Eisner 1994
Eisner 1994
Eisner 1994
Economic Behavior
Journal of Economic Behavior and Organization &Ofganbtion
ELSEVIER 23 (1994) 111-126
Ah&act
1. Introduction
‘See Eisner (1988, 1989, 1991a) for detailed presentation of concepts and measures of extended
accounts for output, capital accumulation and saving.
114 R. Eisner / Journal of Economic Behavior and Organization 23 (1994) III-126
Table 1
Government receipts and expenditures accounts, 1991 (billions of dollars).
Current account
Credits: Revenues 1,747
Charges: Current expenditures 1,771
Goods and services 560
Net interest, transfer 850
payments and other
Capital consumption 361
allowances
Surplus or deficit (-) -24
Capital account
Credits: Capital consumption allowances 361
Charges: Gross capital expenditures 530
Tangible 184
Intangible 346
Education 279
R&D 67
Surplus or deficit: Net capital expenditures (-) -169
Consolidated account
Revenues 1,747
Expenditures 1,940
Total surplus or deficit (-) - 193
Sources: National Income and Product Accounts, Survey of Current Business,
September 1992, pp. 29-30; ‘Federal Government Expenditures by Type and
Function’ and ‘State and Local Government Expenditures by Type and Function’
Tables 3.16 and 3.17; Survey of Current Business, July 1992, Table 3.1, p. 66, 3.7B
p. 69; Bureau of Economic Analysis, ‘Fixed Government Capital (including
residential) historical-cost valuation, gross investment, depreciation and net invest-
ment’, total, National Science Foundation, Science and Technology (1991), Appen-
dix A, Table 4-2, p. 306; U.S. Department of Commerce, Government Finances,
December 1991, Tables 4 and 5, pp. 4, 6. Intangible capital consumption
allowances assumed equal to two-thirds of gross intangible capital expenditures.
Table 2
Gross saving and investment account, 1991 (billions of dollars).
0.0 1
-0.1
-02 0
-0.3
-0.4 -1
-0.5
-0.6 -2
60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90
Fig. 1. Changes in price-adjusted high employment deficit and real federal tangible investment
as percent of GNP.
Table 3
Adjusted deficit and public inv~~ent; AI, = be + brAPAHED6.’
4 be b, R2 D-W rho n
Federal tangible -0.035 0.102 0.250 1.52 0.336 31
(0.049) (0.035) (0.166)
S&L tangible -0.018 0.006 0.274 1.86 0.566 31
(0.040) (0.017) (0.157)
Total.govemment -0.048 0.110 0.302 1.37 0.473 31
tangible (0.079) (0.042) (0.160)
Education & 0.052 0.135 0.347 1.73 0.510 31
training (0.070) (0.035) (0.158)
Government R&D -0.021 0.011 0.29 1 2.08 0.525 30
(0.019) (0.099) (0.163)
Total government -0.056 0.228 0.354 1.70 0.390 30
(0.093) (0.059) (0.176)
‘Ordinary least squares with a first-order correction for serial correlation.
Standard errors are shown in parentheses. Estimated as first differences. Sample
period is 1961-1991 for federal tangible, S&L tangible, and total government
tangible, and education and training and 1962-1991 for government R&D and
total government.
Variable definitions:
APAHED6=price-adjusted 6 percent high-employment deficit as percent of
GNP, first difference.
AI,=real investment of type k - federal tangible, S&L tangible, education 8~
training, government R&D and total government - generally from the BEA,
taken as percent of potential GNP, first differences.
fourth row of Table 3. And all this is reflected clearly in Fig. 4, relating the
Federal deficit to total government investment expenditures, tangible and
intangible, and in the corresponding regression results reported in the
bottom row of Table 3. Each one percentage point change in the ratio of the
inflation-adjusted 6-percent-unemployment deficit to GNP was associated on
average with a 0.228 percentage point change in public investment in the
same direction.
Granted that Federal structural deficits have been associated then with
more public investment and hence real government saving, is it not still
possible that these de&its have absorbed so much private saving and
crowded out so much pr~uate inves~ent that they actually reduced total or
national saving? This would indeed be inevitable, except possibly for foreign
financing of our investment, if national income and product were fixed. For
then more output going to public expenditures would have to mean less
output going to private expenditures or to foreigners. With the presumption
R. Ether / Jownai of Economic Behavior and Organization 23 (1994) 111-126 117
-DGYTNTTP _v_.DPAHEDGP
Fig. 2. Changes in price-adjusted high employment deficit and real total government tangible
investment as percent of GNP.
l- , , / , , , ’ , ’ , ’ , , , , ’ , , ) , , , , , , , (
60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 St
- DEDGVTTP _____DPAHEDGP
Fig. 3, Changes in price-adjusted bigb employment deficit and real total gove~~t expendi-
tures on education as percent of GNP.
118 f Jawnalof Economic Behavior and Organization 23 (1994) III-126
R. Eisner
62 64 66 66 70 72 74 76 78 80 82 84 86 88 90
Fig. 4. Changes in price-adjusted high employment deficit and real total government investment
as a percent of GNP.
