Thomas Schwartz

Congressional Oversight Overlooked: Police Patrols versus Fire Alarms

Scholars have often remarked that Congress neglects its oversight responsibility. We argue that Congress does no such thing: what appears to be a neglect of oversight really is the rational preference for one form of oversight-which we call fire-alarm oversight-over another form-police-patrol oversight. Our analysis supports a somewhat neglected way of looking at the strategies by which legislators seek to achieve their goals.

Agendas and The Control of Political Outcomes

A considerable theoretical literature argues that if everyone votes sincerely, then an agenda setter has near dictatorial influence on final outcomes, whereas if everyone votes strategically, then an agenda setters power is considerably reduced. This literature assumes that all feasible agendas are of a special type called amendment agendas. But actual legislative and committee agendas-notably those found in Congress-often are not of this type. Once we expand the domain of feasible agendas to include all types allowed by common parliamentary practice, the influence of agendas on legislative outcomes expands, even with strategic voting. Besides showing with counterexamples that previous results do not extend to a more realistic domain of agendas, we prove some theorems that specify the limits (such as they are) of an agenda setters power.

Collective Choice, Separation of Issues and Vote Trading

In legislatures and committees, a number of issues are voted on separately, leading to an outcome consisting of positions on each of these issues. I investigate the effects this separation of issues has on collective choices, assuming a very abstract collective choice model, whose assumptions are presupposed by many less abstract models, notably spatial models. Assuming the model, if there exists an undominated outcome (one to which no winning coalition prefers any other feasible outcome), it must be chosen in the absence of vote trading, although vote trading can (perversely) lead to a very different outcome. But vote trading does not necessarily lead to a “voting paradox” situation, contrary to several recent papers. The model enables us to define a natural solution concept for the case where every feasible outcome is dominated. Variations on this concept are explored. The effects of weakening the model are investigated.

Your vote counts on account of the way it is counted: An institutional solution to the paradox of not voting

When we explicitly lay out all its steps, we find that the Paradox of Not Voting (since the chance of one votes making a difference is about zero, why trouble to vote?) rests on a false but hitherto unremarked assumption about the institutional context of elections. My solution to the Paradox is more conservative than others that have been proposed, and it yields a rational-choice model of voting whose consequence accord well with empirical findings on turnout.

The universal-instability theorem

When individual or corporate creatures with partly conflicting preferences cooperate or compromise, form an alliance or make a deal, they exchange support across issues: they support a package of positions, one on each issue involved in the exchange; although none of them favors every position in the package, they agree (in effect) to all the positions because they prefer the package as a whole to the alternative package that would prevail without the exchange. Vote-traders obviously exchange support across issues. So do coalition-government partners; for them the issues concern legislation, portfolio assignments, or both. So, implicitly, do the supporters of a political candidate or legislative leader who builds a winning platform from planks that have too little support to be enacted singly. The participants in an ordinary economic trade also exchange support across issues: if I trade you a banana for a coconut, I have supported your favorite position on the issue of who gets the banana in return for your support of my favorite position on the issue of who gets the coconut. The exchange of support across issues generalized exchange, for short characterizes all political, economic, and other interpersonal activity involving partial conflict of interest. Kadane (1972), Oppenheimer (1972), and Bernholz (1973), building on an insight of Downs (1957), have independently proved that whenever a majority of voters with separable preferences support a package of

The politics of flatland

A version of the median-voter theorem holds for two-dimensional spatial models in which voters regard the two dimensions as economic goods or goodlike activities and in which the set of feasible outcomes is constrained by budget or technology. Although mathematically trivial, this fact has widespread analytical uses. After arguing that our two-dimensional model, with its stability property, fits a number of important and general policy areas, we use our analysis to address some prominent theoretical issues.

Welfare judgments and future generations

The author argues that long-range welfare policies - policies designed to provide significant, widespread, continuing benefits to future generations, remote as well as near, at some cost to ourselves - cannot be justified by appeal to the welfare of remote future generations. He questions whether they can be justified at all. The problem is that the failure to adopt such a policy would not make any of our distant descendants worse off that he would otherwise be, since had the policy been adopted, he would not even have existed. These considerations also bring out a conflict between utilitarian and Paretian principles.

No minimally reasonable collective-choice process can be strategy-proof

Abstract In this paper I will prove that every collective-choice process fulfilling certain mild requirements - notably a version of Nondictatorship and a very weak unanimity-reduction condition - is manipulable: none is strategy-proof. Gibbard and Satterthwaite have each proved a similar result just for resolute processes - ones yielding only single-member choice sets. My theorem does not assume resoluteness, which I criticize as an unreasonable requirement.