1 “Measuring Production in the USSR,” American Economic Review, XLVIII, No. 2 (May 1958), 398.

2 The Real National Income of Soviet Russia Since 1928 (Cambridge, Mass., 1961), p. 3.

3 For example, the Soviet index of gross output in industry for the period 1958-63 coincides exactly with gross output of industry at enterprise prices for the same period (Promyshlennosf SSSR [Moscow, 1964], pp. 35-36). The ruble prices are described as being“wholesale prices of July i, 1955,” but, as nearly as I can tell from a reconstruction of the industry accounts for these years, these ruble values are, in fact, at unadjusted current prices. This is inconsistent with recent price indices published by Soviet sources; see S. G. Stoliarov, O tsenakh i tsenoobrazovanii v SSSR (Moscow, 1963), pp. 111-12.

4 Implying a production function of the form P — aK where a may be interpreted as the output-capital ratio.

5 Although, to the extent that the Second World War destroyed more capital than labor, the productivity of investment probably was increased in the postwar period until the appropriate long-run proportions of capital and labor were restored.

6 For example, Kenneth J. Arrow,“The Economic Implications of Learning by Doing,“ and Solow, Robert M., “Substitution and Fixed Proportions in the Theory of Capital,” in The Review of Economic Studies, Vol. XXIX, No. 80 (June 1962).Google Scholar

7 In the two-factor case, inputs are aggregated using a Kendrick-type production function with linear isoquants or a Cobb-Douglas function.

8 “Agricultural Production,” in Economic Trends in the Soviet Union, ed. Abram Bergson and Simon Kuznets (Cambridge, Mass., 1963), p. 218.

9 Productivity Trends in the United States (Princeton, 1961), p. 136.

10 Cohn's output series, presented in Table 1, are available in Annual Economic Indicators for the U.S.S.R., Joint Economic Committee, U.S. Congress (Washington, 1964), p. 95.

11 Solow, R. M., “A Contribution to the Theory of Economic Growth,” Quarterly Journal of Economics, Vol. LXX (February 1956)Google Scholar, and Swan, T. W., “Economic Growth and Capital Accumulation,” Economic Record, Vol. XXXII (1956).Google Scholar

12 Consider an economy which is producing output, Y(t), with two factors of production, capital and labor, K(t) and L(t), subject to a linear homogeneous production function multiplied by a neutral technological improvement factor A(t). It experiences an exponential growth of its labor force at rate n; it enjoys neutral technical progress at a constant rate g; and it has been saving a certain proportion of its national product, s. We have three equations: (1) Y(t) = A0ef t K(t)* L(t)b (production function) (2) K = sY(t) (investment equation) (3) L(t) = L0ent (labor supply equation)

13 In the limit both output and capital will increase at the rate n + g/b.

14 d log Y _ Y dt ∼ Y

15 World War II presents some particular problems which I will ignore in the present example. The effect of the war was not only to reduce the proportion of national product devoted to capital formation but also to destroy a substantial quantity of capital stock. Such an event, in effect, placed the USSR on a much lower growth path. But, in addition, the remaining capital stock did not have the equilibrium structures which would have been generated by a long-run savings rate appropriate to the postwar size of capital stock. So the initial short-run growth of the Soviet economy after the war should have been higher than the rate which a long-run model would predict.

16 Initially, when an economy increases the proportion of national income devoted to capital formation from some savings ratio, S0, to some new value, Slt the initial rate of growth of output will equal (n+ g/b) where a is the capital coefficient.

17 “The Harrod-Domar Model vs the Neo-classical Growth Model,” The Economic Journal, LXXIV (June 1964), 380-87; and“Fiscal Policy in a Neo-classical Growth Model: An Analysis of Time Required for Equilibrating Adjustment,” The Review of Economic Studies, XXX (1), No. 82 (February 1963), 16-23.

18 The length of time required to adjust k percent of the way from an initial to the limiting value may be estimated as: g + bn where the notation is as before.

19 See John Conlisk,“The Analysis and Testing of the Asymptotic Behavior of Aggregate Growth Models,” unpublished doctoral dissertation, Stanford University (May 1965).