Tuesday, July 26

Evidence on Caste-Based Discrimination

The author carried out a survey which is quite common in the UK and USA to test for situations like discrimination and racism. Basically what you do is to take the same cv, duplicate it, change the name to say something that is "white race" sounding and something that is black race sounding or asian sounding, send to employers, see what the response is, and then try to show that the difference in the reaction to those applications is because of that variable (race, sex, region, location, etc.) being different. So this lady did this to Chennai based entry level white collar jobs and on websites between march and dec 2006. She found that a high caste applicant had to send out 6.2 cv's versus 7.4 for a low caste applicant, a difference of 20%. Male recruiters, Hindu recruiters and large firms showed the biggest difference on these 523 applications. But only 2 resumes were sent out, which is interesting, why didnt they send out 4 to correspond with the 4 categories? (other castes, Other backward castes, SC and ST)? in this study, because of this, the results are interesting but hardly surprising.

Now the difference between these is given by these stats: 2004-05 median per capita consumer expenditure bracket (in Rupees), within urban areas of India, was [580,675] for the SC, [675,790] for the ST, [675,790] for the OBC and [1100,1380] for the Other category. Interesting or what? Given the wide variety, it would have been interesting to see the response rate. Anyway. Some interesting bits. Lower caste Women get lower callbacks. And also for both customer services and front office/administration jobs. Now this is the interesting bit, they saw the recruiter bit from the name given in the advertisement. I am not really sure that the contact person in the advertisement is the actual manager. So this has to be taken with a grain of salt. For all you know, the contact name on the advertisement is the name of the junior most person whose responsibility is to handle advertisements in the HR department. But even without that, makes sense that male and Hindu recruiters would show this discrimination.

Firms without multiple domestic offices and foreign offices favour high caste applicants. makes sense, they are parochial firms but firms with multiple domestic and foreign offices prefer low caste applicants. I am not sure if the reverse explanation works. Bigger firms are, by definition, more amorphous and more bureaucratic. Thus they will end up hiring more low caste employees simply because of the probability distribution rather than any explicit liberal positive discrimination. The author suggests its due to political pressures in Tamil Nadu, the state where Chennai is based. Not clear, sorry. Other researchers seems to have found that there is no difference between high or low caste workers in IT but do find in call centre work. But all these points show up fascinating points on how caste is still in place but one positive thing that i take away from here is that i think this gap is reducing as more urbanisation happens and more firms become big and more education happens etc. etc. Would be good to confirm this with a repeated experiment in say 10 years time.

Here's the abstract.

Caste-based quotas in hiring have existed in the public sector in India for decades. Recently there has been debate about introducing similar quotas in private sector jobs. This paper uses a correspondence study to determine the extent of caste-based discrimination in the Indian private sector. On average low-caste applicants need to send 20 percent more resumes than high-caste applicants to get the same callback. Differences in callback which favor high-caste applicants are particularly large when hiring is done by male recruiters or by Hindu recruiters. This finding provides evidence that differences in callback between high and low-caste applicants are not entirely due to statistical discrimination. High-caste applicants are also differentially favored by firms with a smaller scale of operations, while low-caste applicants are favored by firms with a larger scale of operations. This finding is consistent with taste-based theories of discrimination and with commitments made by large firms to hire actively from among low-caste groups

No comments: