NYU Tax Policy Colloquium, week 13: David Schizer's Energy Tax Expenditures: Worthy Goals, Competing Priorities, and Flawed Institutional Design

Yesterday David Schizer presented the currently above-titled paper at our penultimate session.  The title has changed since he last presented it at another school, and probably will again, as it really isn't on tax expenditures in particular.  Definitely one of the more Graetzean subtitles that I've seen in anything not actually written by Michael Graetz.

While the paper offers a general overview and framework for thinking about the taxation of energy in light of multiple considerations (global and more local environmental concerns, national security, distributional effects, instrument design, etc.), I will emphasize one particular aspect here.  Schizer notes that it might be highly desirable to raise the gasoline tax in the U.S., perhaps significantly.  Even apart from the positive effects on highway funding given our fiscal mechanisms, this might have both national security and environmental benefits.  At present, however, this faces what appear to be, at least in the short run, insurmountable political barriers.  (Plus, he has previously written about the gas tax.)  So what might we do instead?

Here is an idea that the paper sketches out - somewhat preliminarily, as it is an early draft.  Suppose the big political obstacle to increasing the gas tax is that people just don't like being charged for their driving - national security, environmental, and other negative externalities be damned.  In the words of Mary Poppins, just a spoonful of sugar helps the medicine go down.  So if we stapled the gasoline tax to a lump sum transfer, it would look like it was merely reducing a positive amount, rather than creating a liability.

Thus, suppose Congress enacted a $400 "gas-savers' credit."  This would basically be a uniform demogrant going to each household, or individual over the age of 18, or taxpaying unit, or whatever.  (Separate set of questions, obviously, regarding how to define the recipient unit.)  But the credit you would get at the end of the year would be reduced by a charge per gallon of gas purchased or used or deemed purchased or used.  Suppose the average tax charge for the year was about $200.  Then, on average, the recipients would get $200.  If you don't have a car, you presumably end up with $400 - leaving aside questions of how we handle business use, e.g., in the case where I ride on a taxi or bus.

Given the demogrant, the thing would function at the margin as a gasoline tax.  Only, the hope is, people would code it as merely reducing the pat on the back for virtue, rather than as a nasty ol' penalty.

Suppose it is indeed a $400 demogrant per adult, on average $200 net.  Assuming that low-income tax-filers manage to get it (and, note of course, that if netted on income tax returns it would increase "47%" style claims about takers), its cost would depend on the current U.S. population of people over age 18.  This currently stands at about 250 million.  So we are talking an annual budgetary outlay of $50 billion (under this admittedly back-of-the-envelope analysis), enacted so that we can overcome the political unfeasibility of raising the gas tax.

One conceded design flaw is that, once you get to $400 in gas taxes, as high-mileage drivers presumably will, it disappears and there is no net tax at the margin.  (One could of course change that feature, but it would undermine the "spoonful of sugar" presentation.)  To minimize this problem, one has to set the demogrant much higher than the average gas tax that people incur.  Just how much higher presumably depends in part on variance in driving levels within the U.S. population.  But with significant variance, a high net budgetary cost becomes more necessary.

Another design issue is that it requires somehow tracking how much each person drives.  This would be tough to do at the pump.  Other possibilities that I have heard mentioned (i.e., I am not myself advocating them) include using GPS technology, looking at people's odometers when they have mandated inspections, and piggybacking off car insurance companies that use mileage-related fee structures.

Business use would appear to be another big issue.  One doesn't want cab drivers, or for that matter Uber drivers, to get to $400 and then be able to pay no tax.  What about truck drivers and other people driving long distances for business.  What if I have two cars, perhaps one of them in a family member's name, and so forth.

I have to admit, I don't really see this as likely to help sufficiently.  Even leaving aside all the implementation issues, stapling the gasoline tax to what might be in the neighborhood of a new $50 billion net outlay is not where I would look first, or second, in terms of making the gasoline tax more politically feasible.

I also think that political opposition to the gas tax is not quite innate, even within the distinctive DNA of U.S. tax culture.  After all, there are high gas taxes in many other countries.  And it has bipartisan support in the intellectual class.  (Martin Feldstein, for example, favors it with his own proposed spoonful of sugar.)  I get the sense that many responsible Republicans back it, as well as Democrats.  It's just something that can't quite happen just now, but that could happen if the logjam broke sufficiently for proposed alternatives, such as those suggested by Schizer and Feldstein, to be themselves feasible.

Substantively, if we assume political and administrative feasibility, I like the Schizer plan for a reason of my own, which is that I like the $400 demogrant.  Forget the pairing and its optical purposes: insofar as it's workable, this really is the equivalent of separately enacting a $400 demogrant and a gas tax that's capped at $400 per (person or whatever).  I would likely favor the demogrant without the gas tax, just as I would favor the gas tax (though preferably with no per-person ceiling) without the demogrant.  So putting them together isn't inherently bad from my standpoint.  Only, to accept my reason for liking it, you, too, have to like the demogrant, which not everyone will.

If you don't like the demogrant but you do like the gas tax, then, assuming both that the thing works and that you otherwise can't get the gas tax, you have to decide whether you like the package on balance.

While this is a fairly novel proposal (at least so far as I know), it bears a relationship to other ideas that have been posed before.  In the general setting of gas taxes, carbon taxes, and other such instruments, it's sometimes said, for political economy reasons, that we ought to get the incentives right, by having the tax, but avoiding affecting the government's net budgetary position (on the view that this is a separate issue, on which people's political preferences differ) by giving the money back to all the taxpayers in a lump sum, uniform per-person manner.  When you do this, on average everyone pays zero net, but at the margin everyone is paying the tax on extra usage of the polluting commodities, thus satisfying the Pigovian efficiency criterion.

The Schizer proposal, by doing the lump sum payout upfront instead of at the back end, requires overpaying (thus creating a net transfer) if it needs a cushion so that the tax won't disappear at the margin too frequently.  Again, this could either be a feature or a bug, depending on how you like demogrants.  But even in the standard version, only political economy considerations could support choosing that particular payout.  A more general and rational approach would be to set Pigovian taxes as you like without specifying uses of particular tax revenues, and then to figure out the rest of the tax system based on all of the standard considerations, embracing all the usual distributional and efficiency issues.  There is no particular reason, other than making political deals easier to reach, to specify particular outlays with reference to particular revenues.  Money is fungible.