Miguel Luzuriaga (Neu-Ulm University of Applied Sciences) & Oliver Kunze (Neu-Ulm University of Applied Sciences) have posted The Risk-Distributional Preferences and Risk-Aversion Paradox: A Gift-Exchange Experiment on SSRN. Here is the abstract:
In a principal-agent setting, offering maximum levels of incentives, i.e., the compensation is completely variable, is far from optimal because this reduces the principal`s profits, and because the risk-aversion nature from the agent matters. In this paper, we hypothesize that an incentivized agents’ risk-taking about the composition of their own wage structure can offset the typical risk-aversion, and thus lead to an increase in performance. Using a reframed version of the gift-exchange game by Fehr et al. (1998) workers decide, before their effort choice, how much of their wage is variable and fixed. The variable component is limited to max. 30% of the initial wage offer (the rest remains fixed) and it increases/decreases with effort and also with a random factor. We find that risk-averse workers exposed ¾ of their variable component to risk, and outperformed a control group (fixed wage scheme) by half in effort levels. This increase in effort is positively correlated to their risk-distributional preferences. Theoretical explanations (supported by data) are discussed.