%DGD
J
8
5 2
3 1
I 0
-I -1
-2
-3 +* I. I ( 1s I. I. t . , St. I. I ., . , -2
! 64 66 60 70 72 74 76 78 80 82 84 86 88 9[1
YEARS
...-_-.-- -
[z ikDP87PC _____ -4
Fig. 5. Percent changes in real gross domestic product and lagged price-adjusted high-
employment deficit, 1961-1991.
R. Eisner / Journal of Economic Behavior and Organization23 (1994) Ill-126 119
z See Eisner (1986) for an early, comprehensive treatment of some of these issues and, for more
up-to-date treatments, focussing more considerably on national saving, Eisner (1992a and
1992b).
3 See, for example, Eisner and Pieper (1992).
120 R. Eisner / Journal of Economic Behavior and Organization 23 (1994) III-126
3 2
-1
-2
i
-3
62 64 66 66 70 72 74 76 76 60 62 64 66 66 90
I_= DUWPC~-.-~~DPAHES6P(-l)(
Fig. 6. Percent changes in unemployment and IaBged price-adjusted high-employment surplus,
1961-1991.
national saving - the sum of gross private domestic investment and net
foreign investment - might still be negatively related to these deficits.
That, however, is not the case. As shown in the second column of Table 5,
reporting the regression of NSO, after taking into account the effects of real
changes in exchange rates, we find that each percentage point of structural
deficit, expressed as a ratio of GNP, was associated with almost exactly one
percentage point more of subsequent national saving over the period from
1973 to 1991.4
But finally, what about a broader measure of national saving, one that at
least includes government and household investment in tangible capital. Fig.
7, graphing changes in the real structural deficit against changes in NSl, this
expanded measure of national saving, both taken as percentages of GDP,
tells a striking story. It is clear to the naked eye that changes in the deficit
have been positively associated with subsequent changes in national saving
over the years from 1961 to 1991. And this relation is confirmed in the
regression, for the years 1973 to 1991, where the monetary base and
4Earlier years were excluded because of the absence of earlier measures of trade-weighted, real
exchange rates.
R. Eisner / Journal of Economic Behavior and Organization 23 (1994) Ill-126 121
Table 4
Changes in adjusted budget deficits, monetary base,
real GDP and unemployment, 1962-1991; Ax=
b,,+b,APAHED6,_, + b*DMB,_ r.
Variable AY,
or
parameter GDP Unemployment
C 1.922 0.462
(0.570) (0.170)
APAHED6,_, 1.227 - 0.486
(0.317) (0.162)
DMB,_, 8.510 -4.414
(2.493) (1.052)
AR(l) 0.378 -0.010
(0.208) (0.235)
R’ 0.481 0.444
D-W 1.713 1.771
n 30 30
C Constant term
AR(l) First-order autoregressive coelfi-
cient
R2 Adjusted coelkient of determi-
nation
D-W Durbin-Watson coefficient
n Number of observations
APAHEDQ- 1 Lagged price-adjusted, BEA 6%
high-employment budget deficit as
percent of GNP
DMB,_ I Lagged real change in monetary
base as percent of GNP
(December figures divided by 4th
quarter GNP implicit price defla-
tor).
exchange rate variables are also included. With deficits to some extent
financing household and government investment, the relation is even stronger
than that for the narrower measure of national saving. We now note that
each percentage point change in the deficit was associated with an average
move in the same direction of 1.193 percentage points of national saving as a
ratio of GNP. The deficits have apparently contributed to more government
saving and investment and to more total, national saving and investment as
well.
122 R. Eisner / Journal of Economic Behavior and Organization 23 (1994) Ill-126
Table 5
Adjusted budget deficits, changes in monetary base, and real exchange
rates and gross private domestic investment and alternate measures of
real national saving, NSO and NSl, first differences, 1973-1991; Ax=
b0+bIAPAHED6,_I+b2ADMB,_I+b3ADMB,_2+b,AERR,_,.
Variable AE;
or
parameter GDP1 NSO NSl
C - 0.295 -0.376 - 0.430
(0.190) (0.241) (0.169)
APAHED,_ , 1.258 1.004 1.193
(0.274) (0.241) (0.263)
ADMB,_ 1 5.646 6.925 7.605
(1.440) (1.211) (1.318)
ADMB,_, 4.142 5.953 6.223
(1.284) (1.149) (1.180)
XADMB 9.788 12.877 13.828
(2.001) (1.806) (1.834)
AERR,_ 2 - 0.092 -0.122 -0.124
(0.030) (0.026) (0.027)
AR(l) - 0.049 0.282 - 0.062
(0.327) (0.297) (0.322)
R2 0.696 0.746 0.796
D-W 1.866 2.135 1.822
n 19 19 19
C Constant term
AR(l) First-order autoregressive coefficient
R2 Adjusted coeffkient of determination
D-W Durbin-Watson coefficient
n Number of observations
PAHED, _ 1 Lagged price-adjusted, BEA 6% high-employment bud-
get deficit as percent of GNP
DMB,_, Lagged real change in monetary base as percent of
GNP (December figures divided by 4th quarter GNP
implicit price deflator)
DMB,_* Two-year lagged real change in monetary base as
percent of GNP (December figures divided by 4th
quarter GNP implicit price deflator)
ERR,_* Trade weighted, real exchange rate, lagged two years
(March 1973 = 100)
GPDI Gross private domestic investment as percent of GNP
NSO Conventional net national saving= gross saving - capi-
tal consumption allowances with adjustment, as per-
cent of GNP
NSl Change in total fixed reproducible capital plus net
foreign investment, as percent of GNP
R. Eher 1 Journal of Economic Behavior and Organization 23 (1994) Ill-126 123
62 64 6fi 66 70 72 74 76 76 60 62 64 66 66 90
Fig. 7. Percent changes in ratios to GDP of total, real, tangible, national saving and lagged
price-adjusted high-emp. deficit.
