Economists often argue that in their research their engage in positive science, which means that they use models of human behaviour to test the consequences of various policy scenarios. Importantly, they do not engage–in their own contention–in normative analysis, i.e., they do not attempt to formulate prescriptions as to which policies/modes of behaviour are the right ones (except when “right” means “welfare maximising”). There is much to be questioned about that, including whether economics actually deserves being called a science or whether welfare maximisation is or, in fact, can be normatively neutral as a source of guidance for analysis. I will not dwell upon these questions today. Rather, today’s topic is the model of human behaviour conventional economics rests upon and whether one can call the application of this model–the so-called homo oeconomicus–positive analysis.
One of the most famous proponents of the idea that economics is essentially only about positive analysis and that normative questions are beyond the discipline’s realm was the late Milton Friedman, author of Essays in Positive Economics, one of the intellectual leaders of the Chicago School of Economics and a highly controversial figure, indeed.
Friedman and other members of the Chicago School were the creators and propagators of Rational Choice Theory. This theory or, rather, model of human behaviour is still very influential and widely applied in conventional economics today. Its basic assumption–indeed, its essence–it the homo oeconomicus, a representative agent who has stable, unchanging and pre-existing preferences, perfect knowledge and information, knows no uncertainty or ambiguity and acts to maximise a utility function which includes only his self-interested goals and in most applications equals income. Based on this, economists in and beyond Chicago have assessed and evaluated policies, projects etc. for decades.
I won’t dwell here upon the question whether homo oeconomicus represents a moral human-being–while Friedman and others argued that there is nothing morally wrong in being selfish*, this is irrelevant for our discussion here, for we want to talk about positive, not normative analysis. But is economics based on Rational Choice Theory a positive science?
Well, the problem is that Friedman’s claim that economics is and should be positive, and the behavioural model he was a proponent of are contradictory. There is no such thing as homo oeconomicus–psychological research has been showing us for decades that people do not act “rationally”. Indeed, at least two “Nobel Prizes” in economics went to psychologists who proved Friedman and his disciples wrong (Herbert Simon in 1978 and Daniel Kahneman in 2002). We do not exhibit anything like stable preferences. Often, we do not have any, so they have to be “created” in the course of social processes. We do not possess perfect information and knowledge and are not even always able to aquire them if needed (as some more “progressive” models assume). Every day we have to deal with uncertainty and ambiguity–and we are not very good at that. And last but not least, our motivations go far beyond self-interest and money. So, behavioural assumptions based on the vision of a homo oeconomicus are wrong.
But wasn’t it Milton Friedman who in the most famous of his essays in positive economics argued for a kind of Ockham’s Razor, i.e., models as simple (simplistic?) as possible, provided they offer realistic predictions–which he claimed the Rational Choice does? Unfortunately, the work of psychologists, behavioural and experimental economists and, more recently, neuroeconomists has shown that the behavioural predictions of standard economic models have little to do with how people really behave and make decisions. This ranges from micro-environments such as simple games (e.g., the classic Ultimatum Game from game theory) to highly complex macro-environments such as financial markets (see, e.g., this book and that one). Today, it should be obvious that homo oeconomicus is nothing but an illusion, a myth.
What does this mean for Friedman’s and other economists’ claim that their science (?) is positive? Well, it means a lot. An analysis is positive rather than normative as long as it generates predictions showing what will happen when a policy is inforced given how people behave. Economics based on Rational Choice Theory–i.e., most of economics–instead offers predictions based on how “rational agents” behave. Since those hardly exist and cannot be assumed to be a non-distorting simplification (see above), an analysis based on the assumption of their existence becomes in effect normative–“Look, if you would behave like a bunch of homini oeconomici, this policy would be welfare maximising whereas that other one would be Pareto-inferior. Therefore, we all should behave so to achieve this optimal, efficient and welfare maximising outcome.” (the latter sentence being mostly implicit but always present)
So, must economics be normative in the sense of my above argumentation? Or is there a way out of the dilemma? I believe that there is (I leave aside the question of the definition of the goals against which economics has to analyse and assess policies). There are ways to incorporate the insights from psychology into economic analysis (see the work of Herbert Simon, Daniel Kahneman, Amon Tversky, Richard Thaler, Paul Slovic and others). The resulting models are not as unambiguous and simple as those based on the assumption of homo oeconomicus–closer to reality means closer to the mess of our everyday world–, and they surely are themselves imperfect and can’t incorporate all the biases, limitations, heuristics etc.–no model can–, but they are much closer to the ideal of positive economics than what Friedman and his followers are calling it.
* – In this context, the famous baker-example by Adam Smith is being invoked, which is another case of misrepresentation of quotes. As pointed out by Amartya Sen, Smith believed that self-interested action is beneficial for the whole society only under very specific circumstances and that this is not a generalizable rule.
Further Reading:
- Daniel Kahneman Thinking, Fast and Slow, Penguin, 2012.
- Amartya Sen Rational Fools: A Critique of the Behavioural Foundations of Economic Theory, Philosophy and Public Affairs 6(4), 1977.
