Economic Experimentation Lea Oksman • Can experimental economics save social science? January 2003

Maier’s law says that if the facts do not conform to the theory, they must be disposed of. While most do not know who Maier was, many people swear he was a social scientist. The social sciences – psychology, sociology, and the like – are perpetually hovering on the boundary between real science and quack rationalizations. Sociological findings are invariably debatable, and the most interesting research in psychology usually has more to do with other fields – for instance, neuroscience – than with psychology itself.

Problems for social sciences invariably become problems for philosophy. Ultimately, philosophy, particularly natural law, has to have coherent understandings of human nature. How can we claim that someone’s behavior undermines his nature as a human being without being able to explain what that nature is? Hence, if social science cannot explain human tendencies in a coherent manner, then natural law philosophy becomes difficult to do. And if the Right continues to rely on natural law theory as the basis for their ethical claims, then problems with the methodology and claims of social scientists become problems for the Right.

The chief trouble with social science is that studies must target a range of material that is too broad – the complete human personality. Consider psychology, for example. Every theory, by its very nature, must attempt to explain all of behavior: since the personality is a single unit, it is arbitrary to claim that some elements work according to one model while others obey a different one. Yet human behavior is incredibly complex and self-contradictory; the number of psychological patterns manifested in any person is numerous, and the way they are integrated in one person is highly variable. No single theory can encompass this process with enough consistency to make predictions with a satisfying degree of accuracy. Experimental studies do not fare much better than the theoretical work. Their research findings are riddled with unreliable results that are difficult to replicate. To make matters worse, the experimental methods are often biased. It then comes as no surprise that researchers find the exact behavior patterns they were looking for.

One social science, however, does not suffer from these problems. Economists have avoided the multi-factor, multi-theory issue by restricting their domain. They study only a single and seemingly objective aspect of human behavior, financial decision- making. With the simple and largely uncontestable assumption that everyone likes to acquire resources and hates losing them, economics models trade interactions.

This assumption and its consequent research focus, however, pose a new kind of problem. Functional models require presupposing that viewing human behavior in the market is highly rational, completely ignoring a large chunk of human behavior. One aim of economics is to arrive at models that can be used to predict market conditions and behavior. The problem, however, is that economists often attempt to impose the model on reality, rather than the other way around. A new field of economics has set out to resolve these issues.

Experimental economics began when 2002 Nobel Prize recipient Vernon Smith, then a young economist and devout socialist, sought to prove scientifically that markets were inefficient. By running experiments with small groups of students, he demonstrated that interactions between subjects over the prices of items resulted in an equilibrium predictable by the law of supply and demand. This occurred despite the fact that the subjects had neither complete information about the items nor specific trading rules – previously considered crucial assumptions for the law to apply.

Smith subsequently developed a general experimental paradigm. He would divide his subjects into “buyers” and “sellers” of a certain item and ask them to bid for prices for which they would be willing to buy or sell, respectively. Then, he would observe how competitive interaction from both sides resulted in the final setting of a price. Experiments of this sort consistently demonstrated that markets are efficient and predictable from economic models even when seemingly important conditions – such as complete rationality or perfect information – are not necessarily fulfilled.

His research persuaded Smith that economic theory has great predictive power and encouraged him to apply his experimental approach. Smith’s most important innovation in the study of market processes was the introduction of consumer-side bidding as an important force in effective price stabilization. Smith and other experimental economists have used this strategy to design new real-life market models. In an interview in the November 2002 issue of Reason, Smith described a workshop he organized at the University of Arizona with power company executives. They created a wellfunctioning market system in which – to keep prices down – buyers accepted temporary power interruptions. Smith also ran experiments on a computergenerated market system in Australia, which contributed to the government’s decision to deregulate the electricity market.

Smith believes that buyer bidding may lead to more efficient policies in other fields, such as the prioritizing of landing and takeoff times at airports. The implementation of his models requires the deregulation of industries. These investigations, among other motivations, thus led Smith to become a libertarian.

In its methods, experimental economics seems safe from uncertainty. Smith remarked in an interview with Reason that experimental economists “are less interested in what people think than in what they actually do in specific situations”. They avoid the quagmires that trap psychologists and sociologists. The focus is on measurable outcomes: bringing people together and demonstrating their ability to create a productive, rational, and – crucially – predictable system. This tactic, research seems to show, works quite well. It is also confirmed by observation of real-world markets – an important strategy in the field.

Smith’s experiments seem to show that the “rationality issue” is not a significant problem in economics. This is not to say that human beings do not suffer from flaws in reasoning or make bad economic decisions because of emotional factors. Smith’s findings merely suggested that factors balance out, creating efficient markets.

This finding suggests that we don’t quite know what makes markets efficient – the evidence merely shows that they are. This missing information is very important if the findings of experimental economics are to be fashioned into a reliable theory. In the philosophy statement of the Interdisciplinary Center for Economic Science, which Smith directs at George Mason University, he emphasized: “Acknowledging and recognizing the workings of unseen processes are essential to the growth of our understanding of social phenomena, and we must strive not to exclude them from our inquiry […] In this way we at least can attempt to escape the very significant disadvantage of being a human in studying human behavior.”

Another new school of economics – led by Daniel Kahneman, who shared the Nobel Prize with Smith – may be able to contribute on this point. Behavioral economics integrates research in psychology with that in economics; in particular, it focuses on psychological phenomena inconsistent with rational decision-making. An example of these occurrences is the conjunction effect. In an experiment by Amos Tversky and Kahneman, subjects were given a description of a woman that strongly suggested her feminist leanings. A majority then chose “The woman is a feminist bank teller” as being more probable than “The woman is a bank teller.” Yet common sense dictates that a person is more likely to belong to one group than to both simultaneously. Other instances of experimentally verified faulty reasoning include overconfidence and the tendency to “develop” patterns where none exist.

These observations have limited predictive value outside of experimental situations because they cannot at this point be organized systematically. Eldar Shafir and Tversky, leaders in behavioral economics, note that the findings of the field “do not form a unified theory comparable to the rational theory of choice.” This presents a serious problem for behavioral economics as a field; as a supplement, however, to experimental economics the behavioral studies are extremely important for they may be the best approach to understanding Smith’s “unseen processes.”

Behavioral economics, however, still falls into some of the same traps as the other social sciences. Because, like psychology, it attempts to explain complex human behavior, it has a difficult time isolating single factors that affect human thinking. Hence, the merging of behavioral and experimental economics, though desirable, is problematic. Experimental economics has an edge over the other social sciences, particularly classical model-driven economics, because it attempts to observe human behavior rather than explain the hidden processes behind them. Maybe this is the limitation of social science. Until social scientists realize that human beings are not interchangeable cogs with a finite number of physical states, they will continue to produce results that fail to isolate any single factor. It seems, then, that the other social scientists might just have something to learn from Vernon Smith.

Lea Oksman is a freshman in Trumbull College.