Working Paper: CEPR ID: DP2879
Authors: Alison Booth; Mark L. Bryan
Abstract: Economists have, at least since Olson (1965), suggested that there is a free rider problem associated with labour union membership. The reason is that union-set wages are available to all workers covered by unions irrespective of whether or not they are union members, and - given that there are costs to membership ? workers will only join if they are coerced or offered incentive excludable goods. Yet empirical research for both the US and for Great Britain has shown that there is a substantial union membership wage premium amongst private sector union-covered workers. An implication is that the free rider hypothesis is therefore irrelevant, since these studies reveal significant economic gains in the form of higher wages for union members. Using rich data from a new linked employer-employee survey for Britain, we show that this is not the case. While estimates assuming exogenous membership do indeed suggest there is a union membership wage premium of a similar order of magnitude to that found in other studies, we demonstrate that ? with appropriate instruments based on theory and with additional controls ? this wage premium vanishes.
Keywords: Employer-Employee Data; Member-Nonmember Covered Wage Premium
JEL Codes: J51; J52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| union membership (J51) | wage premium (J31) |
| membership endogeneity (D71) | wage premium (J31) |
| unobserved characteristics (D80) | union membership (J51) |
| union power (J51) | wage premium (J31) |
| unobserved influences (D80) | wage premium (J31) |