Tuesday, January 2, 2018

SALT margins

I think most of the debate misses an important point about the state and local tax deduction -- incentives.

Suppose you are in the top, (roughly) 40% marginal federal tax bracket.  If you pay an extra $100 in state taxes, you deduct $100 from income, and pay $40 less in federal taxes. So, you really only pay $60 in state taxes. The federal government effectively transfers $40 to the state from taxpayers in other states.

That's a big incentive to raise money as deductible taxes from high-bracket tax payers! This incentive doesn't work if the state raises taxes from lower bracket taxpayers.

California's tax system, for example, seems to respond heartily to this incentive. California's income tax rate is highly progressive, topping out at 13.3%.  As reported in the Sacramento Bee
Nearly 90 percent of the money [income tax receipts]  comes from one-fifth of the taxpayers – those making $91,000...Forty-five percent of the state’s income tax money comes from the top 1 percent of filers – those with adjusted gross income of at least $501,000.
and, therefore, in the highest Federal tax bracket, and also likely to itemize deductions.

The state of California relies a lot on income taxes. The state of California gets 65% of its revenue from individual income taxes, 22% from sales taxes, and 8% from corporate taxes.

The usual stories about California rest on its progressive, redistributionist politics, and there is certainly much of that rhetoric around. But it also happens to be responding perfectly rationally to a strong incentive.

Conservatives have long objected to governments that spend other people's -- taxpayers' -- money unwisely.  But money raised from taxpayers from other states, who do not vote in your elections, provides doubly bad incentives.

This strikes me as a potent economic argument against deductibility of state and local taxes, that hasn't been made loudly enough. Of course no argument (pro or con) based on incentives has been made loudly enough!

This strikes me as theme of many things needing reform America. The New York Times  report on astounding infrastructure costs rang a nerve. Most infrastructure is directed by state and local officials, who spend Federal money.  Medicaid remains a matching fund -- the state spends a dollar, the federal government chips in a dollar.  The move to turn it in to a block grant, which failed last summer, would have removed this incentive.  Federal money and state control is a common pattern in social programs, and John Cogan explains well the legal and institutional reason for this separation in our history. But it leads to atrocious incentives.

Perhaps a general reform lesson is that states should have to raise the marginal dollar for anything from their own voters.

PS. I was long for the deductibility of SALT on the grounds that it keeps down the total tax rate. If both federal and state charge 50%, with deductibility you keep 25% of your income (Federal takes 50%, state takes 50% of what's left, so you have 25%.) With hundreds of different taxes, it is possible to exceed 100%! However, I've been persuaded these incentives are more important. Moreover, if paying twice is such a problem, dear California, you can make federal tax payments deductible from state income. I do not hear a groundswell for this solution.


A few commenters (and by email) claim that California pays more to the federal government than it gets back. The point is not the overall subsidy, the point is on the margin. A marginal $100 leads to a marginal $40 cross subsidy, no matter who pays who on average.

Also do not confuse margins and levels for people. The loss of SALT deductibility hurts a lot less on average than on the margin. Someone paying 13.3% marginal state tax in California only pays that rate on income over $1,000,000. So their total loss from the loss of SALT is not equivalent to paying federal taxes on the full 13.3%.


The original post contained the following paragraph, which was wrong, so I took it out.
Property taxes in California raise about $60 billion, roughly equal to the total raised from personal income taxes.  Since the state is home to extremes of housing values, thanks to land use and building restrictions, property taxes are also raised largely from people in high tax brackets, and therefore benefiting from the 40 cents on the dollar subsidy from the rest of the country. (This is a guess. If you know of data on property tax by income brackets, I'd like to see it.) 
It occasioned the following updates and the  next post which really shows how wrong it is.


A reader sent me links to two wonderful resources, "real estate taxes paid by income bracket, and by state, available as part of the IRS Statistics of Income:

California here: https://www.irs.gov/pub/irs-soi/15in05ca.xlsx (Lines 74-75)

All states and national summary here: https://www.irs.gov/pub/irs-soi/15in54cm.xlsx (Lines 75-76)

Individual state files and additional data: https://www.irs.gov/statistics/soi-tax-stats-historic-table-2

From the California analysis:

The row "real estate taxes, amount" and following addresses the question I asked in the blog -- to what extent is the apparently flat property tax actually progressive? It's not as much as I thought. The top two categories pay 37% and 14% of real estate taxes, though they bay 72% and 49% of income taxes. (The table verifies the state numbers on how concentrated California's income tax receipts are.) In retrospect, even overpriced real estate is a normal good -- people don't pay larger fractions of their income on real estate as they get more income -- so a flat real estate tax will raise more money from people with higher incomes, but not a higher fraction of income.

The table is fascinating, and a clear mine for blog posts. One big lesson that sticks out to me (ok, confirmation of something I've been mulling for a while) is just how meaningless income is as a social yardstick. Our policy discussions talk about "low income people" as if that is a permanent caste distinction. Yet look how many people with million dollar incomes are taking unemployment compensation! In the next row, just who are people with under $1 of income? You imagine homeless people roaming the streets of San Francisco. Well, of the 286,000 such people in California 51,000 are s-corporation owners who lost money, collectively $5 billion.

Update 2 

Morris Davis sends along the following data on property tax / household income.

Just how much are property taxes as a fraction of income? It's interesting for a lot of reasons. For one, total taxes  matter to the economy. Too many commenters decide that the top federal rate of 42% (last year) is low and we ought to tax people more. They forget state and local income taxes, sales taxes, excise taxes, corporate taxes, etc. etc. And property taxes.

Granted, we're veering off topic here, but here is the table. New Jersey, where Morris lives, has the highest tax rates. California is up there, but proposition 13 and a lot of low-priced inland areas must offset ridiculous house prices in the coastal areas.


  1. Didn't you mean Medicaid, not Medicare?

  2. How is the argument different that companies should be able to deduct state taxes? Moreover, what is the argument that companies should be able to deduct foreign taxes?

  3. If such incentives were so effective in influencing state tax policy, one would expect to find fairly progressive income tax structures at the state level. But one does not. Instead, the common state-level pattern is for generally flat rate structures, with fairly modest top rates. While I agree that, as a general matter, incentives are important in policy making, this particular incentive argument is not compelling.

  4. I don't know the details but my impression was that in the US system taxes from rich states like California are at least partially transferred to poorer states through various mechanisms. If this is true, SALT basically allows California to reduce this transfer. I don't know if there should be higher or lower transfers but your argument that Californians' ability to limit these transfers to poorer states should be restricted is not obvious to me.

  5. I have long been of the assumption that a tenet of conservative political philosophy is that, in the overall scheme of things, the role of states should be maximized and the role of the federal government should be minimized. Government should be closer to where citizens actually reside is part of that argument. If that is accepted, then, from that perspective, the "incentives" associated with the SALT deduction (correctly) cut both ways. (For similar reasons many conservatives will support the charitable deduction under the belief that this provides incentives to reduce the responsibilities of the federal government by supplanting them with private initiatives. I recall, for example, the "public/private partnership initiatives under Bush II).

    I would think that the incentive flowing the other way might be that, all else equal, the federal government collects less money and the state governments collect more. This *should* result in the states spending more and the federal government less than would otherwise be the case and it *should* affect the overall division of responsibilities. Unfortunately, it might also mean that the overall level of government (state plus federal) is higher, but that could mean that these incentives are not strong enough or that there is a lack of coordination. For this reason, I would favor more a more effective system of federal revenue sharing based on state population (combined with a delegation of authority and responsibility). Under such a scheme, the "progressive" result is that richer states effectively transfer to the poorer states (in a much more transparent manner). Eliminating multiple jurisdictional levels of income tax would make the US method of rasing revenue more efficient and equitable. If the residents of particular states want to spend more money than they are allotted, they would be free to raise it via other forms of tax, e.g., consumption and property tax which already exist.

    "Perhaps a general reform lesson is that states should have to raise the marginal dollar for anything from their own voters."

    Yes, but I seem to recall a recent comment (in response to a Peggy Noonan article) to the effect that "...not taking your money is not the same as gifting you money", heartily endorsed here by Professor Cochrane. I guess this is an exception to the rule whereby not taking money from the residents of certain states *is* a gift to those citizens (and the states in which they are resident).

    It's difficult to be consistent from a strictly ideological point of view.


  6. Great point. The New York Times article from 12/31 about Democrats in high-tax states "fighting back" against the loss of deductibility of SALT is a further example of states responding to this incentive to spend money from other tax jurisdictions. The federal government takes away deductibility of state-and-local income taxes, so states are saying ok, let's shift towards still-deductible payroll taxes and corporate taxes. Or even more egregious, we'll allow taxpayers to recharacterize their state income taxes as federally deductible charitable contributions.

    States will continue to look for ways to spend what you call "money raised from taxpayers from other states, who do not vote in [their] elections". Tax policy in a federal setting should be guided by the principle that they will always try this.

  7. So is this the formula for the average price effect for coastal real estate in California? Current Price - The discounted present value for 30+ years of MarginalTaxRate * (4% of 350k (750 to a million 1rst + 100k Heloc mortgage interest) + Current Property Taxes - 10k?)

  8. Note that the IRS tables only include property taxes for those taxpayers who itemize their deductions. For 2018 and after, that number will be a lot smaller than in the past.

    1. That's an important point---not only for the number of those who itemize, but also for the analysis as to whether real estate property taxes are "progressive", i.e., the main topic under discussion.

      My *guess* would be that those persons who are omitted from the table because they don't itemize are weighted more heavily toward the lower income groups. That is because 1) they would tend to buy less costly houses and thus have lower property taxes; and 2) would also tend to have other lower deductions (e.g., state income taxes, mortgage interest and charitable deductions). The effect would be that those lower deductions would result in many of their taxes being left out of the charts because they claim the standard deduction rather than itemizing (if I am correct, the AGI of those taxpayers would be included, but their taxes would be omitted from the charts). We also won't get into those who pay real estate taxes indirectly as renters...

      IRS statistics are very useful for many purposes; however, they also have their limits which should be understood: If the item isn't on a return, it essentially isn't counted. As I said in an earlier comment on a related topic, these really are not "facts", they are estimates.


  9. Excellent post. It is tiresome to hear Californians' arguments about contribution to the national and global economy. It's a subsidy at the margin, created by the Internal Revenue Code. Period.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.