Tax article roundup

This post covers four articles:

• Shaviro, Bittker’s Pendulum and the Taxation of Multinationals Tax Notes Federal (2021)
• Hemel, Holtzblatt, and Rosenthal, The Tax Gap’s Many Shades of Gray
• Doran, The Great American Retirement Fraud
• Gans, Kaestner Fails: The Way Forward, 11 Wm & Mary Bus. L. Rev. 651 (2020)

CIT and the Pendulum

Dan Shaviro, a NYU law professor, is one of my favorite authors, well of the complicated tax article genre anyway. A bonus is that he writes creatively entertaining stuff for tax geeks, like blog posts suggesting Mirrlees optimal tax theory as a solution to MLB’s tanking problem, and short accessible pieces explaining why a problem TCJA purported to address (high corporate tax rates suppressing investment) probably wasn’t much of a problem. His Tax Notes article, Bittker’s Pendulum and the Taxation of Multinationals is a fascinating read for someone like me who has followed corporate income tax (CIT) policy for a long time. The article synthesizes and explains several divergent strands related to the swings in conventional thinking about the CIT (at least as it applies to multinationals) over the last 60+ years, including:

• Changes in the legal and economics academic consensus on its incidence, effects on economic efficiency, and so forth (Shaviro notes similar trends occurred in the thinking about antitrust, the minimum wage, and similar)
• International tax structure and policy – source versus residence, definitions of corporate residence, etc.
• Thinking about optimal tax base (e.g., depreciation versus expensing, effects of taxation based on corporate residence, etc.)
• How proposals, politics (to a limited extent), and enacted policies follow those changes in thinking.

I read the article in November 2021 when it came out. Rereading it now confirmed how much I liked it. It summarizes and explains national trends that as a casual observer I had noticed and only partially understood. He weaves it all together, successfully explaining the how and why of the changing views, both in the legal and economics academe. (One interesting nugget is that both of Shaviro’s parents were PhD economists probably partially explaining his facility with economic analysis and communiating about it.)

His jumping off point, as is apparent from the article’s title, is a 1979 Boris Bittker article that contrasted the views of “old turks” and “young fogies.” It’s a story of how conventional academic CIT wisdom has oscillated between love and hate and back again, partially anyway. Or to put it some less colorfully, from in-fashion (progressive with a low efficiency penalty) and out-of-fashion (regressive and inefficient). In my grossly oversimplified and probably somewhat misleading summary (you should read the article if you’re interested), there are three eras:

1. The era of the Old Turks (e.g., New Deal and later economists; Harberger’s initial incidence analysis, Stan Surrey and similar legal academics) – considered the tax to largely fall on capital with modest negative efficiency effects
2. Rise of the Young Fogies – (e.g., Marty Feldstein, Harberger’s revised incidence, etc.) – assume simple Econ 101 principles govern, pricing taxes into capital assets, globalization of capital markets shifting the incidence to consumption and labor, larger efficiency penalties, suggesting an optimal rate may be zero (“End of History” as Shaviro puts it or that the tax should be integrated with the individual income tax), etc.
3. Current thinking – focusing the tax on excess returns (i.e., above the normal return on capital) by allowing expensing can rescue the tax as a relatively efficient tax on capital.

The article points out how the administration’s proposal is supported by much of current academic thinking about the CIT and nature of capital markets and multinational businesses. His discussion of the implications for international tax rules and the application to TCJA’s changes and Biden’s proposals are particularly interesting.

SALT angle

On first blush, I assumed the article provided little insight for state CIT policy and I wasn’t going to bother to try to memorialize my thinking as a result. (Since states impose source-based CITs using formula apportionment, the discussions of residence-based federal taxation in the international context are clearly irrelevant.) On reflection, I began to wonder if that was correct. That states are small open economies with easy interstate factor mobility has always been a given (unlike the national/international context). That leads to an assumption that any effort to tax capital (other than geographically fixed capital like land and preexisting improvements) is self-defeating. In the long run, capital flows will shift the burden to consumption or labor as they seek states yielding the highest after-tax returns (all else equal, which it is not of course true often, and short run effects are important). 

The swing of the pendulum to what Shaviro calls the “End of History” was based on similar thinking. But more nuanced and sophisticated economic thinking calls that into question, especially for a tax focused on excess returns and firms less bound by locational capital (e.g., companies with much of their income derived from IP). This probably merits more thinking about the incidence of alternative state CIT structures and assessment of how much ability state have to tax capital with CITs.  

