Numerous scholars have considered the relationship between gubernatorial power and political outcomes. In fact, gubernatorial power has been used as a key explanatory factor in analyses of topics such as gubernatorial approval, divided government, regulation, and even individual political behavior. The key to these studies is the precision with which scholars can measure gubernatorial power and many such studies rely on the Formal Powers Index (FPI)-a measure maintained by Beyle. In this article, we reconsider these commonly used power scores in three parts. First, we argue and show that FPI suffers from a key measurement error that is particularly problematic in analyses of time-series data. Second, we present a new approach to estimating gubernatorial power and explain how this approach deals with the measurement errors in the FPI. Finally, we use our new scores to replicate a study that originally relied on the FPI to analyze the effect of gubernatorial power. Given the prevalence of the FPI in the existing literature, our results have key implications for the study of the effects of gubernatorial power. Downloaded from Krupnikov and Shipan 439 create public policy. These differences in gubernatorial power are examples of institutional conditions that can affect outcomes at the state level. Indeed, the governor's formal power-often relative to that of the state legislature-has been identified as a key explanatory factor in numerous behavioral and institutional phenomena, ranging from the way individuals attribute responsibility for political outcomes (Rudolph 2003) to a governor's influence over state agencies (Dometrius 2002).Much of this research relies on the same measure of gubernatorial power: the Formal Powers Index (FPI) that was originally created by Schlesinger (1965) and then streamlined and regularly updated by Thad Beyle. 1 The current version of the FPI (Beyle and Ferguson 2008) details a governor's power over various institutional areas: budget, veto power (VP), separately elected officials (SEP), tenure potential (TP), gubernatorial party control (PC), and appointment power (AP). These ratings, which are widely accepted and used within political science, provide state-level power scores for governors over nearly a 50-year period, from 1960 to 2007. The FPI is ubiquitous, and for good reason: These are useful, simple ratings that draw on legitimate differences in gubernatorial power across states. Nonetheless, this ubiquity is a source for concern when scholars rely on these scores to make longitudinal claims. In this article, we consider the underlying issues with using the FPI in a longitudinal manner. Specifically, we will focus our discussion on the measure of budgetary power, which is a key component of gubernatorial power more generally (see, for example, Alt and Lowry 2000; Barrilleaux and Berkman 2003).Using these ratings to make longitudinal claims is problematic due to the structure of the index. In particular, updates over time to the power scores have included and accounted for differ...