices," Brook-
ings Papers on Economic Activity (3:75), pp. 613-655.
16. Hall, Robert E. "The Rigidity of Wages and the Persistence of Unem-
ployment," Brookings Papers on Economic Activity (2:75), pp. 301-335.
17. Jorgenson, Dale W., and Zvi Griliches, "Issues in Growth Accounting: A
Reply to Edward F. Denison," Discussion Paper No. 215, Harvard Institute of
Economic Research (Cambridge: November 1971).
18. and Calvin D. Siebert, "A Comparison of Alternative Theories of
Corporate Investment Behavior," Working Paper No. 116 Center for Research
in Management Science, Institute of Business and Economic Research (Berkeley:
September 1967).
19. "Optimal Capital Accumulation and Corporate Investment Be-
havior," Working Paper No. 117 (Berkeley: September 1967).
20. Klein, Lawrence R., and Virginia Long, "Capacity Utilization: Concept,
Measurement and Recent Estimates," Brookings Papers on Economic Activity
(3:73), pp. 743-756.


21. Macrakis, Michael S., ed., Energy: Demand, Conservation and Institu-
tional Problems (Cambridge: The M.I.T. Press, 1974), Introduction.
22. Nordhaus, William D., "The Recent Productivity Slowdown," Brookings
Papers on Economic Activity (3:72), pp. 493-531.
23. Okum, Arthur M., "Inflation: Its Mechanics and Welfare Costs," Brook-
ings Papers on Economic Activity (2:75), pp. 351-402.
24. "Potential GNP: Its Measurement and Significance," reprinted
as an appendix in The Political Economy of Prosperty (Washington: The Brook-
ings Institution, 1970).
25. Orcutt, Guy H., Martin Greenberger, John Korbel and Alice M. Rivlin,
Microanalysis of Socioeconomic Systems: A Simulation Study (New York: Harper
and Bros., 1961), oh. 9.
26. Perry George L., "Changing Labor Markets and Inflation," Brookings
Papers on Economic Activity (3:70), pp. 411-441.
27. "Labor Force Structure, Potential Output, and Productivity,"
Brookings Papers on Economic Activity (3:71), pp. 533-565.
28. Simler, N.J., and Alfred Tells, "Labor Reserves and the Phillips Curve,"
Review of Economics and Statistics, vol. XLX, No. 1 (February 1968), pp. 32-49.
29. Sims, Christopher A., "Output and Labor Input in Manufacturing,"
Brookings Papers on Economic Activity (3:74), pp. 695-728.
30. Thurow, Lester C., and L. D. Taylor, "The Interaction Between Actual
and Potential Rates of Growth," Review of Economics and Statistics, vol. XLX,
No. 4 (November 1966), pp. 351-360.
31. Wachter, Michael L., "A New Approach to the Equilibrium Labour Force,"
Economic, vol. 41, No. 161 (February 1974), pp. 35-50.
32. "Primary and Secondary Labor Markets: A Critique of the Dual
Approach," Blirookings Papers on Economic Activity (3:74), pp. 637-680.
33. U.S. Department of Commerce, Bureau of Economic Analysis, "A Study of
Fixed Capital Requirements of the U.S. Business Economy, 1971-80" (Washing-
ton: December 1975).


By GEORGE M. VON. FURSTENBERG, American Enterprise Institute
The paper by Albert Eckstein and Dale Heien greatly broadens
the range of factors that may be considered in estimating potential
output but it is not clear that this necessarily improves the reliability
of the resulting estimates. Processing through a supply-constrained
macroeconomic model, the authors have derived a new set of estimates
of the annual percentage output gap. Until 1968 these estimates
consistently show a somewhat larger gap than the official series pub-
lished prior to the 1975 benchmark revisions. However, the average
annual growth rates of potential output between full-employment
years such as 1953 and 1968 are almost identical. After 1968 Eckstein-
Heien (EH) estimate much larger gaps.
Part of the difference can be explained by the exclusion of the farm
and general government sectors from the EH estimates. Because the
gross products originating in these sectors are cyclically insensitive,
the percentage gap is widened by excluding them from the estimates
of both actual and potential output. Furthermore, the growing im-
portance of these sectors in recent years may have caused part of the
increasing divergence between the official and the EH estimates
shown in Table 1.