Given national pricing by some firms (think Apple or Microsoft) and 100% sales apportionment, I have often wondered about the validity of the assumption that the tax falls on consumers (per Charlie McLure’s classic view that it is a tax on the relevant apportionment factor(s) and most states use 100% sales apportionment now). Assuming they have nexus and rigid national pricing policies, it’s hard to see how the tax is shifted forward to consumers. Prices of iPhones or Windows software do not vary from state to state, although there are more subtle ways to vary real prices. Since payroll and property are not factors, they should be indifferent as to where they locate those. This crucially assumes that they cannot avoid nexus, of course. Whether firms like Meta and Alphabet, which theoretically can adjust their opaque pricing to reflect state CITs, do so is a good question. Moreover, if they are setting monopolist style prices nationally, further adjustments to shift state CITs may be self-defeating? That assumes they already capture every dollar of profit possible and further price increases would just depress net revenues. I hope someone with well more than my expertise and brainpower is at work on those questions. But because SALT issues are something of a backwater for academic economists, I do not hold out much hope.

50 Shades of (tax gap) Gray

Anyone with visions of easy money from closing the tax gap (e.g., by more IRS funding), whether for deficit reduction or to fund new initiatives should read Hemel, Holtzblatt, and Rosenthal, The Tax Gap’s Many Shades of Gray (Sept. 30, 2021). I have posted multiple times on this topic, starting with Summers and Sarin early op-eds, and about the need for adequate IRS funding to the IRS. But I have been skeptical about the potential bonanza of revenue that Summers and Sarin claim would result. This article provides context for those claims and debunks proposals that I did not realize were out there (i.e., the ill-advised idea to mandate IRS close the tax gap by a specified percentage that Elizabeth Warren and Ro Khanna propose). As its title suggests, the article’s main point is that ambiguity clouds measurement of the tax gap, especially substantive tax law ambiguity and complexity, and will thwart efforts to close it solely with compliance measures, such as increased IRS funding or better information reporting. Tax law clarification is also needed.

Holtzblatt and Rosenthal have a short TPC blog post that summarizes the article’s main points. Unfortunately, it does not do it justice. Anyone interested in issue should at least skim the full article.

The article:

• Provides a primer on how the IRS calculates the tax gap. The article has a clear and succinct (well, 7 pages) description of the IRS methodology. (I may not be the best judge of this, since I have read many explanations over the years and already knew most of this. But it seemed to me one of the best descriptions I have read.) A crucial point here is that the baseline measure of the gap is IRS examiners’ conclusions as to the tax owed (subject to statistical adjustments), less what is actually paid. Obviously, the IRS’s initial audit position or notice of deficiency (NOD) is not the actual tax owed; IRS Appeals or a court may reduce it. The IRS position may understate liability too. The article makes it clear that the tax gap does not equal or measure evasion (intentional failure to pay), which some including administration spokespeople have said. The authors show why the tax law’s complexity (hence the shades of gray) makes the measure inevitably imprecise. The PTE world, as anyone who has even casually delved into it knows, is a swamp of complexity that invites aggressive positions and eludes defining and measuring noncompliance, only in part because of the lack of audits and measurement under the NRP used to calculate the tax gap. Nobody knows precisely what the law requires. Valuation of assets, fuzzy rules on who is a contractor or an employee, and the distinction between business or personal expenses are a few more gray areas. More than 50, certainly.
• Explains why the tax gap is a flawed measure of tax owed but not paid (if the standard is a utopia of omniscient understanding of the law and full compliance by all). It’s both too high because of its baseline (mainly reliance on IRS examiners’ assessments) and too low because the calculations ignore income from illegal activity and, as independent research has shown, likely understate PTE and offshore income. As the authors say, “knowing the exact size of the tax gap—an inherently subjective quantity—is less important than understanding what the tax gap does and does not represent.” (p. 25) My personal guess is that gap is likely larger than the estimates, but that’s irrelevant. The reality is that it is massive. The only question is whether it’s a blue whale or just a gray whale.  The real issue is how much could reasonably be collected and what measures Congress and the IRS should adopt to do about it. The authors address that in the last dozen pages of the article.
• Discusses the policy implications. I think you can boil the authors’ policy advice down to one “don’t” and two “dos” – (1) Don’t use the tax gap as a budget target, as some bills have done. That seems so obvious that it is distressing that members of Congress are proposing it. (2) Adequately fund the IRS. Even more obvious. (3) Reduce ambiguity in the substantive tax law. This is the real message of the article (hence, its title). The authors give a thumbnailing listing some of the key areas, including valuation of illiquid assets, partnership special allocations, independent contractor rules, etc.  I will say that this is counter to the popular narrative to fix the problem with just IRS funding (e.g., that’s the core message of this NPR interview of Deputy Treasury Department Secretary Wally Adeyemo). The two “dos” are interrelated. As the authors say, without reining in the ambiguity:

