Hall has shown that, with perfect competition and price flexibility, total factor productivity measured using labor's share either in revenues or in costs will be acyclical regardless of the level of labor hoarding. We show that if firms producing a homogeneous good under constant returns must pick their prices before demand is known, both measures of productivity become procyclical. The model implies that productivity should be more procyclical the more important is labor hoarding. Empirically, productivity is more procyclical in industries and in nations where labor hoarding appears more important.

I. Introduction, 705.—II. The model, 707.—III. Tax incidence, 710.—IV. Conclusion, 715.

This paper uses cross-sectional and longitudinal data to examine differences in pay for equally-skilled workers in different ind ustries. The major finding is that there is substantial dispersion in wages across industries, even after allowing for measured and unmeas ured labor quality, working conditions, fringe benefits, transitory d emand shocks, the threat of union-ization, union bargaining power, fi rm size, and other factors. In addition, evidence is presented demons trating that turnover has a negative relationship with industry wage differentials. These findings suggest that workers in high-wage indus tries receive noncompetitive rents. Copyright 1988 by The Econometric Society.

[This paper estimates the incidence of response errors in the Current Population Survey. It proposes a procedure for adjusting the Bureau of Labor Statistics' gross flows data on labor market transitions to account for these errors. Although the findings are not definitive because the procedure makes particular assumptions regarding the stochastic process generating response errors, they illustrate the potentially substantial effect of response errors on studies of labor market behavior. The adjustment procedure suggests that because measurement errors give rise to spurious transitions between labor market states, the labor market may be less dynamic than previously thought. The results imply that conventional measures may understate the duration of unemployment by as much as eighty per cent, and overstate the frequency of labor force entry and exit by even more.]

James M. Poterba, Lawrence H. Summers, Unemployment Benefits and Labor Market Transitions: A Multinomial Logit Model with Errors in Classification, The Review of Economics and Statistics, Vol. 77, No. 2 (May, 1995), pp. 207-216

Using data from the United Nations Comparison Project and the Penn World Table, we find that machinery and equipment investment has a strong association with growth: over 1960–1985 each extra percent of GDP invested in equipment is associated with an increase in GDP growth of one third of a percentage point per year. This is a much stronger association than found between growth and any of the other components of investment. A variety of considerations suggest that this association is causal, that higher equipment investment drives faster growth, and that the social return to equipment investment in well-functioning market economies is on the order of 30 percent per year.