Past Eckstein-
official Heien
estimates estimates Difference

1952 ----------------------------------------------------- 0.2 -0.4 0. 6
1953 ...-------------------------------------------- -.8 -0 -. 8
1954 ---------------------------------------------------- 4.0 5.0 -1.0
1955- -----------------------.2 .3 -. 1
1956-------- ------------------- 1.8 1.6 .2
1957 ------- --------------------------------------------3.7 4.0 -. 3
1958...----------------------------------..------------------ 8.0 8.8 -. 8
1959...-------------------------------------------- 5.5 5.6 -. 1
1960...----------------------------------------------------- 6.4 6.9 -. 5
1961-.------------------------------------------------------ 7.8 8.5 -. 7
1962 ----------------------------------- ----------------- 5.1 5.3 -.2
1963 ------ ----------- ------ ------4.8 5.1 -. 3
1964....--- ------------------------------------------------- 3.2 3.3 -. 1
1965-----.....------ ------------------------- --- -------------- .8 1.4 -. 6
1966 --------- ------------------------------------- --- -1.7 -1.2 -.5
1967 --------------------------------------------------- -.3 .4 -. 7
1968 ---------------------------------------------------1.0 -.3 -.7
1969 ------------------------------------------------------. .3 1.6 -1.3
1970 ---------------------------------------------------- 4.6 6.0 -1. 4
1971...---...----------------------------------------------- 5.2 7.3 -2.1
1972..-------- ---------------------------------------------- 3.2 5.2 -2.0
1973..--- ----------------------------------------------- 1.5 3.4 -1.9
1974------------------------- 7.3 9.6 -2.3
1975 -- -- ------------------------------------------- 13.7 15.1 -1.4

1 EH use the potential gross private nonfarm product as referent.
2 Based on data for the 1st 3 quarters of 1975 reported prior to the benchmark revisions.
Source of official estimates: U.S. Department of Commerce, "Business Conditions Digest," January 1975, p. 109, and
December 1975, p. 95.


In deriving the past official series last published in the December
1975 issue of Business Conditions Digest the Council of Economic
Advisers assumed that output per manhour grew by 2.5 percent and
potential output by 4.0 percent per annum from the fourth quarter of
1969 to the third quarter of 1975. It has since become apparent that
these estimates may have been too high for the most recent years.1
However, after detailed consideration of the changed energy and
materials supply factors, EH still find higher growth rates of potential
output over this period than are reflected in the past official series. The
counterintuitive result of more than 4 percent potential growth even
from 1974 to 1975 is at variance with the results of other studies. Roger
Brinner has cited several factors supporting his finding that the growth
in aggregate factor productivity has recently been declining and that
this slowdown will persist into the future.2 Hence it may well be
doubted that the particular specifications and data used by EH to
model some of the structural changes of recent years yield results which
prove to be robust under alternative specifications or data selections.
Unfortunately such sensitivity-testing will not proceed since EH use
input data which have already been superseded by the benchmark
revisions. These revisions generally had the effect of making the
cyclical amplitude of real GNP and its major components less than
previously estimated, so that the officially reported gaps have since
been reduced.
Data limitations aside, the authors do not fully explain the con-
ceptual basis for their estimates or the uses to which they might be put.
While EH set out to improve "our understanding of the economy's
potential for absorbing a growing labor force at tolerable rates of
inflation" neither the level of potential employment nor its relation or
lack of relation to tolerable inflation rates is analyzed. Rather EH
accept the traditional four percent unemployment target as the ap-
propriate measure of the economy's potential for labor force utilization
and do not investigate the compatibility of such a low unemployment
rate with tolerable price stability under present conditions.
If 4 percent unemployment approximately represented the "natural"
unemployment rate compatible with non-accelerating inflation in the
mid-fifties, there are several reasons why it can no longer have the
same meaning. Arguments to this effect, which have centered on
changes in the composition of the labor force and the growth of income
maintenance programs have been surveyed in the 1975 and 1976
Economic Reports.3 In recognition of at least some of these factors,
variable-unemployment rate estimates of potential output have been
provided both in the 1974 Report and in other studies 5 on an ex-
ploratory basis.
Potential output estimates have different uses and no single esti-
mate is equally appropriate for all. In some of the most crucial and
least political applications, potential output is treated as no more than
a cyclically-adjusted measure of output, and the degree of excess
1 See the revisions in U.S. Department of Commerce, Business Conditions Digest, April 1976, p. 95.
* Roger E. Brinner, "The Growth of Potential GNP," The Data Resources U.S. Long-Term Bulletin,
winter 1976, pp. 95-98.
3 Economic Report of the President, Transmitted to the Congress, February 1975, pp. 94-124, and
Economic Report, January1976, pp. 106-117.
4 See the Economic Report, February 1974, p. 31.
5 George M. von Furstenberg, "New Potential Output Estimates for Economic Policy," pp. 186-195 in
1974 Proceedings of the Busin6ss and Economic Statistics Section, American Statistical Association, Wash-
ington, 1975; and R. Jeffery Green, "Three Estimates of Potential GNP with Projections to 1980," mimeo.


