Association for Public Policy Analysis & Management

This paper addresses issues of measurement in the current literature regarding public revenue diversity which may, in part, be a driving factor behind contrasting and inconsistent empirical results in the public finance literature. There has been an ongoing debate in the literature about the role of revenue diversity in the public financial system, with one camp holding that it is strategic tool policymakers can employ to create fiscal illusion and inflate public expenditures, and another camp maintaining that it is a sound financial management tool to reduce revenue volatility. After critically assessing the literatures in which revenue diversity has been employed as a key variable and discussing the specifics of diversity measurement therein, additional conceptual issues are addressed. These issues include theoretical concerns, as well as variations proposed and intermittently utilized in the empirical literature. Finally, a host of alternate specifications of revenue diversity are defined and calculated using a standard U.S. state-level dataset, and these measures are compared to one another according to both the rankings of states that result as well as the absolute measures of diversity obtained. These calculations result in substantial variability, both across states and over time, which suggests that the spcecific measurement choices made may be extremely relevant to final empirical results. This is demonstrated by performing sensitivity analysis in a basic fixed-effects regression set up, employing revenue diversity as an explanatory variable for both total public expenditures and revenue volatility. The number of revenue categories defined as well as the breadth of definition of the term "revenue" (i.e. whether only taxes are included, or whether broader revenue sources such as charges and user fees are included) can significantly alter the direction and significance of the results. The paper concludes with a discussion of particular measurement choices that warrant more careful consideration than they are currently afforded in the empirical literature.