Wednesday, January 30, 2013

Tax Policy Colloquium, week 2: Ed McCaffery's "Bifurcation Blues"

Yesterday at the colloquium, we had a very enjoyable and stimulating session on Ed McCaffery's paper, "Bifurcation Blues: The Perils of Leaving Redistribution Aside."  This was a very early draft (Ed had kindly offered to go early in the semester), but even as it stands it raises a number of interesting issues.

Not to presume, but it is possible that the concept of "bifurcation" will disappear from the paper at some point.  Right at the start, the paper mentions the well-known public finance idea (going back to Richard Musgrave more than 50 years ago, but beyond that back to at least Adam Smith, and so far as I know perhaps the ancient Romans or Babylonians) that, at least as a conceptual matter, we should think separately about efficiency and distribution.  AKA, the size of the pie and who gets what.

The paper then mentions the well-known (among academics) argument by Kaplow and Shavell that "legal rules" should aim purely at efficiency, leaving distribution to be handled entirely by the "tax system."  "Legal rules" means such things as tort law, contract law, corporate governance rules, product liability, landlord-tenant law, etcetera, etcetera.  The "tax system" includes the transfer system.

The basic Kaplow-Shavell argument is that any time we tilt a legal rule away from efficiency to serve distributional ends (e.g., favoring consumers who tend to be poor over business owners who tend to be rich), we both (a) distort labor supply, since money is being shifted from higher earners to lower earners, and (b) distort other stuff as well, such as optimal consumer contracts.  By contrast, if we use the tax system to the same end, we just distort (a) but leave (b) alone.  They argue that (b) is an additive distortion that is unlikely to mitigate (a), so you are probably just making things worse without increasing the amount of redistribution that you can accomplish.

McCaffery suggests in the paper that he is willing, at least for argument's sake, to accept Kaplow-Shavell as correct regarding the first-best, but that the "bifurcation" they endorse has failed in practice because we have not gotten nearly enough progressive redistribution from the tax system in actual practice.

As I read the piece, this led me to anticipate that he would start saying later on that we should also redistribute in all sorts of other ways, even though they might be inferior in pure efficiency terms, because otherwise, as a political matter, we will end up with too little redistribution (at least based on Ed's preferences, which I'll admit I generally share in this regard).

But that's not at all where the paper goes, at least in the current draft.  Instead, its complaint about "bifurcation" has more in common with what I decided one might call the "A Funny Thing Happened on the Way to the Forum" problem.

I don't know if you've seen the uproarious film version of this inspired musical farce, starring Zero Mostel (and featuring the talents of Stephen Sondheim, among others), but it has a famous opening number, "Tragedy Tomorrow, Comedy Tonight."  Sub in "efficiency" for "comedy" and "redistribution" for tragedy, and the paper complains that we are being told every night, in effect, "Tragedy tomorrow, comedy tonight."  So it is never time, so far as the political system is concerned, for redistribution - despite, in Ed's view, not only his support for it but that of a majority of the American people.

What makes this arguably "bifurcation" is the idea that the two things are always done separately,  In particular, he has in mind the adoption of tax cuts that are separated in voters' minds from the spending cuts they may ultimately necessitate (this of course is "starve the beast"), and the hijacking (in his view) of tax policy discussion to focus purely on efficiency and hence the case for low rates.  And where this is financed through base-broadening (i.e., 1986-style tax reform) and thus isn't starve-the-beast, it's still limiting the issues on the table to efficiency if everyone agrees that the legislation has to be not just revenue-neutral but distribution-neutral.

I'm not sure how fruitful it is to link this with Kaplow-Shavell or Musgrave's multiple conceptual branches of government as types of "bifurcation."  But this at some point becomes just a semantic question.

The paper also critiques the political economy effects of optimal income tax (OIT) theory, which ostensibly has helped in practice to push the case for relatively flat, rather than graduated rates.  As it happens, on February 26 our speaker will be Peter Diamond, presenting a paper that offers an OIT-based argument for steeply graduated rates that might reach 70 percent or so at the top.  But fair enough, OIT has typically been invoked in support of flattish rates, and for a good logical reason (rates at the low and middle ends are inframarginal for lots more taxpayers than those at the top - for example, if you'll be earning at least $1 million, your marginal incentives are wholly unaffected by the Social Security payroll tax that cuts off at about $110,000).

OIT is definitely not "bifurcation at work," however.  If there is one thing that OIT clearly is, it is an effort to link and integrate efficiency considerations with those of distribution in determining what the rate structure.should look like.

Not to mention that it's far from clear how politically relevant OIT has ever been, even though people such as me, in our ruminations and writings, find it a nice instrument for thinking clearly about issues that have real-world relevance.

Moving beyond all the bifurcation talk, the paper's main themes (with my brief comments on them) include the following:

1) The big inequality problem in our society, it argues, is inequality of wealth.  That is what the tax system should address, in Ed's view, and is not doing so nearly sufficiently.  An interesting question is how this squares with Ed's prior work arguing that saving is good and that we should have a progressive consumption tax.  If wealth is the problem, wouldn't you want to tax it?  And individuals don't hold wealth unless they both (a) acquire it and (b) retain it, i.e.., by saving some of what they have acquired.

