Thaler's Misbehaving is a personal account of the development of modern behavioral economics. It is not the history of behavioral economics. It is a (part of the) history of behavioral economics. Thaler is a contemporary witness, and at the same time one of the major figures in modern behavioral economics.
I like Misbehaving for (at least) two reasons.
First, Thaler establishes very early and often reinforces later that standard economic [consumer / decision] theory, rational choice theory is a normative theory. It describes how people should behave if they were to optimize their utility. It (often) does not describe what they really do. Rational choice theory is based on mathematical axioms, not true human behavior. For many purposes, this is absolutely fine. In many contexts, the observed aggregate behavior is driven more by the institution than the individual. For describing human decision making, for predicting an individual's choices it is not. This is where we need a positive, descriptive theory.
Human cognition is bounded. Full rationality (in its mathematical definition) is, therefore, impossible. Bounded rationality is the best we can hope for. And this is the core of behavioral economics.
Without a pre-existing unifying model to compete with the dominant Rational Choice Theory research had to start with identifying “anomalies.” Thaler did exactly this. He reports many of the initial hostilities and criticisms against his heretics, the abandoning of the dominant doctrine. Sometimes he also reports a researcher's conversion as a result to economics becoming a more empirical science. Nevertheless, still today some colleagues, and even colleagues among the experimental economists, would start to defend Rational Choice and Expected Utility Theory even if I just described it as a normative and not a positive theory.
The still standard normative economic theory approach can serve many purposes well and is often easier than more realistic approaches. As-if utility maximization has its purpose. Yet, as the sole policy analysis tool it may lead to the wrong conclusions and should, therefore, be augmented with the many insights we have gained from neighboring fields and the empirical economic research of the decision maker. A recommendation that, obviously, also Thaler advocates and has already helped to implement on several occasions.
Second, somewhere in the middle of the book Thaler alerts,
Tempering expectations about the magnitude of the sizes of effects that will be obtained is important because the success of […some nudges…] can create the false impression that it is easy to design small changes that will have big impacts. It is not.
It is also crucial to understand that many improvements may superficially appear to be quite small: a 1 or 2% change in some outcome. That should not be a reason to scoff, especially if the intervention is essentially costless. […] A 2% increase in the effectiveness of some program may not sound like a big deal, but when the stakes are in billions of dollars, small percentage changes add up. As one United States senator famously remarked, “A billion here, a billion there, pretty soon you’re talking about real money.”
I believe this statement is more important than its place in the book and its extent of the discussion in the book implies.
In the laboratory, we are used to large effects. Experiments are often designed such as to generate as large an effect as possible. Even though the lab is the real world with real world incentives and real world decision makers, decisions outside the lab are made in a context that matters, after a series of other different decisions that matter, by more heterogeneous decision makers what matters, too. This is not just additional noise. These factors need to be investigated as well. Yet, this means that an effect in the field of maybe 2% when standard theory would predict none is huge.
Of course, this also has implications for research. Experimental results obtained under “clean” conditions with small samples in the laboratory will not always translate to similar effects outside the laboratory. The small samples imply that statistical significant effects may be over-estimated. The “clean” lab environment may lack moderating factors. Hence, large-scale field studies will become more and more important as the basis for evidence-based policies. We have already begun to see this.