ESSAYS ON LABOR CONTRACT DURATION
WALLACE, FREDERICK H.
Doctor of Philosophy
The purpose of these essays is to test empirically a model of wage contract duration developed by Gray (1978). The focus is on wage agreements negotiated between unions and firms. In most contract length models negotiating agents seek to minimize an objective function comprised of costs associated with a contract. This objective function has two components, a fixed cost borne each time negotiations are undertaken and a variable cost which increases with the level of uncertainty confronting the firm and as the bargain becomes longer. The presence of the fixed cost usually makes a spot market suboptimal while the variable component prevents the agreement from being of infinite duration. These models suggest that contracts will become shorter as price or output uncertainty increases. The initial empirical findings, surprisingly, suggest that contracts become longer, not shorter, as output fluctuations, measured by the output forecast standard error, increase; a result opposite to that predicted by the theoretical models. There are several possible explanations for such a result. First, and most obviously, the theory may be wrong. Second, a missing variable such as the degree of wage indexation may affect both contract length and the uncertainty measure thus obscuring the true relationship between the two included variables. Third, econometric problems including bias, heteroscedasticity, and autocorrelation may be affecting the estimated coefficients. An alternative estimation procedure using individual firm data indicates that, for some industries, contract length is positively related to uncertainty caused by productivity (industry-specific) shocks and negatively related to money supply shocks. In other instances the results suggest no significant relationship between contract duration and the uncertainty measures. Using pooled data adjusted to eliminate heteroscedasticity and introducing industry dummy variables into the estimations indicates that contract length is not significantly related to the industry-specific shock measure. At the 95% confidence level contract duration is negatively affected by money supply uncertainty only among durable goods producers.