The book is not long, and it was indeed a quick read, and like Gneezy and List’s The Why Axis the writing is aimed for a more general crowd. This had its merits and demerits, as some found it fascinating and some found it to be “watered-down sociology.” The explicit concept and term `identity economics’ seems to be now incorporated into a higher category of `social economics’, and not addressed separately. (Incidentally, I have no idea why The Handbook of Social Economics does not have a chapter with either Professor Akerlof or Kranton as contributors.)
The ingredients of categories, norms and ideals, and identity utilities were indeed interesting, and identities of course do matter, but I do not wholly agree with the book’s approaches. I do not wish to criticize the book on the basis that here identities and norms are exogenously given, because I see that the dynamic model of their prototype Red-Green model is yet to come. But including identities cannot be a magic bullet for everything (see Arnold Kling’s review), and there are several points which make me doubt them. For instance with identity utilities, it is highly likely that the equilibrium where groups discriminate against each other is the unique equilibrium, or this is as how I understood the book. Is it really? Take Peski & Szentes (2013), i.e., Spontaneous Discrimination. Even though the rationale behind their paper also mostly stems from the sociology literature, and the term `type’ is very akin to `category’ used in Identity Economics, the approach here allows for multiple equilibria where if two groups one group may take color-blind strategies. I think this is more likely, and we will be easily able to find such situations where, for instance, female employers as well as male employers discriminate against female employees.
I did like the part where it pointed out that “The internal consistency of these ethnographies gives a criterion for their validity, though a different one from a statistical test” (Kindle Locations 1723-1728) because I have spend several years arguing that economists should work with anthropologists to cross check whether the optimization problems were indeed framed correctly. Also enlightening was the part about existing theory becoming harder and harder to falsify.
The internal consistency of these ethnographies gives a criterion for their validity, though a different one from a statistical test. Milton Friedman and others have warned us that we should be careful about drawing inferences from what people say: they may misunderstand their own motives, and they may be self-serving. But the best ethnographic studies have a check for this. From the many details they record, and the attention they give to the subtexts of what people say, they construct a consistent picture of people’s behavior. Indeed, the very best ethnographic studies do not just record what people say; they decode what people say and do.
Thus, even in testing a well-specified economic model, the economist has many, many choices regarding how to run her test. Because each of these decisions can be made independently, even the most straightforward test has literally millions of possible specifications. That makes it difficult to follow the dictum to accept the model unless it is rejected; it makes it difficult to falsify any theory. It is rare to find a model, no matter how silly, whose millions of specifications will be uniformly rejected.
If I thought identities did not matter, I would not have spent so much effort trying to heighten the local staff’s morale and bringing them closer to `insiders’ during my days in Malawi. I am still intent on pursuing social economics related topics.