Liquidity Constraints, Fiscal Policy, and Consumption TAX POLICY AND TAX REFORM are important items on the current policy agenda. In evaluating alternative tax policies, decisionmakers must consider both their normative and positive impact; in particular, they must examine the effects of competing policies on the overall well-being of the taxpayers and on various indexes of economic activity, in both the short run and long run. Basic to understanding the impact of tax policy is analysis of the

more »

... elationship between taxation and taxpayers' decisions about consumption, saving, and work effort. Such analysis is especially sensitive to assumptions made regarding individuals' abilities to use capital markets to transfer income across time. By the same token, we think that policy simulation models that ignore "liquidity constraints" result in flawed tax policy analysis. In this paper we analyze the impact of liquidity constraints on consumption functions and use the resulting view of aggregate demand to address two categories of tax and fiscal policy issues. The first issue is developing a tax system that least reduces taxpayer well-being for the amount of lifetime revenue extracted. Recent applications of theoretically based models of individual behavior have facili-We are grateful to John Simpson and Nicolas Williams for excellent research assistance and to Mervyn King, Laurence Kotlikoff, and members of the Brookings Panel and the Faculty Research Seminar at the John F. Kennedy School of Government for helpful comments. Financial support from Northwestern University is acknowledged. 1 2 Brookings Papers on Economic Activity, 1:1986 tated comparisons of the effects of alternative policies. Indeed, much of the present debate over the relative reliance on various tax bases and the optimal degree of progressivity of the income tax has been reflected in studies using policy-simulation models. I These models generally assume that individuals or households maximize well-being over their lifetime, subject only to the restriction that the present value of consumption is no greater than the present value of income. The models then assess how tax policies can alter the rewards to saving and working, thereby distorting intertemporal choices about consumption and work effort. But models with an overall lifetime budget constraint ignore the many restrictions on individuals' ability to shift income in a world of capital market imperfections, restrictions that substantially affect those intertemporal choices.2 Taking into account the effect of liquidity constraints affects the calculation of the welfare costs of taxation substantially; prevailing arguments, based on "perfect market" models, against capital taxation or progressive income taxation and in favor of wage and consumption taxation must be substantially muted and often reversed.3 The second issue that we examine is the importance of liquidity constraints for the debate over the impact of temporary tax cuts financed by debt. We show that liquidity-constraint considerations are quantitatively more important than the frequently discussed "finite-horizon" considerations that focus on how debt shifts the tax burden onto future generations. We also find that when both considerations are incorporated into the consumption function, debt-financed tax cuts of the type observed in the post-World War II era do not substantially alter aggregate consumption. In this light, the plan of the paper is as follows. First, we elaborate on the potential macroeconomic importance of liquidity constraints and their role in tax policy analysis in life-cycle consumption models. We then develop a simple life-cycle model to examine the effects of liquidity constraints on measures of national saving and individual welfare. We focus on the life-cycle model because of its easy applicability to fiscal policy analysis. Use of this model introduces no bias a priori in our examination of the effects of liquidity constraints, since the life-cycle model under rational expectations and with perfect capital markets offers fiscal policy little scope in influencing consumption. We consider three major applications: the relative welfare costs of capital versus labor income taxation, effects of progressive versus proportional income taxation, and effects of temporary tax changes on consumption. Liquidity Constraints and Consumption The effect of liquidity constraints, defined in various ways, on consumer spending has been considered in many studies.4 In response to Robert Lucas's critique of econometric policy evaluation, Robert Hall proposed the "Euler equation" approach to testing the sensitivity of consumption to current income changes.5 In Hall's model, to be consistent with the permanent-income hypothesis under rational expectations (with no borrowing restrictions), conditional on lagged consumption, expected consumption should be independent of other lagged information. Other empirical studies find consumer spending to be sensitive to income changes.6 Findings of excess sensitivity of consumption to changes in disposable income are corroborated in a study of food expenditures by Hall and Frederic Mishkin.7 Our analysis is predicated on the idea that this excess sensitivity can be traced to the operation of liquidity constraints, the aggregate importance of which for the United States is amply illustrated by historical evidence. Bradford DeLong and Lawrence Summers note that from 1899 to 1916, "essentially all consumption was done by liquidityconstrained consumers.' '8 Their findings for the entire pre-World War II period broadly support this conclusion and suggest the possibility, which we consider in our simulation exercises, that forced lifetime saving (that is, underconsumption) by consumers who are liquidity constrained may be an important component of total saving. For example, Alan Auerbach and Laurence Kotlikoff note that personal saving rates in the United States exceeded 20 percent during the 1880s-before the availability of consumer credit and the pursuit of stabilization policy .9 Fumio Hayashi finds that liquidity-constrained consumers accounted for approximately 20 percent of all consumption in post-World War II United States. 10 In a separate effort using microeconomic data, Hayashi 6. Marjorie A. Flavin, "The Adjustment of Consumption to Changing Expectations about Future also notes that the relationship between consumption and income movements differs significantly for "high saving" and "low saving" families.11 Ben Bernanke finds no evidence that the permanent-income hypothesis needs to be amended for liquidity constraints in his examination of individual expenditures on automobiles.12 Automobile loans, however, are self-collateralized, and our focus is on the unavailability of noncollateralized consumption loans. Marjorie Flavin finds that the estimate of the marginal propensity to consume is affected dramatically by the inclusion of proxies for liquidity constraints .13 In her econometric work, Flavin uses the aggregate unemployment rate as a proxy for liquidity constraints and tests "myopia" and liquidity-constraint explanations of the excess sensitivity findings. She reports that the estimated marginal propensity to consume out of transitory income is explained almost entirely by proxies for liquidity constraints. We analyze the liquidity constraint arising from a nonnegativity constraint on net worth. 14 That is, consumers are not permitted to borrow against income to be received in the future; current consumption is percent of personal consumption expenditures could be due to liquidity-constrained households. Fumio Hayashi, "The Effect of Liquidity Constraints on Consumption: A Cross-