Pre-emption: federal statutory intervention in state taxation.

It should be noted that the SSUTA by no means establishes an "economically satisfactory" sales tax regime, even if it does address some of the administrative problems that arise under current policy. In particular, retail sales taxes are generally quite poor approximations to destination-based consumption taxes. On the one hand, their tax bases are too inclusive, in that they tax many intermediate-goods transactions. As Ring (1999) has shown, the states differ widely in the proportion of sales tax revenues derived from taxation of sales to consumers. This figure is as high as 89 percent in West Virginia, but as low as 28 percent in Hawaii, averaging 59 percent for the nation as a whole. (The wide interstate variation in this respect is one important indicator of the extent of state revenue autonomy, and also of the hurdles to be surmounted if states are to achieve significant "base harmonization," a reform that would likely facilitate Congressional action in support of an initiative such as the SSUTA.) For the nation as a whole, then, it appears that more than one-third of sales tax revenue derives from transactions that would not be taxed under a consumption tax. On the other hand, state sales tax bases are not inclusive enough, viewed from the perspective of consumption taxation. Exemptions for food and clothing are widespread and well known. Just as importantly, especially in an economy with a growing service sector, expenditures on health, education, financial, and other services provided to households are also often exempt from sales taxation. (6) (Business services should be exempt from retail sales taxes that are intended to tax consumption, since these services are intermediate inputs, not final consumption.) For both of these reasons, retail sales taxes are far from ideal taxes from an efficiency perspective. Indeed, as noted below, state personal income taxes may approximate a tax on final consumption better than existing retail sales taxes do.

State Taxation of Pension Incomes

As is well known, personal income taxes, as they are implemented in practice, are not taxes on true economic income. Instead, they are "hybrids" of income taxes and consumption taxes. A consumption tax differs from an income tax in that it taxes the uses of income, at the time that it is consumed, rather than the sources of income as it accrues. Federal and state tax treatment of the income from retirement savings, capital gains, and other types of income produces a tax system that diverges substantially from a true income tax and that corresponds in important ways to a personal consumption tax. In particular, when households elect to save a portion of their earnings using tax-sheltered retirement savings in IRAs, 401(k)s, and other similar accounts, the return on their savings within these accounts is not subject to tax until the assets within the accounts are distributed upon retirement. This means that the economic income arising from the return to capital in these accounts escapes taxation. Similarly, employer contributions to employee pension plans as well as the return on these contributions are not subject to tax until they are distributed at retirement. Since the proceeds of these distributions finance retirement consumption, the taxation of distributions from tax-sheltered savings and pension accounts effectively shifts the personal income tax away from a tax on all sources of income to a tax on the uses of income when consumed, that is, to a consumption tax. The taxation of capital gains on a realization basis offers similar opportunities for households to opt out of a tax on economic income and Into a tax on consumption. By electing to defer realization of capital gains until the proceeds from asset sales are needed in order to finance current spending, households are again able effectively to convert the "income" tax to a tax on consumption expenditures.

These features of current federal and state personal income tax systems do not result in true or "pure" personal consumption taxation, since there are limits on the amount of nonwage income that can be sheltered from taxation as it accrues. Furthermore, early distributions from sheltered accounts are often subject to penalties, making them unattractive instruments for non-retirement savings. Consequently, a significant amount of nonwage income is taxed as it accrues, and to this extent the personal income tax diverges from a consumption tax and more closely approximates a true income tax. Existing federal and state income taxes are, thus, "hybrids" of income and consumption taxes.

The interstate mobility of households over the life cycle adds an interesting aspect to the question of consumption taxation at the state level. Suppose that a household earns wage income when residing and working in one state, say state A, directing a portion of this income into tax-sheltered retirement savings accounts. Suppose that the household moves to a different state B upon retirement and then receives distributions from its retirement savings accounts. This household's life-cycle consumption is now spread across two states. Assuming that both states impose taxes on personal income, the question arises as to whether distributions from retirement accounts "should" be taxed in A or B. If these distributions are taxed by state A, then its personal income tax base is the lifetime consumption of households that reside and earn wages there early in the life cycle. If retirement distributions are taxed in state B instead, then state A's personal income tax allows it to tax consumption early in the household's life cycle, while state B's personal income tax base includes the household's retirement consumption expenditures. These two alternatives may be referred to as "source based" and "residence based" taxation of distributions from sheltered accounts, and they result in source-based and residence-based consumption taxation, respectively. (7)

Needless to say, this simple illustrative example abstracts from many nontrivial complications, including the possibility that households may reside in several different states at different stages of the life cycle and may face graduated and time-varying income tax rates in different states. It serves, however, to show that the tax treatment of retirement distributions on a source or destination basis has important implications for state personal income taxation. Both types of taxation have some economic virtues. Source-based taxation allows states (state A in the example) to tax a household's lifetime consumption if it earns wage income there when young and takes advantage of opportunities to shelter a portion of this income from current income taxation. State-level taxes on lifetime consumption may be viewed as desirable on equity grounds, and, if so, the principle of horizontal equity might be interpreted to require that lifetime tax burdens be fixed independently of the state in which consumption occurs. Under this principle, residence-based taxation would be undesirable because it would result in unequal lifetime state income tax burdens for households that remain in one state for their entire lifetimes as compared with other households that relocate at some point in the life cycle, thereby exiting the personal income tax system of the previous state(s) of residence and entering the system of their new state(s) of residence. The implementation and even the conceptual justification for this horizontal equity argument for source-based consumption taxation appears to be quite problematic, however, when applied to households that reside in several different states during the pre-retirement portion of the life cycle.

Destination-based consumption taxation might be preferable to source-based taxation on efficiency grounds, insofar as households (including retirees) impose public-service provision costs on the states where they reside. Location-contingent taxes, such as a destination-based consumption tax, can enhance locational efficiency by serving as indirect congestion tolls when governments provide congestible public goods. (8)

PL 104-85, passed in 1996, has decided this issue of state taxation in favor of the residence principle. Prior to the passage of this law, states had the option of imposing income taxes on distributions from the retirement accounts of former residents and some 16 states did so, at least in principle. (9) By declaring that states may only tax such distributions on a residence basis, this statute has clarified how states may exercise some of their taxing powers, obviating potential constitutional and other legal disputes regarding double taxation and nexus. It also obviates the difficult administrative issues that arise, under the source principle, for individuals who reside in multiple states during their working lifetimes. By settling on the residence principle, the statute equips states that attract older residents with an important policy instrument with which to finance the public services that these households demand, even as it limits the ability of states to impose taxes on their working populations. This is an important policy distinction in an economy with a rapidly aging population and growing amounts of wealth held in tax sheltered accounts. The availability of this tax instrument permits states to shift the burden of government finance to the elderly, if desired. Competition for older workers under these circumstances may result in state-expenditure, regulatory, and other policies that are more favorable to older residents with significant amounts of accumulated tax-sheltered savings.

Corporation Income Taxation