Testing Strategic Bargaining Models Using Stock Market Data

This paper presents three empirical teats of a class of asymmetric information bargaining models using stock market data. The basic idea behind these models is that protracted bargaining can be used to infer information that is privately known by another party to the negotiations. A fundamental implication of these models is that there should be evidence that negotiations result in learning taking place. in the context of union contract negotiations, if bargaining is primarily motivated by the union's uncertainty over the firm's future profitability, then there should be evidence that contract negotiations reduce this uncertainty. This prediction is tested by comparing the variance of the firm's stock price prior to and following a contract negotiation. The data indicate that bargaining results in a significant reduction in this variance. Other predictions of these bargaining models are that the firm's stock price should decline during a strike and increase on the settlement date. The data generally support these predictions with the exception of a decline in the firm's equity value following the settlement of a contract which did not involve a strike.