From a purely practical standpoint, comparable worth is likely to prove a mixed blessing. Reduced to its essentials, comparable worth amounts to a policy of raising the cost of employing low-wage, predominantly female labor. Hence, other things being equal, it will reduce employment of such labor. To the extent that it raises overall labor costs, it may also reduce employment in other categories, for example, predominantly male or integrated jobs. Thus, comparable worth "solves" the problem of women's low wages only to aggravate others. The key problem is to determine the likely magnitudes of both the wage and employment effects of comparable worth.
George Johnson and Gary Solon investigated wage effects. Their results imply that comparable worth would raise the pay of women by only about 6.4 percent, on average, and would reduce the female-male wage gap of approximately 40 percentage points by no more than about 4 percentage points.
However, their estimation procedure probably understates substantially the likely impact of comparable worth on wage rates. That procedure in effect assumes that comparable worth would correct only that portion of the aggregate male-female wage differential among individuals that is associated with the proportion female in one's occupation, other things being equal--where the "other things" include not only such factors as education and job evaluation points (for factors such as skill requirements and physical demands), but also gender.
Ronald Ehrenberg and Robert S. Smith do not attempt to estimate the effect of comparable worth on the aggregate female-male pay gap, but do consider its likely impact on women's wages, using data taken from comarable worth job evaluations of State government employment in Connecticut, Minnesota, and Washington. Their estimates imply that full implementation of comparable worth could be exected to raise pay in predominantly female jobs in these three States by about 15 to 20 percent.
On the assumption that comparable worth might therefore raise women's wages by 20 percent, Ehrenberg and Smith then attempt to estimate the resulting effect on women's employment. Their results imply that, if State and local government budgets remained fixed in the face of such wage increases, female employment would fall by no more than 6 percent. However, they add that it is likely that State and local government personnel budgets would increase somewhat in response to such cost increases. If so, they estimate, the decline in female employment would probably be halved, to about 3 percent.
Ehrenberg and Smith note that these estimates are "surprisingly small," but several caveats are in order. First, the Ehrenberg-Smtih estimates are based on econometric results of somewhat doubtful solidity (for example, of 16 own-wage employment demand elasticities, seven are positive and only two are statistically significant at conventional levels). Second, they refer to broad aggregates (for example, all professionals and managers in local government noneducational employment) and do not allow for differential effects within those aggregates (for example, for a greater employment effect on nurses than on computer programmers). …