On March 31, a book called The Timing of Lawmaking, edited by Frank Fagan and Saul Levmore, will come out via Edward Elgar Publishing. Further information about the book is available here.
One of the articles in the volume is my piece, The More It Changes, The More It Stays the Same? Automatic Indexing and Current Policy. I posted an earlier draft on SSRN here.
Here is an abstract for the volume as a whole: "Legal reasoning, pronouncements of judgment, the design and implementation of statutes, and even constitution-making and discourse all depend on timing. This compelling study examines the diverse interactions between law and time, and provides important perspectives on how law's architecture can be understood through time. The book revisits older work on legal transitions and breaks new ground on timing rules, especially with respect to how judges, legislators, and regulators use time as a tool when devising new rules. At its core. The Timing of Lawmaking goes directly to the heart of the most basic of legal debates: when should we respect the past, and when should we make a clean break for the future?"
The abstract for my chapter says (or said) that it "addresses issues associated with automatically indexing fiscal policies, such as those in the U.S. income tax and Social Security systems. Under indexing, a statistical measure - pertaining, for example, to inflation, wage levels, life expectancy, or income inequality - is used to determine changes to nominal legal rules that then take effect automatically. One possible reason for favoring automatic indexing is that it may keep the underlying policy, by some metric, 'the same' as empirical circumstances change. While indexing often makes sense, from the standpoint of a policymaker whose long-term preferences it would keep in place barring further legislative action, identifying the set of 'current policies' that one might want to perpetuate (or change) can be surprisingly difficult. The paper explores broader conceptual issues pertaining to policy continuity and competing objectives when legislation remains on the books indefinitely, with particular reference to examples drawn from the history of the U.S. income tax and Social Security."
Contributors to the volume, other than the two editors and me, listed in the order in which their chapters appear, are David Kamin, Daniel Farber, Tom Ginsburg & Eric Alston, Jacob Gersen & Jeannie Suk, Adam Samaha, Martha Nussbaum, Anthony Niblett, and Mark Ramseyer.
Wednesday, March 29, 2017
Tuesday, March 28, 2017
Short article just published
In October 2015, I participated in a conference at Brooklyn Law School, entitled "Reconsidering the Tax Treaty." I wrote a short article for the occasion (about 3,500 words), entitled "The Two Faces of the Single Tax Principle."
A Brooklyn Journal of International Law symposium issue, based upon the conference, is now available online, and my piece is available here.
A Brooklyn Journal of International Law symposium issue, based upon the conference, is now available online, and my piece is available here.
NYU Tax Policy Colloquium, week 9: Len Burman & Kim Clausing, "Is U.S. Corporate Income Double-Taxed?"
Yesterday Len Burman presented the above co-authored piece, which is still an incomplete draft hence not posted online. It confirms, through distinct data, the finding in Steven Rosenthal's and Lydia Austin's widely noticed 2016 study to the effect that these days only about 30 percent of U.S. corporate equity is held in taxable accounts - as opposed to more than 80% fifty years ago. The main reasons for this decline are the rise of (1) foreign ownership of U.S. shares - although this remains a bit low, relative to the scenario in which home equity bias had entirely disappeared, (2) the holding of corporate stock in tax-free savings accounts such as traditional and Roth IRAs and 401(k)s, along with 529s, and (3) stock ownership by nonprofits.
Should we think of stock held in a traditional IRA or 401(k), in which contributions are deducted upfront and withdrawals are taxed at the backend, as actually tax-exempt? The answer is yes, given that the present value of one's tax liability from the account is zero, if we assume both (a) that the account just earns normal returns that the taxpayer can't scale up (a la Mark Zuckerberg hypothetically deciding to create two comparably profitable Facebooks rather than just one), and (b) that the taxpayer's marginal rate is the same at the contribution and withdrawal stages. Given the deduction limits for these types of tax-favored savings accounts, along with the fact that most people are just putting in publicly traded stock as to which they have no reason to expect special-opportunity above-normal returns, I'm fine with classifying them as tax-exempt for purposes of the studies, so long as we keep the nuances in mind.
The most obvious takeaway from the Rosenthal-Austin and Burman-Clausing bottom lines is that "double taxation" is less of an issue than many analyses of corporate income taxation tend to assume. This may reduce the appeal and urgency of treating corporate integration as a top priority, although it doesn't rebut the existence of the distortions that proponents of integration emphasize.
A further set of takeaways that I get pertain to how we should think about the existing two levels of taxation. These studies help remind us that, if one eliminated the entity-level tax and shifted corporate tax collection purely to the shareholder level (e.g., by taxing dividends and capital gains more rigorously), we would eliminate the existing indirect taxation of tax-exempts and foreigners via the entity-level tax. But the arguments for structural reform of the corporate tax, and for shifting to the owner level, are distinct from the issues around how we ought to tax (or not tax) tax-exempts and/or foreigners who or that hold U.S. equity. So those issues ought to be addressed if one is changing the entity-level tax. (For my money, there is no reason to extend the effective reach of tax-exempts' nominal exemption by letting them escape the tax that they are now indirectly paying; for foreigners the key question for me is incidence: to what extent are they actually bearing the tax?)
What about the owner level? If we're relying more on taxes on this level, it's important to consider features such as interest charges for deferral plus realization at death (as in a recent Grubert-Altshuler proposal), or taxing unrealized appreciation for publicly traded stock (as in a recent Toder-Viard proposal).
But also, at the owner level we may be changing somewhat the mix of income that is being taxed. A not very important example, just to make the basic point, is that if a corporation earns municipal bond income and then uses it to pay dividends, there is no tax at the entity level but there is a taxable dividend. Likewise, there's capital gain if the bond interest leads to stock appreciation and thus capital gain on sale. Obviously, a more important example than this is capital gain to shareholders who sell the stock of U.S. multinationals that have appreciated by reason of their earning huge unrepatriated foreign profits.
In sum, we have these two levels of corporate tax, each frequently avoided, thereby reducing the amount of actual double taxation, and each, at present, bearing differently from the other on distinct types of persons or income. So there are extra complications to keep in mind when one is revising corporate income taxation at one level or the other. Things would be a lot simpler if the two levels applied less distinctively from each other.
Keeping both taxes, even if their degree of integration is improved, is one possible takeaway. Most (but perhaps not all) people in the biz favor lowering the U.S. corporate rate from 35% - say to something in the 15% to 25% range - if this can be done in a manner that is budget-neutral and distribution-neutral. (Among the issues that would have to be addressed, however, is use of the corporate form as a lower-rate tax shelter, e.g., by owner-employees who underpay themselves from the salary standpoint because they are also benefiting from stock appreciation.) The rationale commonly offered is that this would reduce distortions in multiple dimensions (choice of entity, debt vs. equity financing, profit-shifting within a multinational group, etc.)
Why doesn't this argument suggest lowering the corporate rate all the way to zero? After all, then these distortions disappear, rather than merely being mitigated somewhat. Well, to proponents of replacing the existing corporate income tax with a destination-based cash flow tax, it does. But if, as a political economy or administrative matter, one needs both levels of tax in order to accomplish various distinct things that one values, then on balance one might want to retain both levels of tax, with or without express integration mechanisms - as do, for example, both the Grubert-Altshuler and Toder-Viard proposals.
Should we think of stock held in a traditional IRA or 401(k), in which contributions are deducted upfront and withdrawals are taxed at the backend, as actually tax-exempt? The answer is yes, given that the present value of one's tax liability from the account is zero, if we assume both (a) that the account just earns normal returns that the taxpayer can't scale up (a la Mark Zuckerberg hypothetically deciding to create two comparably profitable Facebooks rather than just one), and (b) that the taxpayer's marginal rate is the same at the contribution and withdrawal stages. Given the deduction limits for these types of tax-favored savings accounts, along with the fact that most people are just putting in publicly traded stock as to which they have no reason to expect special-opportunity above-normal returns, I'm fine with classifying them as tax-exempt for purposes of the studies, so long as we keep the nuances in mind.
