The origins of parallel segmented labor and product markets: A reciprocity-based agency model with an application to motor freight

Why do some workers who apparently perform similar tasks and exercise similar job skills get paid very different wages? And, why do firms have the boundaries we observe; in particular, why do firms using closely related production technologies and serving closely related markets specialize instead of merging? That is, how do labor and product market segments emerge, and why might they persist in a competitive economy? I offer an integrated explanation for the striking case of the emergence of such market segments in for-hire motor freight, after its deregulation in 1980. Using firm-level data, I provide econometric evidence of the survival value of carrier specialization, as a result of either original status or strategic change, into one of two types. I also document the associated bimodal segmentation of the labor market for drivers/freight handlers. I argue that a difference in optimal human resource policies between the two types of firms is an important cause of the parallel segmentations. Differences in how similar production technologies are used to serve the two markets mean that firms have different optimal solutions to the agency problem they face in motivating employees, leading to high powered incentives at reservation wages in one case, and low powered incentives with positive rents in the other. But this difference in compensation schemes sharply increases the agency or transaction costs involved in bringing both types of production under common hierarchical control, due to pay equity effects, while the corresponding benefits are modest, leading most firms to specialize. To formalize this account, I extend a simple version of the standard "risk-sharing" principal agent framework by adding a reciprocity component, producing a new model with endogenous segmentation of the specified type. The new model also provides new hypotheses about the source of union wage differentials, and details a mechanism by which technological change can lead to increasing inequality in labor incomes that is distinct from the usual differential returns to skills account.