The economic consequences of debilitating illness are defined and then estimated for one such illness--multiple sclerosis. Economic losses are defined as the consumption losses to the affected household because of illness. These losses are measured as the change in earnings of all family members plus the increase in gross (not out-of-pocket) medical costs. Earnings models are specified and estimated, and gross medical costs calculated for multiple sclerosis. The average annual loss to the multiple sclerosis household is $5, 336 per year in 1976 dollars; the estimated aggregate decline in consumption for the year 1976 was $0.656 billion. Lifetime costs (discounting by 0.06) total $207, 200 per multiple sclerosis household and $25.50 billion for society as a whole for the current (1976) pool of multiple sclerosis patients, and $30.45 billion for all future multiple sclerosis patients.

In September 1990, the city of Philadelphia went to the municipal-bond market to borrow $375 million dollars for one year, a not unusual event for a large city needing cash to cover weekly expenditures until the quarterly or annual tax revenues are collected. What was unusual for Philadelphia was the size of the loan request; the 1990 proposed borrowing was $187 million larger than the previous year's request of $188 million. Investors rightly suspected that Philadelphia needed money for more than just the typical cover of a few month's cash shortfall. A closer look at the city's books revealed over $73.8 million in accumulated deficits from the previous six budgets. The extra $187 million of short-term borrowing was necessary to help cover the $73.8 accumulated debt and fiscal year (FY) 1991's expected deficit of $132.6 million. Investors rejected Philadelphia's request for the additional borrowing, and the city fell into a three-year fiscal crisis. The city emerged from the crisis with city residents and shoppers paying an additional 1-percent sales tax, city workers facing a two-year wage freeze and a reduction in employee benefits, and residents living with reduced public services. The long-run effects on city employment and property values are sure to be damaging. How does a city get into such a fiscal mess? This paper seeks to provide one answer to this question using the recent (FY 1963-1990) fiscal history of Philadelphia as a case study. I. Accounting for a Fiscal Crisis

The phenomenal growth in our public deficits over the past twenty years is a matter of public and professional concern (see Symposium, 1989). While this concern may indeed be well founded (particularly given current deficit levels), public deficits are not always bad. Certainly in times of deep recessions, short-term deficit financing can stimulate aggregate demand, increase national income, and reduce unemployment in a Keynesian fashion (see Richard Startz, 1989). Second, long-term deficit financing may be needed to sustain a long-term path of optimal consumption (see Peter Diamond, 1965). Third, public deficits (including pay-as-yougo Social Security) can offer welfare gains for significant subsets of consumers who have been liquidity constrained (R. Glenn Hubbard and Kenneth Judd, 1986), who have been unable to purchase indexed private annuities (Alan Blinder, 1988), or who have been unable to sufficiently diversify their investment portfolios away from human capital (Robert Merton, 1983). Finally, Robert Barro (1979) has argued that public deficits may be an important policy instrument to insure intertemporal welfare maximization when public expenditures are stochastic and public taxes are economically inefficient; taxes can be smoothed to reduce the lifetime excess burden of public financing. The concern today is not that we have deficits, but rather that we may be overdoing a good thing. If this is the issue, then we need to ask: Why, and what can be done about it? This is my agenda here. I. The Political Economy of Recent Deficits

In June of 1978, the voters of California approved Proposition 13, restricting the rate of local property taxation to no more than 1 percent of 1975 market values. In November 1978, the voters of Michigan approved the Headlee Amendment, which limited state revenues from own sources to a fixed share of state personal income. In November of 1980, Massachusetts voters followed the lead of California and required all cities and towns to limit their taxation to a rate of 22 percent of full and fair cash value. Numerous hypotheses have been advanced and analyzed to explain the emergence of these new restrictions on fiscal choice. The results point to two explanations: voters feel governments are too big, providing more services than what they prefer, or governments spend too much on redistributive activities to help the poor and, through higher wages, to help public employees themselves.' The response to this perceived failure of government has been to propose-and in California, Michigan, and Massachusetts to approve-absolute limits to government expenditures. The question I wish to address here is whether a compelling economic argument can be advanced for such a policy response, politics and personalities aside. The answer, I think, is yes, but particular preconditions must apply. First, and perhaps most importantly, is the inability of citizens to directly control the provision of public services. Responsibility for providing public goods is delegated to agents of the voter, and these agents-be they bureaucrats or elected representatives-have their own objectives and an ability to act upon them to the possible disadvantage of the voters. Second, without direct control over public outputs, voters must use indirect controls through the manipulation of their appointed agents. These indirect controls take two general forms: price incentives or quantity restrictions. Third, the choice of either of these indirect controls must take place before the agent acts to provide public goods. Further, voters must often make their choice of a price or quantity control before the benefits or costs of the public service are known to them with certainty. Agents, however, can wait to observe true benefits and coststhat is why we use agents-and then provide a level of public services in response to costs, benefits, and the voters' chosen control. Voter uncertainty over benefits and costs can be interpreted as placing the voter within a constitutional perspective where a voter's precise position in society, that is, his or her benefits and costs, are unknown when choosing a control. Geoffrey Brennan and James Buchanan (ch. 1) were the first to really stress this important point. Fourth, if benefit and cost uncertainty is predominantly over the position (rather than slope) of the marginal benefit and marginal cost schedules and if the marginal benefit schedule is relatively steep (loosely, an inelastic demand curve) while the marginal cost schedule is relatively flat (an elastic supply curve), then quantity controls on agents' behavior through tax or spending limitations will be the preferred control. These four preconditions strike me as quite plausible. Rather than just an emotional reaction to a peculiar configuration of political and economic circumstances, Proposition 13, the Headlee Amendment, and Proposition 24 may be perfectly reasoned responses to a failure in the politi*Professor, University of Pennsylvania, and research associate, the National Bureau of Economic Research, Inc. 'An alternative hypothesis which has been tested and generally rejected attributes voter approval to a desire to alter the tax mix away from property taxation and towards other taxes. What approving voters really want, it seems, is increased government efficiency and less redistribution. See the recent studies by Jack Citrin; Paul Courant, Edward Gramlich, and Daniel Rubinfeld; and Helen Ladd and Julie Wilson.

In 1971 the California Supreme Court opened the door to a major reform movement to restructure the present system of decentralized school finance. With the exception of Hawaii, elementary and secondary education in the United States is supported primarily by local property taxation supplemented in part by state funded grants-in-aid. The California Supreme Court, in the now famous Serrano rulings, declared the California system in violation of the state constitution's equal protection clause. Similar rulings have also been handed down by the New Jersey Supreme Court (Robinson vs. Cahill) and the Superior Court of Hartford, Connecticut (Horton vs. Meskill). In addition, ten states have recently enacted major reform bills, and legislation is under consideration in several others. The pressure for reform is strong and continuing. As a review of the recent reform proposals indicates, the legislative search for new means of financing local schools is not simply an incremental tinkering with existing laws.' Major changes, often court required, are at issue. Long-run outcomes are uncertain; each proposal has new winners and new losers. When planning a major reform of local school finance, therefore, past experience from incremental policymaking may not be an adequate guide to choice. Long-run general equilibrium predictive models and a clearly specified evaluation rule will be needed. It is the purpose of this paper to develop such a policy framework and to apply the analysis to one region currently in the midst of school reform, the New York metropolitan area. Six alternative reform proposals are considered: foundation aid, two district power equalization plans, property tax credits, expanded Title I assistance under the Elementary and Secondary Education Act, and centralized financing and spending controls. Preferred reforms are selected under utilitarian (promiddle class), Rawlsian (pro-poor), and equal school spending (Serrano) criteria.