Increasing Distortions and Feeding Leviathan: The Internet Sales Tax
Tuesday, April 30, 2013
The debate over state taxation of Internet sales has brought to the surface an important difference among economists over approaches to tax reform. Many (including some of my colleagues at the American Enterprise Institute) support the traditional argument for policies promoting neutrality, or the elimination of economic distortions caused by tax policies. This perhaps is summarized crudely by the phrase “equal treatment of equals”: if brick-and-mortar sellers (and their customers) must pay sales taxes, then an exemption for Internet sales represents an artificial advantage, and thus a distortion. In this view, the government budget — both its size and composition — essentially is independent of alternative mixes of taxes, and the central question is how to finance that budget at the lowest possible economic cost.
But if the budget is affected by the choices among taxes, then the neutrality approach is too limited. There is an alternative view, dominant among public finance economists oriented toward the political (“public choice”) analysis of government behavior: the size and composition of the budget are affected by the tax system. Alternative tax regimes lead to different spending outcomes, as driven by the pressures that are created by political competition under democratic institutions. Moreover, government itself is an interest group with powerful incentives to display Leviathan-like growth, in particular in the context of the wealth transfers among interest groups that result from choices among alternative spending and tax policies.1
Accordingly, a more useful perspective for the Internet taxation question is this: would rational individuals striving to maximize their well-being choose to tax Internet sales? The answer turns out to be no, if we view the relationship between the citizenry and the state as contractual (or constitutional): taxes are prices paid for the aggregate costs of delivering public services, and those prices should reflect the differing valuations placed on those services by taxpayers. In this orientation, the citizen/state relationship is analogous to that of individuals engaged in voluntary market transactions. That is the only view consistent with the preservation of liberty.
Local customers do not demand services tied to their out-of-state Internet purchases.
And that is why the economic efficiency of a shift in tax policy — the analysis of distortions — cannot be evaluated without a discussion of the values of the public services being provided. If the elimination of a (purported) distortion results in a government budget that provides less value for taxpayers, then the policy shift might be perverse. That value is determined by the underlying preferences of individual taxpayers, just as those preferences determine prices in private-sector markets. And just as individuals make choices among the myriad goods and services produced in the private sector, they (implicitly) make choices among different bundles of public goods, particularly in the context of their location decisions, and also through their voting behavior.
Specifically, if an individual does not consume a given private-sector good, he does not pay for it. If a taxpayer in the context of an ordinary market transaction does not demand services from a given governmental unit, and thus does not impose costs upon that government, there is nothing “efficient” about a tax payment mandated from that taxpayer to that government.2
Brick-and-mortar sellers and their customers demand local government services. Obvious examples are police and fire protection. The same is not true for out-of-state sellers and their in-state customers, who do not demand services from the local governments in the context of these particular transactions. In other words, local customers do not demand services tied to their out-of-state Internet purchases.
That is the issue not addressed by my AEI colleagues Alex Brill, Alan Viard, and James Pethokoukis — all brilliant, rigorous, knowledgeable, learned, scrupulously honest, and highly collegial, but all focused too narrowly on the conventional neutrality view of efficient tax policy. Again, that view largely ignores the issue of the value of public services by essentially treating the size and composition of the government budget as independent of tax prices.
It is no accident that it is the big sellers — brick-and-mortar Wal-Mart and Internet Amazon — who hope to derive a competitive advantage from this proposal.
Alex Brill argues, quite correctly, that the federal income-tax deduction for state and local taxes distorts the fiscal choices of those lower levels of government by allowing them in effect to shift some of the costs of their spending onto federal taxpayers writ large — that is, taxpayers in other states. This is particularly the case for high-spending states that impose “higher income-tax rates on high-income individuals … most likely to itemize.”
So far, so good. But Alex then argues that Congress should eliminate this tax preference and replace it with legislation allowing state taxes on interstate Internet sales, as a substitute tax instrument with which state revenue losses could be mitigated while reducing the purported distortion created by the taxation of brick-and-mortar sales but not of Internet transactions. But in the context of Alex’s application of the neutrality principle, there is a larger reality: both the deductibility of state taxes from federal income taxes and state sales taxes imposed upon out-of-state Internet transactions create the same kind of distortion. They impose the costs of state government services on individuals not demanding those services (on the margin); that is, both weaken the relationship between tax prices and the value of public services as perceived by taxpayers.
Alex might respond that the deductibility of state and local taxes from the federal income tax base shifts the costs of state services onto taxpayers elsewhere, while the in-state taxation of Internet sales merely “permit[s] states to collect the sales tax they are owed” from in-state residents, just as in-state residents pay sales taxes at brick-and-mortar businesses.3 Not so fast. Because out-of-state sellers and their in-state customers demand fewer (or no) services from local governments in the context of Internet transactions, in effect they impose fewer costs upon the local governments, and so it is appropriate that less tax be collected. Alex argues that the Internet “doesn’t deserve a tax advantage.” Yes it does.
Alan Viard, in a detailed and rigorous analysis, argues that “the real economic cost of the failure to collect use tax on Internet sales … comes from the distortion that arises when some sales are taxed and others are not.” Again, this distortion argument depends crucially upon the implicit assumption that all such sales impose similar costs upon the local public sectors, or, equivalently, that the respective groups of sellers and customers have similar demands for the services provided by the state and local governments. For the reasons discussed above, this implicit assumption is deeply problematic.
