//Briefing: Is Economics a Science? Explaining the resurgence of Experimental Economics

Briefing: Is Economics a Science? Explaining the resurgence of Experimental Economics

In this post, our Economist on the ground, Jeremy Kamil, brings us a background piece on Experimental Economics. Which reminds me, have you ever heard the one about the guy who asks you to push a button or pull a lever in the name of ‘science’ (pic on website):

Source: http://www.economics.com/
Source: http://www.economics.com.au/

Is economics a science?

That question, although not the central thesis of our analysis today, is an important starting point. A science, loosely defined, is a discipline that builds knowledge through observation and experimentation to explain naturally occurring phenomena. Now pick up any introductory Economics textbook, and they will tell you that modern economics is concerned with the “efficient allocation of scarce resources”. As most of you will attest to, this relies on a market mechanism, and mutually beneficial trade. So far, economics looks nothing like science.
Indeed, unlike students of physics, chemistry or biology, the economic student never sits in the lab to test hypotheses in an attempt to explain naturally occurring phenomena. Milton Friedman, the famous economist, wrote in 1953 (Friedman, 1953:p.10):

“Unfortunately we can seldom test particular predictions in the social sciences by experiments explicitly designed to eliminate what are judged to be the most important disturbing influences. Generally we must rely on evidence cast by the ‘experiments’ that happen to occur.”

Friedman is confirming what we suspected: in economics we cannot create controlled experiments, thus we cannot test a hypothesis in the lab to explain the real world. Thus, lab experiments in economics were believed to be artificial, and so the results could not be trusted. Ergo, since lab experiments could not be used to test economic theory, economics had no place as a science. Economic theory throughout the 50’s and 60’s relied on untested assumptions, which were justified on the grounds that they allowed economists to produce useful models in predicting human behaviour.
It was not until the 1980’s that perceptions began to change. Economics was being treated more and more like a science. Economists were finding themselves in laboratories, musing over experiments, trying to examine the fundamental assumptions that built up their theories. In fact, economic theory always attempted to explain naturally occurring phenomena. Where it broke down as a science was the inability of economists to test these hypotheses in a controlled environment (‘the lab’). But now, with the rise of what is now known as ‘experimental’ or ‘behavioural’ economics, economists were able to test their theories in the lab, and slowly began entering the scientific fraternity.

Experimental Economics Defined

But what was it that allowed such a drastic rethink of economic study? Before answering this question, it is more useful to first ask ‘what is experimental economics?’ Experimental economics can be broken down into 3 parts, as detailed below.

The first area of ‘experimental economics’ is lab experiments. Economists can test economic theory by placing a group of people in a laboratory, and testing how they respond to various incentives, under varying circumstances. This has the advantage of allowing human behaviour to be tested in a controlled environment: the tester can alter the conditions slightly, to see what the key motivations behind an individual’s response are. With the rise of lab experiments there is no need to assume how people behave; economists can accurately test how people behave. As long as the incentives are real the results are transferrable to the real world. Lab experiments help economists make predictions as to how people will react in different conditions, based on actual observation of human behaviour, not basic assumptions of human behaviour. This has lead to a strengthening of many economic theories, as well as improved policy development, as will be outlined later.

A second area of ‘experimental economics’ is field experiments. Field experiments are very popular in development economics. For a field experiment, the tester must first design a hypothesis. He must then decide on a set of interventions to test the hypothesis. A classic example was conducted in India, where it was felt that school teachers were not coming to school, as there was insufficient monitoring of the school system. (The hypothesis.) The interventions decided upon were that teachers were to be paid on appearance, and they had to confirm appearance via a photograph. To test whether or not the interventions were successful, the tester must set up two different groups: the treatment group, and the control group. The interventions are only enacted in the treatment group. After a period of time, the tester can compare results between the two groups, and if there are significant differences, then the interventions can be determined to be successful. The results, as with lab experiments, can strengthen economic theory as well as help determine policy.

A third and last area of experimental economics is that of natural experiments. These are used rarely, as they are difficult to find. They work similarly to field experiments, except that in natural experiments, the two groups are identified to occur naturally in history, there is no intentional allocation to a treatment or control group. Essentially, what is required is a change of the circumstances experienced by a subpopulation, within an entire population. We can use any economic differences that then emerge between the groups to determine future policy, based on the inherent differences between these groups. The classic example of a natural experiment is the divide of Germany into East Germany and West Germany. The progress of West Germany, compared to the poverty of East Germany, creates the obvious conclusion that the capitalist regime of West Germany was preferable to the communist regime of East Germany. For the economist, the Berlin wall created a perfect, naturally occurring, division of the German population into ‘treatment’ and ‘control’. The real art of natural experiments is to identify creatively when a glitch of history, or circumstance, or policy has formed two comparable populations.

The resurgence

So why the resurgence of experiments in Economics? An important aspect to realize is that neo-classical, or modern economists, have long ignored the importance of diversity in human behaviour in their models. They have simply assumed that all people always act perfectly rationally, regardless of circumstance, despite the obvious evidence to the contrary. If we delve a little deeper into economic history, we discover human behaviour has not always been ignored by economists; in fact quite the opposite is true. Lord Keynes, considered the father of macroeconomics, wrote that people most definitely acted irrationally in time of fear, which (partly) explained the persistence of the great depression. He writes about human behaviour in his famous book: ‘A general theory of employment interest and money.’ Going back even further, Adam Smith, the man regarded as the father of economics, wrote ‘the theory of moral sentiments’, which is a book attempting to explain human behaviour, before he wrote his most famous piece, ‘The Wealth of Nations’. Both these men, two of history’s greatest economists, were acutely aware of the importance of variable and complex human behaviour. They just lacked the means to test the validity of their claims, so human behaviour was never explored fully, and neo-classical economists and their assumptions of perfect rationality took over.