The most obvious takeaway from the Rosenthal-Austin and Burman-Clausing bottom lines is that "double taxation" is less of an issue than many analyses of corporate income taxation tend to assume. This may reduce the appeal and urgency of treating corporate integration as a top priority, although it doesn't rebut the existence of the distortions that proponents of integration emphasize.
A further set of takeaways that I get pertain to how we should think about the existing two levels of taxation. These studies help remind us that, if one eliminated the entity-level tax and shifted corporate tax collection purely to the shareholder level (e.g., by taxing dividends and capital gains more rigorously), we would eliminate the existing indirect taxation of tax-exempts and foreigners via the entity-level tax. But the arguments for structural reform of the corporate tax, and for shifting to the owner level, are distinct from the issues around how we ought to tax (or not tax) tax-exempts and/or foreigners who or that hold U.S. equity. So those issues ought to be addressed if one is changing the entity-level tax. (For my money, there is no reason to extend the effective reach of tax-exempts' nominal exemption by letting them escape the tax that they are now indirectly paying; for foreigners the key question for me is incidence: to what extent are they actually bearing the tax?)
What about the owner level? If we're relying more on taxes on this level, it's important to consider features such as interest charges for deferral plus realization at death (as in a recent Grubert-Altshuler proposal), or taxing unrealized appreciation for publicly traded stock (as in a recent Toder-Viard proposal).
But also, at the owner level we may be changing somewhat the mix of income that is being taxed. A not very important example, just to make the basic point, is that if a corporation earns municipal bond income and then uses it to pay dividends, there is no tax at the entity level but there is a taxable dividend. Likewise, there's capital gain if the bond interest leads to stock appreciation and thus capital gain on sale. Obviously, a more important example than this is capital gain to shareholders who sell the stock of U.S. multinationals that have appreciated by reason of their earning huge unrepatriated foreign profits.
In sum, we have these two levels of corporate tax, each frequently avoided, thereby reducing the amount of actual double taxation, and each, at present, bearing differently from the other on distinct types of persons or income. So there are extra complications to keep in mind when one is revising corporate income taxation at one level or the other. Things would be a lot simpler if the two levels applied less distinctively from each other.
Keeping both taxes, even if their degree of integration is improved, is one possible takeaway. Most (but perhaps not all) people in the biz favor lowering the U.S. corporate rate from 35% - say to something in the 15% to 25% range - if this can be done in a manner that is budget-neutral and distribution-neutral. (Among the issues that would have to be addressed, however, is use of the corporate form as a lower-rate tax shelter, e.g., by owner-employees who underpay themselves from the salary standpoint because they are also benefiting from stock appreciation.) The rationale commonly offered is that this would reduce distortions in multiple dimensions (choice of entity, debt vs. equity financing, profit-shifting within a multinational group, etc.)
Why doesn't this argument suggest lowering the corporate rate all the way to zero? After all, then these distortions disappear, rather than merely being mitigated somewhat. Well, to proponents of replacing the existing corporate income tax with a destination-based cash flow tax, it does. But if, as a political economy or administrative matter, one needs both levels of tax in order to accomplish various distinct things that one values, then on balance one might want to retain both levels of tax, with or without express integration mechanisms - as do, for example, both the Grubert-Altshuler and Toder-Viard proposals.
Sunday, March 26, 2017
A contrast in styles
On both Thursday night and Saturday night, I saw concerts in Town Hall at West 43rd Street (a very nice smallish setting, apart from being a bit low on bathrooms). However, despite the respective artists' having been born just 12 years apart and having some overlap in influences, the two shows and bands were pretty different.
On Thursday night, it was Yo La Tengo performing a show entitled "And Then Yo La Tengo Turned Itself Inside Out." Despite the reference to their album, "And Then Nothing Turned Itself Inside Out," only three or so of the tracks they played were from the album, most notably "Autumn Sweater." While they rocked out occasionally, a lot of it was jazzy and improvisational, with the aid of a 4-piece horn section plus extra guitarist, drummers, and harpist. Very intense and New York cool - a lot of frowning concentration and interplay between the musicians, but not much gab to the audience apart from introducing the guest performers.
On Saturday night, it was the Zombies of 1960s quasi-fame. The first set was a potpourri, including a couple of songs from their 2015 album. The second set consisted of "Odessey and Oracle"(sic), their subsequently celebrated (though at the time ignored) baroque pop classic from early 1968. Broadly smiling, happy to be there, at this late date, with a good-sized and receptive audience, earnestly recounting stories about particular songs with apologies to those who had already heard them, expressing pride in their old work but wishing people were more eager to hear their new material, etc.
The Zombies also have a range of styles, from ballad to more rave-up, but they're very much in McCartney's 1960s territory. However, they don't sound like they're imitating him (apart from a track on Odessey, called A Rose for Emily, that might be a bit of an Eleanor Rigby rewrite). Instead, they just seem to have similar DNA, including the tunefulness, use of piano-based songwriting, rich harmonies that also show a Beach Boys influence, multiple musical ideas in the same song, and wistfulness without bitterness. Although my technical musical knowledge is extremely limited, they also seem to use more odd chords (III and VI for starters, and with plenty of minor keys) than anyone else of the era apart from the Beatles.
Their only really famous songs are She's Not There, Tell Her No, and Time of the Season (plus Hold Your Head Up from keyboardist Rod Argent's 1970s band). But Odessey and Oracle is well worth a listen if you're the sort who likes Sgt Pepper and Pet Sounds, and who wishes that Satanic Majesties were as good as the Stones' psychedelic singles (Dandelion, She's a Rainbow, Ruby Tuesday, Paint It Black, Lady Jane, etc.).
On Thursday night, it was Yo La Tengo performing a show entitled "And Then Yo La Tengo Turned Itself Inside Out." Despite the reference to their album, "And Then Nothing Turned Itself Inside Out," only three or so of the tracks they played were from the album, most notably "Autumn Sweater." While they rocked out occasionally, a lot of it was jazzy and improvisational, with the aid of a 4-piece horn section plus extra guitarist, drummers, and harpist. Very intense and New York cool - a lot of frowning concentration and interplay between the musicians, but not much gab to the audience apart from introducing the guest performers.
On Saturday night, it was the Zombies of 1960s quasi-fame. The first set was a potpourri, including a couple of songs from their 2015 album. The second set consisted of "Odessey and Oracle"(sic), their subsequently celebrated (though at the time ignored) baroque pop classic from early 1968. Broadly smiling, happy to be there, at this late date, with a good-sized and receptive audience, earnestly recounting stories about particular songs with apologies to those who had already heard them, expressing pride in their old work but wishing people were more eager to hear their new material, etc.
The Zombies also have a range of styles, from ballad to more rave-up, but they're very much in McCartney's 1960s territory. However, they don't sound like they're imitating him (apart from a track on Odessey, called A Rose for Emily, that might be a bit of an Eleanor Rigby rewrite). Instead, they just seem to have similar DNA, including the tunefulness, use of piano-based songwriting, rich harmonies that also show a Beach Boys influence, multiple musical ideas in the same song, and wistfulness without bitterness. Although my technical musical knowledge is extremely limited, they also seem to use more odd chords (III and VI for starters, and with plenty of minor keys) than anyone else of the era apart from the Beatles.
Their only really famous songs are She's Not There, Tell Her No, and Time of the Season (plus Hold Your Head Up from keyboardist Rod Argent's 1970s band). But Odessey and Oracle is well worth a listen if you're the sort who likes Sgt Pepper and Pet Sounds, and who wishes that Satanic Majesties were as good as the Stones' psychedelic singles (Dandelion, She's a Rainbow, Ruby Tuesday, Paint It Black, Lady Jane, etc.).
