Experimental Design
I plan to enroll around 500 casual daily wage workers across labor stands in Chennai, India. Workers are asked to rate employers they have previously worked with on dimensions such as timely payment, respectful communication, micromanagement, provision of meals or transportation, and feelings of respect at the workplace. First, the study establishes that workers in low-wage, unskilled labor markets have preferences over employer quality and workplace amenities such as respect, even at the cost of higher wages. To test this, I will conduct an incentive-compatible choice experiment in which workers select between contracts with higher- and lower-rated employers offering varying wage levels. This approach tests whether workers are willing to forgo higher wages to avoid employers perceived as "bad" (lower rated by fellow workers at the stand).
Once worker preferences are established, the project examines the flow of information about employer quality within labor stands.
H1: Workers are less likely to share negative experiences with employers than positive ones. To test this hypothesis workers are asked to recall either a positive or negative experience (the balance is randomized) with an employer and decide whether to share it with a peer outside their social group. First, I ask if the worker would like to share information with the assigned peer. For respondents, who say no, I ask them to indicate their willingness to accept (WTA) payment to share this information. My hypothesis predicts that negative information will be less frequently shared due to potential social or professional consequences.
I then run experiments to test the reasons behind asymmetrical flow of information at the stand:
H2: Workers fear retaliation from employers for sharing negative information. This hypothesis examines the role of anonymity in information sharing. The randomization described in H1, will be cross randomized with anonymity of the disclosing worker to test this hypothesis. If workers are reluctant to share negative experiences when their identity is disclosed, it suggests a fear of employer retaliation.
H3: Workers avoid sharing negative information because they are looking out for fellow workers. To test this, I rely on the variation in willingness to disclose positive versus negative information for workers who are identifiable and further randomly vary whether the worker receiving the information has received an economic shock in recent times. Workers are asked whether they would share negative or positive information (the balance is randomized) about an employer with a peer who has undergone an economic shock versus a peer where I provide no such information. The hypothesis predicts that altruistic motives may prevent workers from sharing negative information if it could discourage peers, particularly those in vulnerable situations, from pursuing available jobs.
H4: Workers avoid sharing negative information because they interpret their own signal as weak (i.e. workers may have different perceptions of employer quality). To explore this channel, I ask workers to share information about employer behavior that they have either experienced themselves, or our research team collected at an aggregate level from the stand. I conduct this experiment under both anonymous and non-anonymous settings, and believe that if workers worry about their own signal as being misinterpreted then they would be more willing to share aggregated information on an employer and not their own personal experiences.