On setting the agenda for Pennsylvania school finance reform: An exercise in giving policy advice

The matter of reforming the fiscal structures of metropolitan school economies is a pressing yet extremely complex policy issue now confronting most state legislatures in the United States. New, often radically different strategies for the funding of local schools are being sought. Good economic policy advice is at a premium. This paper has sought to outline a methodology for providing that advice which is sensitive to both the complexities of the policy problem and to the needs of busy decision-makers who must make a choice. An agenda-setting strategy which balances the gains from flexibility against the costs of policy deliberation has been proposed and then applied to the task of recommending a property-tax credit (λ) for local school finance. The best single recommendation(λ=.25) and the best agenda (λ=.2, λ=.4, plus the status quo) were calculated for the Philadelphia metropolitan school economy. Numerous qualifications and extensions suggest themselves. First, our alternative objective functions consider only private income and education. Yet one of the central issues in reform of local finance is the potential efficiency losses associated with the effects of local property taxation on housing consumption. To confront this matter, housing services must be introduced explicitly into the household utility function. But if a wage tax is to be used to finance the reform, then, to avoid the trivial result of fully substituting an efficient for an inefficient tax, the allocative consequences of wage taxation must be made endogenous as well. A four-commodity household utility function over housing, leisure, education, and other goods would therefore be an important extension of this work. The parameterization of such a four-commodity utility function through our MacCrimmon-Toda indifference curve procedure is feasible but probably suspect. This fact suggests a second extension. While we have relied upon interviews to provide the relevant data to estimate constituent preferences, a more attractive option would be to estimate such functions econometrically from revealed household behavior for each constituent group. While such estimates will no doubt be statistically less precise than those presented in Table 2, they are likely to be a turer approximation of actual constituent preferences. Further, the uncertainty of such statistical estimates can be included in our analysis, as can the uncertainty attached to our behavioral models' predictions for the effects of policies on the valued outcomes of income, education, housing, and leisure. This work would constitute a useful third extension. There remains an important final question, however, which has been purposefully pushed into the background. It's answer is fundamental to all the results developed here. What is the correct agenda-setting criterion? We have assumed for this analysis that the correct criterion is to maximize the expected value of agendas over alternative (winning) coalition preferences. That criterion has some appeal. From behind a veil of ignorance, a ‘universal legislator’ not knowing what coalitions he or she might fall within on any issue — yet wanting to insure that if they are a member of the winning coalition, there is an agenda which offers an attractive range of choices — might well choose our criterion of maximizing expected welfare across coalition preferences as the preferred agenda-setting rule. Yet this criterion is not the only candidate. The universal legislator might wish to design agendas which provide equal protection to all possible winning coalitions in which case our criterion becomes maximize the simple sum, rather than the probability weighted sum, of coalition preferences. From this extension it is easy to see the next step, the step which places the problem of finding a criterion for optimal agenda-setting within the larger context of designing an attractive social choice process. That is the arena in which we must look for an answer to our fundamental question of what is an optimal agenda-setting rule. 34 Hopefully the analysis here will convince others of the importance of such a search. Copyright Martinus Nijhoff Publishers 1981