Soviet growth slowdown: econometric vs. direct evidence
比较了使用柯布-道格拉斯和CES生产函数分析苏联增长放缓原因的研究，发现计量结果不一致，无法确定放缓主因。
Studies using the Cobb-Douglas specification blame growth retardation on total factor productivity growth slowdown which set in after 1958. The most comprehensive accounting for the impact of the observable growth determinants (Abram Bergson, 1983) left the decline in total factor productivity growth rate mostly unexplained, and arbitrarily attributed it to technological progress. A number of studies found that the CES production function with a constant rate of growth of the residual fits Soviet industrial data better than the Cobb-Douglas production function (for example, Martin Weitzman, 1970). In this case, postwar growth slowdown is explained (at least for part of the period) by decreasing returns on capital under elasticity of substitution below unity and a rapidly increasing capital-labor ratio. But the estimates of elasticity of substitution and other parameters vary widely across the studies; the implied rate of return on capital in the early 1950's is implausibly high (Bergson, 1979, pp. 117-20; Norman Cameron, 1981, p. 26). Some scholars found elasticity of substitution significantly lower than unity up to the mid-1960's, and close to unity after that (Cameron, p. 36; Ryan Amacher and Darius Conger, 1977, p. 318). Others, in contrast, did not find evidence of a structural break in the sample period (Weitzman, 1983). In some of the latest work, the CES production function with lessthan-unity elasticity of substitution and a constant rate of growth of the residual was found to fit the data no better, or even worse, than the Cobb-Douglas production function with slowing growth of the residual (Weitzman, 1983; Padma Desai, 1985). Thus, production function analysis of the causes of the growth slowdown is inconclusive.