The IRS—however well-resourced it might be—will struggle to fight back. For that reason, investments in tax enforcement must be complements to, not substitutes for, legal reform.

The Tax Gap’s Many Shades of Gray, p. 30.

Unfortunately, Congress has not only refused previous entreaties to do that but instead merrily goes along enacting provisions that compound the problem. The recent poster child is TCJA’s baking QBI’s mindlessly complexity into the law unleavened by any credible policy rationale.

SALT Angle

All of the authors’ recommendations will automatically redound to the benefit of states with income and corporate taxes by enhancing and protecting their tax bases. That would not apply to some tax expenditures, like QBI which almost all states have rejected and exclusively federal credits. But, in general, adequate long term funding of the IRS and clarification of ambiguity would strengthen the independence of states and further the goals of federalism. My second wish is that Congress would assess the impact on state tax systems when it considers federal tax changes, which provides a segue into the next article.

The Wolves of K Street

Michael Doran, The Great American Retirement Fraud, is a good read for someone (like me) who’s interested in federal retirement policy. Doran is a University of Virginia law professor and former Treasury staffer. I wish I had been able to read his article before I wrote Roth aversion, my ponderous post on why I don’t like the Roth structure. Doran’s paper compelled me to correct my post regarding the paternity of the Roth structure.

Doran provides a chapter-and-version account of what I alluded to in my post: i.e., a pattern of congressional changes that appear heavily driven by the financial services industry, large employers, and the plan industry. The changes made it possible to contribute ever larger amounts to retirement plans and are a factor in the mega-IRA phenomenon, the subject of media stories (such as ProPublica’s on Peter Thiel’s Roth IRA), congressional hearings, and eyepopping numbers from JCT. Undervaluation of in-kind contributions, as Doran suggests (fn. 4), is undoubtedly a factor but less nefariously, if one makes maximum contributions and earns market returns very large accounts can be amassed (see Hemel and Rosenthal testimony for the congressional hearing). The Hemel-Rosenthal example of garden variety mega-IRAs illustrate just how unnecessary the changes Doran describes were.

Doran’s article consists of three parts:

1. A history and general description of the American private retirement system, including what he calls “The Retirement-Reform Project” – this is an excellent introduction to how this complicated system evolved and works (pp. 8 to 39).
2. How the post-1995 changes deviated from prior policies and practices and why he considers them fundamentally flawed (a “bipartisan sellout”) – he describes the myriad of changes made to the retirement plan rules by Congress in the last 25 or so years (pp. 39 – 62).
3. An account of some effects the changes have had on participation, retirement account balances and similar (pp. 62 – 82).

The most interesting part of the article is its description of the post-1995 changes and how they were sold. It’s a story of Senators Rob Portman (R-Ohio) and Ben Cardin (D-Maryland) selling the changes as good policy, while gaining bipartisan support for them. Doran disagrees and makes the case that they failed to achieve their primary objective, materially increasing retirement savings by the people with average incomes, while nevertheless costing a lot. (His tone is remarkably direct in criticizing members of congress for a law review article. I do not doubt that they deserve it. An interesting question is why this happened without Democrats resisting much. I speculate about that below.) Both of them started in the House and moved to the Senate, no doubt aided in part by the reputation they gained as bipartisan dealmakers.

Just like the byzantine structure of the American retirement system, it’s a complicated story that resists a succinct or simple summary. The following is a major oversimplification. Read the article for the full account.

The main policy problem Portman and Cardin (“P&C”) purported to address was inadequate private retirement savings to supplement social security benefits. (Note, as Doran points out, this is not a problem at the top of the income distribution. Almost all those folks save no matter what and the public does not need to worry about their retirement security, in any case.) The problem occurs in the context of a retirement system that almost wholly depends on voluntary saving by individuals (beyond social security), as a result of the 401(k)-revolution that began in the early 1980s.

