In this paper, we provide an explanation for why risk taking is related to optimism. Using a laboratory experiment, we show that the degree of optimism predicts whether people tend to focus on the positive or negative outcomes of risky decisions. While optimists tend to focus on the good outcomes, pessimists focus on the bad outcomes of risk. The tendency to focus on good or bad outcomes of risk in turn affects both the self-reported willingness to take risk and actual risktaking behavior. This suggests that dispositional optimism may affect risk taking mainly by shifting attention to specific outcomes rather than causing misperception of probabilities. In line with this, in a second study we find evidence that dispositional optimism is related to elicited parameters of rank dependent utility theory suggesting that focusing may be among the psychological determinants of decision weights. Finally, we corroborate our findings with process data related to focusing showing that optimists tend to remember more and attend more to good outcomes and this in turn affects their risk taking.

Working time autonomy is often accompanied by output-based incentives to counterbalance the loss of monitoring that comes with granting autonomy. However, in such settings, overprovision of effort could arise if workers are uncertain whether their performance suffices to secure the output-based rewards. Perfor-mance feedback can reduce or eliminate such uncertainty. We develop an exper-iment to show that overprovision of costly effort is more likely to occur in work environments with working time autonomy in the absence of feedback. A key fea-ture of our design is that it allows for a clean measurement of effort overprovision by keeping performance per unit of time fixed, which we achieve by calibrating subjects' productivity on a real effort task ex ante. This novel design can serve as a workhorse for various experiments as it allows for exogenous variation of perfor-mance certainty (i.e., by providing feedback), working time autonomy, productivity, effort costs, and the general incentive structure. We find that subjects provide significantly more costly effort beyond a level necessary to meet their performance targets in the presence of uncertainty, i.e., the absence of feedback, which suggests that feedback shields workers from overprovision of costly effort.