Tuesday, March 21, 2017
NYU Tax Policy Colloquium, week 8: Daniel Hemel's "The Federalist Safeguards of Progressive Taxation"
Yesterday we resumed after spring break, and dipped a toe into the deep blue waters of constitutional law. Daniel Hemel's paper looks at three recent developments in federalism doctrine, viewed by many as efforts by a conservative Supreme Court to kneecap liberal legislation, and argues that they increase the progressivity of the overall U.S. tax system (counting all levels of government).
First, a couple of brief thoughts on federalism doctrine and practice. Leaving aside some constitutional law professors and perhaps judges, my Rule 1 about federalism arguments in the political process - i.e., discussions of whether state governments should be able to set policies distinct from or even antithetical to those being pursued at the national level - is as follows: Almost everyone is almost always hypocritical.
I don't mean this to impugn people (much though it may sound like it). The thing is, almost everyone cares far more about substantive issues than about which level of government should get to decide things. So, when their side controls the national government, they want the states to have to fall in line, and when the other side controls the national government, they want "their" states to be able to diverge, So people are hypocritical about federalism, not because they're hypocrites, but because it isn't really what they care about.
Modern examples, of course, include the Medicaid expansion, gun or abortion or immigration policies, etc. But I learned the basic lesson about this before I was 20 years old, as an American history major in college. When New England Federalists controlled the national government under Washington, the Virginians were all about "states rights." Then, when the Virginians took over, New Englanders started to talk that way during the Embargo / War of 1812 era. Then of course we all know about the South during the run-up to the Civil War, except that they wanted broad and aggressive federal powers to require enforcement of the Fugitive Slave Act. So one doesn't even need the modern history regarding civil rights and "states' rights" to know the basic score.
But back to the Hemel paper. He notes three constitutional doctrines in the area of federalism that recent Supreme Court jurisprudence has emphasized, all viewed by many as conservative moves, and argues that they will have progressive effects. I'll just discuss two of the doctrines here: anti-commandeering and anti-coercion (the third pertains to sovereign immunity).
Anti-commandeering featured in a recent case, Printz, that barred the federal government (through the Brady Handgun Violence Protection Act) from requiring state and local officials to run background checks on prospective handgun purchasers. The idea was that the feds can't requisition local officials to perform federal tasks.
Anti-coercion featured, I gather not very coherently, in the ACA case (NFIB), where the Roberts opinion said the feds couldn't tell states that they'd lose all their Medicaid funding if they didn't extend coverage to people up to 133% of the poverty line, even though the expansion was funded for a number of years. This apparently was too much a "gun to the head." Some gun, I'd say, and the lack of sharp intellectual contours becomes clearer when one asks: What if Medicaid had included the higher coverage as part of its package from the very start? This of course led to the extraordinary spectacle of Republican governors turning down what was in effect free money for many of their poorer residents. (This, of course, is a drama that continues to play out, with states increasingly accepting the expanded coverage, and the current excitement about Paul Ryan's effort to ... well, if I simply described what he is trying to do it would sound partisan.)
Anyway, the constitutional literature has apparently noted that what all this means, in terms of the allocation of powers, is not that the feds can't use the state governments to advance their own ends in either scenario, but that they need to get consent. So two big favorites of modern legal scholarship - the Coase Theorem and the distinction between property and liabiltiy rules - have been deployed to describe the doctrines' effects.
The Coase idea is that the entitlement to use state government services belongs to the states, not to the feds. So we have an assignment of property rights that can be followed (depending on transaction costs) by bargaining to achieve the net-optimal outcome for the parties. Suppose the only issue raised in both of the above settings - rather than pertaining to ideology and politics around handguns, health care, etc. - was administrative cost. E.g., suppose the federal government could do something at the state and local level at a cost of $12, but the state government could do it for only $10. Then one might expect the feds to offer the state, say, $11 to do it. But if the feds could command the state to do it, the only immediate and direct difference (in the world of this little hypothetical) would be that the state doesn't get the $11. But if it cost the state $12 and the feds could do it for just $10 and the feds could command the state to do it, the deal would go the other way - the state would pay the federal government $11 for permission to beg off.
I don't actually find this an enormously illuminating or realistic way to think about either the handgun checks or the Medicaid expansion. But that's the literature as Hemel finds it. His point is that the anti-coercion and anti-commandeering rules, if these Coasean deals take place, make the federal government "poorer" and the state governments "richer" than under the opposite rule. More on that in a moment.
Turning to property rules vs. liability rules, the paper notes that the states have a "property" right since the feds wouldd need them to agree voluntarily to "sell" it (as in the case where they accept the Medicaid expansion or agree to do gun checks conditioned on the feds, say, paying for the services provided). It's not inalienable - the state can agree to do it - and it's also not a "liability" rule, as it would be if the feds could "commandeer" but then could be sued and required to pay compensation.
The liability rule version of this, though fanciful, offers a useful thought experiment. Suppose the feds could have required the handgun checks, subject only to compensating states financially for the cost of in effect lending out their personnel to do federal jobs. I imagine that the officials in pro-gun states would have remained unthrilled - presumably, political views about gun policy were central to this dispute. Likewise, in the Medicaid example, the states actually were being substantially compensated, and I doubt that the prospect of having to pay more down the road was really central to the reluctance (how often do governors of any party think that far ahead when they have a chance to get current big $$ for their constituents?). So the "liability" rule, if it defined compensable damages in terms of the administrative costs to the state and local governments, would be friendlier to federal power than the "property" rule. (But if we imagined other compensable harms, such as from some measure of the state officials' subjective reluctance to pursue policies they disliked unless the state was given a whole lot more money, it would start to look more similar to the property rule.)
Anyway, the paper's main argument is that, since federal tax financing is significantly more progressive than state tax financing, requiring the feds to buy off the state governments might be expected to increase revenue-raising by the former and lower it by the latter. Hence, absent other adjustments, which the paper does extensively consider, the result would be to make overall tax financing more progressive. This is a nice point. But it is offset to a degree if the big-money policies that the feds want to pursue but can't due to their inability to purchase state cooperation, would tend to be progressive ones a la Medicaid expansion.
First, a couple of brief thoughts on federalism doctrine and practice. Leaving aside some constitutional law professors and perhaps judges, my Rule 1 about federalism arguments in the political process - i.e., discussions of whether state governments should be able to set policies distinct from or even antithetical to those being pursued at the national level - is as follows: Almost everyone is almost always hypocritical.
I don't mean this to impugn people (much though it may sound like it). The thing is, almost everyone cares far more about substantive issues than about which level of government should get to decide things. So, when their side controls the national government, they want the states to have to fall in line, and when the other side controls the national government, they want "their" states to be able to diverge, So people are hypocritical about federalism, not because they're hypocrites, but because it isn't really what they care about.
Modern examples, of course, include the Medicaid expansion, gun or abortion or immigration policies, etc. But I learned the basic lesson about this before I was 20 years old, as an American history major in college. When New England Federalists controlled the national government under Washington, the Virginians were all about "states rights." Then, when the Virginians took over, New Englanders started to talk that way during the Embargo / War of 1812 era. Then of course we all know about the South during the run-up to the Civil War, except that they wanted broad and aggressive federal powers to require enforcement of the Fugitive Slave Act. So one doesn't even need the modern history regarding civil rights and "states' rights" to know the basic score.
But back to the Hemel paper. He notes three constitutional doctrines in the area of federalism that recent Supreme Court jurisprudence has emphasized, all viewed by many as conservative moves, and argues that they will have progressive effects. I'll just discuss two of the doctrines here: anti-commandeering and anti-coercion (the third pertains to sovereign immunity).
