This past Monday, the JCT issued a pamphlet taking on an important conceptual issue, tax expenditure analysis, and making what I think are significant strides in how to rescue the analysis, and its important informational content, from sterile debates about what constitutes a "normal" tax base. The JCT's main move is to distinguish between (a) narrow subsidies, which on their face depart from the usual treatment of a broader category of items, evidently reflecting Congressional intent to affect resource allocation, and (b) what the pamphlet calls "tax-induced structural distortions," such as quirks arising by reason of the realization requirement. The JCT also proposes to measure negative tax subsidies (i.e., allocatively-minded penalties) as well as positive ones.
My writings on tax expenditures (such as in chapter 8 of Taxes, Spending, and the U.S. Government's March Toward Bankruptcy) are similar in spirit although different in some details (e.g., how to classify the earned income credit). So I am delighted to see the JCT taking up the cause of making the analysis more useful and less mired in pointless debates, such as that between income and consumption tax advocates, that are orthogonal to its informational content.
I first got interested in writing about the topic some years ago, when I was a commentator at an AEI event in which Bruce Bartlett criticized tax expenditure analysis because he saw it as a device used by income tax advocates to peddle their side of the ongoing debate. I agreed with him that it had been used this way, going back to Stanley Surrey, but argued that it has more general informati0nal content, and can advance agendas such as his (favoring small government and identifying departures from it) no less than Surrey's. Plus one need not have an agenda in order to favor more crisply identifying cases in which Congress conceals what seem clearly to be allocative policies (e.g., favoring a particular type of investment) by embedding them in a seemingly distributionally motivated instrument (such as a general income or consumption tax).
My favorite example of the core point made by tax expenditure analysis remains one that I got from David Bradford. Let's cut both taxes and spending by $50 billion, David pretended to urge, by zeroing out $50 billion of military spending (to buy advanced weapons) and enacting instead $50 billion worth of "tradable tax credits" that would go to the very same weapons suppliers for the very same weapons. At the end of the day, everything would be exactly the same, but taxes and spending would each be reported as $50 billion lower. Without a tax expenditure concept, it is hard to show as crisply that nothing in this scenario has genuinely changed.