seems to suggest that this undervaluation is diminishing.
Finally, the set of PITCHER, CATCHER, and OUTFIELDER attempts to measure the impact of relative risk aversion on players' behavior in the arbitration process. These are dichotomous variables that equal one for the selected group and zero for the rest of the field. In examining the OUTFIELDER row first, it can be seen that the risk variable implies that outfielders make more conservative salary demands in arbitration. This suggests that there is a greater risk of performance deteriorating or career-ending injury in this position. The difficulty with this interpretation is that pitchers and catchers are more frequently injured, compared to outfielders and infielders. However, the coefficients on these variables are insignificant. This result implies that pitchers and catchers are not significantly different from infielders. 5 One possible explanation for this is that the data set includes only those players who went through the arbitration process. It may be that players in risky positions tend to settle before arbitration in order to secure longer-term contracts that would help mitigate some of their risk.
Results generated by the empirical tests of this chapter provide strong evidence that: (1) players going through major league baseball's salary arbitration procedure for a second time behave differently from first-time entrants by lowering the spread between their final salary offer and the final offer of their team, (2) small market teams increase the spread presumably because these teams value their assets based upon local economic conditions, and (3) collusion among owners had the effect of lowering player expectations due to the apparent lack of free play in the market. These three results are consistent with the findings in Frederick, Kaempfer, and Wobbekind ( 1992).
In addition, we find that players of color and players occupying some injury- prone fielding positions behave differently from other players going through the salary arbitration procedure. This study's foray into the race and ethnicity dimension yielded interesting and somewhat surprising results. We had anticipated that racial discrimination might possibly lead to greater spreads in the salary arbitration process. In fact, spreads among players of color are lower. This finding suggests that nonwhite players seem to systematically undervalue themselves. Our test of this relationship over time implies that the undervaluation is decreasing. Second, it was found that outfielders, occupying an injury- prone fielding position, attempted to mitigate the higher risk of their position by submitting lower bids. However, pitchers and catchers apparently did not, although this may be the result of these players settling before arbitration.