Knowledge that Transforms

In recent years there has been a systematic exploration of the implications of relaxing the strict assumptions of the textbook version of the Heckscher-Ohlin theory of international trade. Some of the bestknown propositions have turned out to be highly vulnerable to changes in assumptions. Thus if the assumption of constant returns to scale is relaxed, one must be prepared to sacrifice the Stolper-Samuelson and Rybczynski theorems (see Ronald Jones, 1968, and Kemp), and one must accept the possibility that commodity prices and outputs are negatively associated (see our 1969 paper). And if factor-market distortions are admitted, one must be prepared for the possibility that the ranking of industries by value factor intensities may be opposite to the ranking by physical factor intensities, implying in turn that outputs and prices may be negatively associated and that the Stolper-Samuelson and Rybczynski theorems must be rephrased and then cease to be dual to each other (see our paper with Stephen Magee; Jones, 1971). Against the background of traditional results, the new findings have been viewed as perverse and paradoxical.

While contemporary rhetoric often highlights differences between races, the data show that blacks are becoming less distinguishable from whites in at least one relevant index of performance-market earnings. Relative to white males, black male earnings have gradually increased, and the rise during the 1960's and the early 1970's is larger than that observed earlier. (See Table 1.) Yet, it is clearly the contrast between white and black females that is extraordinary. Twenty years ago the average black woman employed full time was earning approximately half the wage of a similarly employed white woman. By 1975, almost complete racial parity among women had been achieved. In a recent article (1977), Finis Welch and I argued that the advance in the relative income of black males between 1960 and 1970 was due mainly to converging educational distributions by race and a narrowing in wage differentials between regions. Skill levels were relatively constant within cohorts and convergence was accomplished as increasingly similar racial cohorts entered labor markets while other less similar cohorts retired. Finally our test of affirmative action pressures indicated that before 1970 they had little impact. My first objective is to update our previous research to determine if the events of the last decade for males have continued unabated into the mid1970's. Since a complete understanding of the dynamics of blackwhite changes necessitates explaining the patterns for females, my second goal is to expand the wage comparisons to include women. The major explanations for narrowing in racial wage differences can be placed under four general categories. The central idea of the vintage hypothesis is that more recent black cohorts begin their job experiences with larger initial stocks of human capital, relative to whites, than previous cohorts. The second explanation involves migration. The rural-South to urban-North migration has partly been superceded by southern blacks moving to what are by now economically vibrant southern cities. The third category involves the effects of government affirmative action. Since 1970, it is alleged that a series of court cases imposing severe financial penalties on firms for noncompliance with affirmative action goals have added sharp teeth to government jaw boning. Finally, changes in other aspects of market work may be important in narrowing relative wages. This factor is more relevant for women than men and includes the choice of partor full-time work, unique characteristics of certain occupations, and biases due to limiting comparisons solely to working women.

In a recent article in this Review, David lIevhari and Don Patinkin (hereafter noted L-P) develop an equilibrium growth model for a simple economy in which money is treated as a productive factor, entering an aggregate production function. In their policy section, they are unable to determine the effects of an increase in the exogenously determined rate of inflation on the equilibrium capital and real-balance intensities. They are also unable to establish whether or not steady-state equilibrium is stable. The purpose of this note is to extend their dynamic analysis and to show that 1) the effects of a change in the rate of inflation are more or less predictable, and 2) steady-state equilibrium can be expected to be stable. The L-P model can be briefly summarized. Assume a growing neoclassical economy where Y = G(K, M/P, N), and where Y, K, M/P, and N represent real output, the stock of physical capital, the real money stock, and the labor force, respectively. Assume further that G is linear homogeneous and twice continuously differentiable. The function can therefore be written y = g(k, m), where y, k, and m are Y/N, K/N, and M/PN, respectively. Assume g is wellbehaved such that gi > 0, gii 0. Let the labor force grow at some exogenously determined exponential rate n. Money is costlessly produced by the government and injected into the economy via transfer payments to the public. In order to avoid stability problems noted by Miguel Sidrauski, it is assumed that the rate of nominal expansion is altered by the government in order to maintain a constant target rate of inflation.' The rate of nominal expansion is u = DM/M, where D denotes the time derivative of the variable which follows it. The rate of inflation is p = DP/P. It follows that D(M/P) = (u p)M/P. From this and the labor force growth assumption, it follows that the rate of growth of the real per capita money stock is