George A. Akerlof

This was a quick read with some important ideas. The core idea is that we can integrate identity into standard economic models by formalizing identities as sets of social norms and assigning costs and benefits for both following and not following those norms.

Let's expand on that. Following the authors, we define identities as sets of norms that people generally expect others to follow. These can be the usual identity categories like gender or race. They can also encompass identities such as whether or not an employee has an "insider" or "outsider" identity -- that is, do they see their job as just a job or as part of their personal identity. People decide which norms apply to them by observing others. This is necessary even if one does not want to conform to the norms -- because not following norms generally imposes costs which are important to understand.

Norms and identity as a general category are neither good nor bad. In this model, what makes an identity good or bad is the cost of nonconformance. Although the authors do not go into it (that I remember), in this model I would define as stereotype as sets of norms that are applied to people who would prefer not to be held accountable to those norms (especially, but not only, when those norms make inaccurate factual claims in addition to normative claims).

To formalize the notion of the costs and benefits of identity, the author apply a four step process:

1. Associate individuals with social categories.
2. Specify prevailing norms for these categories.
3. Posit individual gains and loses from different decisions given particular identities and norms.
4. Combine these with other forms of economic analysis to determine what people will do.

Note that this is a descriptive process, not a prescriptive process. Thus, these steps are not about attempting to define and enforce norms. They are about describing the categories and norms that exist in the world and influence people's behaviors.

That's the model. The rest of the book applies the model to a number of case studies including organizational identities, student association with education, gender, and race.

The authors also spend time discussing why this is a useful extension to the economic model, especially given that many of the case studies have "well, duh" sort of conclusions. E.g., people who identify with an organization will be more honest with the resources of that organization or people who work in counterstereotypical professions will experience negative costs, including discrimination or harassment from those who want to maintain the existing norms.

But that "well, duh" is kind of the whole point. These conclusions seem obvious because they're such an important part of how the world works. Other economic models -- including other types of non-strictly-monetary utility such as those deriving from taste, imperfect information, or cognitive biases -- don't predict these obvious outcomes. The authors do not claim that identity is the only or even the dominant form of utility if most decisions. As step 4 of the process says, "Combine [identity factors] with other forms of economic analysis to determine what people will do." However, identity based norms have an influence on our lives, and if we want our economic models to be useful tools for society, we need to take them into account.… (more)