Wednesday, March 05, 2014

NYU Tax Policy Colloquium, week 6: Hines and Logue, Delegating Tax

Yesterday at the NYU Tax Policy Colloquium, Jim Hines and Kyle Logue presented their paper, “DelegatingTax.”  The paper’s main arguments are: (1) Congress does less delegation of lawmaking authority in tax than in other areas of law, such as environmental regulation, (2) the grounds for favoring significant delegation in other areas also apply to tax, and therefore (3) it might be desirable for Congress to increase the degree to which it delegates lawmaking authority in the tax area.

Possible examples of greater delegation that the paper discusses include (a) giving the Fed authority to make tax rate changes for countercyclical reasons, (b) giving greater discretion, with regard to the design of tax preferences either to the Treasury or to subject-matter experts, and (c) empowering a commission, a la the military base-closing commission of some years back, with the authority to make broad determinations with regard to which tax preferences might be eliminated to fund either lower rates (in a 1986-style reform) or else long-term deficit reduction.

This was an interesting paper and I am generally sympathetic to its line of argument.  To be sure, it is hard to pin down what exactly one means by greater versus lesser delegation (although there clearly is some underlying content there).  Moreover, if it could be defined crisply enough, one might want to try to examine empirically the relative delegation in tax versus other areas in the U.S., and as between the U.S. and other countries.

Further thoughts that I had upon reading the paper are contained in the following, which is an expanded and reorganized version of notes that I prepared for myself in order to be ready for the session:

1) What do we mean by delegation of tax lawmaking authority, and where do we tend to find it?
(a) It has something to do with the choice between rules and standards, where the latter might be viewed as involving greater delegation, but is not exactly the same thing.  (E.g., one could delegate either more or less under either a rules-based or a standards-based approach.)

(b) If we are talking about tax delegation, we should keep in mind delegation to the courts, not just to the Treasury Department.  Most judicial decisions on tax in the U.S. involve statutory interpretation, and thus not only could be overturned by Congress ex post, but in many cases could have been headed off ex ante if Congress had wanted to specify more precisely what it meant.  (Often it doesn’t, however – apparently preferring to delegate.)

(c) Congress appears highly inclined to delegate in cases where it is applying very broad legal concepts that have murky or unclear underlying economic content.  Examples include:

            (i) Section 482, which authorizes the Treasury to reallocate taxable income between related parties based on the principles of clear reflection of income and preventing evasion.  Under the regulations, this is generally done under an “arm’s length” standard.  One asks the counterfactual question: “What would the related parties have done if acting at arm’s length?” and tries as best as one can to finesse the problem that this question commonly lacks a coherent, much less discernible, answer.  My point here is not that doing this is a bad idea (not doing it would probably be even worse), but that the difficulty of devising workable principles has apparently induced Congress to punt / delegate.

            (ii) Debt versus equity – This is the re-delegation that failed.  As is well-known, there is no coherent basis on which the tax law can distinguish debt from equity, at least in close or mixed cases, and there also is no particular policy reason for treating them so distinctively.  Congress for decades delegated the whole issue to the courts.  However, in 1969, it passed Code section 385,  instructing the Treasury to issue regulations sorting out the mess.  The Treasury tried for a while, issuing two sets of proposed regulations (if I recall correctly) that it then decided not to finalize.  So delegating / punting the issue to the courts remained the prevailing norm.

            (iii) Economic substance requirement – Case law has long held that tax-motivated transactions will be ineffective if they lack requisite economic substance and/or business purpose.  This is of course a “standard” developed by the courts and for decades left alone by Congress.  When Congress finally codified the economic substance approach, for use in applying penalties, it bent over backwards to make clear that it was not adjusting or revising the prior common law approach in any way, apart from one distinct matter: it provided that taxpayers must meet both the “objective” and “subjective” aspects of the prevailing test, whereas case law had been divided as to whether one needed to satisfy both, or just either one.

2) Is there less tax delegation, and if so why?
(a) Presumably, Congress will only delegate lawmaking authority (or delegate “more” rather than “less” when it realistically has the choice) where the relevant decision-makers conclude that the political benefits of doing so exceed the political costs.

Making a consequential political decision often has both political benefits and costs.  One gets the credit but also the blame; one pleases the winners but also angers the losers.  Given, however, that politicians both often like to exercise power and benefit from doing so (e.g., reputationally and in fundraising), presumably “extra” delegation that was not practically necessary tends to reflect some particular motivation.

Here are 3 representative examples:

            (i) Public interest reasons for delegating: If, say, an independent agency can actually do a demonstrably better job than Congress, and members of Congress believe that on the whole this will be to their political benefit, they may delegate for entirely “good” reasons.  A classic example is delegation of power over monetary policy to the Fed.  (Admittedly, this example risks becoming obsolete in the era of Rand Paul.  I suspect that, if were necessary to pass new legislation retaining the Fed’s current authority over monetary policy, it would fail in the House and be filibustered in the Senate.)

