The Sales Tax Puzzle – Has the law of demand been repealed?

TRA86 removed the deductibility of state and local sales taxes from U.S. personal income taxes. This effectively increased the price to state governments of sales tax revenue relative to income tax revenues. This implies that state reliance on income taxes relative to sales taxes should have increased after TRA86 was implemented. Leading public finance economists investigated this in the period shortly following the reform and invariably found that the predicted substitution of income for sales taxes did not take place. In fact, several studies noted that state dependence on the sales tax increased relative to income taxes – hence the Sales Tax Puzzle. Several experts tried to rationalize their non-findings. These rationalizations are unconvincing. This paper asserts that the puzzle may be resolved in two different ways. The first, seemingly trivial, is that the analysis must incorporate sufficient time so as to allow for relatively complex adjustments. Our contributions here are to anchor this “obvious” point with a theoretical model related to earlier work of the authors, and an explicit empirical examination of the lag structures of individual states’ reactions to the comprehensive tax reform. The second demonstrates that the analysis must take into account regional shifts that were taking place in the U.S. during this period. When such compositional shifts involving political and deductibility patterns are explicitly introduced into the model, the paradoxical findings are resolved both in the long and in the short run. Our contribution here is a demonstration that the use of detailed state-level data, unlike any of the earlier work in this area, allows for the inclusion in the analysis of inter-regional shifts in various parameters. An additional contribution is the explicit use and emphasis of the propensity to itemize as an independent variable. The use of this variable ties this area of analysis of post tax reform behavior into a hitherto untapped strand of public-finance literature. The variable is both theoretically interesting, and proves to have experienced compositional shift patterns which help resolve the sales-tax puzzle paradox. Copyright Springer-Verlag 2003