The 401(k)-revolution dropped old fashioned defined benefit (DB) pensions in favor of defined contribution (DC) plans. DC plans require individuals to fund (potentially with help from employers) and manage their own retirement savings accounts. Businesses like 401(k)s because the structure relieves them of responsibility for investing and ensuring the adequacy of the funding for their employees’ retirements. But it requires voluntary private savings. To overcome the natural human reaction against deferring gratification (saving), the system relies on economic incentives, sometimes once or twice removed (if business owners want to get tax benefits, they need to figure out how to get their employees to save, for example). Rube Goldberg would have liked the system.

To make it more explicit (read Doran for the full-bore explanation; he does a good job), the system tries to get people to save in two ways:

• Providing tax incentives for retirement saving. This is done by exempting investment returns on the DC retirement accounts from income taxation. The progressive income tax rates inherently skew that incentive to high income folks. Recall that almost half of the population does not pay income tax. So, this incentive has little application to that group. You need to move to the world of refundable credits to incent them.
• Encouraging employers to offer plans to their employees and to adopt measures that get their employees to contribute. This is done by limiting business owners’ and their highly paid employees’ access to the tax benefits unless regular employees also benefit to some degree. There are two components: (1) Restricting the tax benefits of non-employer sponsored plans (IRAs), so high-income business owners and managers do not just use IRAs and forgo offering employee plans. (2) Imposing regulatory rules on employer plans (nondiscrimination rules) that ensure minimum participation by lower-paid employees. The nondiscrimination rules are very complex, especially since employers can offer both DB and DC plans and they have an incentive to game the rules to benefit themselves and their top employees.

Up until the 1990s, Congress emphasized keeping a lid on tax benefits for budget and fairness reasons by limiting tax-deductible contributions to IRAs and employer plans and maintaining tight nondiscrimination rules for employer plans. That worked when employers provided traditional DB plans that automatically covered many employees. It did not work well after the 401(k)-revolution shifted the onus to individual employee saving. Evidence showed inadequate retirement savings; again, the main policy problem P&C were addressing.

This is a challenging problem to fix if your only options are incentives (all carrots, no sticks). I assume that was a given for P&C since Republicans controlled one or more houses of Congress or the presidency during this period and their governing philosophy is anti-mandate to put it mildly. There are three main policy levers and none of them work well or at all.

1. More tax benefits for employer plans – allowing larger contributions does not work for most employees. They were already not contributing the maximum amounts, so increasing the maximums is irrelevant. A high-rate credit might have helped. It’s possible that more generous benefits would induce more employers to offer plans. But the impact would be limited because many employees with access to plans do not participate or do so at low levels. Of course, the cost is high because all employers that already offer plans benefit needlessly. Giving one-time tax benefits to firms without plans to establish them would have been the best policy and some of that was done. But the main approach was to make the benefits permanently richer for all plans. Expensive and ineffective.
2. Tax benefits for individual plans (i.e., IRAs) – you could argue that average- and lower-income self-employed and employees of firms without plans may participate if the tax benefits are richer. But most of these folks were not contributing at the lower levels allowed before the P&C started on their efforts, so that seems unlikely. And making them generous will discourage firm owners from establishing plans (they can just contribute to their own IRAs and forgo forming plans for their employees), cutting against lever #1. Making these benefits richer is expensive and ineffective but has the patina of fairness to the uninformed because it can be spun as helping those without access to employer plans. Doran points out how even supposedly sophisticated observers (writers of law review articles!) fell for this. I notice it regularly in the comment sections of online articles on how to fix retirement policy on sites like Morningstar.
3. Nondiscrimination rules that encourage wider distribution of benefits to all employees – the Republican’s anti-mandate ethos prevented tightening them. But employers and financial service industry argue that the complexity and cost of the rules discouraged employers from setting up plans. So, loosening them could expand access by employees to plans. Again, that is an instance where the potential effects are ambiguous, because loosening them up will likely reduce participation by employees with current access (some employers will dilute the efforts they use to increase participation to meet looser nondiscrimination tests) offsetting the effect of encouraging employers to establish plans.

So, what policy solution did P&C (or the trade organizations and their lobbyists who were the likely real movers involved) come up with? Starting in 1996, as Doran puts it:

“Congress raised or repealed every limit on the tax incentives for retirement savings.”

