Restrictive rules reconsidered

Theory: Three classes of theories-partisan, informational, and distributive-yield a diverse set of predictions about legislative organization and legislative outcomes. A specific area in which they are amenable to empirical tests is the House of Representatives' use of rules governing the amendment process. Dion and Huber (1996) propose and test a formal model of the choice of rules that lies solidly within the partisan or majority-party-leadership class of theory and also comports well with conventional wisdom about rules and their strategic uses by the majority party. Hypotheses: The Dion-Huber theory identifies a necessary condition for open rules: that the House median voter lies interior to the Rules Committee median and the standing committee median. This condition defines a rule profile that Dion and Huber show helps to predict rule assignments. Informational and distributive theories yield additional hypotheses, only one of which Dion and Huber test. Methods: (1) Replication and extension of Dion and Huber's tests using their committee-level data set, OLS, GLS, and fixed-effects estimations. (2) Joint tests of the rule-profile hypothesis with informational and distributive hypotheses using a rule-level data set and probit analysis. Results: In the committee-level data set, the rule-profile effect is sensitive to model specification. In the rule-level data set and in the presence of both informational and distributive variables, the rule-profile effect is barely or not at all distinguish able from zero. Information- and distributive-theoretic findings are significant and consistent with previous studies.