Because it concentrates on the co-movements of jointly determined endogenous variables, the traditional analysts of labor productivity does not directly address the question of the causes of productivity change.This problem is solved by a modelling approach in which productivity and other choice variables are assumed to respond optimally to five broad classes of exogenous (causal) shocks.Although these shocks are unobservable to the econometrician, maximum likelihood estimates of their relative importance in the determination of productivity change are obtained.

This paper is an econometric analysis of the on-the-job training (OJT) decisions of a group of white American males during 1975. The data are obtained from the Panel Study of Income Dynamics, which asked a very careful series of questions concerning the individual's OJT status. Each individual's internal rate of return is estimated and used as an explanatory variable to predict the probability of taking OJT. The individual's marginal tax rate is also entered in the equation. The results suggest that income taxation has tended to increase the probability of being involved in OJT. I conjecture that this is because income taxation makes investment in physical capital a less desirable vehicle for carrying consumption into the future, and hence increases the attractiveness of human capital.

This paper examines the conventional monetary equation of exchange rate determination. Under certain exogeneity conditions, one can write the price level, at home and abroad, as the ratio of the nominal money supply to the demand for real money balances. Then, since the exchange rate is the domestic price of foreign exchange, one can equate the exchange rate to the ratio of domestic to foreign prices. This then allows one to write, and estimate, the exchange rate as a function of the money supply differential, income differential and interest rate differential. If the domestic and foreign money demand errors are autocorrelated, and if deviations from purchasing power parity are autocorrelated, tests based on the above model may be invalid. Only if all autoregressive parameters are equal will test results be valid. A full information maximum likelihood procedure is used to estimate and test the assumptions necessary for the conventional procedure to be correct. Finally, two alternative models of exchange rate determination are considered to illustrate the importance of introducing the error terms at the beginning of the analysis.

Using a new set of directly observed wage expectations among firms, this paper finds that in general firms' forecasts fail the unbiasedness and efficiency requirements of weak-form rational expectations. These market participants consistently underestimate the wages they actually end up paying, and their expectations do not efficiently utilize the information in past realizations. The mean absolute forecast error of two percent compares with an error of only five percent if static expectations were held. The major source of wage fore-cast error seems to be errors in predicting demand, rather than in predicting supply or the general price level. Wage forecast errors are positively correlated across fields with distinct supply patterns, and are positively correlated with quantity forecast error. The properties of stochastically weighted expectations and the effectiveness of the wage and price controls of the early 1970's are also discussed.

S ince the breakdown of the par-value system there has been a dramatic increase in the variability of exchange rates. Often the question is raised whether this variation is rational or whether it is the outcome of disorderly markets which are bereft of stabilizing speculation or-even worse-are dominated by destabilizing speculation. The considerable research stimulated by this issue has yet to yield conclusive results, and whether the foreign-exchange market is rational or efficient is still an open question.' Of course, the market is not an abstraction but is comprised of flesh-and-blood participants whose behavior may show varying degrees of rationality. This paper utilizes a new body of data to undertake a preliminary investigation into the market behavior of one major group of participants: U.S. firms whose foreign-currency positions are regularly reported to the U.S. Government.

[Excerpt] This paper presents new evidence on the determinants of place-to-place migration in the United States. For understanding the causes of differential migration rates into and out of labor markets, knowledge of place-to-place migration functions is of interest for a number of reasons. Given a thorough understanding of gross place-to-place flows, one can proceed to calculate net flows; the reverse, of course, is not possible. There are also other advantages of place-to-place studies: parallelism to microeconomic behavior, opportunity to investigate specific 'origin-destination match-ups, recognition of the number and location of alternative opportunities for persons residing in different origins, and exploration of possible asymmetries. Following a large body of economic literature, the analytical approach adopted regards migration as a form of human investment. Economic variables used in the empirical-work exhibit effects in the hypothesized direction and explain up to two-thirds of the variance in intermetropolitan migration rates. However, this high degree of explanatory power is achieved only for certain functional specifications involving particular independent variables. Thus, the empirical results confirm the usefulness of the human investment approach to place-to-place migration, but they show too that the economic factors used as explanatory variables must be carefully specified and measured.

Work supported by the RANN Division of the National Science Foundation under Grant #GSF SIA 75-00379.

Prepared in association with the Sloan School of Management and the Dept. of Economics