All this changed with Vernon Smith’s 1976 article, Experimental Economics: Induced Labour Theory. He defended the rights of economists to test behaviour in the lab. He contended that as long as economists follow certain guidelines, such as ensuring the incentives on offer are real, there is no reason the results won’t be transferable to the real world. Secondly, he argued that economic experiments add value to economic research, since the initial paradigm of any experiments can be reset, and a hypothesis can be retested, until the key factors in determining the hypothesis can be uncovered. This can’t be done in the real world. With these two basic points, there was an exponential rise in the number of experiments being conducted by economists. These results were being published, and were being used to build up new theory and policy.

Experiments and some sacred cows …

So, what were these new theories being developed? Experimental economics has challenged the basic assumptions of neo-classical economics. From experimental economics, behavioural economics has risen. Behavioural economics relies on experimental results to test human behaviour, which helps make more accurate models. These models contribute to both economic theory, and economic policy. Below is a brief discussion of some of the changes that have occurred to economic theory.

People are no longer considered to be perfectly rational. Rather, they have what is known as ‘bounded rationality’. Their ability to make a rational choice is limited by the information they have, and the time they have to make a decision. Continuing in this vain, the research has determined that people are creatures of habit. People will make a choice, not because it is the most rational, or the ‘utility maximizing’ choice (as Economic models assume), but because they have done so in the past. The Neo-Classical notion of ‘revealed preference’ is another assumption which is not supported in the lab. People are not consistent in their choices or rankings over goods or services; these can change depending on how they are presented to the decision maker (the so-called ‘framing effect’, like how a picture frame alters the way we see reality). The moral from these lessons can be linked to the clearly incorrect assumption of perfect rationality. People do not sit down, and perform calculations, and cost benefit analysis on every decision they make, for a number of reasons. They may lack the time; they may lack the information, or maybe they value something other than monetary gain, such as happiness which doesn’t depend simply on being more wealthy. Whatever the reason, it has become clear that people do not act as neo-classical economists expect them to.

Furthermore, the value of experimental economics can be used to settle a classic debate in economics: Keynes VS Friedman: Does spending your way out of a recession really work? Keynes believed that people would spend a constant faction of their income, thus government transfers to the people will be spent accordingly, raising demand and ending the recession. Friedman’s permanent income hypothesis, which is opposed to Keynes’s theory, states that when income increases temporarily, people will not increase their consumption significantly, as they spread this income over their lifetime. Friedman’s theory has an important assumption: that people are forward looking, and make decisions in the best interests over their lifetime. Experimental economists beg to differ. They find, on evidence produced in a lab, that people put too much emphasis on present consumption, and too little emphasis on the future. In other words, most people will spend one off increases in income; the additional income won’t be saved as people don’t value spending in the future as much as the permanent income hypothesis would predict. Experimental economics has brought into question the assumption of perfect rationality, and with it, Friedman’s permanent income hypothesis. Now this kind of work becomes critical if you are a government faced with a deep recession with its associated unemployment and dramatically reduced spending. The experimentalists tell us that if a government puts some more money in their citizen’s pocket in a one-off ‘stimulus’ plan, they will in all likelihood head to the shops and spend it, thus holding up demand and saving not just a whole lot of jobs, but most likely the government of the day.

Some criticisms

Of course, as with any new branch of economics, experimental and behavioural economics has received some criticism. Some of these critiques include the selection process through which those who participate in the experiments are recruited. Most participants are students, and so it is argued, are not representative of the whole of society. Other issues raised include the fact that lab and field experiments are most often conducted in the short run (on the scale of hours or days), thus ignoring long run effects; or they are tested only local populations; and when introduced into the economy as a whole, there may general equilibrium effects that weren’t considered in the initial experiment (e.g. complex feedbacks between micro- and macro- variables that could never be replicated in the lab). In other words, experiments that test results in one market, don’t consider how changes in that market will then affect other markets. However, most experimental economists point out that this isn’t a criticism of the use of experiments per sae, only the current methods being used. As the field develops, the methods will become more refined, which will only strengthen the field.

Looking ahead

Here at Monash, there is strong research in the area of behavioural and experimental economics. Some of the issues currently under the microscope include decision making under uncertainty, corruption, and an examination into what leads people to act dishonestly, and the economics of happiness, which aims to uncover whether or not money is the only incentive people respond to. The field of experimental economics has grown significantly over the last 20 years or so, and with good reason. The ability to test behavioural assumptions has only strengthened the field of economics: theories and policy are now based on actual human behaviour, not expected human behaviour. As a result, the field of economics is in a far stronger position to better explain naturally occurring phenomena, and gain that elusive acceptance into the field of science.


  • Friedman, M. (1953), “The Methodology of Positive Economics: Essays in positive Economics”, University of Chicago Press.

Dr SIMON ANGUS is a computational and complexity scientist and member of the Department of Economics, Monash University. With a background across the physical and social sciences, he has diverse interests including complex systems science, data-science, networks, systems biology, evolutionary game theory and the study of technology.