Some weeks ago, a shocking video showing a South Carolina policeman shooting down an unarmed man as he was running away came to light; needless to say it created a furore all over the US. It reminded people of a similar incident that occurred last August in which an 18-year-old boy, again unarmed, was shot dead by a police officer in Ferguson, Missouri.
There have been other such incidents — the killing of a 12-year-old boy in Cleveland, Ohio and the death of a 43-year-old man from a chokehold applied by a police officer in New York City last year. The list goes on.
Quite apart from the issue of police brutality, this has brought to the fore the issue of racism and racial discrimination in the US. Because in all the instances the victims were black and the police officers white. However, the answer, as economists will argue, is not so simple.
For us it is not about race, but bias of one form or the other. India grapples with gender prejudice, caste prejudice and religious prejudice, at the very least.
Economists have proposed two different notions of discrimination — statistical discrimination and taste-based discrimination — and they suggest different explanations for such phenomena. For example, assuming that these are instances of discrimination, is it the same as the behaviour of the Alabama police officer who threw down and grievously injured the 57-year-old Sureshbhai Patel from Gujarat?
Statistical discrimination The idea of statistical discrimination is usually credited to the Nobel Prize-winning economist Kenneth Arrow. It says, statistically speaking, if a black individual is more likely to be involved in a criminal incident than a white individual, then, when a police officer faces a black individual, the officer is more likely to have a prior belief that he is dealing with a criminal.
Consequently, a police officer is more likely to stop a black person for a traffic infraction, more likely to search and investigate a black individual under suspicious conditions, and be more cautious and more likely to use force when dealing with a black person. It is also the same reason why a person who appears to be of West Asian origin may be subject to extra scrutiny at an airport.
Of course while statistical discrimination does not justify discriminatory behaviour against a particular individual, it relies on data — which creates certain beliefs based on which the officers may have, arguably, a reason to behave in a certain way in order to ensure their own safety as well as the safety of others.
Taste-based discrimination Taste-based discrimination, propounded by the 1992 Nobel laureate, Gary Becker, on the other hand, is based on the idea that the members of a majority or influential group treat the members of the minority or less influential group unfavourably.
Applied to the above situation, a black individual is more likely to be treated with excessive force by the police simply because of his skin colour and not because he is more likely to be involved in a crime. And quite possibly an Indian man might face police brutality simply because he is Indian! Closer home this would imply that the upper castes treat the lower castes unfavourably simply because of caste prejudice.
Another example of taste-based discrimination is the so-called ‘son preference’ which, as suggested by Nobel laureate Amartya Sen, can be measured using the number of missing women. This measure captures the difference between the actual male-female sex-ratio and the expected sex-ratio if there were no sex-selective abortions, and female children were provided adequate medical care and nutrition. Bear in mind that though somewhat far-fetched, it would also be possible to construct a statistical discrimination argument for son preference.
Clearly, one cannot provide justification for taste-based discrimination using any kind of reasoning except that it is, as the name suggests, based on the preferences of a group.
In practice, however, things are murkier, and it is often not possible to distinguish between the two types of discrimination. For instance a Washington Post - ABC News poll carried out in December 2014 suggests that only 20 per cent African-Americans believe that blacks and other minorities are treated like whites by the police. The corresponding number for whites on the other hand is about 60 per cent. These beliefs are aside from all crime statistics which makes it difficult to distinguish between the two types of discrimination.
In the labour market Taste-based discrimination makes very stark predictions about the labour market. It says that black workers will receive a negative premium; that is, a black worker with the same productivity and ability as a white worker will have to accept a lower wage than his white counterpart in order to get an identical job.
In an influential study published in the American Economic Review , economists Marianne Bertrand and Sendhil Mullainathan sent identical fictitious resumes in response to help-wanted ads in Boston and Chicago newspapers. Their results, though striking, were not unexpected: A white-sounding name (like Justin or Emily as opposed to Jamal and Lakiesha) received 50 per cent more callbacks for an interview.
In a similar recent study, the same authors along with Abhijit Banerjee and Saugato Dutta sent identical fictitious resumes in response to 371 software and call-centre job openings advertised in newspapers and online job websites in and around Delhi.
While there was no evidence of discrimination against scheduled castes (SC), scheduled tribes (ST), and other backward castes (OBC) for software jobs, non-upper caste members (SC, ST, and OBC) were found to be at a disadvantage for call centre jobs. Also notable was the finding that Muslims are not systematically discriminated against in either of the two sectors.
These findings are somewhat encouraging — it seems that credible skill certification, which is possible for software jobs (but not for call centre jobs), reduces gaps in job opportunities. Thus, the observed discrimination seems to be statistical and not taste- or preference-based. Caste-based discrimination in India, therefore, can be eliminated by providing skill certification.
Yet the news is not so good on the missing girls/son preference front. According to the 2001 census, there were 933 females per 1000 males in 2001, whereas the actual number should have been 1020 as in economically comparable sub-Saharan Africa.
By 2011, this number had gone up to 943, an improvement no doubt, but clearly we still have miles to go.
Sarangi teaches microeconomics and game theory at Louisiana State University; Jha is a doctoral student in economics there