Over the last decade or so, we have seen a long list of excellent books making economics more accessible to a general audience and showing how its insights can be applied in unexpected ways in policy and in daily life. Some of the authors (Tim Harford, Dan Ariely, Steven Levitt, to name just a few) have almost reached superstar status.
One can thus wonder what Gneezy and List could possibly add to this list. A lot, it turns out.
Gneezy and List provide us with a highly accessible survey of their own work, which is, in many ways, revolutionary - they are pioneers in the applications of Randomized Controlled Trials (RCTs) in economics.
Just in the case you don't usually discuss RCTs during breakfast, let me explain in a few words what they are and why their application to economics is so revolutionary.
For a long time, empirical economics had to rely on data that were collected anyway by the government: gross domestic product, public debt, unemployment, inflation, changes in the legal context (for instance, changes in unemployment benefits or in minimum wages, trade treaties) etc. This had a very unpleasant consequence: whenever an economist tried to examine the impact of a policy change (say, for instance, the introduction of minimum wages) on a variable of interest (unemployment), he could never be sure that the observed effect was indeed due to the policy variable of interest. For instance, if a minimum wage is introduced around the same time as a reduction in tariffs, it may well be that the higher unemployment of low-skilled workers is due to cheap imports of products manufactured by low-skilled labour abroad. Or, conversely, if the introduction of the minimum wage goes hand in hand with a large fiscals stimulus, unemployment could actually decrease (even though it may have increased without the stimulus).
Econometricians have developed a lot of extremely clever tricks to disentangle these issues. To give just one example: sometimes there is a variable that is not under the control of anyone involved, and that does allow you to identify clear causal links. For instance, in a classic paper that would have repercussions far outside academia, Steven Levitt showed how the timing of changes in abortion laws in different US states could serve as an indicator of the number of unwanted children that would not be borne, and that this would be a good predictor of crime rates two decades later. (In this case, the causal link is that unwanted children are more likely to commit crimes later in life).
However, such natural experiments are not not always available for the questions that have the highest policy relevance. As a result, the fundamental problem remained that many economic variables change simultaneously, and that it is very hard to attribute causal links. When economists had to explain these issues to other scientists such as, for instance, biologists or chemists, they would typically answer: "What do you want us to do, we cannot set up laboratory experiments like you guys?"
Except that you can. This was the revolutionary contribution of Daniel Kahneman and Amos Tversky on the one hand, and Vernon Lomax Smith on the other hand: they would recruit some people (usually students) willing to participate in an experiment that would simulate some real world setting of interest to economists. This approach has led to new insights, some of which have led economists to question some of the received wisdom of their profession.
However, the laboratory approach is not without problems of its own. One central issue is the artificiality of the setting: people know that they are part of an experiment, and that they are being observed. And this can create a new class of problems. For instance, some experiments showed that people tended to be much more generous than economists would assumed, even if (in the setting of the experiment) they could get away with purely selfish behaviour. What researchers oversaw, was the possibility that the participants behaved as they thought the experimenters wanted them to behave. In other words, the volunteers behaved altruistically in order to please the researchers rather than the participants.
This new insight can be attributed to John List, one of the authors of this book. In short, what Gneezy and List propose, is to continue performing experiments, but in the field. In their research, they would typically split a sample of people in groups (a reference group and one or more treatment groups), but without telling them that they are subject to an experiment (exactly as what happens when new medical treatments are being tested). The treatment groups would then be subject to specific incentive schemes in a natural experiment, where the stakes are real.
The book gives an overview of some of the questions that were tackled by Gneezy and List, some of them in areas that you would not expect in an economics textbook: how do you make sure that parents will be in time to fetch their children at daycare, what are the determinants of discrimination, to what extent are differences between male and female behaviour due to nurture rather than nature, is there any proven way to help very young children from vulnerable family backgrounds, is it even possible to save teenagers living in poor and violent neighborhoods, how can we motivate people to donate more, etc.
Of course, I will not tell the answers here. Go and read the book. It will not disappoint you, and, if you are a policy maker , a manager or a social worker, you will find ideas that you may want to apply.