Friday, March 05, 2010

NYU Tax Policy Colloquium of March 4, 2010

Yesterday, Rebecca Kysar presented her paper, Lasting Legislation. In large part, this paper responds to George Yin's paper (presented at a past NYU colloquium) arguing that the widely reviled sunsets of the 2001 and 2003 tax cuts were actually a good thing. I must confess to a lack of sympathy with George's thesis. thus creating general agreement with Rebecca though there nonetheless was plenty to discuss.

BTW, I should note that it's occurred to me this year that, as the sessions are off the record, I perhaps shouldn't discuss them here as such - rather, I discuss my thoughts about the issues raised by the paper we discussed, so that there will be no possible problem with this. (Perhaps in past years it had never occurred to me that I actually have any readers here.)

Anyway, I see George's argument as twofold. First, given the horrendous U.S. fiscal situation, isn't it good ex post that all those tax cuts formally expired and thus will disappear (or have already done so) unless and until they are reinstated through a new act of Congress. Second, isn't there a nice match between not applying the legislation to "out" years (past the official 10-year budget window) and not having revenue estimates for those years formally before the Congress.

But in thinking about the 2001 and 2003 tax cut sunsets - which I would group, actually, with the front side of crazily complex and varying deferred implementation that these bills also featured - I think 3 points are paramount.

First, they were faux temporary - no one actually was claiming they should apply only temporarily (in the manner, say, of the 2004 dividend tax holiday, which raises its own set of issues concerning time consistency but is fundamentally different).

Second, the sunsets were optional - chosen by the proponents of the tax cuts to serve those individuals' purposes. So the question presented here is not whether tax cuts should generally have sunsets, but whether proponents should (without encountering too much condemnation, as we're discussing a soft "norm" here) have the sunsetting option in their toolkit. Having this option, which one can use or not as one likes, can only empower and aid the proponents of tax cuts. (Which I consider wildly reckless under current fiscal circumstances, when financing isn't provided, given the long-term fiscal gap - this is about tax-shifting, not high versus low taxes over the long run.)

Third, the whole point of the sunsets, particularly in 2001, was to evade budget rules that are meant to restrain fiscal irresponsibility of this kind. (In particular, the Byrd Rule in the Senate concerning revenue loss past the 10-year budget window.) This, of course, is rather ironic given the healthcare debate today - it involved aggressive Republican gamesmanship to avoid needing 60 votes and get what they wanted with a simple majority. Odd, how majority rule apparently wasn't a terrible thing back then.

I think that's enough to make the case that the sunsets were terrible policy unless one takes the view that tax cuts should generally be enabled and budget rules evaded. That might be a plausible stance if (a) the tax cuts had been financed, (b) there was no long-term fiscal gap, and (c) there was no particular reason to think Congress errs in the direction of enacting unfinanced policies that create a long-term fiscal gap. But for any of those points to hold, we would have to travel to an alternative Bizarro Universe (albeit, a better one than ours).

The delayed implementation rules in 2001 and 2003, none of which made any sense substantively, were also budgetary gamesmanship, meant to create much bigger tax cuts than the budget agreements (or legislative understandings) about their permissible 10-year cost would otherwise allow. Delayed implementation makes a further mockery of 10-year budget windows, by causing them to be much less representative than otherwise of the true size over time of the new programs being enacted. This was exploited once again with the Medicare prescription drug benefit, a wholly unfinanced $20 trillion entitlement (that being its infinite horizon cost as subsequently estimated) that supposedly cost only $350 billion (?) over 10 years due to extremely delayed start-up, solely to game this number, plus dishonest squelching by the Bush Administration of revised estimates showing that it was costlier even inside the window.

Is the Obama Administration doing the same thing now with delayed implementation of the healthcare plan? Presumably so, but I confess to not knowing the details here, plus it is supposedly financed over the long term.

My main takeaways from thinking about all these issues were as follows. First, there really is no substitute for infinite horizon budget forecasting that disregards claimed future tax increases or benefit cuts that are discontinuous (i.e., not being phased in gradually within the short term) and not otherwise entirely credible. Second, such forecasting would have to be used in relatively binding budget rules to make any difference. These would have to be independently estimated, non-waivable, and would have to use a combination of mandatory adjustment rules to move back towards balance and/or super-majority rules. (Though note that super-majority rules only help in this setting if you have today's hyper-partisanship - in other situations they can simply lead to bigger and costlier logrolls.) Third, none of this is going to happen.

Why do I support super-majority rules for making increases to the fiscal gap harder to enact but not for, say, healthcare? The only difference is whether one wants to systematically discourage a particular type of legislation. I see a pathology in the form of increasing the fiscal gap through non-sustainable policies. An argument that this reasoning applies as well to "big" legislation generally - if applied symmetrically to Democratic and Republican initiatives - could only be rebutted by challenging that particular application of the discouragement principle - not by reference to majority rule versus requiring a larger consensus as such.