II. A GENERAL PROGRESS PROGRAM: ON GOVERNMENTAL COORDINATION
I start with the notion of a "General Progress Program" that I developed with the help of Robert Conlan in 1995. I explain this concept as follows in a treatment of "intergovernmental coordination" in a section extracted from the aforementioned essays.
The Nation needs more effective decision-making mechanisms for intergovernmental coordination of expenditures and regulations. Efficient coordination can be achieved in an incentive-compatible way where the effects of spending by any one subunit on another subunit can be more effectively taken into account as compared with current decision-making procedures.
Effective coordination also relates to an important national economic policy concern which is efficient and equitable spending and deficit reduction. The need for better coordination is evidenced by fractious Congressional debate (i. e. the Congressional debate over welfare reform) over how to implement acceptable reductions in intergovernmental transfers, accompanied by reductions in Federally mandated spending requirements.
Given the existence of decision-making mechanisms (i. e. demand revealing) that would motivate decision-makers to accurately reveal the positive and negative effects of spending and regulatory decisions on one another, why not use these mechanisms to achieve more effective coordination? The Federal Government can establish an initial level of entitlements (called a distributional status quo) and then, with appropriate provisions for agenda-control, let the political subunits decide what levels of actual spending, including Federally mandated spending, is required. In effect, Congress determines the initial distribution but allows more implementation flexibility to the subunits, adjusting the financial flows to reflect external effects.
To illustrate, I will assume we are applying the process to all Federal discretionary programs (about $500 billion annually, and including National defense expenditures), later confining it to about $55 billion in programs (transportation, communications and the environment) financed by Federal excise taxes. Assume first that, we take the President's budget and allow a citizen's advisory committee to propose substitutes or eliminations designed to "maximize the likelihood of minimizing harm" (a criterion elaborated by Bailey, 1997). Subject to oversight by a commission, it will allocate cost shares among political subunits (in this case Congressional districts) so as to equalize per capita (or household) net benefits among the jurisdictions.
Suppose that the allocation of beneficiary tax-shares resulted in pressures for spending cuts that were nonuniform. For example, spending may be cut 10%, from $5,000 per household on average to $4500. Lower spending districts with lower cost shares might end up with an average of $4,000 while higher spending districts remained close to the original $5,000 average. Under the system, compensating transfers of $500 per household would be made from the higher to lower spending districts. If the advisory committee recommended prudent changes that also generated large net benefits and the compensating transfers were designed to achieve equal or proportional per capita net benefit increases for each district (that, say also averaged $500 per household), then the proposed package of changes would be accepted unanimously. The result is that, relative to the original status quo, the lower spending districts would be loaning $500 per household to the higher spending districts and sharing in the social surplus created by more productive investments.
Although this attractive, but seemingly impossible, result might characterize the package of changes, particular elements of the package may result in positive net benefits for some jurisdictions and negative benefits for others. In this paper, following Bailey, I present an approach combining demand revealing with Thompson insurance to help resolve this difficulty, and while the sharing of gains works out a little differently, these differences do not appear very important. Under the standard demand revealing approach, take for example, a program element where all but two equally populated districts are indifferent to a particular program change. The net benefits of the change are $60 per household for District B but District A would be harmed by the change in an amount of $50 per household. (This might arise when each is equally sharing $100 per h (household) in project costs and District B, with $110 per household in benefits, reaps all the gross benefits). In this case, the incentive compatible "pivotal mechanism" (i. e. demand revealing) described below would result in the change being made but District B would pay $50 a VCG tax that would be refunded to all the other 434 districts on a equal per capita basis (leaving about 12 cents in remaining surplus). Although I will elaborate on the point later, it is also important that the net surplus be sufficiently positive-sum ($60/$50 = 1.2) for the proposed change even to be successfully carried on the agenda.
Under Thompson insurance, where the best effort to reach a perfect Lindahl solution has left District A with $50 per h in harm, A and B would be asked to express their WTP (willingness to pay) for insurance against material harm (resulting from choice of a less preferred alternative), and a statistical group under the aegis of the Electoral Commission would assess the probabilities that A's preferred option would be selected over B's. If these were equiprobable, and the districts expressed their preferences solely in terms of material harm, A would pay $25 (.5 x $50) per h in Thompson insurance, and receive back $50 if A was selected. A total of $55 per h (including B's purchase of insurance) would be collected, and after payment of compensation of $50 to A a surplus of $5 per h from A would remain that would also be distributed among the other districts in the manner described above (leaving about 1.2 cents in undistributed surplus).
One of many advantages of the combined use of the pivotal and Thompson insurance mechanism is that, relative to the pivotal standing alone, the combination is more individually rational for A and B taken together. As indicated above, B for example obtains $25 (1/2 of his loss in compensation) and both A and B are better off by this amount plus a small refund relative to the refund from the pivotal standalone. The advantages are much stronger with respect to coalition and informational incentives, however. In addition to demonstrating important synergy between the pivotal and Thompson's mechanism, Bailey's recent work shows how we get around certain technical limitations of demand revealing (Groves and Ledyard, 1977). In this paper, I try to show how these principles might work as applied to the budget and regulatory processes of a large country while elaborating on other aspects of the relation between the pivotal mechanism and Thompson insurance, including further aspects of the incentive effects of the simple per capita surplus distribution or refund described above, in more detail. The important effects of combining the pivotal mechanism and Thompson insurance is the way in which it corrects for potential misrepresentations when preferences reflect both material and nonmaterial benefits and/or harms, or when participants believe the government's probability estimates are wrong. In addition, through combining the mechanisms, coalition problems associated with the pivotal mechanism are eliminated while giving significant added incentives for citizens to ensure an accurate representation of their preferences, i. e., overcoming the problem of rational ignorance.
In the following section, I explore these concepts in the context of particular distributive Federal programs affecting transportation and the environment. Based on Bailey's work, I will posit use of the pivotal mechanism for agenda setting and combined use of the pivotal mechanism and Thompson insurance in making final budget and regulatory decisions in a manner that will achieve efficiency while minimizing redistributive harm. The added feature here is the integration of Bailey's approach with the Limited Fund Mechanism (LFM) which, like Thompson insurance, uses information generated by incentive-compatible means to compensate losers and minimize redistributive harm, through a process I will now call "compensated incentive compatibility," or CIC.
The second added feature is the combining of the agenda setting process with incentive compatible Congressional voting, at least initially, at the referendum stage. Rather than voting by individuals through direct democracy (explored in a following paper), I rely on representative voting in Congress or by Congressional committees representing regions or states in the process of agenda setting. I shall describe first an agenda setting stage where the system is operating under the aegis of one (monopoly) advisory committee which establishes one set of options for departures from the status quo. The committee works with executive officers (EOs) who essentially represent the interests of constituent regions, i. e., the executive officers substitute for competitive regional committees. In terms of the composition of the committees, following Bailey, I assume about 300 citizens representing a complete cross section of the population with about 30 citizens each drawn from the ten regions of the country. The competitive alternatives may be drawn from 3600 citizens operating in regional committees (300 in each of the regions) with a similar number operating locally. I then turn to cases where there are competitive alternatives, e. g., a national committee's preferred alternatives competing with those of the executive officers or regional committees, where the decisions are made by Bailey's process in the context of Congressional voting via the VCG-Thompson mechanism. While the idea of such a method used in present legislatures may seem initially farfetched, one is usefully led into investigations of how it could be implemented in a second best setting.