Anti-commandeering featured in a recent case, Printz, that barred the federal government (through the Brady Handgun Violence Protection Act) from requiring state and local officials to run background checks on prospective handgun purchasers. The idea was that the feds can't requisition local officials to perform federal tasks.
Anti-coercion featured, I gather not very coherently, in the ACA case (NFIB), where the Roberts opinion said the feds couldn't tell states that they'd lose all their Medicaid funding if they didn't extend coverage to people up to 133% of the poverty line, even though the expansion was funded for a number of years. This apparently was too much a "gun to the head." Some gun, I'd say, and the lack of sharp intellectual contours becomes clearer when one asks: What if Medicaid had included the higher coverage as part of its package from the very start? This of course led to the extraordinary spectacle of Republican governors turning down what was in effect free money for many of their poorer residents. (This, of course, is a drama that continues to play out, with states increasingly accepting the expanded coverage, and the current excitement about Paul Ryan's effort to ... well, if I simply described what he is trying to do it would sound partisan.)
Anyway, the constitutional literature has apparently noted that what all this means, in terms of the allocation of powers, is not that the feds can't use the state governments to advance their own ends in either scenario, but that they need to get consent. So two big favorites of modern legal scholarship - the Coase Theorem and the distinction between property and liabiltiy rules - have been deployed to describe the doctrines' effects.
The Coase idea is that the entitlement to use state government services belongs to the states, not to the feds. So we have an assignment of property rights that can be followed (depending on transaction costs) by bargaining to achieve the net-optimal outcome for the parties. Suppose the only issue raised in both of the above settings - rather than pertaining to ideology and politics around handguns, health care, etc. - was administrative cost. E.g., suppose the federal government could do something at the state and local level at a cost of $12, but the state government could do it for only $10. Then one might expect the feds to offer the state, say, $11 to do it. But if the feds could command the state to do it, the only immediate and direct difference (in the world of this little hypothetical) would be that the state doesn't get the $11. But if it cost the state $12 and the feds could do it for just $10 and the feds could command the state to do it, the deal would go the other way - the state would pay the federal government $11 for permission to beg off.
I don't actually find this an enormously illuminating or realistic way to think about either the handgun checks or the Medicaid expansion. But that's the literature as Hemel finds it. His point is that the anti-coercion and anti-commandeering rules, if these Coasean deals take place, make the federal government "poorer" and the state governments "richer" than under the opposite rule. More on that in a moment.
Turning to property rules vs. liability rules, the paper notes that the states have a "property" right since the feds wouldd need them to agree voluntarily to "sell" it (as in the case where they accept the Medicaid expansion or agree to do gun checks conditioned on the feds, say, paying for the services provided). It's not inalienable - the state can agree to do it - and it's also not a "liability" rule, as it would be if the feds could "commandeer" but then could be sued and required to pay compensation.
The liability rule version of this, though fanciful, offers a useful thought experiment. Suppose the feds could have required the handgun checks, subject only to compensating states financially for the cost of in effect lending out their personnel to do federal jobs. I imagine that the officials in pro-gun states would have remained unthrilled - presumably, political views about gun policy were central to this dispute. Likewise, in the Medicaid example, the states actually were being substantially compensated, and I doubt that the prospect of having to pay more down the road was really central to the reluctance (how often do governors of any party think that far ahead when they have a chance to get current big $$ for their constituents?). So the "liability" rule, if it defined compensable damages in terms of the administrative costs to the state and local governments, would be friendlier to federal power than the "property" rule. (But if we imagined other compensable harms, such as from some measure of the state officials' subjective reluctance to pursue policies they disliked unless the state was given a whole lot more money, it would start to look more similar to the property rule.)
Anyway, the paper's main argument is that, since federal tax financing is significantly more progressive than state tax financing, requiring the feds to buy off the state governments might be expected to increase revenue-raising by the former and lower it by the latter. Hence, absent other adjustments, which the paper does extensively consider, the result would be to make overall tax financing more progressive. This is a nice point. But it is offset to a degree if the big-money policies that the feds want to pursue but can't due to their inability to purchase state cooperation, would tend to be progressive ones a la Medicaid expansion.
Wednesday, March 15, 2017
AMT nostalgia
As per this NYT op-ed, the 2005 Trump tax return is causing some people to salute the alternative minimum tax (AMT) and argue that it should be retained after all.
I actually wrote about this topic nearly 30 years ago, when I was just starting out in academics and reflecting on my legislative staff experience, which had included being on the team that made sure the AMT worked technically as intended. (It did.) But it has had few defenders since, and they haven't included me.
That said, I would want its repeal to be accompanied by adjusting overall revenue and distribution appropriately, which the Congressional Republicans obviously wouldn't do. But let's take a quick look at the relevance of the Trump tax return to the AMT issue.
In his 2005 tax return, Trump had about $150 million of regular adjusted gross income (AGI), reduced to about $50 million by what may have been the last of the late-1990s net operating losses (NOLs), and then to about 2/3 of that by itemized deductions. With predominantly capital gain remaining, he would have owed only about $5 million of tax, but the AMT raised this to about $36 billion total. This implies that his AMTI (alternative minimum taxable income) might have been in the neighborhood of $112 million or so.
Why was it (a) so much lower than AGI pre-NOLs, yet (b) so much higher afterwards? We don't know. So how in this case it was performing the backstop role, and how common that is for a broader range of circumstances, we also don't know.
A key point here is that there often is nothing wrong with allowing NOLs to reduce current year taxable income. If you genuinely lose $100 million in Year 1, and make $100 million in Year 2, that's a net of zero. But in Trump's case, of course, if that $100 million of NOLs was the last residue of his late 1990s gambit that enabled him to deduct more than $900 million of other people's losses that they also were deducting, then his NOL deduction was bad in a policy sense for that reason - but not relatedly to its being hit by the AMT (which generally allows NOLs, as computed for its purposes, anyway, albeit with a rule that stops their current year allowability 90% of the way to zero).
Certainly a bit random if, this time around, the AMT caused Trump to pay closer to the amount to what he arguably "ought" to have been paying, and not enough to induce one (or at least me) to endorse the darned thing, especially if it could be traded in for something with a better-designed impact on people at Trump's income level.
UPDATE: I'm quoted as follows in an article about the AMT:"The current AMT doesn't address sophisticated modern tax avoidance techniques - such as Larry Ellison ... getting a $1 salary and borrowing against the value of his appreciated stock, or companies such as Apple directing their global profits to tax haven subsidiaries."
I actually wrote about this topic nearly 30 years ago, when I was just starting out in academics and reflecting on my legislative staff experience, which had included being on the team that made sure the AMT worked technically as intended. (It did.) But it has had few defenders since, and they haven't included me.
That said, I would want its repeal to be accompanied by adjusting overall revenue and distribution appropriately, which the Congressional Republicans obviously wouldn't do. But let's take a quick look at the relevance of the Trump tax return to the AMT issue.
In his 2005 tax return, Trump had about $150 million of regular adjusted gross income (AGI), reduced to about $50 million by what may have been the last of the late-1990s net operating losses (NOLs), and then to about 2/3 of that by itemized deductions. With predominantly capital gain remaining, he would have owed only about $5 million of tax, but the AMT raised this to about $36 billion total. This implies that his AMTI (alternative minimum taxable income) might have been in the neighborhood of $112 million or so.
Why was it (a) so much lower than AGI pre-NOLs, yet (b) so much higher afterwards? We don't know. So how in this case it was performing the backstop role, and how common that is for a broader range of circumstances, we also don't know.
A key point here is that there often is nothing wrong with allowing NOLs to reduce current year taxable income. If you genuinely lose $100 million in Year 1, and make $100 million in Year 2, that's a net of zero. But in Trump's case, of course, if that $100 million of NOLs was the last residue of his late 1990s gambit that enabled him to deduct more than $900 million of other people's losses that they also were deducting, then his NOL deduction was bad in a policy sense for that reason - but not relatedly to its being hit by the AMT (which generally allows NOLs, as computed for its purposes, anyway, albeit with a rule that stops their current year allowability 90% of the way to zero).
Certainly a bit random if, this time around, the AMT caused Trump to pay closer to the amount to what he arguably "ought" to have been paying, and not enough to induce one (or at least me) to endorse the darned thing, especially if it could be traded in for something with a better-designed impact on people at Trump's income level.
UPDATE: I'm quoted as follows in an article about the AMT:"The current AMT doesn't address sophisticated modern tax avoidance techniques - such as Larry Ellison ... getting a $1 salary and borrowing against the value of his appreciated stock, or companies such as Apple directing their global profits to tax haven subsidiaries."
Wrong road for anti-Trump protestors
I tell a Village Voice reporter that, no, refusing to pay your federal income taxes isn't a good way to show opposition to Trump.
When the going gets tax, the tax profs go on TV
So yesterday, as I was innocently enjoying a snow day at home (and to a degree getting things done, as I was prepared for it), this Trump tax story broke. I heard about it early on, and ended up saying a few words on the Lawrence O'Donnell show after 10 pm.
I was then invited to appear on more TV shows today at 6:30 am and 10 am. The latter I turned down because enough is enough and I do have other things to do, but the former I accepted, despite the 5:30 am car service pickup. Then they called me at 5:15 am, while I was getting dressed, to tell me the gig was cancelled. Ah, the joys of live TV.
I suspect they cancelled because it turned out not to be that big a story, in reality / substance. The released information was too limited and anodyne to tell us much. So I am open to the view that is being widely expressed that Trump may have leaked this himself. (In support of that theory, note how promptly and indeed efficiently the White House responded to word of the leak, essentially by confirming the numbers - is that how this White House typically operates?)
There's not a vast ocean of things that we learn from the leaked 1040, but here are a few quick thoughts:
--He seems to have used up the OPM (other people's money) $900 million NOL that he got in the 1990s, even assuming, as seems reasonable but not certain, that the $100 million loss on the return is the last dregs of that.
--Since he did pay some federal income tax in this year, his reason for hiding his tax returns isn't that they would show he never paid any federal income tax. So the secrecy theories about his business dealings, associates, etc. gain relative force. But of course we don't know if 2005 was a typical tax year for him.
--The AMT liability ($31M out of the $36M total liability) potentially triggered an AMT credit against the regular tax in subsequent tax years. Insofar as AMT liability is attributable to timing differences (e.g,. the rate of depreciation) rather than permanent differences (e.g., an income exclusion) betweeen the two tax bases, such liability generates a tax credit that in later years can be used to reduce regular tax liability to AMT liability. So this may have been mainly a timing thing, albeit without contradicting that he was paying something at some point in time.
--Probably not much by way of charitable deductions (which comes as no surprise). His $17M of itemized deductions might be expected to include significant state and local income tax liability, especially if he was meaningfully subject to the NYS and NYC income tax.
--$31 million of AMT liability, 28% AMT rate (after full phase-in) suggests that he might have had, say, $112 million or so of AMT income. That is less than his pre-NOL regular taxable income for 2005, although we don't know if there was an NOL used in the AMT part as well. (It would have been a separately computed AMT NOL - the two systems run on separate tracks.) Unclear why the AMTI would have been less than the regular taxable income in 2005, although who knows given multiple timing differences between the systems (e.g., you get higher AMT than regular tax depreciation deductions in the later years of an asset's useful life). But maybe the last dregs of AMT NOL were also used in 2005.
--Even so, it seems as if he must have used up the AMT NOL, assuming he got it in roughly the same amount, faster than the regular tax NOL. Again, it's unclear why. Note that the AMT NOL generally can only offset 90% of AMTI, whereas the regular tax AMT can actually be used faster, all else equal, since it can be used to offset 100% of regular taxable income.
--Only $5 million of regular tax liability on $31 million of regular taxable income, presumably reflecting that what was left after the NOLs was mainly capital gain.
--If really interesting tax return stuff from Trump comes out, it won't be the tax people (like me) who get the main insights from it. Rather, the main value will come from investigative reporters who can trace the information about business associates, etc.
I was then invited to appear on more TV shows today at 6:30 am and 10 am. The latter I turned down because enough is enough and I do have other things to do, but the former I accepted, despite the 5:30 am car service pickup. Then they called me at 5:15 am, while I was getting dressed, to tell me the gig was cancelled. Ah, the joys of live TV.
I suspect they cancelled because it turned out not to be that big a story, in reality / substance. The released information was too limited and anodyne to tell us much. So I am open to the view that is being widely expressed that Trump may have leaked this himself. (In support of that theory, note how promptly and indeed efficiently the White House responded to word of the leak, essentially by confirming the numbers - is that how this White House typically operates?)
There's not a vast ocean of things that we learn from the leaked 1040, but here are a few quick thoughts:
--He seems to have used up the OPM (other people's money) $900 million NOL that he got in the 1990s, even assuming, as seems reasonable but not certain, that the $100 million loss on the return is the last dregs of that.
--Since he did pay some federal income tax in this year, his reason for hiding his tax returns isn't that they would show he never paid any federal income tax. So the secrecy theories about his business dealings, associates, etc. gain relative force. But of course we don't know if 2005 was a typical tax year for him.
--The AMT liability ($31M out of the $36M total liability) potentially triggered an AMT credit against the regular tax in subsequent tax years. Insofar as AMT liability is attributable to timing differences (e.g,. the rate of depreciation) rather than permanent differences (e.g., an income exclusion) betweeen the two tax bases, such liability generates a tax credit that in later years can be used to reduce regular tax liability to AMT liability. So this may have been mainly a timing thing, albeit without contradicting that he was paying something at some point in time.
--Probably not much by way of charitable deductions (which comes as no surprise). His $17M of itemized deductions might be expected to include significant state and local income tax liability, especially if he was meaningfully subject to the NYS and NYC income tax.
--$31 million of AMT liability, 28% AMT rate (after full phase-in) suggests that he might have had, say, $112 million or so of AMT income. That is less than his pre-NOL regular taxable income for 2005, although we don't know if there was an NOL used in the AMT part as well. (It would have been a separately computed AMT NOL - the two systems run on separate tracks.) Unclear why the AMTI would have been less than the regular taxable income in 2005, although who knows given multiple timing differences between the systems (e.g., you get higher AMT than regular tax depreciation deductions in the later years of an asset's useful life). But maybe the last dregs of AMT NOL were also used in 2005.
--Even so, it seems as if he must have used up the AMT NOL, assuming he got it in roughly the same amount, faster than the regular tax NOL. Again, it's unclear why. Note that the AMT NOL generally can only offset 90% of AMTI, whereas the regular tax AMT can actually be used faster, all else equal, since it can be used to offset 100% of regular taxable income.
--Only $5 million of regular tax liability on $31 million of regular taxable income, presumably reflecting that what was left after the NOLs was mainly capital gain.
--If really interesting tax return stuff from Trump comes out, it won't be the tax people (like me) who get the main insights from it. Rather, the main value will come from investigative reporters who can trace the information about business associates, etc.
Tuesday, March 14, 2017
Caught in the act
Monday, March 13, 2017
2018 NYU Tax Policy Colloquium
Things are now pretty much official. A year from now at NYU, I will be co-teaching the Tax Policy Colloquium with Lily Batchelder. Subject to both of our schedules, this may indeed become our norm in the future.
But next year we will be switching the day of the week on which we meet from Mondays, as it was this year, back to Tuesdays. We will meet on every Tuesday afternoon, from 4 to 5:50 pm, from January 16 through May 1, with the exception of (1 February 20 (legislative Monday) and (2) March 20 (spring break).
But next year we will be switching the day of the week on which we meet from Mondays, as it was this year, back to Tuesdays. We will meet on every Tuesday afternoon, from 4 to 5:50 pm, from January 16 through May 1, with the exception of (1 February 20 (legislative Monday) and (2) March 20 (spring break).
Friday, March 10, 2017
NYU-KPMG 17th Annual Tax Symposium at NYU Law School
Just now at NYU, I participated in the mid-day debate session at this year's NYU-KPMG Tax Symposium. This year's session was entitled "U.S. Tax Reform - A Perfect Storm."
We had four interesting propositions to discuss (or five, depending on how you count), but not only did the debaters tend to agree on them all - so did the audience. They have people vote after each proposition through little hand-held devices, and each time one side won the support, not merely of all the panelists, but also of close to 80% vote of the audience.
The propositions, and what I had to say about each, are as follows:
1) Should the Blueprint's destination-based cash flow tax with border adjustments replace the corporate income tax?
The NO side carried this by what would prove to be the usual huge margin. Our panel met right after Michael Graetz had discussed his proposal to enact a VAT and use the revenues to lower the individual and corporate income taxes (including through a $100,000 exclusion amount for the former). Graetz had argued that his proposal pretty much accomplishes everything the DBCFT is meant to accomplish, and I decided to use that as an entry point for my brief remarks. I said: It's kind of as if they had decided to enact Graetz's plan, but only as gratuitously modified to make a horrible mess of it. In particular:
a) Graetz would keep the corporate income tax, but lower the rate to 15%. The Blueprints, by repealing it, would in effect go all the way to zero. For time reasons I just said: I agree with Graetz not the Blueprints on this one, but there are arguments on each side and it's a legitimate debate.
b) Due to the wage subsidy, the DBCFT runs into big problems with taxpayer losses that (unlike a typical VAT) it would make nonrefundable. Thoughtful commentators have argued that the interest charge is likely, in practice, to fall well short of fully addressing these problems.
c) The VAT Graetz has in his plan would be WTO-compliant. The DBCFT is highly likely to violate the WTO, and thus risks leading to a trade war or requiring prompt repeal.
d) The DBCFT strips out all of the features in the Graetz plan that aim to address revenue and distributional issues.
e) The DBCFT is a novel instrument that it will be tricky to get right, and I lack confidence that the people designing it will be able to get it right within the time they will have. This is not a knock at the staff, but at the leadership and how I gather they plan to rush it through.
1-a) If NO above, should the plan be adopted without border adjustments?
Once again, NO by acclamation. The specific question proposes increasing the proposed 20% rate to make up for the short-term revenue effect of dropping border adjustability. So I suppose it would mainly amount in practice to a smaller rate cut, still with expensing and net interest nondeductibility.
I mainly addressed the rate cut, saying there's a wide consensus about cutting the statutory U.S. corporate rate within a broader set of changes. I agree that this should be done in the right way, but do not think that this Congress or Administration would do it anywhere close to right.
2) Should the U.S. adopt a VAT and use revenue therefrom to pay for significant personal and corporate income tax rate cuts? (A la, via the Graetz plan.)
YES by acclamation. I said yes, depending on how the VAT revenues are used in terms of the full budget package. VATs in isolation are regressive but are commonly parts of fiscal systems more progressive than that in the U.S. I differ from Graetz in that I would use only some of the revenues to cut income taxes, while using others to fund things that I believe the government should be doing (e.g., with regard to healthcare, education, and infrastructure). But I consider Graetz's proposal, not merely a great deal better than present law, but probably also a whole lot better than anything that's at all likely to pass, be it now or in the foreseeable future.
3) Should FATCA be repealed as part of U.S. tax reform?
All panelists, and apparently the audience, agreed that this would be crazy, after all that the U.S. has done to get global coordination around FATCA. Since it was genuinely uncertain, when FATCA was enacted, that it would work out as well as it did, I compared repeal today to putting your chips on red at the roulette wheel and then, when the wheel comes up red, pulling them back off without collecting your winnings. But I and others did agree that the burdens the U.S. income tax system generally - not just through FATCA - places on nonresident citizens ought to be addressed. Whatever one's ultimate views on citizenship taxation, there's currently way too much burden being imposed relative to revenue collected.
4) Should Chevron deference be given to Treasury regulations that construe the Internal Revenue Code?
This would mean that the courts would evaluate each legal issue resolved by a tax regulation de novo, rather than giving affirmative weight to what the Treasury decided. Under Chevron, the administrative agency's interpretation of an ambiguity in the statute must be reasonable, even if the court thought some other interpretation was more reasonable still. Without Chevron, the court would simply decide what interpretation it felt was the most reasonable one.
Once again, YES (for retaining Chevron deference in tax) by acclamation. Other panelists noted the Treasury's superior tax expertise in interpreting and fleshing out complex statutes than one would expect from generalist federal judges. I noted that taxpayers often place a lot of value on certainty, which would likely be sharply reduced by this change. So they might be unhappy about it on balance even if they were glad that, in particular cases where they dislike a regulation, it would strengthen their hands.
I think of the move to overturn Chevron as part of the contemporary Republican war against the administrative state. Even if one is more sympathetic to that war than I am, tax may be an example of an area in which the people who are subject to the regulation would be far less enthusiastic about it than they might be in some other areas. I noted by analogy that, when Justice Scalia was waging his holy war against the use of legislative history, it was my impression that tax legislative history is widely viewed by people in the field as serving a far more benign role than Scalia saw it as having.
Tuesday, March 07, 2017
NYU Tax Policy Colloquium week 7: Scott Dyreng's (et al) Trade-offs in the Repatriation of Foreign Earnings
Yesterday at the colloquium, Scott Dyreng presented the above-named paper, coauthored by Kevin Markle and Jon Medrano. It examines how having U.S. net operating losses (NOLs) affects U.S. multinationals' repatriation of foreign earnings. Where the repatriated earnings are IRFE (indefinitely reinvested foreign earnings) for financial accounting purposes, repatriation can (a) reduce the present value of the taxpayer's expected long-term U.S. tax liability, and also (b) free up cash to be used more easily as the company prefers. However, this comes at the cost of (c) a negative effect on financial accounting income, relative to keeping the funds abroad (with an assumed future repatriation tax of zero) and acquiring a tax asset in the form of the NOL. The authors' premise, therefore, is that the companies should be expected to trade off (b) against (c), taking (a) as given.
Here are some quick thoughts on the issues that this interesting, although at this point still preliminary, research raises. The first set of issues concerns tax policy in relation to repatriation decisions; the second, financial accounting's effect on behavior by publicly traded companies.
1. Taxes and repatriation
(a) Time value, NOLs, and FTCs - One way of thinking about the underlying issue of how to minimize present value U.S. taxes is as follows. NOLs don't earn interest. Hence, they lose value in present value terms if they must be carried forward. (Not a problem for carrybacks, and subsequent drafts of the paper may try to ease this aspect more.)
Likewise, foreign tax credit (FTC) carryforwards don't earn interest. Note, by the way, that even low-taxed, though not zero-taxed, foreign source income can generate FTC carryforwards if used to wipe out NOLs. For example, suppose Acme would have a $10 million NOL but for its repatriating $9 million in cash that is associated with $1 million in foreign taxes paid. This is treated as a $10M repatriation, raising taxable income to zero, so there is no U.S. tax liability for the $1M in FTCs to offset. Hence, one has a $1 million FTC carryforward even though this was low-taxed income by U.S. standards. One issue that the paper examines is whether companies were especially keen to avoid turning NOLs into FTC carryforwards in the era when the latter had a much shorter carryforward period than the former.
