The stability of logrolling: An expectations approach

This essay provides an explanation for the existence of stable logrolling arrangements. The model presented assumes that issues are voted on one at a time, that people are risk averse, and that each person forecasts future issue decisions with a random variable whose mean is linear in the alternatives of the present issue. We show that under this model, the expectation that logrolling will take place stabilizes the logrolling process. For this expectation to be stable, it must be symmetric, which means that revoting of issues must be possible. Since revoting is possible in institutions such as Congress, we conclude that an important set of real world voting bodies satisfies the conditions for stable logrolling. Copyright Martinus Nijhoff Publishers 1986