The Great American Retirement Fraud, p. 11

More specifically Congress in bills sponsored by P&C (see p. 31ff):

• Dramatically increased the amount that could be contributed both to employer and non-employer plans (IRAs), including authorizing Roths, which is a de facto increase in allowed contributions (but much worse than that in my view).
• Allowed the money to remain in the plans longer – i.e., loosened the required minimum distribution (RMD) rules. As I have blogged about before, this is totally unnecessary as a retirement policy; it’s an inheritance enhancement measure. What’s worse is that the current Congress (controlled by Dems – you can’t blame the GOP for this) appears poised to further dilute RMDs. This is part of package with some good features, like an auto-enrollment mandate (but see here suggesting small effects), but the bad highly likely outweighs the good.
• Weakened the nondiscrimination rules, adopting a trade-off that favored encouraging employers to establish plans by making it simpler and easier to favor owners and highly compensated employees, rather than measures that distributed benefits of the plans more broadly to all employees.

Any disinterested party would not expect these changes (or maybe any set of changes relying on tax incentives and eschewing mandates) to have much effect on retirement savings by workers outside of the top income quintile. And any effect would come at a high cost. Top income people were already saving adequately but much/most of the tax benefit would inevitably go to them.

I can imagine well-designed tax credits, matching grants to employers or other mechanisms that would have minimized those costs somewhat and might have been modestly more effective at much lower cost. But the core problem is the proverbial paradigm of pushing on a string. It is extraordinarily difficult to increase savings by a population that is barely able to pay day-to-day expenses living in a consumption-obsessed society and your only tool is to provide financial incentives.

The third part of the article details some effects of these changes. As one would expect, the changes dramatically increased the tax expenditures for retirement, making it the costliest category (about $380 billion/year) and equal to 40 percent of the outlays for Social Security. The benefits are skewed to the top of the income distribution.

The goal was to increase retirement savings. Did they do that? Yes, but primarily for the upper income strata. As Doran points out, proponents cite how much the averages increased. But averages are skewed by the large amounts at the high end; median participation rates and amounts increased but by much less. The bottom half of the distribution changed little, as one would naturally expect. They don’t pay much income tax, why would one expect them to respond to income tax incentives?

Although I agree with Doran’s conclusions, participation and account values did increase materially for all income levels. (Even the bottom half of the distribution shows increases in his tables.) For those in the 60 – 89 percentiles, participation rates increased by 6 to 7 percentage points and account values (in inflation-adjusted terms) more than doubled (much of that is simply attributable to the stock market, of course). See his Tables 4 and 5 (pp. 65, 67). Those are meaningful improvements. The real point is the changes were both immensely expensive and undermined progressivity, while failing to provide retirement security for the target population. Given the policy goal, a failed effort is the only fair characterization and about what one should expect given the policy instruments that they used.

The changes in participation and account amounts, of course, do not tell the real story – i.e., whether the more generous rules induced or caused the increases. Insight into that requires sophisticated statistical analysis that controls for other relevant factors. The obvious point is that many people simply shift existing assets into retirement accounts. Economists who have analyzed the issue, it is worth noting, do not agree as to whether retirement tax incentives increase savings much. To the extent they do, the effect is mainly at the top. Doran’s real point is clearly valid – any change is very expensive and tilted toward the high end, failing to help those who actually need it.

The misleading way P&C sold the changes is especially irritating. As Doran describes it:

Portman and Cardin showed that favorable perceptions about legislation framed as expanding retirement-plan coverage and enhancing retirement security insulated them from criticism for rewarding well-heeled supplicants. In almost any other context, increasing tax subsidies for affluent individuals would draw condemnation, possibly from the center and certainly from the left. Whatever the merits of the various economic justifications that might be offered, proposals to reduce tax rates on long-term capital gains, increase exemptions for the estate tax, and reduce the top marginal tax rates on ordinary income are normally highly divisive. But the retirement-reform project has escaped such controversy. Portman and Cardin studiously positioned themselves as promoting private retirement savings, a policy goal that sounds as though it does as much for lower-income and middle-income earners as it does for higher-income earners. And remarkably, the more the growth in the retirement savings of higher-income earners has outstripped the growth in the retirement savings of lower-income and middle-income workers over the past twenty-five years, the more attractive the retirement reform project seems to become, not only to Portman and Cardin but to other legislators as well.

