Yesterday at the Tax Policy Colloquium, Heather Field presented a draft of her paper, Tax Elections and Federal-State Conformity. The paper extends her earlier work regarding explicit tax elections in U.S. federal income tax law, by asking to what extent states with income taxes should seek to bind taxpayers to the elections that they have made for federal income tax purposes. The elections at issue might cover anything from married couples' choice between joint returns and married-but-separate filing, to claiming the standard deduction versus itemized deductions, to corporations' electing under Code section 338 to have a stock purchase of another company treated as if it were an asset purchase.
Once again, rather than writing a fresh comment on the issues raised by the paper, why don't I offer here an expanded version of the outline that I prepared to help guide discussion at the session:
2 topics: (1) What is electivity & why does it matter, (2) federal-state conformity, in light of broader issues of fiscal federalism.
1. What is electivity & what does it matter?
Paper rightly recognizes that you can have electivity without explicit elections.
E.g., realization of loss versus gain assets with sufficiently cheap financial engineering, no constructive sale rules, no loss nonrecognition rules.
I’d add: you can also have explicit elections that really aren’t what we have in mind.
E.g., consider standard vs. itemized deduction. Arithmetically equivalent to everyone gets the standard deduction, itemized only allowed above a floor that equals the standard deduction.
What do we really have in mind here? Call it taxpayer self-sorting. You choose between Regime A and B, want to minimize the sum of tax liability & deadweight loss (DWL).
Revised itemized deduction with floor still has an element of TP self-sorting if the question is whether to try to get above the floor. E.g., don’t bother with record-keeping or base charitable decisions on expecting to get the subsidy.
How should we analyze the merits of TP self-sorting? While taxpayers presumably aim to minimize the sum of taxes paid & DWL, suppose for simplicity it was JUST one or the other (though in practice, think in terms of ratio).
We’re glad if they self-sort to minimize DWL. But if to minimize tax liability, that’s often bad. In principle, a right amount to pay given relevant attributes (e.g., ability to pay).
Nonetheless, in some cases we might be OK with electing into lower tax liability.
--Federal married filing separately in practice (but a harder question if could elect single filing).
--§338 elections, given lack of a good rationale for treating stock & asset sales differently.
--Domestic check-the-box – Note lack of any evident purpose or rationale. But this reflects other rules in the area. E.g., C corporation status is unavoidable if shares are publicly traded, prior multi-factor test couldn’t impede individuals from using limited partnerships in tax shelters since the test had previously been rigged by the Treasury the other way (to block the self-employed from claiming self-paid fringe benefits), other means (such as the passive loss rules) are used instead to impede tax sheltering.
--International check-the-box – rightly controversial due to large tax stakes. But is it actually a bad thing? Distinguish U.S. MNE avoiding U.S. tax vs. German tax.
SO: ability to elect is good if mainly about DWL or higher taxes that we don’t like, much more questionable if avoiding taxes we might want to impose.
This is the same analysis that we should use for cheaper versus costlier electivity. E.g., how much economic substance, in terms of last week’s paper how hard we fight against regulatory arbitrage. And the fact that there’s so much of it reflects how much arbitrary line-drawing we have in our system.
E.g., in a pure Haig-Simons income tax, with no admin or compliance costs, household issues, etc., the only “elections” are how much to work & save, & how much to invest. Can only reduce income through a “costly” election (such as working less), though how costly determines optimal rates.
But in our actual system we have realization, the cubbyholes for different financial instruments, etc. – so costly vs. cheap electivity issues are pervasive.
Often don’t want electivity to be too cheap (e.g., the Miller equilibrium & debt; realization electivity and strategic trading). This is why, for lots of things, we wouldn’t think explicit elections make sense – their key feature may be cost of zero (apart from the need to decide and keep records).
But by even raising the question of whether state elections should have to follow federal, the paper demonstrates that, just because we have an explicit election, that doesn’t necessarily mean that zero is the right exercise price. One key argument for tying elections together, or requiring that they be consistent, is to make them costlier than otherwise.
The paper discusses cross-border tax arbitrage, which is semantic arbitrage or inconsistent effective electivity. E.g., dual resident companies.
U.S. rule against it in effect says, you can’t deduct affiliate’s losses in the U.S. unless you aren’t suspected of deducting it somewhere else.
Likewise, consider tax vs. accounting semantic arbitrage (e.g., the income measure or debt vs. equity). Separate systems, but may want to require consistency simply to make electivity costlier (my argument on the subject).
4 main propositions before we get to distinctive federal vs. state issues:
(a) Is the right category TP self-sorting, as per the point that an explicit election may be thought of otherwise. (Universal, even under H-S though not lump sum ability tax)?
(b) Is there indeed nothing special about explicit electivity other than its being at the low end, & potentially easier to avoid (by just repealing the election)?
(c) Is tax vs. DWL, & what we think about the tax, the right organizing principle?
(d) Is requiring consistent elections mainly about making elections costlier?
2. Federal-state conformity
As when considering federal-state conformity generally (rather than just re. elections), it’s worth considering 2 perspectives here: state residents acting unilaterally, & all states’ residents acting collectively.
State residents acting unilaterally – I want to distinguish this from “sovereignty” as sometimes used. Expressing political preferences (e.g., how high should taxes be, how allocate among residents) has value. But keep in mind the agency problem with state political actors.
E.g., state officials may value their ability to extract rents from lobbyists.
This makes me want to second-guess the value of state-level discretion, based on voter salience. E.g., cost recovery for particular assets vs. rules for allocating tax burdens between different kinds of households.
Paper notes a dilemma that state political actors face with piggybacking & costly adjustment. Congress imposes revenue externalities that may lead voters to misallocate responsibility. But properly assigning responsibility is already a huge mess, so unclear how much mileage we get from this.
Value of simplification vs. complexity: If just borne by voters, fine to have the state decide for itself.
Residents of all states acting collectively – Here we add externalities that all should in principle agree to eliminate. E.g., harmful state tax competition (incentives for football stadiums & factories, rather than taxes that are sufficiently linked to benefits that there is Tiebout competition).
Consider also complexity externalities. E.g., suppose a state with market power (such as California) had its own debt vs. equity rule, the complexity from which would be borne partly by out-of-staters.
Also tax exportation, such as the NYC hotel tax. One could imagine this being done through the income tax instead, & blocked by a federally imposed conformity requirement (or by a bribe, a la highway funds).
OK, let’s try some examples.
Standard vs. itemized deduction: Once we see this as a universal standard deduction that’s part of the federal rate structure, plus an itemized deduction floor, analysis is simpler. No reason for states to require consistency, apart from their own (internalized) interest in being able to free-ride on federal enforcement for itemized deductions claimed federally.
Joint versus separate marital filing: If state wants a different policy, this is salient to voters. No reason to mind joint filing even if federally separate (including, e.g., same-sex spouses). For separate when federal filing is joint, the only issue is the ability to free-ride on federal-level divisions of income.
Other: For a wide range of tax base things (338 election, sub S election, definition of debt & equity, business cost recovery rules, etc.), there will often be a reason for requiring or rewarding uniformity. E.g., if we think it’s state-level political agency costs & rent-seeking, or bad tax competition (e.g., lower tax rates for sports teams or companies with mobile capital), or tax exportation (e.g., hotel tax rules, even if not in the income tax), there may be good reason for Congress to do more to block states’ distinctive rules than it does under current law.