Experiments on reciprocity, social comparisons and overconfidence.
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
This thesis consists of three chapters focusing on negative reciprocity, transparency and job assignment, and overconfidence. To test my hypotheses I use experimental economics methods, which offer control over the data generating process by motivating people financially. Experimental methods are used to identify causal processes and motivations that can be confounded in field settings. Economics experiments reduce response noise, e.g. extreme outliers probably caused by thoughtless, unmotivated subjects (Smith and Walker, 1993). The central theme of my first two experiments is reciprocity – a kind reaction to kind and generous behaviour and retaliation to hostile behaviour. The first chapter discusses reciprocity in connection with violation of the property rights and the source of the endowment on an individual level. In the second chapter the reciprocity is studied in an employer-worker setting, in the presence of social comparisons. The last chapter deals with firm’s overconfidence when deciding about entering the market. All of the chapters study aspects of economic behaviour in social contexts, which have implications for the field. Each of the chapters is described briefly below. My first chapter experimentally explores the impact of the strength of property rights on retaliation decisions. I induce strong property rights by having experimental subjects earn money by performing a real effort task and weak property rights by endowing them with windfall gains. I ask whether people are less likely to respond to a hostile behaviour with retaliation when earned money as opposed to windfall money is at stake. My experimental design identifies two reasons why property rights might influence the size and frequency of retaliation. The first reason is that retaliation might be perceived to be more costly when using earned as opposed to windfall money to pay for retaliation. The second reason is related to the violation of property rights. If another person decreases a decision-maker’s endowment and the endowment consists of earned money rather than windfall money, the decision-maker might consider it to be a stronger violation of his property rights, which in turn could trigger stronger retaliation. The purpose of this experiment is to separate these two effects. While I find support for the fact that subjects retaliate more because of the violation of their property rights, I also find that participants actually retaliate more with their earned money than with windfall. This suggests that participants do not perceive such retaliation to be more costly but rather that their behaviour is driven by violation of property rights. The second chapter focuses on the fairness perceptions of the job assignment process in an employer-worker relationship. In reality, employers have at their disposal jobs of different importance, which have to be assigned to different workers. Workers in more important jobs usually get offered higher wages and workers in less important jobs get offered lower wages. If the interpersonal concerns were absent, the employer would provide a higher wage to the worker in the more important job. When a worker decides what wage to accept, he may compare his wage to the other workers’ wage. An employer anticipating this might adjust the wage policy in order to avoid unnecessary losses or to maximise profits. I experimentally study the fairness perceptions from the workers’ and the employer’s point of view. I ask the following questions: 1. Do workers react to the wages paid to their co-workers and does the job assignment procedure affect workers’ wage rejections? 2. Do employers react to the fact that workers compare themselves with their co-workers and do they compress wages when the job assignment procedure is perceived less fair? These questions have implications for labour market, in particular a firm’s wage policy. If an unfair assignment elicits more wage rejections (i.e. zero profit for the employer and zero wage for the worker) due to social comparisons, it can have detrimental effects on the performance of the firm and thus firms might choose to practice wages secrecy. I examine the impact of job assignment in the presence of social comparisons. In order to test for social comparisons, the worker is only informed about the wage that the employer offered to his co-worker and which job he has been assigned to. Only after workers state their minimum acceptable offers, which determine if the offered wage is accepted or rejected, they come to know their own wage. I posit that a worker’s reaction to a specific job assignment depends on the procedure by which they are allocated to the jobs. An assignment to a less important job will not be perceived as unfair if it arises from an unbiased procedure, for example random assignment with equal probabilities. It will, however, be perceived as unfair if workers think that the employer favours some workers over others for reasons that are unrelated to efficiency concerns. This experiment does not provide evidence on social comparisons or employers compressing wages when the assignment to jobs is perceived unfair. My third chapter is a replication of the high-impact overconfidence and excess entry experiment by Camerer and Lovallo (1999). The topic of overconfidence is crucial for understanding business failures. Camerer and Lovallo were first to directly test overconfidence by measuring economic decisions and personal overconfidence at the same time. Camerer and Lovallo test whether managers’ overconfidence about their skills could predictably influence economic behaviour when entering into markets. I implemented Camerer and Lovallo’s experiment with modifications reflecting the technological progress of economic experiments of past 15 years. While Camerer and Lovallo run their experiment with male participants (who have been shown to be more overconfident than females), my experiment studies the effect of overconfidence of both genders, making it a more conservative test. The thesis employs cutting edge techniques from Experimental Economics to study economic decision-making. My research provides empirical evidence on violation of property rights, fairness considerations in labour markets and impact of overconfidence on market entry decisions.