The Random Coefficient Approach for Estimating Tax Revenue Stability and Growth

Abstract

The issue of tax revenue stability and growth has been of concern to policy makers and economists for many years. One important focus of the literature is the optimal tax portfolio, which assumes that revenue variance is entirely unpredictable. However, as evidenced by Fox and Campbell, some revenue variance arising from changes in economic conditions is predictable. The purpose of this study is to revisit Fox and Campbell's work. They studied revenue growth and stability with a fixed coefficient model (FCM). This study uses a random coefficient model (RCM). The RC Mapproach appears to provide improved estimates and confirms the conclusions of their earlier work. The response of short-run elasticities to the business cycle appears both strong and variable across commodities, and no single commodity dominates revenue growth or stability. Although this study supports the design of an optimal tax portfolio, it emphasizes the need to explicitly model for economic conditions and to continually adjust the tax portfolio. However, given the political and budgetary process, these adjustments may not be feasible.