Finally, we may bring some of our data to bear on the critical question of
the productivity of public tangible investment, of which infrastructure
investment is of course a signifrcant part, a part that we may expect will rise
with the decline of military spending and the policy commitments of the new
Clinton Administration. We have estimated a log-linear production function
from data on total labor hours worked and the end-of-previous-year stocks
of private and public tangible capital. 5 Parameter estimates indicate the
expected major role of labor and suggest the usually observed increasing
returns to scale. Most interesting for our purposes, however, are the
significantly positive coefficients to both private and public tangible capital.
We may indeed note that, while the differences are hardly statistically
significant, the coefficient of public capital is larger than that of private
capital. One percent more of private capital was associated with 0.19 percent
more of GDP, while one percent more of public capital was associated with
’ We lack useful series on stocks of government-financed capital stocks of education and R&D.
We are further inhibited from entering even crude approximations into our production function
by the unreasonableness of expecting changes in such stocks from having any effect on output
o&r than with long and variable lags.
124 R. Eisner / Journal of Economic Behavior and Organization 23 (1994) 111-126
Table 6
Production function: gross domestic product and
labor hours and private and public tangible capital;
LGD87, = b0 + b,LLBRHRS,+ b,LKPRVTTAN,_ ,
+ b,LKPUBTAN,_ I.
0.27 percent more of GDP. Inasmuch as the measured public capital stock
was much less than the measured private capital stock, this suggests a much
greater marginal productivity of public capital. If we could take the numbers
seriously - I would caution that they be viewed as just one more bit of
evidence in a very difficult, tricky and controversial set of estimation
R. Eisner 1 Journal of Economic Behavior and Orgmization 23 (1994) 111-126 125
68ee, for a not entirely representative sample, Aschauer (1989), Munnell (1990) and Eisner
(1991b).
‘The mean of the observations of LKPRVTTAN, the natural logarithm of private tangible
capital stock, is 8.8850, corresponding to 7,222.7 billions of 1987 dollars. The mean of
observations of LKPUBTAN, the natural logarithm of private capital stock, is 7.7377,
corresponding to $2,293.1 billion. The ratio of (geometric) means is thus 3.15. The mean of the
observations of LGDP87, the natural logarithm of gross domestic product is 8.0852, correspond-
ing to 3,245.9 billion in 1987 dollars. The implied gross marginal product of private tangible
capital is thus 0.0855358. The implied gross marginal product of public tangible capital is
0.3817155, or 4.46264 times the marginal product of private capital.
126 R. Eisner / Jou?na/ of Economic Behavior and Organization 23 (1994) Ill-126
References
Aschauer, D., 1989, Is public expenditure productive?, Journal of Monetary Economics, March,
117-200.
Eisner, R., 1986, How real is the federal deficit? (The Free Press, New York).
Eisner, R., 1988, Extended measures of national income and product, Journal of Economic
Literature 26, December, 1611-1684.
Eisner, R., 1989, The total incomes system of accounts (University of Chicago Press, Chicago).
Eisner, R., 1991a, The real rate of U.S. national saving, Review of Income and Wealth 37, no. 1,
March, 13-32.
Eisner, R., 1991, Infrastructure and regional economic performance: Comment, New England
Economic Review, September/October, 47-58.
Eisner, R., 1992a, Deficits: Which, how much and so what?, American Economic Review 82, no.
2, May, 295-298.
Eisner, R., 19924 U.S. national saving and budget deficits, For The political economy of
investment, saving and finance: A global perspective, Gerald Epstein and Herbert Gintis, eds.,
A project of the World Institute for Development and Economic Research (WIDNER), The
United Nations University, Helsinki, Finland, manuscript copy, 6/20/92.
Eisner, R. and P.J. Pieper, 1992, National saving and the twin deficits: myth and reality, in: The
Economics of Saving, James H. Gapinski, ed. (Kluwer Academic Publishers).
Lange, O., 1938, The rate of interest and the optimal propensity to consume, Economica 5,
February, 12-32. Reprinted in American Economic Association, Readings in Business Cycle
Theory (Blakiston, Philadelphia) 169-192.
Munnell, A.H., 1990, How does public infrastructure affect regional economic performance?,
New England Economic Review, September/October, 1l-32.