The Great American Retirement Fraud, p.53

Doran is the boy in the fable, The Retirement Reform’s New Clothes, who points out the King’s nudity. He plans to write an article on what he would recommend doing instead. I will be interested to read it. After spending my career working for legislators, I suspect it may not pass the all-important political feasibility test – especially given the status quo (thanks to P&C), which gives entrenched interests (the financial services industry and affluent households) the relatively easy task of playing defense in a gridlocked Congress. For example, waiting six years to curtail mega-IRAs, a more politically compelling problem, in the House’s BBB bill is revealing – even a Democratic Congress is reluctant to end retirement tax benefits for the undeserving who they claim to want to raise taxes on. Of course, worrying about political feasibility is not in an academic’s job description.

My observations

Political reality imposes limits. In a world where the policy levers are limited to tax incentives, good solutions are scarce. Moving the needle on retirement savings will be expensive and the results underwhelming. This is a weak defense of Cardin. Doing nothing would have been better but legislators are innately oriented to doing something and there is no denial there is a problem to address (still is). Given these limitations, Cardin probably felt he was doing the best he could. That’s wrong but hard to see in the moment.

Why not more opposition from liberal Dems? A puzzle, of course, is why many liberal Dems agreed to this. The Clinton administration got the ball rolling and/or acquiesced. One possibility lies in the support in the economics profession (including many liberal economists) for replacing the income tax with a consumption tax. There is little doubt that economic advisors carry a lot of clout in sophisticated national policy circles (like the Clinton White House). That preference for consumption taxation creates potential policy ambivalence about expanded retirement savings provisions because they can be viewed as a transition to a consumption tax. One structure would allow unlimited contributions to tax favored accounts and impose tax when those amounts are withdrawn and spent/consumed. The Nunn-Domenici USA Tax Proposal in the 1990s (contemporaneous with P&C actions) was along those lines. Given that frame of reference, P&C’s changes may not have looked so bad from a policy perspective to more centrist Dems who were listening to economists. Just a thought. Another possibility is that they were considered tradeoffs to get other stuff the Dems wanted from Gingrich’s Congress. I don’t think there is any evidence for that.

RMD changes were gratuitously bad. Even in my most charitable moments, it is hard for me to see the justification for changes allowing more money to stay in qualified plans for longer periods. There wasn’t an policy problem here to addres. The latest increases in the RMD age, initially to 72 and now (based on legislation that passed the House and will probably be enacted) to 75, is apparently based on the rising life expectancy trend. Yes, life expectancies have been increasing (well, COVID has reversed the trend, temporarily one hopes), but that is merely an abstract point. A policy problem would be if the RMD rules were compelling individuals who depend on the accounts for living expenses to prematurely empty them. There is no evidence of that; the real problem is the opposite. Life circumstances require many to empty their accounts too fast. Crafting narrow rules (to address close to an almost nonexistent problem) would have been easy – e.g., increasing the starting age for those with low balances (< $500K, say) or who are still working and have modest balances. Loosening up RMDs is an inheritance enhancement device, pure and simple. Same is true for the catchup provisions. It is revealing to me that when confronted with the problem of mega-IRAs, a Democratic Congress put off imposing RMDs on them until 2029 (based on revenue provisions in the Build Back Better that passed the House – see JCT estimate p. 6). Given the quest for revenues and their “tax the rich” rhetoric, imposing those rules in 2023 would seem to be no-brainer. Typically, delayed effective dates are an invitation for later Congresses to further delay and/or totally repeal the provision (see ACA revenue provisions as an example).

SALT angle

The article is about federal tax policy, but the effects flow down to state income taxes. Because it is not practical for state income taxes to deviate much, if at all, from federal retirement provisions, the changes Doran describes had nasty long-term effects on state income taxes. The immediate effects are obvious, moving current compensation into retirement accounts. But most states tax retirement income much more lightly than wages or regular investment income. Iowa’s recently tax cut exempting ALL retirement income is the latest extreme example. So aside from the benefits of deferral, converting more wage compensation to retirement benefits implicitly reduces most state tax bases. A second effect is to pass more of that compensation down to heirs, which will tend to shift the jurisdiction to tax it away from the state where it was earned. Heirs often live in other states than their benefactors. Under federal law, retirement account distributions are only taxable in the recipient’s state of residence.

