WHEN I STARTED studying economics some 17 years ago, I was motivated by a strong desire to learn about the mechanisms governing the economy, society and the state. I wanted to know why some people and some countries are rich and others are poor, and why some goods are expensive while others are cheap. I assume the same is true for most polit students today. (Ok, I also thought that studying economics will earn me a higher salary later on.) I was then initiated in standard economics which explains a wide range of phenomena ranging from consumer choice to market interaction, from fertility choice to drug addiction, by assuming that all economic agents are rational and strictly egoistic.
Along with awe and excitement about the elegance of standard economics, I also felt an uneasiness with my field of study that I found hard to articulate at the time. It appeared to me that the models I learned are highly abstract and unrealistic, and that economics is more about formal models rather than about what real people do. It should be obvious – at least to any non-brainwashed person – that real people are neither fully rational nor strictly self-interested as standard theory assumes. Instead, there is considerable heterogeneity: some people are more rational than others, some people are altruistic, some are reciprocally motivated, some are spiteful, and some are just plain egoists. On the other hand, I also recognized that standard economics is a powerful tool to predict aggregate behavior, for example how markets respond to exogenous shocks. I was puzzled: How can it be that an obviously unrealistic theory provides predictions that seem to fit the facts? This puzzlement was at the origin of my interest in behavioral and experimental economics – and it is still on mind today.
“First wave” behavioral economics
Behavioral economics attempts to explain what real people like you and me do, and its proponents believe that using a more realistic model of man will make the predictions of economics more accurate. In what I call “first wave” behavioral economics, researchers provided evidence demonstrating that real people in fact are boundedly rational and that they do have social preferences. Much of the evidence on what seems like “silly” behavior (this is called “biased” or “anomalous” behavior in the literature) from the perspective of standard economics was provided by psychologists like Daniel Kahneman and Amos Tversky, and they mainly used hypothetical questionnaires to demonstrate the many ways how Homo Sapiens differs from Homo Oeconomicus.
I just mention a few objections to “first wave” behavioral economics that were common until recently: “We are not interested in what people say they would do in some hypothetical situation, but what people actually do when exposed to economic incentives.” And: “Yes, there are people making silly choices, but they will quickly learn to make smart choices.” Or: “A few silly people will not move the world. Instead, there will be smart people who correct (and exploit) their mistakes. Therefore, the silly people will have a negligible impact on aggregate (e.g. market-level) outcomes, and we as economists can safely ignore their silly ways.” These are important objections, and the debate about their relevance is still going on.
“Second wave” behavioral economics
In my course on behavioral and experimental economics (starting in September), I focus on how institutions interact with “silly” behavior, and I call this “second wave” behavioral economics. The question is how real institutions (like markets, but also bargaining, voting, and communication) transform various biases. The message is that “anomalous” behavior can matter much in some environments, but matter very little in other environments. It is therefore possible that the same people will bring about outcomes that look as if everyone was silly in one institution, and that look as if everyone was fully rational and egoistic in another institution. Experimental methods serve to identify these institutional conditions, and to test when standard economics predicts real behavior well, and when it does not (students in my lectures are surprised how well standard theory can predicts behavior in some cases).
Experimental economics
In the field, it is very difficult to identify why people behave in sometimes surprisingly silly ways. Often, there are alternative explanations for observed behavior. People may, for example, behave in ways that look silly because they have limited information rather than actually being boundedly rational.
To illustrate how difficult it is to identify the exact reason of “silly” behavior in the field, consider the following example. The summer 2004 was lousy everywhere in Europe, but the summer of 2003 was excellent. In Switzerland, for example, the average temperature in June to August in 2004 was 3.1 °C lower than in 2003. Within this one year, the number of newspaper reports in Swiss newspapers on “global warming” fell by 57% and on “glacier dying” (the retreat of Alpine glaciers) fell by 82% (Weltwoche 35/2004: 90). How can such a dramatic short-term swing in reporting on very long-term issues be explained? One explanation refers to silly journalists. They may be subject to cognitive biases putting too much importance on dramatic events and things that happened recently (“its so hot his week, that must be a sign of global warming”).
Another explanation is that journalists are fully rational but must write what people want to read, and their subscribers are subject to these biases (“our subscribers don’t want to read about global warming when we freeze to death in July”). Another explanation is that neither journalists nor subscribers are biased, but that journalists are under time pressure and just copy what other journalists write about (this property is called “strategic complementarity” in the literature. In this context, it means: “If they write about it, it must be a ‘hot topic,’ and we must report about this too.”) Should we conclude from these erratic reporting patterns that Swiss journalists and their subscribers are extremely silly people? Not necessarily. Suppose that journalists and subscribers are fully rational but that journalists (erroneously) believe that subscribers only want to read about global warming when it is hot. The journalists’ over- and underreporting then becomes contagious because of strategic complementarity. This example illustrates that silly social outcomes are possible even if no-one is silly (but some have biased beliefs and incentives to act on them). While it is hard to tell whether this is a reasonable explanation for what happened in the Swiss media, we can test related hypotheses in the laboratory by measuring and controlling relevant factors (like beliefs and the degree of strategic complementarity).
So, how do I deal today with the puzzlement I felt when I was student? What do I think does behavioral and experimental economics imply for way we should teach economics? For the way we do theory? For the policy advice we give? That could be a topic of my next contribution to Polit’en.
Martin Nø...
I skriver, at ‘tunge viden...
2