SYSTEM EFFECTS AND THE PROBLEM OF PREDICTION
ABSTRACT: Robert Jervis's System Effects (1997) shares a great deal with game theory, complex-systems theory, and systems theory in international relations, yet it transcends them all by taking account of the role of ideas in human behavior. The ideational element inserts unpredictability into Jervis's understanding of system effects. Each member of a “system” of interrelated actors interprets her situation to require certain actions based on the effects these will cause among other members of the system, but these other actors' responses to one's action will be based on their own perceptions of their situation and their interpretations of what it requires. These ideas are fallible, but we cannot predict the mistakes people will make if the errors are based on information we do not have or do not interpret in the same way they do. Not only members of a system but social-scientific observers and policy makers are ignorant of others' information and interpretations, and therefore are as likely to err in their behavioral predictions as are members of the system. Thus, Jervis's book raises serious questions about how to evaluate policies directed toward producing positive system effects. The questions are unanswerable at this point, but they might be susceptible to analysis by an ambitious form of political theory.
THE COMPLEXITY OF SYSTEM EFFECTS
Andrea Jones-Rooy and Scott E. Page
ABSTRACT: Complexity science has witnessed a number of advances since the publication of Jervis's System Effects. These advances better allow us to untangle the messy elements in a system and predict sets of likely outcomes. However, just because a system is complex doesn't mean that all the ideas relating to complexity—such as agent-based modeling, path dependency, tipping points, between-class versus within-class effects, and networks—are necessarily relevant. One of our tasks is to determine whether they are—and, if so, their implications. As examples, we use China's role in the formation of the United States housing bubble; the federal government's bailout of AIG and Bear Stearns but not Lehman Brothers; and China's failure to experience a regime change such as the Middle East's Arab Spring.
WE CAN NEVER STUDY MERELY ONE THING: REFLECTIONS ON SYSTEMS THINKING AND IR
Nuno P. Monteiro
ABSTRACT: Robert Jervis's System Effects was published just as systems thinking began to decline among political scientists, who were adopting increasingly strict standards of causal identification, privileging experimental and large-N studies. Many politically consequential system effects are not amenable to research designs that meet these standards, yet they must nonetheless be studied if the most important questions of international politics are to be answered. For example, if nuclear weapons are considered in light of their effect on the international system as a whole, it becomes clear that they have obviated the need for a global balance of power by allowing states to counterbalance threats by acquiring nuclear weapons rather than investing in massive conventional balancing efforts. Similarly, systems thinking should inform our understanding of the impact of a “unipolar power” such as the United States, which has enjoyed an overwhelming preponderance of conventional military power since the fall of the Berlin Wall. A unipolar power is likely to become involved in frequent conflicts because it is not restrained by the presence of a peer competitor.
JERVIS ON COMPLEXITY THEORY
Richard A. Posner
ABSTRACT: The correct solution to complex problems, such as those involved in international relations, can generally be discovered ex post but not predicted ex ante. Economics and game theory attempt to model such complexity, but have difficulty taking into account psychological subtleties, the myriad factors that each agent considers when making a decision, and cultural differences. And understanding that one is dealing with a system—that is, with interacting factors instead of with insulated monads—may not make the questions any more amenable to prediction, particularly because the more unique an event, the less likely it is to be foreseen. Jervis's analysis of complex systems may therefore be more of a contribution to the historical sciences than to predictive social science.
SHOULD “SYSTEMS THINKERS” ACCEPT THE LIMITS ON POLITICAL FORECASTING OR PUSH THE LIMITS?
Philip E. Tetlock, Michael C. Horowitz, and Richard Herrmann
ABSTRACT: Historical analysis and policy making often require counterfactual thought experiments that isolate hypothesized causes from a vast array of historical possibilities. However, a core precept of Jervis's “systems thinking” is that causes are so interconnected that the historian can only with great difficulty imagine causation by subtracting all variables but one. Prediction, according to Jervis, is even more problematic: The more sensitive an event is to initial conditions (e.g., butterfly effects), the harder it is to derive accurate forecasts. Nevertheless, if awareness of system effects can help forecasters better calibrate their probability estimates of whether or not certain events will come to pass, systems thinkers who are pessimistic about prediction are diluting their confidence too much. The challenge is a meta-cognitive one: thinking systematically about when to engage in systems thinking; and weighing the costs and benefits of using simple or complex heuristics in policy environments that can shift suddenly from quiescence to turbulence.
SYSTEM EFFECTS REVISITED
ABSTRACT: System effects often stand in the way of attempts to come up with simple explanations of politics. Systems are often characterized by nonlinearities, where an effect is more than the sum of the effects of the actions taken by multiple actors. Another system effect is feedback, where the effect of actions is to amplify the problem the actions are intended to solve. There may also be indirect effects, where an incidental aspect of an action becomes more important (to other actors) than the primary intention; contingencies, such that an effect is not inevitable but depends on idiosyncratic or even anti-strategic initial actions; interaction effects, where the behavior of an actor changes the environment of action, so that other actors do not respond as anticipated; and unintended consequences, where the long-term or secondary effects of an action differ from the intended effect. Each of these system effects can frustrate scholarly attempts to understand political behavior using simple models of action, and, even more, can frustrate the attempts of political decision makers to predict the effects of their actions.