Academic journal article Economic Inquiry

Pensions and Wage Premia

Academic journal article Economic Inquiry

Pensions and Wage Premia

Article excerpt


Despite the widespread acceptance of the theory of compensating differences, empirical efforts to estimate the magnitude of these differences have met with only mixed success.(1) With the exception of fatal risks, economists have not been able to consistently find a significant relationship between wages and other job characteristics. The failure to find consistent evidence on compensating differentials has been attributed to either poor measures of non-wage characteristics or inadequate controls for worker productivity.(2) Studies by Gunderson, Hyatt and Pesando [1992], Montgomery, Shaw and Benedict [1992], and Moore [1987] which have been able to explicitly address these issues have had greater success in finding significant compensating differentials.

In this paper, we focus on an alternative explanation for the variation in estimated compensating differentials - that the magnitude of the compensating differential varies systematically across different sectors in the labor market. The theory of compensating differentials predicts that the implicit price or cost of providing a job attribute (such as pensions) will be a function of firm or industry technological differences and that the value attached to that attribute will also vary across workers depending on their tastes. If workers and firms are not randomly assigned to sectors, this will lead to systematic variations in the size of the estimated compensating differential across sectors or markets. The result of this variation is that studies using different subsamples of the labor force will yield very different estimates of the size of the compensating differential.

The few studies that have addressed sources of heterogeneity in compensating differentials have concentrated on testing for individual differences in tastes for risk.(3) In contrast, our emphasis is on firm-side differences in technologies or market power. We estimate the compensating differential between annual pension accruals and wage rates, and let this tradeoff vary by firm size and union status (with corrections for selectivity bias).

The implicit cost of pensions (and hence the size of the compensating differential) may vary by finn size or union status for a number of reasons: greater ability to spread fixed administrative costs across a large work force, more difficulty monitoring worker effort, or the presence of market power. Thus, the cost of fringe benefits to the firm may be lower in large or unionized firms than in small nonunion firms. We find empirical evidence supporting this view - the estimated wage-pension compensating differentials are smaller in unionized and large firms than in small nonunion firms.

The results from this paper are significant for several reasons:

1. They provide evidence that compensating differentials vary across markets. This evidence could account for some of the differences in estimates of tradeoffs across studies that have used different nonrandom samples.(4) More importantly, these sectoral differences serve as a reminder that "the labor market" is in fact a multitude of markets and that technological differences that vary across markets should be considered when addressing issues ranging from estimating the returns to human capital investment to the impact of the minimum wage.(5)

2. These results also deepen our knowledge of the relationship between pensions and wages, a relationship that has important implications for the shape of workers' wage earnings profiles (or for the mechanisms for motivating workers and for the returns to human capital) and for the distribution of income. Regarding the latter, pensions are a large component of compensation, and work by Even and Macpherson [1990], McCarthy and Turner [1983], and Moore [1987] suggests that coverage rates and returns are smaller for women and blacks; our analysis suggests that an element of the differences could be sectoral choices. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.