Large Stakes and Little Honesty? Experimental Evidence from a Developing Countr

We experimentally study the extent to which individuals are honest when lying can result in a gain of several months’ worth of income. Randomly selected individuals from villages in Bangladesh participated in a sender-receiver cheap talk game. We varied the potential benefits from providing false recommendations. While we find that individuals are more likely to provide false recommendations when stakes are very large, we still observe that almost half of the senders refrain from lying. Receivers are generally suspicious and approximately half of the times do not follow recommendations received. In addition, we observe that one-fifth of the senders do not send any message if they can remain silent and that the option to remain silent crowds out honesty but not dishonesty. These findings provide novel insights on the prevalence, robustness, and motivation of honesty over large stakes from a developing country.