Mobile Edition
Print
Get the Mag
Weekly UpdatesInstitutional Foundations of Public Finance: Economic and Legal Perspectives. Edited by ALAN J. AUERBACH and DANIEL N. SHAVIRO. Cambridge, MA and London: Harvard University Press, 2008, pp. 282.
INTRODUCTION
This book contains the proceedings of a May 5, 2006 conference held at the NYU Law School in honor of the late David E Bradford. The book has seven chapters: Alan J. Auerbach writes on the choice between income and consumption taxation, David A. Weisbach on implementation of the two tax systems, Louis Kaplow on the transition to consumption taxation, Wallace E. Oates on fiscal federalism, Roger Gordon and Martin Dietz on dividend taxation, and Jerry Green and Laurence J. Kotlikoff on the "language" of fiscal policy. Each chapter is followed by comments by two discussants.
The book is exceptionally well suited to serve as a tribute to David Bradford. The authors and discussants address the questions that Bradford studied and they do so at the same broad and conceptual, but policy-relevant, level at which he addressed them.
FISCAL FEDERALISM
The chapter by Oates on this topic is usefully organized around the Oates (1972) decentralization theorem, which concludes that, under certain assumptions, local decisions on the supply of a public good are at least as good as a centralized decision. The key assumptions are perfect decision making by local governments, no differences in the costs at which the two levels of government can provide the good, no spillovers across localities from the public good, and no ability of the central government to vary output levels across local areas. Because the assumptions are clearly restrictive, Oates focuses on surveying the robustness of the theorem.
Oates sensibly concludes with a general presumption in favor of decentralization, but aptly observes that "a delicate balancing act" between central and local governments is required. Harvey S. Rosen's comment provides additional support for the decentralization presumption, noting that centralized decisions can be flawed for various reasons and therefore need not improve on local decisions even when the Oates (1972) assumptions do not hold.
Of course, the merits of decentralization vary across policy issues, a point unintentionally reinforced by two examples of allegedly excessive centralization that Oates cites. He is on firm ground when he refers to an Environmental Protection Agency rule restricting arsenic in drinking water as "an especially dramatic case of welfare losses from centralization," noting that the rule imposes the same standards on all localities, even though the per-person costs of attaining the standard are hundreds of times larger in small localities than in large ones. But he is on weaker ground when he criticizes efforts to reduce cross-district variations in per-pupil public school spending within metropolitan areas. As he acknowledges, and as Bradford and Oates (1974) discuss at greater length, there may be valid arguments for equalization of spending across rich and poor districts, making this a dubious choice to showcase the drawbacks of centralization.
Oates devotes some attention to the problems that arise if localities face soft budget constraints due to transfers from the central government. But neither he nor the discussants address the related topic of countercyclical intergovernmental transfers. The February 2009 stimulus package gives states increased Medicaid matching funds, as the March 2002 stimulus package had done on a smaller scale. Because the additional funds are not targeted to states that have encountered financial problems, they do not raise the classic soft-budget-constraint problem. Because the grants are linked to Medicaid program size, though, their availability allows states to run larger programs while letting the federal government absorb the costs of preventing undesirable cyclical variations. And, while the grants are conditioned on states not adding to their rainy-day funds during the recession (to avoid diluting the desired stimulative effect), they are not conditioned on states adding to, or even maintaining, their funds during economic expansions. More work remains to be done on the proper design of countercyclical intergovernmental transfers, a project that will require insights from both macroeconomics and public finance.
DIVIDEND TAXATION
Gordon and Dietz tackle the perennial question of the proper model of dividend taxation. The authors provide a comparison of the "new" and "traditional" views of dividends. While they predictably and appropriately conclude that neither view fits all of the evidence, they find somewhat greater support for the traditional view. Their conclusions are sensitive, however, to their specifications of the two views. Like the discussants, William D. Andrews and George R. Zodrow, I think that the new view holds up somewhat better than Gordon and Dietz suggest.