demand or supply that is held constant in that measure is of no par-
ticular significance. The full-employment surplus, for instance, has
traditionally been measured with reference to the official estimate of
potential output but it has been emphasized that changes in the full-
employment surplus, possibly scaled by full-employment GNP, and
not the level of that surplus are most relevant for fiscal policy analysis.
Similarily, econometric estimates in which potential output gaps have
been used to represent changing demand pressures on particular
sectors are generally invariant to scalar multiplicativee) transforma-
tions of potential GNP. Hence, for these uses, the level of potential
output is immaterial provided the percentage changes from any
starting level are estimated consistently. In that case potential output
serves merely as an analytical tool and not as a norm of performance.
By the same token, the usefulness of this tool would be unaffected by
anyone showing, for instance, that potential is consistently one percent
larger or smaller than officially estimated.
Traditionally, measures of potential output have attempted to hold
the degree of labor force utilization constant at some level without
explicitly incorporating capacity utilization. This was done in the
belief that there can be no permanent mismatch between capital and
labor if factor and product prices are free to vary and if resources can
shift and be substituted for each other if it is profitable to do so. For in
that case similar degrees of excess supply or demand would probably
prevail for both major factors of production over a period of years.
Both Brinner and Eckstein and Heien attempt to remedy this impre-
cision by specifying the fixed degree of capacity utilization that is
believed to be consistent with 4 percent unemployment. Brinner sets
the Federal Reserve Board capacity utilization rate in manufacturing
at 87 percent, while EH pick 83 percent at potential. However, the
desired degree of capacity utilization is clearly a function of relative
factor prices as well as of long-run expectations and it is not clear
whether any postulated pair of unemployment and capacity utilization
rates that appeared feasible in the past will be any more consistent in
the future than past unemployment-inflation trade-offs. Hence, each
new argument or factor that is introduced into potential output
analysis brings with it the need to impose new constraints on the com-
binations of variables for which the model is solved at potential.
Elaborate supply-oriented models, such as the one provided by
Eckstein and Heien are most valuable for improving our understand-
ing of the structure of the economy. However, they have not hitherto
yielded estimates of potential output which are demonstrably more
reliable and useful in my view than the existing official estimates. The
note to the CEA estimates published in Business Conditions Digest,
emphasizes that the official estimates are subject to a margin of error.
They are also intended to reflect long-term trends rather than annual
wobbles. Even though these estimates thus involve a great deal of
averaging and recognize only the most durable trends in the growth of
productivity, of the labor force, and in average hours worked, there is
nothing in the EH paper to suggest that the official series has greatly
misrepresented trend movements until 1968. For the 1969-75 period
however, the EH estimate of more than 4.3 percent average annual
rates of potential growth is likely to be wider off the mark than the past
official estimate of 4 percent.

Sti t111111 t 0111 113 36l1111 1111111 111111ill tll II
3 1262 09113 3636