Jonathan Bendor

The Evolution of Norms1

Social norms that induce us to reward or punish people not for what they did to us but for what they did to other members of one’s group have long been thought as sine qua non sociological and thus impossible to explain in terms of rational choice. This article shows how social norms can be deductively derived from principles of (boundedly) rational choice as mechanisms that are necessary to stabilize behaviors in a large class of evolutionary games.

Rethinking Allison's Models

The ideas in Graham Allisons Essence of Decision (1971) have had an enormous impact on the study and teaching of bureaucracy and foreign policy making. While Allisons work has received considerable critical attention, there has been surprisingly little examination of the content and internal logic of his models. We subject each of Allisons three models to a systematic critical analysis. Our conclusion is that the models require substantial reformulation.

When in Doubt...: Cooperation in a Noisy Prisoner's Dilemma

In the last decade, there has been a resurgence of interest in problems of cooperation, stimulated largely by Axelrods work. Using an innovative tournament approach, Axelrod found that a simple strategy, tit-for-tat (TFT), was most successful in playing the repeated prisoners dilemma (PD) in a noiseless environment. However, recent analytical work has shown that monitoring problems caused by noise significantly impair TFTs effectiveness. The primary purpose of the present research is to discover whether there exist alternative strategies that perform well in noisy PDs. To investigate this question, the authors conducted a computer tournament. The results of the tournament demonstrated that, consistent with analytical work, TFT performed rather poorly. In contrast, strategies that were generous (i.e., cooperated more than their partners did) were quite effective.

Spatial Models of Delegation

Although a large literature on delegation exists, few models have pushed beyond a core set of canonical assumptions. This approach may be justified on grounds of tractability, but the failure to grasp the significance of different assumptions and push beyond specific models has limited our understanding of the incentives for delegation. Consequently, the justifications for delegation that have received recent scrutiny and testing differ from some of the more plausible justifications offered by informal studies of delegation. We show that surprisingly few results in the literature hinge on risk aversion, and surprisingly many turn on the ignored, though equally canonical, technological assumption that uncertainty is fixed (relative to policies). Relaxing the key assumptions about dimensionality and functional forms provides a clearer intuition about delegation—one that is closer to classical treatments. The theory allows us to relate different institutional features (commitment, specialization costs, monitoring, multiple principals) to delegations observable properties.

Institutional Structure and the Logic of Ongoing Collective Action

Work by Axelrod, Hardin, and Taylor indicates that problems of repeated collective action may lessen if people use decentralized strategies of reciprocity to induce mutual cooperation. Hobbess centralized solution may thus be overrated. We investigate these issues by representing ongoing collective action as an n-person repeated prisoners dilemma. The results show that decentralized conditional cooperation can ease iterated collective action dilemmas—if all players perfectly monitor the relation between individual choices and group payoffs. Once monitoring uncertainty is introduced, such strategies degrade rapidly in value, and centrally administered selective incentives become relatively more valuable. Most importantly, we build on a suggestion of Herbert Simon by showing that a hierarchical structure, with reciprocity used in subunits and selective incentives centrally administered, combines the advantages of the decentralized and centralized solutions. This hierarchical form is more stable than the decentralized structure and often secures more cooperation than the centralized structure. Generally, the model shows that the logic of repeated decision making has significant implications for the institutional forms of collective action.

An Adaptive Model of Bureaucratic Politics

In this article we outline a new framework for the formal analysis of bureaucratic politics. It departs from standard neoclassical approaches, notably those of Niskanen (1971) and Peltzman (1976), in several important respects. First our approach explicitly models a system of three-way interaction among bureaus, politicians, and interest groups. Second, it allows for institutional features of each type of participant. Third, it is a model of dynamic process. Fourth, participants make choices adoptively rather than optimizing. Fifth, participants are only minimally informed.The result is a dynamic model of adaptive behavior, very much in the spirit of Simons (1947) behavioral tradition, that offers a new perspective on political control, bureaucratic power, and the “intelligence of democracy.â€

Bureaucratic Expertise versus Legislative Authority: A Model of Deception and Monitoring in Budgeting

The empirical literature on the control of bureaus notes that politicians have difficulty observing bureaucratic output, but this insight is rarely represented informal models. To analyze how bureaus use this uncertainty strategically, we develop a model of expertise-based agenda control, building on the Niskanen (1971) and Miller and Moe (1983) tradition. We show that under some plausible conditions, bureaus will underestimate the benefits, and overestimate the costs, of their programs. In the model, politicians are neither passive nor omniscient: they anticipate the bureaus strategic behavior and establish a monitoring system to counteract it. This possibility of detection changes the bureaus behavior: even imperfect monitoring reduces the bureaus deception of the legislature, whether or not the legislatures demand for the bureaus services is concealed. Moreover, uncertainty by itself matters: if the legislature makes it harder for a risk-averse bureau chief to predict demand or penalty, the bureau will restrain its deception.

The Evolutionary Stability of Cooperation

Is cooperation without central authority stable? If so, how robust is it? Despite what might be the conventional wisdom, The Evolution of Cooperation did not solve this problem deductively. In fact, results obtained later by others seem to have contradicted the books main message. Reexamining this exceptionally influential work yields a new picture. Part of Axelrods evolutionary story turns out to be false. But the main intuition, that retaliatory strategies of conditional cooperation are somehow advantaged, proves correct in one specific and significant sense: Under a standard evolutionary dynamic these strategies require the minimal frequency to stabilize. Hence, they support the most robust evolutionary equilibrium: the easiest to reach and retain. Moreover, the less efficient a strategy, the larger is its minimal stabilizing frequency; Hobbesian strategies of pure defection are the least robust. Our main theorems hold for a large class of games that pose diverse cooperation problems: prisoners dilemma, chicken, stag hunt, and many others.

Uncertainty and the Evolution of Cooperation

It is well known that inferential errors can induce nice but provocable strategies to engage in vendettas with each other. It is therefore generally believed that imperfect monitoring reduces the payoffs of such strategies and impairs the evolution of cooperation. The current literature, however, only scrutinizes specific strategies, either analytically or in particular tournaments. This article examines in a more general way how monitoring uncertainty affects the fate of cooperation in tournaments of the iterated prisoners dilemma (IPD). The first set of results shows that imperfect monitoring does create a sharp trade-off between cooperativeness and unexploitability. The second set examines how random shocks affect the tournament payoffs of several large classes of strategies in the IPD, and shows how noise can help certain nice strategies. The third set analyzes how imperfect monitoring can facilitate the emergence of cooperation based on a population of non-nice strategies. Thus the idea that inferential uncertainty always harms nice strategies and always impairs the evolution of cooperation must be sharply qualified.

A Model of Muddling Through

As arguments about the effectiveness of “muddling through” have proven frustratingly inconclusive, incrementalism—once a major approach to the study of boundedly rational policy processes—has gone dormant. In an attempt to revitalize the debate, I present a formal model of muddling through. The model, by clarifying the logical structure of the informal theory, presents a clearer target for criticism. More importantly, it establishes numerous deductive results. First, some of Lindbloms less controversial conjectures—about the benefits of seriality (repeated attacks on the same policy problem) and redundancy (multiple decision makers working on the same problem)—turn out to be correct if conflict across policy domains is absent or takes certain specified forms. But given other empirically reasonable types of conflict, even these claims are wrong. Second, the advantages of incremental (local) policy search (Lindbloms best-known and most controversial claim) turn out to be still less well founded: in many empirically plausible contexts the claim is invalid.