CAN'T BUY ME LOVE? A FIELD EXPERIMENT EXPLORING THE TRADE-OFF BETWEEN INCOME AND CASTE-STATUS IN AN INDIAN MATRIMONIAL MARKET

Abstract

I. INTRODUCTION The relationship between economic motive and observance of discriminatory social customs is hardly new to economists. This is implicit in the writings of Becker (1957, 1973, 1974) on discrimination and other social customs, and explicit in Arrow's essay (1972) on discrimination and Akerlof's work (1980) on the interactive relationship between economic incentives and discriminatory social customs. As scholars like Becker, Arrow, and Akerlof have investigated the relationship between economic incentives and the prevalence of custom-based discrimination (of which status-based discrimination is an example), it is natural to consult the related empirical evidence on the impact of economic incentives on the extent of status-based discrimination. Surprisingly, the extant empirical literature is mostly silent on how economic incentives may directly interact with status-based discrimination. (1) This is atypical as one would be hard-pressed to find an issue as weighty as status-based discrimination, yet our knowledge in this aspect remains oddly limited. Our study aspires to rectify this gap. This study investigates whether and what magnitude of economic incentive can diminish the incidence of status-based discrimination by designing a controlled field experiment in an actual marketplace--a reputable arranged marriage market--that is conducted via newspaper advertisements. The existing literature suggests that this market might be especially prone to strong discriminatory mate-selection behavior based on the social identity/status of an individual, where caste identity of an individual is an easily measured indicator of social status. The genesis of this discrimination can be attributed to a high-status female's concern for a decline in her social status if she engages in a marital relationship with a low-status male. Against this backdrop, we ask whether a low-status male could attempt to weaken this enduring form of discrimination by providing systematic economic incentive (in the form of high monthly groom incomes) to a high-status female so that the former's marriage proposal receives preliminary consideration by the latter, conditional on both individuals' participation in this market. The marketplace that we utilize is a well-functioning arranged marriage market, conducted via matrimonial advertisements in a widely circulated Bengali language newspaper. Marriageable males in this market typically advertise their caste and income information, among other attributes. After these advertisements appear in the newspaper, interested families contact the advertiser, mostly in form of detailed letters, expressing their interest in that advertisement. This provides a confidential and reliable measure of intent of the interested party, in this case, females. Therefore, the conjunction of an exceptionally strong status-based mate selection behavior, a reputed market for arranged marriages, and an exclusive correspondence process furnish us with a unique opportunity to investigate the main research hypothesis of our study. To the best of our knowledge, this is the first field experiment that exclusively focuses on how economic incentives can influence status-based discrimination in an actual market. (2) To operationalize the experiment, we create nine unique matrimonial advertisements, each representing a fictitious groom type. (3) These potential grooms belong to one of the three blocks where each block contains three grooms from a given caste group. We vary the caste-status across these blocks. We consider three different caste groups--high, middle, and low, strictly ranked so in the study region. We vary the income features of grooms within a block. We consider three different monthly income levels (i.e., high, medium, and low) that are considered feasible for each caste group in the study region. Furthermore, we randomly assign other characteristics relevant for this market, to all the potential grooms so that these features, on average, do not affect the behavior of responders. …