Yesterday afternoon, we held the first public session of the 2026 NYU Tax Policy Colloquium, now in its 32ndconsecutive year. The paper we discussed was Daniel Hemel’s and Adam Kern’s Separation of Bases and the Fiscal Constitution, which is forthcoming in the Texas Law Review.
For some years I used to write blogposts regarding each paper that we discussed in the Colloquium – based on the papers themselves, not the public sessions, because the latter are off-the-record.
I stopped doing this a couple of years ago because I was finding the time demands of posting difficult to meet. This reflected that, while my comments here are somewhat casual and off-the-cuff, and by no means require or receive the standard of care that I would demand of myself if I were writing a publishable piece, they also aimed at being thorough and fair – meaning that they took a bit of time to do properly.
This year, I expect to find that I’m able to post after each public session. The difference is that I am once again (alas, for what I think is the last time) co-teaching the colloquium with Lily Batchelder. With her sharing the responsibilities, I get just enough relief to be able to add this to my list of things to do.
So anyway, back to Separation of Bases and the Fiscal Constitution. Let me note up front that the authors are not only friends whom I know well, but rightly prominent in tax law academia, each ranking among the very best scholars in (at a minimum) his particular age cohort. That said, I had some problems with the article. But there is something institutionally of note about this. The article’s flaws, if that’s not too strong – or, to put it more neutrally, the ways in which it was not written to my personal taste – reflected the incentives for tax and other legal scholarship that arise because of the role that student law reviews play in the publication process.
How do you “sell” an article to leading law reviews? Perhaps the best formula there is, which this article follows, is to say something like the following: “Everyone thinks X is the right way to think about this set of issues. But that’s all wrong. Actually, the right framework is Y. By showing that Y, rather than X, is correct, this article fundamentally changes the conventional wisdom.”
And it’s better still if one can show that adopting Y in lieu of X has lots of important applications, including to a few of the hot legal topics of the day.
The article aims to do all this, in my view not always entirely convincingly. In doing so, it potentially obscures the valuable contribution that it does in fact make, which is as follows:
In a federal system with overlapping national and sub-national governments, a question arises as to how the different levels should coordinate their tax systems. Suppose, for example, that they each might tax income, consumption, real property, wealth generally, cross-border trade, and the like, Under a “separate the bases” view, distinct tax bases should be assigned to each level. (For convenience, let’s just say this concerns the national level and top subnational level, such as states in the U.S. or provinces in some other countries, amalgamating “local” with “state” for convenience.) Thus, we might have only the feds taxing income, and only the states taxing real property, to give one possible implication.
The paper rejects the “separate the bases” generalization, which it characterizes (with some hedging) as the conventional wisdom. It urges instead that tax base assignment to the different levels turn on what it calls the “relative externalities principle.” Here the claim is that, when both levels tax the same base, there is both a positive externality and a negative one. The negative one (emphasized by “separate the bases” devotees) is that, by “overgrazing” the same base, the two levels both cost each other revenue and increase the efficiency costs of the tax, given that taxpayers will presumably respond to the overall marginal rate they face. But there is also a positive externality from enforcement synergies that arise from both levels’ administering the same thing. For example, if one level finds an erroneously excluded wage item, the other level will presumably benefit as well.
There’s lots more – for example, a historical review of the U.S. Constitution’s Direct Tax clause, and a game theory section showing that, say, the U.S. national government might conceivably increase, rather than reduce, its overall power and influence versus that of the states by binding itself, say, not to directly tax real property. This part I won’t discuss here, but I found the setup a bit artificial to support applying the conclusion to such real world settings as the Direct Tax clause’s relevance today to the constitutionality of an unapportioned federal wealth tax (or a tax on unrealized income from real property).
Okay, back to the sales pitch of “The conventional wisdom says separate the bases, but that’s wrong and we should instead follow the relative externalities principle.” I have an objection to each.
For the claim about the CW, it’s that I don’t think “separate the bases” is indeed the conventional wisdom in tax law scholarship. If it were, there would be lots more talk about how, say, since the federal government taxes income the states should get out of the business and cease doing so. But OK, as the paper shows, there is an economics sub-literature or two in which this may indeed be the CW – even though, as the paper notes, such still-relevant landmark economics work in the field as that of Richard Musgrave and Wallace Oates expressly rejects the “separate the bases” principle. So here it’s just a bit of overselling that the student law review editors might be expected to love. (Sorry, I mean no disrespect of these talented, earnest, and hardworking individuals who are quite reasonably trying to position their journals as full of significant must-reads. It’s just that they’re reading in dozens of distinct fields which they can’t possibly know well. The problem is the system, not the individuals.)
Perhaps more importantly, I have a core problem with some of the paper’s discussion regarding how best to implement the relative externalities principle. In a word, or rather three, I discern a problem that I’ll call tax base essentialism. To some extent, the paper treats legally distinct types of formal tax bases – for example, an income tax and a real property tax, although one might also add consumption taxes, wealth taxes, tariffs, and the like – as if they were, in effect, distinct fundamental subatomic particles. By analogy, this is what a standard optimal income tax analysis does with such fundamental subatomic particles as having a wage rate and choosing between work/market consumption and leisure. But those are truer subatomic particles than the income tax, the real property tax, and the like, which are complex multipart instruments with substantially overlapping, and often widely varying, incentive effects and administrative / enforcement features.
To apply what I call the “battle of the externalities” as between discrete tax bases, rather than to all choices generally, one would need each tax base to be unique and distinctive in 2 ways:
(a) what it disincentivizes for the negative externality,
(b) the enforcement overlap for the positive externality.
Each of these preconditions can be questioned.
The paper agrees that taxing one base can crowd out revenues from another base, not just from the same one. But this is not just a side-problem – it’s fundamental.
In the abstract, what activities do these tax bases, as considered in the abstract, deter or burden or disincentivize?
A consumption tax, such as a retail sales tax or a value-added tax, deters work (which generates earnings that can be used for market consumption).
An income tax deters work and saving.
A wealth tax deters work and saving.
A real property tax deters work and saving and the use of real estate for consumption or investment.
A tariff: deters work and saving (the latter, since it applies to business inputs), and cross-border activity.
Given these overlaps, I think the over-grazing problem can’t best be analyzed at the tax base level. Rather, it calls for looking more narrowly at particular applications. For example, income taxation of home ownership overlaps more with real property taxation than does income taxation of financial assets, since work and saving are common to all but the real estate aspect arises in the one income tax application but not the other.
Let’s turn now to administrative overlaps. Here the issue is that positive (and perhaps also negative) administrative and enforcement spillovers may apply between different tax bases, not just the same ones. For example, in a country with a VAT, some of the same information may aid with both income tax and VAT enforcement. This may also happen to a degree in the U.S. as between income taxes, and either or both of retail sales taxes and real property taxes.
A further problem that I’d attribute to tax base essentialism pertains to the question of which levels of government are good or bad at administering particular tax bases. Consider the classic mixed case of wealth. States and localities may be better than the feds at valuing and taxing local real estate. But surely the feds are better at, say, tracking down financial assets. So, for a wealth tax, while the feds might do it better overall, one of its elements might be done better at the state and local level.
This is of interest in relation to the paper’s analysis of why it might make sense for the feds to renounce real property taxation, leaving it to the states so as to influence their tax base choices in ways that the feds might conceivably care about. With wealth taxation having multiple parts, some of which are better left to the states than others, one gets a mixed verdict as to whether federal renunciation makes sense here. Whether or not a federal wealth tax is a good idea (which turns on issues wholly apart from federalism), it shouldn’t be renounced for federalism reasons even if there is a piece of it that might wisely be renounced if considered in isolation.
There’s lots more, well worth reading even though I don’t agree with it all, but perhaps this is enough discussion for here.
