Statistical discrimination

From Wikipedia, the free encyclopedia

Statistical discrimination is an economic theory of inequality based on group stereotypes. In its simplest version, individuals are discriminated against because stereotypes are held against the groups they are associated with. This type of preferential treatment is labeled "statistical" because stereotypes may be based on the discriminated group's average behavior. When this is the case, individuals from different groups are treated unequally because these groups, on average, differ in behavior.

The theory is based on lack of information. For instance, if you don't know an applicant's working ability, how do you know if you should hire him? Here the theory says that the rational thing to do is to base the decision on this applicant's visible features, deciding which group he belongs to, and estimating the average working ability his group has.

Another way statistical discrimination might be applied is due to differing amount of knowledge about different groups. For instance, a large amount of data may be available for group A in comparison to group B. But say these two groups have the same distribution and average ability. If two groups, A and B, have average test scores well above the average for the entire population, but group A's estimate is considered more reliable, then if two people, one from A and one from B interview for a job, using statistical discrimination, A is hired, because it is perceived that his group score is a good estimate, but the group B's group score more likely to be "luck". Conversely, iff the two groups are below average, B is hired, because group A's negative score is believed to be a better estimate.

It is difficult to prove when statistical discrimination is really occurring. Many studies have been done by sending out applications for jobs where the only difference is in their names. Their names have been given based on typical stereotypes of different skincolour (for example Jamal and Yoland are perceived to be African American , while Greg suggests the person is white). The applicants are then tallied based on how many are being called for an interview. James Heckman has criticized these studies by saying that people might be discriminated by single or even many firms, but on aggregated level this might have no effect (see Heckman (1998)).

The variation in distribution between different groups might also lead to discrimination. Consider as an example a black and white's ability to "jump" a basketball. Assume a white's average jumping height ability is the same as for blacks. But assume the blacks have higher variance (normally distributed assumed). If we set a floor that is higher than the average ability for both groups, the blacks able to jump that high will begin to be disproportionate to the general population. Here comes the important part: What if we set the hoop higher for blacks than for whites? For high enough variance for blacks, the blacks will still outperform whites and still be discriminated! This proves the hardness to detect discrimination (example from Heckman (1998)).

Statistical discrimination is often used and tolerated for behaviors that are volunatary, for example, charging smokers more for life insurance, or requiring a college diploma for a job (because it is believed that statistically college graduates perform better in the job at question). Some well documented instances of statistical discrimination for non volunatary group membership also do exist and are tolerated. For example, many countries allow auto insurance companies to charge men and women with identical driving records different rates (or factor in gender when deciding whether to deny coverage).^[1] The same society may not tolerate statistical discrimination when it is applied to protected groups. For example, it has been suggested that home mortage lending discrimination against African Americans, which is illegal, may be partly caused by statistical discrimination. ^[2]

[edit] Footnotes

[edit] Bibliography

• Arrow, K. J. (1972), "Models of Job Discrimination", in A. H. Pascal (ed.), Racial Discrimination in Economic Life, Lexington, MA: D. C. Heath.
• Arrow, K. J. (1972), "Some Mathematical Models of Race Discrimination in the Labor Market", in A. H. Pascal (ed.), Racial Discrimination in Economic Life, Lexington, MA: D. C. Heath.
• Arrow, K. J. (1973), "The Theory of Discrimination", in O. Ashenfelter and A. Rees (eds.), Discrimination in Labor Markets, Princeton, NJ: Princeton University Press.
• Coate and Loury
• Glenn Loury, The Anatomy of Racial Inequality, Princeton University Press. Informally illustrates the theory in the context of United States' racial differences.
• Phelps, Edmund S. (1972). "The Statistical Theory of Racism and Sexism". American Economic Review 62: 659-661. 
• Heckman, James J. (1998). "Detecting Discrimination". Journal of Economic Perspectives 12 (2): 101-116.