Work incentives and salary distributions in the national hockey league

Abstract

Students of efficiency wage models of the labor market hold that contracts which base rewards on an ordering of employee performance will induce employees to work harder [Lazear and Rosen, "Rank-Order Tournaments as Optimum Labor Contracts," JPE, 89, October 1981, pp. 841-64]. But perhaps there are strong work disincentives associated with such contracts. Empirical support for this counterfactual may be found in professional team sports. Players (and, more vociferously, their fans) complain about the extraordinary sums of money team owners spend to sign international stars. The size of these superstar contracts may not only anger veterans on the team but adversely affect overall team performance. To determine whether or not the structure of compensation has any effect on worker shirking, Gird coefficients were calculated for all 26 teams in the National Hockey League. Player salary data for 1996-97 are from www.nhlpa.com/nhlpa/comp/compteam.html. (Salaries of some Canadian players, reported in Canadian dollars, were adjusted at the rate of $1 Canadian = U.S. $0.74, the foreign exchange quote for 1996IV in the 1997 Economic Report of the President.) The larger a team's Gird coefficient, the more unequal is the salary distribution. A Gird coefficient of 1 would represent complete inequality (one player earns all of the team's payroll) and a Gird coefficient of 0 would represent complete equality of the salary distribution. Season-ending point totals (two points for a win, one for a tie, and no points for a loss) by team were regressed on the team's Gird coefficient and average team salary. The reason for including the average team salary in the regression is the belief that players care more about their relative team position than about their salary relative to the league average. The regression results were as follows (t-values in parentheses):