Group Identity and Agency Frictions: Evidence using Big Data

Standard neoclassical models in finance posit that corporate decisions are made virtually in a vacuum and independently of an individual's social identities. But since Akerlof and Kranton's seminal work, there has been increased interest in social identity research in economics and finance, yielding new insights into phenomena which standard economic analysis on individual-level incentives proves unable to explain. Even after a plethora of work, the literature is silent on what role social identities such as native language, native place, or caste play in the board room, including as to the capability of these identities to change managerial compensation, firm value, and to affect the relationship between a manager and the board (agency frictions). My recent paper seeks these answers.

Using a theoretical model, the paper begins by providing an intuition on how similarity of social identity (or group identity) between a manager and the board members can change managerial compensation and firm value. The model exhibits that the manager gets higher compensation if most board members and a manager share the same identities. The firm value increases as the manager exerts extra effort due to the same reason.

As theoretical predictions can differ from empirical findings, it is essential to check whether the predictions withstand empirical analysis. To do that, primarily, it is crucial to check whether the biases based on these identities exist in corporate decisions or not. As it is hard to detect these biases, I chose a setting where stereotypes related to these identities are well established, and the identities can easily be learned from the last name (or surname).  A surname in India can provide information such as family lineage, native language, place of origin, and caste, about a person. As a result, surnames are likely an essential basis of social ties and a source of favoritism or discrimination.

As there is no readily available database to provide the identities (native language, native place, and caste) of managers and directors, a surname dataset was developed using the micro census data from socio-economic caste census (SECC) along with a linguistic survey of India (LSI) data to know the identities of managers and directors.

The main empirical results show that the manager gets 6-8% higher compensation if the manager speaks the same native language or belongs to the same native place as most board members. It is the cost of in-group favoritism. I also find that the firm value (as proxied by the market to book ratio) is 11-12% higher for such firms than their counterparts. Even excluding the cost of in-group favoritism, the residual firm value (a proxy for manager's reciprocation to in-group favor) is 8-9% higher for such firms implying that the manager deploys extra effort to the group identity, which leads to a reduction in agency frictions. These results are robust to endogeneity concerns, firm fundamentals, and corporate governance characteristics. The result that the individual puts extra effort for the members who share the same identities as him or her is in line with theoretical prediction and the literature on group identity in social psychology and social neuroscience.

Jitendra Aswani is a doctoral student in Finance at Fordham University, New York.