The specification of the traditional view constructed by Gordon and Dietz is both sophisticated and restrictive. They posit a signaling model in which management has more information than shareholders about earnings prospects and in which dividends are a more effective signal of profitability than share repurchases precisely because they are more costly (due to their unfavorable tax treatment). The model includes a number of restrictive assumptions about such matters as the extent to which managers hold stock and the timetable on which they sell it and the extent of the information revealed by new share issuance. The resulting complexity necessitates a two-period, rather than an infinite-horizon, formulation; even then, the presentation of the model is not fully transparent.
The Gordon and Dietz specification of the new view is restrictive in a quite different way; in its simplicity, it omits key features of more realistic versions of the new view. To begin, their version of the new view completely precludes share repurchases, leading them to treat the existence of repurchases as evidence against the new view. The argument is misplaced because broader versions of the new view incorporate share repurchases. As Sinn (1991) and Auerbach and Hassett (2003) demonstrate, the economic implications of the new view continue to hold so long as reductions in repurchases comprise the same fraction of investment financing as repurchases comprise of the distributions of returns.
Also, because Gordon and Dietz exclude the possibility of debt finance, their version of the new view predicts that dividends are a volatile residual equal to the difference between profits and new investment, leading them to cite the observed stability of dividends as evidence against the new view. In the presence of debt finance, however, firms can borrow to smooth dividends even when the new view holds. Like the presence of repurchases, this evidence against the new view vanishes when that view is specified in a more general form.
Gordon and Dietz also cite the responsiveness of dividend payout rates to dividend tax rates as evidence against the new view. As Zodrow observes, it may be possible to reconcile such responsiveness with the new view, particularly if changes in the dividend tax rate are transitory. Still, this evidence probably favors the traditional view to some extent. On a more subtle note, Gordon and Dietz also cite evidence about cyclical variations in the market valuation of dividend payouts as supporting their signaling model over the new view.
As Gordon and Dietz ultimately conclude, neither view is fully satisfactory. Surprisingly, the participants devote little attention to firm heterogeneity. Auerbach and Hassett (2003) find that some mature dividend-paying firms behave in a new-view manner while others behave in a traditional-view manner. Further investigation of such heterogeneity may be the most promising direction for future research on dividend taxation.
TAX TERMINOLOGY
The chapter by Green and Kotlikoff posits the general relativity of fiscal language, providing a mathematical demonstration that net taxes and transfers cannot be defined independently of the initial assignment of property rights. In his comment, Daniel N. Shaviro crisply summarizes their result in the following terms: a fiscal system in which a worker initially owns her lifetime earnings and pay taxes to the government equal to 30 percent of earnings can equally well be described as one in which the government initially owns the lifetime earnings and makes a net transfer payment to the worker with a present value of 70 percent of earnings.
Although Green and Kotlikoff's result is mathematically correct, Shaviro and Kent Smetters are right to emphasize its lack of practical implications in their comments. Noting that "little may turn" on the authors' result, Shaviro (p. 263) reasonably concludes, "I will continue using the term 'net taxes' until I become aware of problems with it that are not currently evident to me."
The chapter is most useful as a springboard for examining deeper questions about the role of economic terminology. There is broad recognition, for example, that the term "deficit" is problematic for reasons set forth in the generational accounting literature. While that literature often recommends that the term should be abandoned as meaningless, most economists continue to use the term. The persistence of deficit terminology reflects an appropriate understanding of the practical use of language.
It is true that two policies that feature identical deficits may have different impacts on agents' budget sets while two policies that have identical impacts on budget sets may feature different deficits; such outcomes readily arise, for example, if one policy features back-loaded IRAs and the other involves front-loaded IRAs. Clearly then, deficit terminology is not completely informative. But it can still play a necessary role in an economically meaningful description. The most natural way for an economist to describe such policies includes both their effects on the deficit and the type of IRA involved, thereby providing substantial information about how the policies affect agents' budget sets. In such a description, the discussion of the deficit is meaningful because the impact on budget sets cannot be identified from the type of IRA alone.
FEED