I know Ed disagrees about the view that there is a tension between his diagnosis (wealth inequality is the problem) and his prescription (don't tax wealth until it's spent), but I am not the only reader to believe that the tension was there, at least in this early draft.

2) The paper is extremely critical of the American Taxpayer Relief Act of 2012 (ATRA), which resolved the fiscal cliff showdown, on several grounds.

In particular, it argues that ATRA's distributional effects were blatantly misrepresented by both sides to the political negotiations.  Everyone in Washington depicted it as purely a tax increase for upper income individuals, while middle class and poor taxpayers were held harmless.  But by far the biggest tax increase that it contained in revenue terms, relative to the rules applying to 2012, was the expiration of the payroll tax holiday.  Politicians threw sand in our eyes about this (or perhaps pulled the wool over our eyes and then started throwing sand) by scoring ATRA against a baseline in which the Bush tax cuts continued but the payroll tax holiday expired, which seems logically inconsistent.

As it happens, there is a decent logical argument for this.  The payroll tax holiday was always meant by everyone to be temporary.  By contrast, the Bush tax cuts were intended by proponents to be permanent, and indeed the Obama Administration's support for letting the high-end tax cuts expire was based on its view about good policy, not about what is current policy.  Also, if one wants to condemn the Obama Administration for not fighting to extend the payroll tax holiday (which would have been good macroeconomic policy, while the labor market remains so suppressed), then one is blaming them for something that wouldn't have existed in the first place if they hadn't fought for it in prior go-rounds with the Congressional Republicans.

But on the other hand, yes, I agree that ATRA had big effects relative to the 2012 rules, which certainly offer a relevant datum, and that the "current policy" baseline was certainly all too convenient.for the parties in a way that may have misled millions of Americans and left them quite surprised when they noticed (if they noticed) that their paychecks in January 2013 were suddenly smaller than those from December 2012, all else equal.

3) A very powerful point in the paper concerns the decision in 2012 to extend or restore the estate tax, albeit with exemption levels that will keep 99.7 percent of decedents from being subject to it, while doing absolutely nothing about the tax-free step-up in asset basis at death.  If transferring assets by gift or bequest were treated as a realization event, at least over some threshold income level or amount, games and tricks akin to those that clever planners continually deploy against the estate tax would be largely unavailing.  Depending on the parameters, the tax liabilities imposed might be far greater.  Moreover, at least in some key respects deadweight loss would go down, not up - there would be less tax motivation to hold appreciated assets indefinitely, rather than selling them when otherwise convenient.  I agree with Ed that this was a bad outcome, taken as a unit, and that it reflects the interests of rich and powerful taxpayers in a manner that one can, if one likes, choose to find suspiciously convenient.

1 comment:

blissex said...

«By contrast, if we use the tax system to the same end, we just distort (a) but leave (b) alone. They argue that (b) is an additive distortion that is unlikely to mitigate (a), so you are probably just making things worse without increasing the amount of redistribution that you can accomplish.

McCaffery suggests in the paper that he is willing, at least for argument's sake, to accept Kaplow-Shavell as correct regarding the first-best, but that the "bifurcation" they endorse has failed in practice because we have not gotten nearly enough progressive redistribution from the tax system in actual practice.
»

I am rather entertained by this argument because it is based on purely *assuming* that the legal rules are inded neutral, when they have been designed ever more in the past decades to give an enormous effective advantage to wealthy or high income market participants.

The rules matter enormously because they largely determine who has got more leverage in market transactions, which can usually happen equally plausibly in pretty wide price bands.

Lots of money and effort is spent by powerful vested interests in Washington to ensure that the rules are designed in a way to favour the leverage of those vested interested with respect to their suppliers, including labour, and their customers, because these are far more valuable than even tax credits or government subsidies.

Those who accept the «argument by Kaplow and Shavell that "legal rules" should aim purely at efficiency» are accepting the ridiculous premise that it is possible to aim for aiming at pure efficiency for the rules, or that it is in fact the case that they are already efficient, so only the "tax" side needs discussing.

How can one measure the «efficiency» of rules, without a theory of value and an even more difficult theory of the impact of rules on leverage in transactions?

I suspect that the answer is that proponents of «"legal rules" should aim purely at efficiency» have in mind some tarted up version of Neoclassical Economics which claims to define precisely what «efficiency» means but that is purely propaganda because it is based on "assuming away" difficult issues like capital, money, time, and assuming that no market participant has any leverage, and even so it is in practice entirely irrelevant to policy issues because the Second Best Theorem, which is now 60 years old, and never mind accepting purely for a laugh that «Kaplow-Shavell as correct regarding the first-best».

Instead looking carefully at which vested interests pay lots of money for which rules shows that there is no reasonable expectations that rules can be or are in fact designed purely for «efficiency» and I strongly suspect that aiming at putting the debate only on the "tax" side is designed to draw the attention away from the far more important and valuable "rules" side, also indeed because «the "bifurcation" they endorse has failed in practice because we have not gotten nearly enough progressive redistribution from the tax system in actual practice.».

But the big deal is that the "legal rules" determine in a largely political way a very large part of the "distribution" of income and wealth, and that "redistribution" via the "tax"/"spend" side is far less important.