surance. In contrast, little is known about the effects of incentive∕insurance arrangements
on lower-ranking workers’ compensation. Yet incentives within the firm are not confined
to managers. As is remarked by Lazear (1999), “there are many workers that a firm wants
to be motivated”. The use of bonuses and premiums related to the general conditions of
the firm, possibly with different intensities for different types of workers, is common, as are
more traditional payment mechanisms tied to individual output (such as piece rates).
In this paper we study the role of the firm as an insurance provider. We test the
insurance-incentive trade-off in a context where various types of workers and firms interact,
relying on matched firm-employee data available over a long period of time for Italy.
We start from the observation that if incentive∕insurance considerations are important
in shaping the relationship between workers and firms, then they offer testable implications
for the observed compensation scheme. First, the model predicts that firm performance
and workers’ compensation move in parallel. Second, and more interestingly, it predicts
that the amount of insurance varies with firms and workers characteristics in ways that are
typical of agency models and are not shared by other schemes of workers compensation.
The wealth of information of our data set allows us to measure most of these characteristics
(e.g., workers’ risk aversion or firm performance variability), and to verify if their effects on
the compensation scheme are consistent with the predictions of the model. As we shall see,
our results lend support to the agency model as a better representation of the compensation
scheme than alternative models, such as the standard spot model with or without frictions
and the implicit contract model.
Our study contributes to the empirical literature in several respects. First, since we
base our analysis on a representative sample rather than a specific type of workers and
firms, we can draw more general conclusions on the relative importance of insurance and
incentives. Second, we can study how insurance coverage varies with types of firm and
worker; this allows us to test some direct implications of the basic incentive∕insurance
model and help discriminate between the agency model and the competitive model with
or without frictions. Third, since our data cover a span of years, we can study whether
insurance provision is sensitive to the temporary or permanent nature of shocks. This issue