Dishonesty and trust in finance

In this thesis, I study dishonesty and trust in finance by observing individuals’ behavior in the cheap talk sender-receiver game (Gneezy, 2005). I expect finance majors to be more dishonest and exhibit lower levels of trust than other business majors. These hypotheses are mainly based on prior empirical findings associating economists with heightened self-interest and greed. Relying on the moral identity theory, I predict professionals to be more honest and exhibit higher levels of trust when compared with students. My unique sample consists of 132 participants of whom 46% are finance majors and 12% are professionals. The data was collected during the fall of 2016. Following the lines of prior research, I employ non-parametric tests to compare relative frequencies and implement series of logistic and OLS regressions to control for individual characteristics across the sample. I find that finance majors are less dishonest and exhibit the same level of trust than other business majors. When relative frequencies are employed, 26% of the finance students send a deceptive message when compared to 41% of the non-finance students. In the baseline regressions, the probability of sending a deceptive message decreases more than 20% if the sender is a finance major. The regression results are robust with respect to different specifications of the baseline model. I find weak support for both the selection and indoctrination hypotheses, the latter being more important. The difference in honesty between finance and non-finance majors widens when only senior students and professionals are compared. Based on these results, I conclude that education in finance seems not to promote self-interested or unethical behavior. This study adds to the literature on within-country differences of dishonesty and trust. I find strong support for students being more dishonest and exhibiting lower levels of trust toward their counterparts than professionals. When relative frequencies are used, 32% of the students send a deceptive message when compared to 6% of the professionals. In the baseline OLS regression, the probability of sending a deceptive message decreases about 34% if the sender is over 26 years old. The results are very robust and lead to the same conclusion regardless of the definition of deception used. Furthermore, I find that professionals’ higher income does not drive the results. In contrast, my results support the view of moral values being a source of the intrinsic cost of lying. These findings suggest that students regardless of their field of study would benefit from exposure to moral dilemmas that could be incorporated into course work. These findings also imply that relying on students as research subjects can limit the generalization of the results to a broader population