Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
The theoretical public-choice literature suggests that vote trading is an important determinant of congressional voting behavior. Yet empirical voting models do not allow for vote trading. These models recognize that observed ideology may influence legislative behavior but do not correct for unobserved ideology. This study devises new tests for logrolling and ideology. The empirical model controls for logroll agreements and unobserved ideological interest via the correlation of unobserved variables. The results reflect the presence of vote-trading coalitions on some votes but not on others. The results cast doubt on the importance of personal ideological interests of legislators. Copyright 1992 by American Economic Association.