            (ii) Symbolic politics: Here an example is the creation of the Environmental Protection Agency (EPA).  To be clear, I certainly agree that the creation of the EPA was important substantive policymaking, with real effects that I believe were on the whole decidedly good. But to legislators at the time, I would think that it had an important symbolic element: By creating a new agency, one takes the credit for “doing something” about environmental concerns.  But since the EPA is the party actually making decisions that involve tough tradeoffs (or winners versus losers), it gets the blame on the implementation end.

            (iii) Prisoner’s dilemma: Suppose there is broad consensus that we all lose overall from an array of parochial benefits.  But no one wants to give up his or her narrow benefit, unless enough others are doing the same.  Congress’s creation of a military base-closing commission (with Congress to vote the final recommendations up or down, with no amendments permitted) is a classic example.

(b) In practice, agreement to delegation appears to require broad underlying consensus regarding the general policy approach that Congress   However, while broad delegation’s reliance on the existence of underlying consensus appears clear empirically (at least based on casual observation), it is less clear why, as a theoretical matter, this should be so.  One could, for example, imagine political players who strenuously disagree but are mutually uncertain of success handing things over to an arbiter.  But my impression is that this doesn’t generally happen.

(c) The conditions for broad delegation are generally missing in current tax politics.  Politicians appear not to view the political benefit as exceeding the political cost, and an important reason is the lack of underlying consensus.  With respect to particular examples:

            (i) The Fed and tax rates – Even when there was greater consensus than we have today regarding countercyclical fiscal and monetary policy, why would Congress hand over tax rate authority to the Fed?  Isn’t it more fun to get the credit oneself for cutting taxes during recessions?  And while Congress could reasonably have anticipated that it would be less eager to clamp down when there was inflation risk, at least the Fed was still there as a backstop via monetary policy.

            (ii) Design of tax preferences – Here is where one could most readily imagine it happening.  Indeed, there are a couple of small areas in the law (e.g., with regard to low-income housing credits and state agencies) where delegation has indeed expanded a bit.

            (iii) Repealing tax preferences to fund a rate cut or reduce the fiscal gap – This would be just like military base-closing if there were sufficient underlying consensus, along with willingness to lose one’s own favored items so long as others shared in the haircut.  But there simply is no such underlying consensus. What is more, many of the underlying items are just too big for their proponents to be philosophical about losing them even if lots of other stuff is nicked as well.

3) Would more tax delegation be beneficial?
(a) Analogy between delegation to an independent agency and the existence of nonprofit firms There is a rich law and economics literature concerning when and where we observe the use of nonprofit firms – for example, in charity work, higher education, and the fine arts.  The dominant account, for example from Henry Hansmann’s work, emphasizes that it is a response to the concern (by consumers and/or donors) that here, unlike in many other areas, the standard profit motive would be over-powered and partly misdirected.

Hansmann emphasizes, for example, the role of ambiguous outputs (e.g., higher educational quality, which may be hard to observe, apparently leading to the surmise that tamping down the profit motive may lead to better results.  In a very small piece that I once wrote in this area, I made the additional point that the substitute motivation is somewhat of a black box, which we fill in based on our empirical beliefs concerning the motivations and utility functions of likely nonprofit actors.  This may be why, for example, we don’t observe nonprofit auto repair shops.  Their being nonprofit might mitigate concern that they were using asymmetric information to fool us into overpaying for stuff we don’t need, but there is no theory of altruistically minded nonprofit auto repair workers who would love and respect the process as much as we academics try to respect effective teaching and good scholarship.

Substitute direct political accountability for the profit motive, and independent experts or professionals for nonprofit actors in a charity, and you get a parallel scenario analytically.

(b) Analogy to the Financial Accounting Standards Board (FASB) and the specification of generally accepted accounting principles (GAAP) – Suppose Congress instructed the Treasury to define taxable income, basing it on economic income but subject to reasonable administrative concerns, including use of the realization requirement where Treasury thought it appropriate.  Full stop, subject only to whatever provision Congress separately made for the design and use of tax preferences.

Would this be a good thing?  I am inclined to think so if the Treasury in turn delegated the task to an insulated professional agency.  But admittedly this is just a surmise – if Congress did this (which it won’t), then we would find out how well or poorly it worked.

One reason for thinking that it might be a good thing (leaving aside that it will never happen) is the analogy to FASB and GAAP.  FASB is a somewhat independent and politically insulated agency under the itself-somewhat-independent Securities and Exchange Commission, and it uses its GAAP rules to define financial accounting income.

Most accountants whom I know – and there also are arguably supportive empirical studies – believe that GAAP, financial reporting, and the quality of the information provided to investors by reported earnings would be a lot worse if Congress meddled a lot more than it does in the process.  On the few occasions when Congress has intervened (e.g., when Senator Lieberman, in the late 1990s, successfully bullied them into dropping plans for treating executive stock options as deductible), there is a view that generally or typically it has made things worse.

This naturally inclines one to speculate that similar delegation with respect to defining taxable income would likely be a good thing, if it was done right.  But obviously we don’t know for sure, and an experiment permitting us to find out remains highly unlikely.  I could only see it happening if the U.S. federal income tax became far less important and prominent than it is today (e.g., as tariffs were once a central concern of Congressional policymaking and then ceased to be such).

No comments: