Yesterday at the colloquium, David Albouy presented the above article, applying optimal tax theory to the question of how consumption via home occupancy should be taxed. This was welcome diversification of the balance in our overall portfolio of articles for the semester.
I also got an update that I shouldn’t have needed (i.e., I ought already to have been up-to-date) regarding how Albouy has, in prior work (The Unequal Geographic Burden of Federal Taxation), modified or corrected the view taken by Louis Kaplow (in Regional Cost of Living Adjustments in Tax/Transfer Schemes, Tax Law Review, 1995) and Michael Knoll with Thomas Griffith (in Taxing Sunny Days: Adjusting Taxes for Regional Cost of Living Adjustments, Harvard Law Review, 2003) who concluded that generally the income tax system ought not to take into account regional cost of living differences.
By taking a fuller view of how a national labor market with regional wage and price differences operates, in the presence of a tax system that hits wages but not amenities that are accepted in lieu of wages, Albouy finds that “workers in cities offering above-average wages— cities with high productivity, low quality of life, or inefficient housing sectors—pay 27 percent more in federal taxes than otherwise identical workers in cities offering below-average wages. According to simulation results, taxes lower long-run employment levels in high-wage areas by 13 percent and land and housing prices by 21 and 5 percent, causing locational inefficiencies costing 0.23 percent of income, or $28 billion in 2008. Employment is shifted from north to south and from urban to rural areas. Tax deductions [that take account of regional cost-of-living differences] index taxes partially to local cost of living, improving locational efficiency.”
Thus, without impugning the logic employed by Kaplow, Knoll, and Griffith under their assumptions, it may be that one should adopt an opposite conclusion regarding the bottom line question of whether the tax system ought to address regional wage and price differences. Arguably, the better view is that it should so adjust, due to the distortion that results from taxing wages while not taxing the imputed income (in a broad sense) that a low-wage region may offer at equilibrium in lieu of cash.
Anyway, on to the current article, which is closely related to the earlier one in focusing on the regional distortions that result from taxing cash wages but not their in-kind substitutes such as good weather. Here is an expanded version of my own thoughts about the article. It addresses 3 topics: (1) housing and the work-leisure choice, (2) other inputs to how we should tax housing, and (3) political economy and fiscal federalism considerations.
(1) Housing and the work-leisure choice
Since income (and other related) taxes hit work but not leisure, it’s theoretically agreed that, while we should generally tax all commodities equally, this is subject to the proviso that we should tax those that are leisure complements at a higher rate, and those that are work complements at a lower rate, in cases where we can identify such commodities. This has the efficiency benefit of somewhat offsetting the tax system’s underlying discouragement of work relative to leisure.
Less well-settled is the question of whether there is much to gain practically by looking for leisure complements and work complements. But housing is a very important element of overall consumption that clearly might have systematic relationships to this question. So not looking there would be foolish, yet little has been done on this question in previous work.
As background to this inquiry, the current federal income tax system heavily favors home ownership. But, on the other hand, state and local real property taxes tend to burden home consumption relative to other consumption. The net balance is probably pro-home consumption, but one should keep both pieces in mind.
The article identifies several dimensions to locational choices that might affect how we might like to tax home consumption, in view of its interaction with the work-leisure relationship.
(a) “Hawaii versus Manhattan” – To typify this distinction simplistically for clarity’s sake, Hawaii offers nice weather and beach access, which are nontaxable amenities albeit built into housing prices. Manhattan instead offers two distinct kinds of taxable amenities that are also built into housing prices. The first (earning amenities) is that you may be able earn a lot more if you live in Manhattan than in Hawaii. The second (consumer spending amenities) is that you may be better situated to buy nice things for daily consumption if you live in Manhattan than in Hawaii. For example, consider all our restaurants.
(b) “Westchester versus Manhattan” – Second, even if you work in a high-wage area with low nontaxable amenities, you can either live near work, or else some distance from which you commute. As I’ll discuss in section (2), there are several reasons outside the simple optimal tax model why we might take an interest in commuting. But even just within the basic model, the paper offers evidence suggesting that long commutes tend to crowd out marginal work, more than marginal leisure.
The basic argument is to tax Hawaii housing because its nontaxable amenities are a leisure complement and a work substitute, along (perhaps more contingently) with Westchester housing because the commuting also operates at the margin as a work substitute, while subsidizing Manhattan housing because its two types of amenities are work complements / leisure substitutes. You earn more instead of choosing nice weather and the beach due to the earning amenities, and you use restaurants instead of cooking due to the consumer spending amenities.
This argument makes good sense to me – perhaps no surprise, given that I am a Manhattanite – and of course it dovetails nicely with Albouy’s earlier work concerning the unequal geographic burden of federal taxation. One point I might add, however, is that Manhattan may differ from Hawaii and Westchester with respect to the marginal effect on work versus leisure of increasing one’s house size, given that one lives in a given location. In Manhattan, all you need is a roof over your head to have a shot at realizing the earning amenities. But once you actually have a kitchen, along with enough space to restrain the urge to go out all the time, you may start substituting away from the consumer spending amenities. So possibly the basic optimal subsidy for the Manhattan housing location should be supplemented by a larger marginal rate of tax on increasing house size in Manhattan than in the other two locations.
(2) Other inputs to how we should tax housing
The paper notes 4 main inputs to how we might want to tax housing, other than those involving the work versus leisure choice.
(a) Positive externalities to urban agglomeration – These might also support a Manhattan subsidy.
(b) Henry George case for a land tax on site value – The famous Henry George argument for taxing site value (as distinct from improvements), because land is relatively fixed and yet taxing it tends to be progressive, is one of those things that is potentially important, generally accepted theoretically, and yet generally ignored. It’s more important than ever in a Piketty era – especially when it’s been argued that a lot of Piketty’s finding reflect real estate value hikes, rather than a generalized r > g. Even taxing site value plus the value of improvements (i.e., housing) may retain some of the lump sum tax elements of the pure Henry George land tax, despite its discouraging the improvements.
(c Other commuting issues – One may also want tax housing that is associated with commuting if commuting imposes other social costs that for some reason cannot be taxed more directly. For example, if we fail to adopt proper Pigovian taxes on the pollution associated with car travel, and also don’t adopt proper congestion pricing for rush hour traffic, taxing the housing in communities associated with such travel may be better than nothing.
In this regard, it’s worth noting that the U.S. income tax system actually does discourage commuting a bit. We don’t allow commuting costs for going to one’s primary place of work, whereas some other countries (e.g, Germany do). Purely from the standpoint of measuring income, neither approach is fully correct.
To illustrate, suppose Person A has a job and is choosing between two places to live, one of them downtown with higher rent and lower commuting costs, and the other in the suburbs with lower rent and higher commuting costs. In this scenario, the U.S. income tax system properly disallows commuting costs as, in effect, rent substitutes.
But suppose Person B has a fixed home and is choosing between two jobs, one near home with a lower wager and lower commuting costs, and the other some distance away with a higher wage and higher commuting costs. In this scenario, the German system, rather than the U.S. one, rightly frames the taxpayer’s presumed marginal choice.
While I suspect that the U.S. approach is empirically better on balance from an income measurement perspective, the fact that the Person B scenario sometimes exists suggests that we are “inefficiently” (all else equal) disfavoring commuting a bit, as compared to getting it right all the time. But this may be a good feature, rather than a bug, if we also independently have grounds for tax-disfavoring commuting.
(d) Housing consumption by the poor – There is probably little or no good reason for generally favoring home consumption relative to other consumption. But if homelessness has negative externalities (in addition to being very bad for those who face it), we may want to subsidize housing consumption by the poor, in lieu of just giving them enough aid to fend off homelessness.
(3) Political economy / fiscal federalism
The issues that the paper presents regarding how to tax housing in Manhattan versus Hawaii versus Westchester are singularly those of interest to a national-level decision-maker – not one who is setting tax policy (even optimally from the standpoint of residents) for any of those localities. This is of especial interest given that the relevant tax instruments include real property taxes that are set at the state and local levels. Obviously, these issues belong in a wholly different paper, but given the Albouy paper they are worth noting.