Correction

Doran’s article compelled me to make a correction in my Roth aversion post on the origins of the Roth structure. I mistakenly assumed it originated with the late Senator William Roth, because of the name. (My statement: “Politically, there was probably no way around it for Clinton and the Dems to agree to adding Roth IRAs with the chair of the GOP Finance Committee, the guy the accounts are named after (William Roth), likely insisting on an eponymous legacy.”) Doran’s account reveals I was wrong:

Senator William Roth of Delaware put his weight as chair of the Senate Finance Committee behind what was then known as the “backloaded IRA,” a policy proposal of the Clinton Administration. After immodestly renaming the backloaded IRA after himself, Roth included it in the Taxpayer Relief Act of 1997.

The Great American Retirement Fraud, p. 5

I’m not sure which of the two men’s reputations dropped more in my view (I already held low opinions of both): Clinton by proposing what is a fundamentally bad idea (systematically regressive on a lifetime income basis among multiple other flaws) or Roth who hijacked credit for someone else’s (bad) idea and codified his theft into federal law.

Catch me if you can (you probably can’t)

Article: Mitchell Gans, Kaestner Fails: The Way Forward, 11 Wm & Mary Bus. L. Rev. 651 (2020).

Due process limits states to taxing people, entities, and things with some minimum contact with the state (presumably during the tax period). Trusts are purely legal constructs, and their assets are intangibles or can be easily converted to intangibles by putting ownership in corporations, LLCs, or similar. Thus, neither the trust itself nor its property are subject to physical world limits; they can easily be located anywhere. State jurisdiction, thus, must be derived from the location of one or more of the following people involved with a trust:

• The founder and funder of the trust but by making the trust irrevocable, this link can be time limited.
• Its beneficiaries but contingencies or trustee discretion can easily create ambiguities as to the who, how much, and when of beneficial entitlement, which may be fatal to states asserting contacts based on the residence of beneficiaries for a tax period.
• Its managers and administrators but these can be located anywhere, including in a handful of no-tax and rule-against-perpetuity-free states that are happy to host them without much of a benefit to the host state.

This is a longstanding problem for states heavily relying on income tax revenues. But changes in federal tax rules and aggressive actions by a few states to host trust administration and management have exacerbated the problem. By conveying assets to a qualifying trust domiciled in a no-tax state, a grantor can effectively exempt the investment income from state income tax. To qualify, the trust must be a nongrantor trust and the beneficiaries (technically at least) must not be legally entitled to receive the income. In the past, the federal gift tax limited the attractiveness of this option and still does to a certain extent. But the higher exemptions under EGTRRA and TCJA (> $12m single/$24m married in 2022) reduce that barrier for the modestly affluent. The IRS’s blessing of transfers to certain nongrantor trusts as incomplete gifts for gift tax purposes has a similar effect for the rich.

In 2019 SCOTUS reaffirmed the ongoing vitality of the due process limits in North Carolina v. Kaestner. Some of us thought the Court taking Kaestner signaled an intent to reduce due process limits (like Quill’s holding on that point) and to rely on more flexible Commerce Clause rules. Otherwise, why take a case in which the state court had held its state tax unconstitutional? That proved wrong and the Court reaffirmed its old precedents with little clarification. The Minnesota Supreme Court reached the same result in Fielding, which as Gans points out had somewhat more favorable facts than Kaestner for allowing state taxation. The state petitioned for cert and the Court denied the petition after deciding Kaestner.

Gans is a Hofstra University law professor. As the article’s title implies, it discusses options for states to address the problem after initially parsing the Kaestner opinions. Since Minnesota has not changed its law after losing Fielding, the article is relevant to an issue on the legislature’s plate. Spoiler alert: there is no good solution in my opinion.

Gans discusses three options:

1. Wealth tax – aside from other problems (hello, valuation issues), he concedes that this could be defeated in trust drafting giving the trustee discretion over distributions, negating much or all of the value to the resident beneficiary (part of why North Carolina lost in Kaestner – the resident beneficiary was not entitled legally to the income).
2. Founder’s principle – basing taxing on the residence of individual who created and funded the trust when it became irrevocable. This is the approach Minnesota took and was struck down by Fielding. So, it won’t work here. I doubt it would pass muster with SCOTUS in the unlikely event that it takes up another case. Gans optimistically observes that whether it is a viable solution “remains unclear.” (p. 680).
3. Taxing the grantor – this approach ignores the trust and federal income tax rules (that treat the trust as a nongrantor trust) and makes the grantor taxable on the income. New York adopted this approach in 2014 in response to IRS private letter rulings blessing trust transfers as incomplete gifts for gift tax purposes. I agree with Gans (pp. 684-88) that the issue is how much control the grantor must retain to provide sufficient contacts to satisfy due process. The problem is that this requires a fact-intensive review of the trust terms and potentially the course of conduct under the trust. Messy and impractical is my take, other than maybe for big, sophisticated states like New York and California.

