ABSTRACT
This chapter studies the response of point spreads to a readily observed event: the absence of a key player due to injury. The analysis is thus similar to an event study, with the added feature that the mean price response is compared with the mean effect of the injuries on actual outcomes (game scores). The analysis in this chapter can thus be viewed as a test of event study methods using a market where the simplicity of the financial contract makes such a test feasible. Yet, though the contract is simple, the injuries themselves create problems, since many of them are partially anticipated events. In the case of basketball injuries, an empirical model of the probability of player participation can be estimated and used in conjunction with a model of efficient pricing to interpret the relation between point spreads and scores. The pricing model yields numerous implications that are consistent with the data. Hence, the good news is that the relation between point spreads and scores during injury events is consistent with efficient pricing. The exercise tests and lends credence to the importance of partial anticipation as an important factor in interpreting abnormal returns when the ex ante probability of an event differs substantially from zero.