Moreover, unless a new tax simply replaces an existing one without an effect on the budget, we cannot evaluate the efficiency of the policy shift without an examination of the value of the public services being provided. At this point a conceptual experiment is useful: suppose that the new revenues yielded by the taxation of Internet sales allow state and local governments to acquire additional resources that then are simply squandered. We would have eliminated a “distortion,” but with a perverse result: a smaller private sector and a government spending more but providing no more services. Aggregate wealth would be reduced. “Distortions” notwithstanding, it is difficult to see the source of an economic gain in that world. More realistically, suppose instead that the government does not simply waste the new resources, but that its use of them creates less value than had the resources remained in the private sector. The application of the sales tax to Internet sales again would eliminate a “distortion,” but would result in the replacement of more valuable private-sector services with less valuable public-sector services. Yet again: what is the precise nature of the economic gain?
The federal income-tax deduction for state and local taxes distorts the fiscal choices of those lower levels of government by allowing them in effect to shift some of the costs of their spending onto federal taxpayers writ large — that is, taxpayers in other states.
Alan might respond that the goal is simply to replace one tax with another, less distorting one. But then we return to the central question: such distortions cannot be evaluated without analysis of the relationship between the burdens of the respective taxes and the demands for government services from those bearing the taxes. If that relationship changes, then the government’s budget will change; it is not predetermined. Even the distortion argument itself is less than obviously correct, since brick-and-mortar sellers under the Internet-taxation legislation now under consideration would be responsible for collecting only the in-state sales tax regardless of the domiciles of their customers, while Internet sellers would be left with the administrative burden of collecting taxes imposed by the vast array of states and localities in which their customers reside. It is no accident that it is the big sellers — brick-and-mortar Wal-Mart and Internet Amazon — who hope to derive a competitive advantage from this proposal. Would that outcome reduce “distortions”?
Jim Pethokoukis offers a related argument: that this broadening of the sales tax base would allow for a reduction in state marginal income tax rates, thus moving the overall tax system toward a consumption base as opposed to an income base. Assume for the moment that this tax shift would be revenue neutral. We are still left with the problem of a tax shift reducing the correlation between tax prices and the marginal values of public services, even under the broader assumption that a consumption tax would reflect such demands more closely than an income tax. It is far more dubious to assume that this particular consumption tax applied to out-of-state transactions would improve that correlation.
And the revenue neutrality assumption itself is dubious; instead, the new revenues are far more likely to make the government bigger and the private sector smaller. Jim criticizes this objection to the Internet tax as an assumption that “the pro-growth argument can’t win and thus we should never do tax reform.” Again, not so fast. Under the (reasonable) assumption that the taxation of Internet sales decreases the degree to which the demanders of public services bear the attendant costs, it is not hard to predict that the beneficiaries of government spending will demand (vote for) more of it. Perhaps those paying the new tax — Internet buyers — would demand less government, but it is equally plausible that they will demand compensation from other interests through the bargaining processes of representative democracy. If the median voter is an Internet buyer, it is not hard to construct a simple model in which, for example, the majority (50 percent of the voters plus one additional voter) cuts the provision of collective goods by one dollar per voter, and then transfers to itself two dollars per member of the majority.4 Obviously, many other outcomes are possible, but a dismissal of changing political equilibria under an altered tax regime is not useful in terms of a policy discussion. If “reform” yields a government sector that is bigger, more wasteful, and less constrained by taxpayer preferences, perhaps “reform” is the wrong characterization of the policy shift.
The politics of wealth redistribution are deeply corrosive of the relationship between the citizenry and the state. A policy shift that weakens the link between taxes imposed and benefits received by definition yields wealth transfers, and the vast array of perverse incentives attendant upon them.5 Prominent among those are strengthened incentives for rent-seeking and perhaps weakened incentives for work and investment. The traditional neutrality view of efficient taxation largely ignores this dimension of the Internet taxation issue, notwithstanding its very great value in terms of the general analysis of the effects of policy changes on market behavior. Those analytic findings are not the same as “efficiency.” Government behavior is a crucial parameter also, a reality that all economists — including yours truly — must strive never to forget.
Benjamin Zycher is a visiting scholar at the American Enterprise Institute.
FURTHER READING: Zycher also writes “Earth Day and Four Decades of Fear” and “Would a Carbon Dioxide Tax Be ‘Efficient’?” James V. DeLong discusses “Protecting Property on the Internet,” Roger Bate asks “Do Special Enterprise Zones Undermine Capitalism and Our Safety?,” and Max Eden contributes “The Internet Sales Tax: How Congress ‘Not Working’ Works Very Well.”
1. Adam Smith offered a keen insight, to wit, that a growing market increases the number of interest groups: “That the Division of Labour is Limited by the Extent of the Market,” from An Inquiry Into the Nature and Causes of the Wealth of Nations, Book I, Chapter III.
2. For an early discussion of the implications of the contract model of the citizen/state relationship for tax policy, see Geoffrey Brennan and James M. Buchanan, “Towards A Tax Constitution for Leviathan,” Journal of Public Economics, Vol. 8, No. 3 (December 1977), pp. 255-273. For a lengthier treatment, see Geoffrey Brennan and James M. Buchanan, The Power to Tax: Analytical Foundations of a Fiscal Constitution, Cambridge: Cambridge University Press, 1980, esp. pp. 1-12 and 55-82. For an excellent nontechnical summary of the basic arguments, see James M. Buchanan, “The Constitution of Economic Policy,” Prize Lecture to the Memory of Alfred Nobel, December 8, 1986.
3. Brill is assuming that the incidence (burden) of the state sales tax imposed upon Internet sales would be borne by in-state consumers rather than out-of-state sellers. This is likely to be correct in the long run under a reasonable assumption about supply conditions (“elasticities”).
4. If the median voter is an Internet purchaser, one might ask why the state would choose to tax Internet transactions. One answer might be that the tax itself changes the identity of the median voter. Another might be that the tax is enacted as a result of side payments (or compensation) of some kind made to a subset of the majority.
Image by Dianna Ingram / Bergman Group