Market Structure, Union Rent Seeking, and Firm Profitability

Considerable attention has been given to studies concluding that monopoly profits associated with industry concentration provide the primary source from which labor unions capture rents [Salinger (1984); Karier (1985); Freeman (1983)]. This conclusion has obvious appeal. It implies that the impact of concentration on profitability has been systematically understated in empirical studies owing to omitted variable bias (concentration and unionization are positively correlated), and that deleterious union effects on firm survival are mitigated since union rent seeking is restricted to firms realizing supra-competitive profits. The conclusion that concentration-related profits provide the principal source for union gains has been attacked by Hirsch and Connolly (1987). They first note that evidence on this issue is mixed. Domowitz, Hubbard and Petersen (1986) find little evidence that union effects on price-cost margins are more negative in highly concentrated industries (they find this relationship to vary over the business cycle), while Clark (1984) obtains the surprising result that union profit effects are most deleterious among lines-of-business with low market shares. Evidence from the labor market, moreover, reveals union wage premiums that are typically smaller in highly concentrated industries [Lewis (1986, pp. 154-155)], just the opposite of what is implied by the profitability studies cited at the outset. Hirsch and Connolly also provide firm-level evidence indicating that results from Salinger, in particular, are fragile. Previous studies, however, have been hobbled by serious data limitations. This note presents empirical evidence on the relationship between profitability, market structure, and union coverage using a unique data set intended to overcome deficiencies inherent in past studies. Specifically, we employ firm-level measures of both market value and accounting profitability, utilize companyspecific measures of union coverage collected by the author, and use sales-weighted measures of industry concentration and firm market shares.