- Milton Friedman Essays in Positive Economics, 1953.
The Sceptical Economist 
The first time I do not quite agree with your post:
– the issue of imperfect information is no falsification of the homo oeconomicus model: if you make a decision under imperfect information your decision is – within this range – still rational. It is rational to make decisions under the condition of imperfect information as gathering more information usually implies high transaction costs that will potentially neutralize the benefits of a better informed information. Given the uncertain benefts of making decisions under (near) perfect information (often the rule of the thumb is already the best decision) it is quite rational to decide without looking for more information as already available;
– rational choice does not imply a decision based on self-interest (in the strict sense) or money – altruistic or ethics based decisions are rational as well or can at least be described within the boundaries of the model. This is simply another type of rationality but by far no irrational behavior (remember the statements of Dawkins on the egoistic gene – he uses the rational choice model to explain altruistic behavior all the time);
– unstable preferences are less a problem for the rational choice model as usually short term behavior is analyzed by it. On short terms preferences usually are stable. There is, however, a preference-based conflict with the model which consists of cases where people exhibit contradicting preferences at the same time;
– the contradictions to psychology are often limited to single cases. By contrast, the rational choice model applies to aggregate (collective) behavior. Its assumptions and often predictions are similar to the value-expectations theory from psychology. There are, however, some psychological biases on an aggregate level like i.e. risk averse behavior under certain conditions. But even regarding these biases you will find explanations in the literature moving within the boundaries of the model;
– the main problem with the homo oeconomicus or rational choice model in my mind is not its simplicity and that it can be easily falsificated but rather that the model is so broad and vague that it tends to become a tautology. actually it cannot be falsificated at all. if there is any behavior which seems to contradict the model, apologists can simply claim that we just have not fully understood the rationality applied or the conditions of action.
– I do agree with your normativity-issue: economics or economic thinking can shape the world according to its assumptions. Callon (1998) demonstrated that people educated in economics tend to redefine formerly non-economic areas into “markets”, thus making economic laws apply where they did not apply before (the famous strawberry market example). Yet the normativity here is not the tendency to make people behave rational but to behave according to a very special rationality which is the efficiency rationality of orthodox economics. However, as far as I know, Callon’s findings are not uncontested and sometimes viewed as ‘not wrong but neither representative’ (i.e. MacKenzie/Millo (2001) in American Journal of Sociology).
notes:
Michael Callon (1998): The laws of the markets, Oxford.
regarding the topic in general I like the book by Gebhard Kirchgässner (2008): Homo Oeconomicus, 3rd edition, Tübingen (in German).
“The first time I do not quite agree with your post”
Well, this moment were bound to come someday;-) Although I do not think that we really disagree, rather, we are defining the terms differently. In fact, you defined rationality in a very broad manner–which is not how economics defines it. The very notion of imperfect information is, as I briefly mentioned in the text, an enhancement of the original model of a rational agent, which lacked this idea altogether. It was Neokeynesians who, following Herbert Simon, introduced the notion of imperfect information in their models. Friedman & Co., including the founders of the Rational Choice Theory (if I recollect rightly, it was Lucas who developed the formal model, while others–Friedman, Becker etc.–have applied it), did not include this in theirs.
Look at all the typical economic models, especially macroeconomic models. While economists may say that it is not only self-interest that drives our choices etc., when it comes to formal modelling, utility mostly depends on income only. I do not criticise the rhetoric here, but actual applications.
I mentioned unstable preferences because they are crucial in my own area of research: one of the premises of the more “orthodox” economic valuation theory is that preferences for environmental public goods only have to be elicited, since they are somehow pre-existing. Which is absurd, of course. Also, related to the notion of stable preferences and, in fact, Revealed Preferences Theory in general, is the problem of consumer sovereignty, which again rejects the ideas of both imperfect information and second-order preferences (a smoker might have second-order preferences for a healthy life-style, but she still smokes, because that reflects her first-order preferences within the life-style she is able to lead currently, given all kinds of internal and external limitations).
You are right that when we try to relax the assumptions of rational choice, the whole models looses any predictive power. Which may be the reason why economists stick to the simple, flawed model when it comes to application.
Callon’s argument sounds much like Polanyi’s from his “Great Transformation” (somewhere on my to-read list…), who argued that market economy/capitalism is a sociological driving force leading towards the “incorporation” of what economics would define as rational in our psyche. But that’s not so much about economics, in my view, but about the economy (of course, the structure of the economy is shaped by what economists tell politicians and the public when asked for advice). The normativity of economics results, rather, from its models that implicitly tell us how we should behave to achieve efficiency and narrowly defined welfare (my point from the post), and, which I see as more problematic actually, from the definition of “welfare” it uses. Even the supposedly normatively neutral Kaldor-Hicks criterion isn’t. I could imagine that even classic Pareto improvements (without Kaldor-Hicks) are not necessarily always morally defensible. But that’s what economics claims. Amartya Sen wrote a nice book on that subject–On Ethics and Economics (the UFZ library has it, if you are interested;-).