Working papers

"Does unethical behavior affect choice of profession: public vs private?", Banerjee, Ritwik, Baul, Tushi and Rosenblat, Tanya.{Job Market Paper}

We contribute to a growing literature in experimental labor economics that studies the interaction between workplace culture and worker selection(Dohmen and Falk(2011), Niederle and Vesterlund(2007)). Using a laboratory experiment with university students in India who are either private sector job aspirants or public sector aspirants (wanting to join Indian bureaucracy), we study the effects of labor market sorting in the choice between public vs private sector careers based on one's propensity to engage in unethical behavior. We divide the subjects into two groups: workers and supervisors. The workers perform in a real effort task and the supervisors evaluate the performance of workers and send them their earnings in terms of tokens, which is a medium of exchange in laboratory. The supervisors have to claim the required number of tokens needed to pay the workers. The measure of cheating that we use in our experiment is the difference between the number of tokens claimed by the supervisors and the number of tokens sent to the workers. We then link this cheating behavior to their future career choice. We find that there is a significant difference in cheating behavior between public and private sector aspirants. The proportion of public sector aspirants who engage in unethical behavior is not statistically significantly different than that of private sector aspirants. However, the amount of cheating of public sector aspirants is significantly more than private sector aspirants. Public sector aspirants cheat 51.5% relatively more than those planning to pursue careers in the private sector.

"Social networks and non-cognitive skills", Baul, Tushi, Mobius, Markus and Rosenblat, Tanya.

Abstract: Recent research in psychology and economics has documented the importance of non-cognitive skills for labor market outcomes. In the second essay (joint with Markus Mobius and Tanya Rosenblat), I focus on a particular non-cognitive skill- self-confidence about one's ability. We use economic experiments to measure participants' beliefs about their expected relative performance in a real-effort task using an incentive-compatible mechanism for belief elicitation. We combine self-confidence measures with a unique dataset on the social network of the participants to investigate whether there are peer effects in confidence. Indeed, we find that more confident participants tend to have more confident friends. Peer effects might be an important transmission mechanism for acquisition of self-confidence. We therefore explore whether our finding is due to selection (confident participants choose to have confident friends) or social interaction (one's confidence is influenced by friends' confidence). This paper shows that our friends not only affect the various real outcomes like choosing a major, academic performance, smoking behavior but they also affect our confidence level. We find a significant positive correlation between friend's confidence and individual confidence. We use instrumental variable approach to separate the treatment effect from the selection effect. I also try to separate the treatment effect by finding the correlations separately for different years at college. We find that there is significant correlation between friend's confidence and individual confidence for the seniors and juniors. Both these evidences suggest that there might be treatment effect.

Papers in progress

"Self-selection and creation of corporate cultures: mergers and acquisitions", Baul, Tushi, Rosenblat, Tanya and Strasburg, Chris.

The third essay (joint with Tanya Rosenblat and Chris Strasburg) build on labor market sorting and corporate culture-clash. I propose to highlight that one important reason which might lead to failure in mergers is 'culture-clash'. We want to define incentive schemes as one manifestation of culture. For example, relative performance pay might provide strong incentives but also generates negative utility for workers who strongly dislike competition. Such workers will then have to be paid a compensating differential to prevent them from leaving the company. Hence, relative performance pay is not necessarily a profit-maximizing incentive scheme for these types of workers. Instead, companies might be better off providing team incentives where the profit loss from free-riding is offset by lower retention costs for staff. Similarly, a risk-tolerant employee might enjoy working for a startup with a high degree of autonomy, skewed incentives and a high degree of failure but become unhappy once the startup is acquired by a more risk-averse partner. In this essay I want to study how sorting amplifies corporate culture formation and how different incentive schemes create two different work cultures. I also want to see how the merger between these two firms affects the post merger performance. I will use a series of lab experiments to study this. The design will allow us to measure the compensating differentials that are necessary to retain workers after a merger. In particular, I can compare these differentials to (a) worker types and (b) sorting incentives when they enter the labor market. A new and important feature of our design is that we can attach a monetary value to these differentials which allows us to assess their economic significance. (I am yet to run this experiment and therefore I do not have the results yet to discuss.)

"Social norms regarding bribe giving vs bribe taking: private vs. public sector in India", Banerjee, Ritwik, Baul, Tushi and Rosenblat, Tanya.

We conducted an experiment (joint with Ritwik Banerjee and Tanya Rosenblat) in India which measures different social norms following regarding bribe giving and bribe taking in India following Krupka and Weber, 2011. We give the workers seven different situations is which a person is faced with different alternative actions. For each situation, they rate the extent to which each alternative action is 'socially appropriate and consistent with moral or proper social behavior or 'socially inappropriate and inconsistent with moral or proper social behavior.' Participants are informed that one of the situations for which they provide appropriateness ratings will be selected at random at the end of the experiment, and that one of the possible action choices in that situation will also be randomly selected. If, for this action choice, the participants appropriateness rating is the same as the modal response in the session, then that participant will receive a certain amount at the end of the experiment. The different situations are related to real-life bribe giving and bribe faced by a citizen in India for instance a candidate asked to pay a bribe after he fails in a driving test.

"Social networks and altruism", Baul, Tushi, Bhalla, Manaswini and Rosenblat, Tanya.

I am working on possible extensions of the second paper(joint with Manaswini Bhalla and Tanya Rosenblat). I want to analyze to what extent social preferences are an intrinsic characteristic and to what extent they are acquired through social interaction. I distinguish between two acquisition channels: (1) agents interact repeatedly with socially close neighbors which creates directed altruism (2) Agents are influenced by and imitate the social preferences of those with whom they interact. To evaluate the relative importance of intrinsic versus acquired altruism, we need to control for selection effects. We are conducting this experiment at Indian Institute of Management, Bangalore, India. We will follow a cohort of 350 management students starting May 2013 over a two year period and track their social preferences and social networks through a series of laboratory experiments and surveys conducted in each semester over this period. We follow the design of Leider, Mobius, Rosenblat and Do (2009) to get the social network data of this cohort. Simultaneously, we will also conduct a series of laboratory experiments to capture altruistic preferences, incentive choices, and voluntary social service participation preference in the baseline period and then conduct experiments in each semester with different treatments. The main identification strategy is to exploit the panel nature of our data - in particular, it allows us to measure students baseline altruism, their propensity to find friends with similar social preferences and the treatment effect of these friends. A second identification strategy exploits the fact that at IIMB, the class sitting arrangement is allocated through random assignment. This creates an exogenous instrument that allows us to identify the treatment effect of being seated next to altruistic neighbors. We will run repeated fund-raising events for a real charity during the course of the study to understand the connection between measured altruism and charitable giving. We will also revisit other measures of peer effects in the literature such as choice of specialization (finance, HR), evolution of risk aversion between entry into the program and at the
end of the 2-year degree and how this affects success in the labor market. (We are running the pilot now.)

"Hind-sight bias: does monitoring increase effort?"

This paper looks at the implications of hind sight bias on labor market. The bias says that the ex-post realization of ex-ante belief of the event will be biased in the direction of realization of the event. Hindsight bias is the inability to correctly remember one prior expectation after observing new information or outcome causes trouble in processing information. I want to study the implications of hindsight bias on labor market monitoring. This paper is in its preliminary stage of experiment design.