Gans also discusses the solution the Court pointed to and that is a component of the New York law – a throwback rule (pp. 689 -91). This would tax the beneficiary on accumulated income when it is distributed. Because the distribution will not match when the feds tax the income (they will have taxed the trust when the income was received or realized), this is messy (requires multi-year record keeping), has compliance issues (no current year Form 1099s to rely on), and will not always work (e.g., if the beneficiary is no longer a resident when the distribution occurs).

As usual, workable solutions depend upon Congress. A statutory scheme with apportionment or tie breaking rules to prevent double taxation while allocating tax base to an appropriate state or states is what is required.

[As an aside, my theory on what happened with SCOTUS taking Kaestner ties in with the Court realizing the complexity of any workable solution to the problem. As usual, this is wild speculation on my part and likely wrong since I’m a rank amateur Court watcher at best. But my theory is the Court grant cert because four justices perceived there was a real problem (obviously correct) that they thought they could fix similarly to the Wayfair approach. But once the case was briefed, it became apparent that in too many instances there would be opportunity for two or more states to tax the same income in one or more different years. As a result, the only prudent course they could see was to punt and reaffirm their old murky precedents. Gans makes the case, probably correctly, that for the liberal justices reaffirming stare decisis as they girded for the coming assault on Roe was a factor. For them that is much more important than a minor league protection or restoration of state tax bases.]

On the surface, Congress cannot negate the due process clause. That seems like a minor problem to me. Congress could make a state an agent of the federal government when it imposes and collects its a trust income tax satisfying federal statutory parameters. Because the states would be acting on behalf of the national government, due process issues disappear. Better yet, the IRS could collect the taxes on behalf of participating states and remit the revenues to the states in return for states exempting the trusts from their taxes.

An alternative solution like the old federal estate tax credit for state death taxes could be devised. The federal income tax on trusts could be increased slightly and a credit allowed for state taxes (e.g., a flat 4% tax on income over $50,000) with the overall parameters providing federal revenue neutrality. If the experience with the credit for state death taxes is instructive, that would induce all states to impose a trust tax to take up the federal credit and most states would limit their trust taxes to the federal credit amount. The incentive to domicile trusts in states like South Dakota to accumulate income free of state tax would disappear (the incentive provided by no rule against perpetuities would still be a factor). None of this will happen or be considered in the real world. So, it’s not worth spending time sussing out a workable proposal.

To be more specific, the current political environment in Congress is nothing like 1920s America when the credit for state death taxes was enacted. Instead, we have a Congress that repealed the state death tax credit in EGTRRA and capped the SALT income tax deduction in TCJA. And when pressed by private interests. Congress will limit, rather than protect, state tax bases (think Internet Tax Freedom, Pub. L. 80-272, RRRR Act, etc.; see Wildasen, Preemption: Federal Statutory Intervention in State Taxation for a longer list). I’m hard-pressed think of a case when Congress has helped states, although I’m sure there must be one or two exceptions proving the rule. Congress ignored SCOTUS’s implicit invitation in Quill to solve an obvious problem that favored one category of commerce, creating deadweight loss, economic distortion, and state revenue losses.

That leaves states without a good solution. In the mid-1990s, Minnesota adopted a residence trust definition that relied exclusively on Gans’ founder principle to protect in-state financial institutions competing for trust administration and management businesses with out-of-state firms. The changes were thought to have that effect because they made location of trust management and administration irrelevant. (I know that because I was involved in the legislative discussions. The changes were worked out jointly by the administration and the trust management industry.) The definition made the location of beneficiaries irrelevant under the statutory definition. But, of course, DOR was still free to argue their location matters in determining the state’s constitutional authority to tax a trust (negating the superficial simplicity of an exclusive look at the founder). Given the attraction of locating trusts in no-tax states without rules against perpetuities, that is now a pipe dream. Whether it is worthwhile to craft a set of rules that will impose tax on a few trusts based on some sort of multifactor test is unclear to me.

Posted: April 21, 2022