Working Paper: CEPR ID: DP12955
Authors: Asadul Islam; Debayan Pakrashi; Liang Choon Wang; Yves Zenou
Abstract: In order to determine the relative size of taste-based and statistical discrimination, we develop a simple model to distinguish these two theories. We then test the model's predictions of caste-based discrimination by conducting a field experiment that elicits patients' rankings of physicians of different castes and years of experience in the healthcare market in India. We also run a survey and conduct lab-in-the-field experiments to measure patients' attitudes towards different caste groups. We find that 47 to 80 percent of patients statistically discriminate physicians. The overwhelming size of statistical discrimination has important implications for the use of Affirmative Action policies in India.
Keywords: field experiment; caste discrimination; health
JEL Codes: I15; J15; O12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Patients' decisions reflect statistical discrimination (D91) | Patients rank doctors based on caste (I11) |
| Patients statistically discriminate physicians based on caste (J15) | Patients' rankings are consistent with beliefs about average quality associated with different castes (I14) |
| Patients exhibit strong homophily (C92) | Patients prefer their own caste (J15) |
| The percentage of statistical discriminators increases (C52) | Majority of patients statistically discriminate against physicians (I11) |
| Patients statistically discriminate against physicians (I11) | Implications for affirmative action policies in India (J78) |