A reply to Andrew's OPE-L:3945:
This is an extremely important issue, and the evidence is complicated,
depending on the exact definitions one uses of the capital stock, the
methods of deflation used to estimate capital input, and especially the
time period used.
Dumenil and Levy's book "Economics of the Profit Rate" has a lot of
information on this question. Their figures (which are confirmed by the
Penn World Tables data for the post-1960 period) indicate a rise in the
capital-output ratio from 1869-1919 or so, and again from 1949-1989. In
these periods the U.S. appears to undergo "Marx-biased", that is,
labor-saving and capital-using, technical change. The period 1919-1949
manifests a quite different pattern of "Hicks-neutral" technical change in
which the capital-output ratio falls at about the same rate as the
labor-output ratio. This suggests that the 1919-1949 period (or some
subperiod therein) marks a structural change in the patterns of
accumulation of capital in the U.S., breaking the Marx-biased pattern
characteristic of both the earlier and later periods.
Adalmir Marquetti, a graduate student at the New School, has supplemented
the Penn World Tables capital stock data by his own measures using the
Perpetual Inventory Method. Marquetti's figures show a Marx-biased pattern
of change for a majority of country-year observations (52%), though not by
any means for all of them. This also suggests that the Marx-biased pattern
is the result of some common but not universal structure of capital
accumulation, and raises the interesting question of explaining the
non-Marx-biased cases.
I hope to have this data and a paper based on them on the WWW sometime this
Spring.
Duncan
>As I noted in an earlier post, Frank Thompson presented a paper, "The
>Composition of Capital and the Rate of Profit: A Reply to Laibman," at the
>ASSA conference in New Orleans. In addition to showing that simultaneism
>implies that a rise in the value composition of capital will tend to RAISE the
>equilibrium rate of profit (given profit-maximization, viable technical
>change, and a direct relation between labor demand and the real wage), he uses
>BEA data to show that there's no upward trend in the aggregate physical
>capital/output ratio in the U.S, which tends to support the point that John's
>been making.
>
>Frank looks at 6 measures of the capital/output ratio. All have GDP as their
>denominator. The numerators are (1) Gross Fixed Private Nonresidential
>Capital, (2) GFPNC + Gross Government-owned Fixed Capital, (3) Gross Total
>Fixed Reproducible Tangible Wealth, and (4), (5), and (6) are the same as (1),
>(2), and (3), respectively, except that net figures are used instead of gross
>figures. Note that (3) is the sum of (2) plus gross fixed private residential
>capital plus gross durable goods owned by consumers.
>
>During the period 1929-1994, the average annual percentage changes in these
>ratios are -0.32, -0.23, -0.24, 0.01, -0.12, and -0.12.
>
>There is a sharp downward fall in all of them from the early 1930s to the
>mid-1940s, having to do with the rise in capacity utilization, and so Frank
>also computed the annual percentage changes for the period 1950-1994: 0.49,
>0.09, 0.07, 0.61, 0.12, and 0.11, so there is a slight rise, but the plots of
>the data look basically flat.
>
>Finally, for the period 1982-1994, the average annual percentage changes
>-1.28, -1.42, -0.93, -1.74, -1.78, and -1.23.
>
>Taking the evidence as a whole, my reaction is that the aggregate
>capital/output ratio is very stable except for the Depression years,
>essentially constant.
>
>Andrew Kliman
Duncan K. Foley
Department of Economics
Barnard College
New York, NY 10027
(212)-854-3790
fax: (212)-854-8947
e-mail: dkf2@columbia.edu