We present evidence suggesting that the widely used regression method for testing budgetary incrementalism (Davis, Dempster, and Wildavsky, 1966a, 1966b, 1971) is not suited for U.S. budgetary data that appear to be nonstationary. The method, moreover, cannot detect a nonincremental period following (or preceding) an incremental period. We offer an alternative method that is valid even in nonstationary cases. Our method exploits both the crosssectional and time-series characteristics of the budgetary data to identify statistically the occurrence of incremental decisions (counts) and to estimate incremental cycles for various agencies. More important, the method lends itself to explanatory hypotheses testing. We formulate a set of hypotheses about how various political and economic factors may affect incremental budgeting. We test these hypotheses using the estimated counts in a Poisson regression context. Our results suggest that the Democrats’ control over the political process, a switch in the party controlling the White House or Congress, and presidential election year promises (and political vulnerabilities) all cause departures from incremental budgeting. The public pressure resulting from a persistently large deficit has a similar effect. This work may contribute to our understanding of legislative choice.