Learning in sequential wage negotiations: Theory and evidence

When union-firm pairs bargain sequentially, and when unobserved components of firms' abilities to pay are subject to correlated shocks, unions that ba rgain later in a sequence can acquire valuable information by observing pre vious bargaining outcomes in their industry. We derive the implications of this kind of learning in an asymmetric information model of wage negotiatio ns and argue that the most robust implication is a lower incidence of strik es among "followers" than "leaders" in wage negotiations. Considerable empi rical support for this implication is found in a long panel of Canadian con tract negotiations.