Shirking or stochastic productivity in major league baseball: comment.

I. Introduction

In an earlier article in this Journal[1], Anthony Krautmann tests the hypothesis that a major league baseball player has a greater tendency to shirk during the season immediately following the signing of a long-term contract than he does during the season immediately preceding the signing of a long-term contract. His method was very straightforward. He observed performance data on 110 non-pitchers who signed contracts of 5 years in length or greater from 1977 to 1983. By comparing each player's career slugging average (SA) with the Sa's for the years immediately after and before the player signed a contract, he could not reject the null hypothesis that no shirking took place.

The basic thrust of this comment is that the SA is not the performance measure most likely to reveal shirking if it does exist. Krautmann relates that in[2] Lehn found a positive correlation between the number of years remaining on a player's contract and the number of days spent on the disabled list. However the slugging average used in Krautmann's paper would not necessarily be affected by the number of days a player spends on the disabled list. A measure more closely correlated with time spent on the injured reserve list would be total bases.

II. Model

To test the shirking hypothesis using total bases for the 110 players observed paper, the following regression equation results are presented:

[TABULAR DATA OMITTED]

where [BASES.sub.it], is the number of total bases for player i in season t, [LT.sub.it] is a dummy variable equalling 1.0 if player i signed a long-term contract immediately before season t, 0.0 otherwise, [AGE.sub.it], is player i's age at the start of season t and [Mathematical Expression Omitted] is a random error term.(1) The figures in parentheses represent the standard errors of the coefficient estimates. The value of [R.sup.2] for the regression is .448.

III. Conclusion

Since the sign of the coefficient for [LT.sub.it] is negative and significant, we can reject the hypothesis that shirking does not occur. These results indicate how sensitive Krautmann's results are to the choice of the performance measure.

References

[1.] Krautmann, Anthony C., "Shirking or Stochastic Productivity in Major League Baseball?" Southern Economic Journal, April 1990, 961-68. [2.] Lehn, Kenneth, "Property Rights, Risk Sharing, and Player Disability in Major League Baseball." Journal of Law and Economics, October 1982, 273-79. [3.] Thorn, John and Peter Palmer, eds. Total Baseball. New York: Warner Books, 1989.

(*) The author thanks Dave Denslow, Stephen Donald, Larry Kenny and an anonymous referee for helpful comments. The author also thanks Anthony Krautmann for providing a list of the players used in his original research[1]. The usual caveats apply. (1.) The data were found in Thorn and Palmer[3]. The total bases observations were "deflated" using an annual index which is available from the author upon request. Since each player represented 2 observations, the total number of observations is 220.