The Economics of Discrimination: What is Statistical Discrimination?

An Examination of the Economic Theory of Statistical Discrimination

African-American businesswoman working on laptop in airport
Jose Luis Pelaez Inc/Blend Images/Getty Images

Statistical discrimination is an economic theory that attempts to explain racial and gender inequality. The theory attempts to explain the existence and endurance of racial profiling and gender-based discrimination in the labor market even in the absence of overt prejudice on the part of the economic actors involved. The pioneering of statistical discrimination theory is attributed to American economists Kenneth Arrow and Edmund Phelps, but has been further researched and expounded upon since its inception.

Defining Statistical Discrimination in Economics Terms

The phenomenon of statistical discrimination is said to occur when an economic decision-maker uses observable characteristics of individuals, such as the physical traits that are used to categorize gender or race, as a proxy for otherwise unobservable characteristics that are outcome relevant. So in the absence of direct information about an individual's productivity, qualifications, or even criminal background, a decision-maker may substitute group averages (either real or imagined) or stereotypes to fill the information void. As such, rational decision-makers use aggregate group characteristics to evaluate individual characteristics that may result in individuals belonging to certain groups being treated differently than others even when they are alike in every other respect.

According to this theory, inequality may exist and persist between demographic groups even when economic agents (consumers, workers, employers, etc.) are rational and non-prejudiced.This type of preferential treatment is labeled "statistical" because stereotypes may be based on the discriminated group's average behavior.

Some researchers of statistical discrimination add another dimension to the discriminatory actions of decision-makers: risk aversion. With the added dimension of risk aversion, statistical discrimination theory could be used to explain actions of decision-makers like a hiring manager who shows preference for the group with the lower variance (perceived or real).

Take for example, a manager who is of one race and has two equal candidates for consideration: one who is of the manager's shared race and another who is of a different race. The manager may feel more culturally attuned to applicants of his or her own race than to applicants of another race, and therefore, believe that he or she has a better measure of certain outcome-relevant traits of the applicant of his or her own race. The theory holds that a risk-averse manager will prefer the applicant from the group for which some measurement exists that minimizes risk, which may result in a higher bid for an applicant of his or her own race over an applicant of a different race all other things equal.

The Two Sources of Statistical Discrimination

Unlike other theories of discrimination, statistical discrimination does not assume any sort of animosity or even preference bias toward a particular race or gender on the part of the decision-maker. In fact, the decision-maker in statistical discrimination theory is considered to be a rational, information-seeking profit maximizer.

It is thought that there are two sources of statistical discrimination and inequality. The first, known as "first moment" statistical discrimination occurs when the discrimination is believed to be the decision maker's efficient response to asymmetric beliefs and stereotypes.

First moment statistical discrimination may be evoked when a woman is offered lower wages than a male counterpart because women are perceived to be less productive on average.

The second source of inequality is known as "second moment" statistical discrimination, which occurs as a result of the self-enforcing cycle of discrimination. The theory is that the individuals from the discriminated group are ultimately discouraged from higher performance on those outcome-relevant characteristics because of the existence of such "first moment" statistical discrimination. Which is to say, for example, that individuals from the discriminated group may be less likely to obtain the skills and education to equally compete with other candidates because their average or assumed return on investment from those activities is less than non-discriminated groups.