The tradeoff between the two types of carryforwards should be a wash insofar as they have the same carryforward period and the chances or convenience of using one or the other look about the same. Neither carryforward bears interest. But what about implicit or unripe FTCs, associated with taxes paid on foreign earnings, where the earnings haven't been repatriated yet? Here, while there is not literally interest being earned on the deferral of the foreign tax credit claim, they may be growing in an interest-like way.
Suppose I have $X of unrepatriated foreign earnings, associated with $Y of unripe FTCs. If I am continuing to earn a normal rate of return and am paying foreign tax on that return, then Y may be growing at a normal rate of return. Subject to a lot more analysis that I won't attempt here, this might mean that the present value of unripe FTCs is remaining constant. Or, if there is a disparity between relevant rates of return, it might be growing or shrinking in present value terms once one has determined the proper discount rate for a given taxpayer.
I will not attempt further analysis here. It's the type of topic that might appeal, say, to Al Warren or Fadi Shaheen - writers whose recent work has often looked closely at equivalences and incentive effects under particular rate-of-return assumptions, I also agree that one has to think, not just about the FTCs, but also about the earnings and the residual U.S. taxes with which they may be associated. But insofar as taxpayers can still engage in "splitter"-type transactions, despite section 909 of the Internal Revenue Code, there may be some point to thinking about the FTCs in particular, albeit keeping in mind their connection to all the rest.
For now, however, it's enough to note that the unripe FTCs may be preferable to both NOL and FTC carrtyforwards because (or if) they are growing at a positive r of some kind.
(b) FTC carryforwards now vs. in the past - One point of empirical interest in the paper is that, pre-2005 but not post-2005, FTC carryforwards were much shorter than those for NOLs. My quick thought is that FTC carryforwards have probably become much less important over time than they used to be. Not only have foreign corporate tax rates declined relative to that in the U.S. over the last twenty-odd years, but foreign-to-foreign tax planning has improved, e.g., by reasom of companies' ability to take advantage of the check-the-box rules.
(c) 2005 tax holiday - Recent empirical work appears to confirm the ongoing importance of the 2005 tax holiday for foreign dividends, which enabled companies to repatriate earnings at a greatly reduced tax rate for a limited period. If companies had believed that it was a one-time-only event, never to recur, then it might not have affected expectations going forward. But of course speculation has been rife for years that Congress might do it again.
When companies anticipate the possible enactment of a new tax holiday, this both reduces the expected level for them of the ultimate U.S. repatriation tax that they might pay, and increases its volatility. These are distinct effects, each meriting analysis, although in one respect they push in the same direction - both discourage further repatriations before the next holiday.
2. Accounting and behavior
This is our second paper of the semester (Lily Batchelder's was the first) to look at financial accounting's effects on companies' behavior. Here the issues center around the accounting treatment of deferred repatriation taxes.
(a) Two accounting rules, both of them wrong - Say that Acme, a U.S. company, has $10 million of foreign earnings, stashed in a tax haven, with no associated FTCs. Under financial accounting rules, in the standard case Acme will book the deferred U.S. tax liability of $3.5 million as a current expense. This is wrong because Acme surely doesn't expect to pay that much tax in present value terms. Even assuming that there is a taxable repatriation at some point, Acme's managers surely (and reasonably) anticipate that they will get to pay a much lower repatriation tax rate than 35% - for example, due to another a tax holiday, a reduction in the U.S. corporate rate, a shift to territoriality in which any deemed repatriation tax applies at a much lower rate, etc. So the standard accounting rule here likely overstates the expected liability in realistic present value terms.
Now suppose instead that Acme persuades its auditors (which may not be too hard) that the funds have been indefinitely reinvested abroad, making them IRFE. Now the assumed repatriation tax is zero. This is also probably wrong. There surely is some chance of a taxable repatriation, even if at less than a 35% rate, and also the company may be paying an implicit tax, reducing the value of the funds, to put off the repatriation. (This is why it might be willing to pay a repatriation tax above zero, even if it truly anticipates being able to keep the funds abroad indefinitely.)
(b) What would the "correct" accounting rule do? - I would say that, building on the model that I gather financial accountants use for, say, uncertain litigation that has big bottom-line effects, the right answer would involve making a presenr value estimate for the deferred tax liability. Now, this is admittedly hard, as it depends not only on the company's long-term plans but also on future legislation. But if the grounds for the calculation were disclosed, mightn't this be better than using either of the existing methods? It might not only be more accurate on average, but would eliminate the discontinuity between the existing methods.
Okay, I'm not an accountant, but I suspect that this approach merits more attention than it has been getting. The fact that accountants might dislike it, because it forces them to do something that is inherently rather tuncertain, does not prevent its being potentially better borh for investors and for the tax system - the latter, because once companies have chosen IRFE treatment that may create lock-in wholly apart from the tax analysis.
(c) What if the FASB simply repealed the IRFE rule? - From the standpoint of informing investors, this is no panacaea, given the overstatement in many or most cases of actual expected U.S. liability. While we often hear that financial accounting should be "conservative," under-valuing companies can be as harmful to investors as over-valuing them, and can also create undesirable special advantages for the better-informed. But it would certainly be interesting to observe how greatly (or not) managerial behavior would change if keeping funds abroad, despite creating a likely true economic benefit (if the U.S. repatriation tax rate might be significantly lower in the future), created no accounting benefit.
BTW, no colloquium next week - it's our spring break. We will resume on Monday, March 20 with Daniel Hemel's The Federalist Safeguards of Progressive Taxation.
Here are some quick thoughts on the issues that this interesting, although at this point still preliminary, research raises. The first set of issues concerns tax policy in relation to repatriation decisions; the second, financial accounting's effect on behavior by publicly traded companies.
1. Taxes and repatriation
(a) Time value, NOLs, and FTCs - One way of thinking about the underlying issue of how to minimize present value U.S. taxes is as follows. NOLs don't earn interest. Hence, they lose value in present value terms if they must be carried forward. (Not a problem for carrybacks, and subsequent drafts of the paper may try to ease this aspect more.)
Likewise, foreign tax credit (FTC) carryforwards don't earn interest. Note, by the way, that even low-taxed, though not zero-taxed, foreign source income can generate FTC carryforwards if used to wipe out NOLs. For example, suppose Acme would have a $10 million NOL but for its repatriating $9 million in cash that is associated with $1 million in foreign taxes paid. This is treated as a $10M repatriation, raising taxable income to zero, so there is no U.S. tax liability for the $1M in FTCs to offset. Hence, one has a $1 million FTC carryforward even though this was low-taxed income by U.S. standards. One issue that the paper examines is whether companies were especially keen to avoid turning NOLs into FTC carryforwards in the era when the latter had a much shorter carryforward period than the former.
The tradeoff between the two types of carryforwards should be a wash insofar as they have the same carryforward period and the chances or convenience of using one or the other look about the same. Neither carryforward bears interest. But what about implicit or unripe FTCs, associated with taxes paid on foreign earnings, where the earnings haven't been repatriated yet? Here, while there is not literally interest being earned on the deferral of the foreign tax credit claim, they may be growing in an interest-like way.
Suppose I have $X of unrepatriated foreign earnings, associated with $Y of unripe FTCs. If I am continuing to earn a normal rate of return and am paying foreign tax on that return, then Y may be growing at a normal rate of return. Subject to a lot more analysis that I won't attempt here, this might mean that the present value of unripe FTCs is remaining constant. Or, if there is a disparity between relevant rates of return, it might be growing or shrinking in present value terms once one has determined the proper discount rate for a given taxpayer.
I will not attempt further analysis here. It's the type of topic that might appeal, say, to Al Warren or Fadi Shaheen - writers whose recent work has often looked closely at equivalences and incentive effects under particular rate-of-return assumptions, I also agree that one has to think, not just about the FTCs, but also about the earnings and the residual U.S. taxes with which they may be associated. But insofar as taxpayers can still engage in "splitter"-type transactions, despite section 909 of the Internal Revenue Code, there may be some point to thinking about the FTCs in particular, albeit keeping in mind their connection to all the rest.
