Tax policy colloquium on Dorothy Brown's "Shades of the American Dream"

Yesterday we had a lively session, terminated by the clock when it was still going strong, on Dorothy Brown's "Shades of the American Dream."

The paper argues that the income tax rules for homeowners (exclude imputed rent, allow mortgage interest and real property deductions, disallow losses on sale but also exclude most gains) unfairly disadvantage African-Americans, who take less advantage of the rules than whites. They own comparatively fewer and less valuable homes, on average have lower marginal rates for the deductions, less frequently itemize deductions, and have disallowed losses a higher proportion of the time.

Home ownership differences remain even if one adjusts for income, though this appears to be mostly because, as between average African-American and white families with the same income for a given year, the latter is likely to have greater wealth.

There is a strong pre-2008 financial collapse flavor to the article (a problem for many of us if our articles are overtaken by events while in gestation), since it treats home ownership as a good thing that ought to be encouraged. I imagine that many of the lower-income people, both white and African-American, who bought homes in the last 10 years despite being pressed financially are now sorry that they did so.

Among the main issues that I raised was tax capitalization of the tax subsidy. If homes are more expensive by the full expected value of the tax benefits, then (a) all the whites who are getting these benefits actually aren't benefiting after-tax at all, (b) adding new tax benefits that non-affluent African-Americans could claim after buying a home might simply boost the price they would have to pay, potentially making them worse off if they faced liquidity problems (e.g., the need to pay 10 or 20% of the purchase price in cash).

I noted that marginal tax rate differences (along with itemizing versus non-itemizing) could lead to clientele effects that might change the analysis. For example, if I don't itemize or am in a low tax bracket and am bidding for a home against someone who gets more marginal value for the deductions, then all else equal that person will be willing to pay a higher price than I would. Given that homes are non-fungible assets being sold in what are often thin markets, this could lead to a reduction in the consumer surplus enjoyed by lower-bracket prospective home purchasers. E.g., one loses out on a home that otherwise one could have purchased at a favorable price, or alternatively one gets the home but has to pay a bit more. I thought Dorothy should consider and emphasize this issue a bit more, but I noted that housing markets may be sufficiently segmented by income to reduce its importance. E.g., even if people in the 15% tax bracket are buying homes, they probably aren't, much of the time, bidding on the same homes in the same neighborhoods as people in the 35% bracket.

In a similar vein, Alan Auerbach discussed how expected lack of home appreciation (e.g., buying housing stock in Rochester or Detroit when the city is well-known to be in long-term economic decline) may lower the purchase price, and permit one to receive a higher annual imputed rental value relative to the price than would have been available if appreciation were expected.

I was hoping Dorothy would engage more with these arguments than she did. They strike me as not implausible and also as pretty fundamental to the validity of the claims being made.

We also discussed a bit at the session what the financial meltdown tells us about the tax rules for home ownership. I think we've learned quite a lot. For decades, tax policy types thought we knew how bad the home ownership rules are. We recognized that they encourage enormous inefficiency in the form of substituting home consumption for other consumption simply because it's tax-favored by the exclusion of imputed rent plus the allowance of deductions for items (home mortgage interest and real property taxes) that are tax-arbitraged against it.

But we didn't know the half of it. We didn't fully realize how, when you throw in exclusion of nearly all sale gains plus a strong inducement to adopt maximum leverage, the tax system contributed to the real estate bubble and financial collapse of 2008. The wreckage is strewn everywhere, from the macro-economy to financial institutions to all the individuals who bought over-priced homes they can't afford and now have to deal with the back-end mess.

It's pretty clear to me how the tax rules ought to change, but unfortunately there is zero chance of this happening. Congress wants to shower new benefits (such as the first-time homebuyer credit) on homes that will keep prices too high and prevent rational adjustment. Obviously I don't want to exert strong downward pressure on homes right now, but that's the direction in which we ought to go with sufficiently deferred effective dates to permit adjustment for existing mortgage default problems.

The first big point is that, while there may be a modest rationale for encouraging home ownership relative to rental, there is absolutely no rationale for encouraging bigger or more expensive homes. The rationale for doing anything at all relates to positive externalities, e.g., caring more about the neighborhood and about upkeep that affects neighbors, although there also are arguments the other way (e.g., reduced mobility may increase adjustment costs partly borne by others when jobs migrate). The current tax rules fail to focus on just ownership because the benefits keep rising with home value until one reaches the $1.1 million ceiling on the debt principal that gives rise to deductible interest.

Step 1 would be to greatly lower the $1.1 ceiling. Step 2, convert it to a refundable fixed-percentage credit so that marginal rate differences between prospective buyers are eliminated. (Refundability aside, this is an old proposal - it was in the 1984 Bradley-Gephardt tax reform bill, and I believe it was also in the 2005 Tax Reform Commission's recommendations.) Step 3, sever the link to leverage, by making it a fixed dollar amount for each home (or for a percentage of purchase price up to a low ceiling) without regard to home mortgage interest paid.

But this leaves one other huge problem that the financial collapse reminds us is really important. A home is both a consumer asset and an investment asset. In theory, while a decline in home value from its being used ought to be non-deductible if we don't tax imputed rent (just as you don't to deduct the purchase price of a car for personal use that you eventually sell for scrap), investment gains and losses ought in principle to be fully taxed.

Thus, suppose you buy a home in 2006 for $200,000. Ignoring for simplicity its economic depreciation due to being used, suppose you end up selling it either for $300,000 in 2007 or for $100,000 in 2009. These $100,000 swings from the original purchase price are economic returns to investment that ought to be included, in the 2007 case, and deducted in the 2009 case.

It's common wisdom among tax policy types that, in some settings, having the income tax reach risk by including gains and deducting losses is potentially irrelevant. Thus, suppose I want to place a $100,000 bet with some counter-party on whether oil prices will rise or fall this year. It might make no difference whether the system took account of this bet or not. Thus, suppose we really want to bet $100,000, but that the gain is taxable and the loss deductible at 50%. Our response to exposing the risky outcome to the tax system might be that we simply double the nominal bet, from $100,000 to $200,000, so that after-tax we end up having the bet we really wanted.

Whatever one makes of this line of argument in other settings, it's largely inapplicable to home ownership. Here the problem is that lower and middle-income homeowners have a grossly under-diversified investment portfolio given how large a share they have sunk into home equity and the fundamental difficulties of hedging or diversifying that stake. So the insurance that the tax system could provide by including gains and deducting losses, rather than being irrelevant because people can achieve optimal hedging and diversification by themselves, actually responds to a market failure or missing market (or mistake in investment strategy by millions of people).

For a long time, this seemed less important because the real estate market was rising anyway. But now we have forcefully been reminded that there is downside as well as upside risk to home prices.

Allowing losses on home sales to be deducted while gains are not taxed strikes me as a huge mistake that could encourage new real estate bubbles in the future (or at a minimum continuing over-investment in home ownership). But taxing gains and allowing losses to be deducted upon the sale of a primary residence has a lot to be said for it. Given the lock-in problem for gains (and inducement to realization for losses), one could argue for doing it, on both sides, at capital gains rather than ordinary income rates, although this would reduce the insurance provided. One also might exempt it from the capital loss limitation (and also from having other capital losses allowed against it if a gain), given that it doesn't entirely fit the rationale for the capital loss limitation, which is that, without it, people would sell all the losers while holding all the winners in their asset portfolios. Finally, although this reintroduces some undesirable non-neutrality, one could consider permitting gain rollover when the sale price is promptly reinvested in a new home - the approach that the income tax law took before 1997, when it was changed to simply exempt up to $500,000 of gain. This in turn would increase the desirability of preventing the tax-free step-up in asset basis at death.

As a final detail, Richard Epstein, in his early days as a tax scholar rather than broad-ranging libertarian, wrote a Stanford Law Review article proposing to tax home gains and allow home losses with one adjustment, to capture the personal use element. This was to have home basis decline each year by the amount of tax depreciation typically allowed for real estate, only without the depreciation actually being deducted (given the tax arbitrage against excluded rental income). This annually declining basis would then be used to compute the fully allowable gain or loss on home sale.

There you have it: what I consider a pretty sound set of proposals for taxation of home ownership. I'd enact it today but with a deferred effective date or phase-in given the huge problems we now face with bad mortgages. The fact that it has so little chance of ever being seriously considered tells us something, I think, about the state of our political system and the broader likelihood that it is able to adopt sound policies that would bode well for our economic future.