The interactive terms make it difficult to clearly see the impact of rents and union status on wages and hours. To facilitate the discussion of these interaction terms, Table 3.11 displays the estimated impact of a one-percent increase in union participation and in industry rents separately for a labor-intensive industry and a capital-intensive industry. The impact of union status on weekly hours is very small in magnitude but on wages is significant. Not surprisingly, the presence of unions tends to raise wages in the capital-intensive sectors more than in the labor-intensive sectors. But up until 1985 unions seemed to have a negative effect on wages in labor-intensive industries.
This result is possibly caused by another potential endogeneity problem, this time between capital-intensity and unions. When unions successfully raise wages firms naturally become more capital-intensive. The overall decline of union presence from 39% in 1975 to 20% in 1993 may explain why this odd result has dissipated. Another interpretation is that unions seek preferred contracts in terms of both wages and effort level. In the capital-intensive sectors, where the cost of cutting effort is very high, unions opt for higher wages and perhaps not much change in effort. In the labor-intensive sectors, unions pursue the effort dimension of the contract more aggressively and end up opting for a contract which has greatly reduced effort and also somewhat reduced wages. (Here we are speaking about effort in the form of pace rather than in the total number of hours.)
We are uncomfortable with both the theory of rent-sharing and also our measurement of rents, and we consequently do not place a great deal of faith in the rent results in Table 3.11. According to these estimates, industry rents seem to reduce hours in labor-intensive sectors but raise hours in capital intensive sectors. Rents have a mixed effect over time on wages in the labor-intensive sectors but consistently raise wages in the capital-intensive sector, reflect the potential for profit sharing within an industry which is not fully competitive.
The point of this lengthy discussion of additional variables is primarily to determine if our initial estimate of the wage-effort offer curve is substantially contaminated by the omission of all these effects. Since we are now treating human capital as self-financed, in principle we want to trace out the wage-effort offer curve after aggregating human and physical capital and after removing from wages the implicit rental cost of human capital.
This is not easily done, and what we do instead is to Figure 3.10: Derived Wage-Effort Offer Curve, 60-93 trace out the curve in the same way as before, using the new coefficients on the capital-intensity variables and holding fixed all the other variables at their sample averages. The results are display in Figure 3.10. 24 hour payday loans