For now, however, it's enough to note that the unripe FTCs may be preferable to both NOL and FTC carrtyforwards because (or if) they are growing at a positive r of some kind.
(b) FTC carryforwards now vs. in the past - One point of empirical interest in the paper is that, pre-2005 but not post-2005, FTC carryforwards were much shorter than those for NOLs. My quick thought is that FTC carryforwards have probably become much less important over time than they used to be. Not only have foreign corporate tax rates declined relative to that in the U.S. over the last twenty-odd years, but foreign-to-foreign tax planning has improved, e.g., by reasom of companies' ability to take advantage of the check-the-box rules.
(c) 2005 tax holiday - Recent empirical work appears to confirm the ongoing importance of the 2005 tax holiday for foreign dividends, which enabled companies to repatriate earnings at a greatly reduced tax rate for a limited period. If companies had believed that it was a one-time-only event, never to recur, then it might not have affected expectations going forward. But of course speculation has been rife for years that Congress might do it again.
When companies anticipate the possible enactment of a new tax holiday, this both reduces the expected level for them of the ultimate U.S. repatriation tax that they might pay, and increases its volatility. These are distinct effects, each meriting analysis, although in one respect they push in the same direction - both discourage further repatriations before the next holiday.
2. Accounting and behavior
This is our second paper of the semester (Lily Batchelder's was the first) to look at financial accounting's effects on companies' behavior. Here the issues center around the accounting treatment of deferred repatriation taxes.
(a) Two accounting rules, both of them wrong - Say that Acme, a U.S. company, has $10 million of foreign earnings, stashed in a tax haven, with no associated FTCs. Under financial accounting rules, in the standard case Acme will book the deferred U.S. tax liability of $3.5 million as a current expense. This is wrong because Acme surely doesn't expect to pay that much tax in present value terms. Even assuming that there is a taxable repatriation at some point, Acme's managers surely (and reasonably) anticipate that they will get to pay a much lower repatriation tax rate than 35% - for example, due to another a tax holiday, a reduction in the U.S. corporate rate, a shift to territoriality in which any deemed repatriation tax applies at a much lower rate, etc. So the standard accounting rule here likely overstates the expected liability in realistic present value terms.
Now suppose instead that Acme persuades its auditors (which may not be too hard) that the funds have been indefinitely reinvested abroad, making them IRFE. Now the assumed repatriation tax is zero. This is also probably wrong. There surely is some chance of a taxable repatriation, even if at less than a 35% rate, and also the company may be paying an implicit tax, reducing the value of the funds, to put off the repatriation. (This is why it might be willing to pay a repatriation tax above zero, even if it truly anticipates being able to keep the funds abroad indefinitely.)
(b) What would the "correct" accounting rule do? - I would say that, building on the model that I gather financial accountants use for, say, uncertain litigation that has big bottom-line effects, the right answer would involve making a presenr value estimate for the deferred tax liability. Now, this is admittedly hard, as it depends not only on the company's long-term plans but also on future legislation. But if the grounds for the calculation were disclosed, mightn't this be better than using either of the existing methods? It might not only be more accurate on average, but would eliminate the discontinuity between the existing methods.
Okay, I'm not an accountant, but I suspect that this approach merits more attention than it has been getting. The fact that accountants might dislike it, because it forces them to do something that is inherently rather tuncertain, does not prevent its being potentially better borh for investors and for the tax system - the latter, because once companies have chosen IRFE treatment that may create lock-in wholly apart from the tax analysis.
(c) What if the FASB simply repealed the IRFE rule? - From the standpoint of informing investors, this is no panacaea, given the overstatement in many or most cases of actual expected U.S. liability. While we often hear that financial accounting should be "conservative," under-valuing companies can be as harmful to investors as over-valuing them, and can also create undesirable special advantages for the better-informed. But it would certainly be interesting to observe how greatly (or not) managerial behavior would change if keeping funds abroad, despite creating a likely true economic benefit (if the U.S. repatriation tax rate might be significantly lower in the future), created no accounting benefit.
BTW, no colloquium next week - it's our spring break. We will resume on Monday, March 20 with Daniel Hemel's The Federalist Safeguards of Progressive Taxation.
Wednesday, March 01, 2017
Voice of the people
Too tired to read, I was watching the GSN game show Divided, which requires multiple contestants, none of whom know each other, to agree on the answers to various questions, before at the end they must agree on a 60-30-10 split of their winnings. (That's the most interesting part, of course.) In fairly short succession, there were two questions about the U.S. fiscal system - one subjective (what did survey respondents say?), and the other objective. Each was interesting in its way.
Subjective question: 100 people were asked, which is the most unfair tax system? The available choices were (A) gasoline taxes, (B) income taxes, (C) excise taxes on cigarettes and alcohol.
All 3 of the contestants agreed that income taxes would be called the most unfair - they disagreed on whether gas taxes or cigarettes/alcohol would be ranked #2.
Correct answer, in terms of predicting what the survey respondents said: A - B - C (gas taxes deemed most unfair, then income taxes, then cigarette and alcohol taxes).
Making this harder to interpret: We have 2 Pigovian taxes wrapped around an ability to pay tax. Maybe the survey respondents' collective rationale is: externality plus "sin" should be taxed, but externality without "sin" shouldn't be taxed.
Or perhaps it is that taxing "necessities" is worse than taxing ability to pay (since the respondents may feel they need their cars), but taxing "luxuries" (or sins) is better than taxing ability to pay.
Then again, maybe I am trying to make too much of this, when you consider how Divided contestants dealt with the objective fact question about the fiscal system. They were asked: On which of the following items does the federal government spend more than 5% of tax revenues? (A) the Pentagon and the military. (B) Health care programs, (C) science research?
For this one, they would have gotten $25,000 had they immediately answered it correctly. But instead, they let the time tick down for a while - reducing the value of a correct answer - and then got it wrong by answering: A only.
They were right, of course, about military spending, and also about science research's missing the cut (the host said it's at about 1%), but oh so wrong about healthcare. (I mean, c'mon - Medicare? Medicaid?)
I guess they'd never heard it said that the U.S. federal government is an insurance company attached to an army.
Subjective question: 100 people were asked, which is the most unfair tax system? The available choices were (A) gasoline taxes, (B) income taxes, (C) excise taxes on cigarettes and alcohol.
All 3 of the contestants agreed that income taxes would be called the most unfair - they disagreed on whether gas taxes or cigarettes/alcohol would be ranked #2.
Correct answer, in terms of predicting what the survey respondents said: A - B - C (gas taxes deemed most unfair, then income taxes, then cigarette and alcohol taxes).
Making this harder to interpret: We have 2 Pigovian taxes wrapped around an ability to pay tax. Maybe the survey respondents' collective rationale is: externality plus "sin" should be taxed, but externality without "sin" shouldn't be taxed.
Or perhaps it is that taxing "necessities" is worse than taxing ability to pay (since the respondents may feel they need their cars), but taxing "luxuries" (or sins) is better than taxing ability to pay.
Then again, maybe I am trying to make too much of this, when you consider how Divided contestants dealt with the objective fact question about the fiscal system. They were asked: On which of the following items does the federal government spend more than 5% of tax revenues? (A) the Pentagon and the military. (B) Health care programs, (C) science research?
For this one, they would have gotten $25,000 had they immediately answered it correctly. But instead, they let the time tick down for a while - reducing the value of a correct answer - and then got it wrong by answering: A only.
They were right, of course, about military spending, and also about science research's missing the cut (the host said it's at about 1%), but oh so wrong about healthcare. (I mean, c'mon - Medicare? Medicaid?)
I guess they'd never heard it said that the U.S. federal government is an insurance company attached to an army.
Subscribe to:
Posts (Atom)
