Essay 1 (with Oskar Nordström Skans): This paper provides field evidence on the causal impact of past successes on future performances. Since persistence in success or failure is likely to be linked through, potentially time-varying, ability it is intrinsically difficult to identify the causal effect of succeeding on the probability of performing well in the future. We therefore employ a regression discontinuity design on data from professional golf tournaments exploiting that almost equally skilled players are separated into successes and failures half-way into the tournaments (the “cut”). We show that players who (marginally) succeeded in making the cut substantially increased their performance in subsequent tournaments relative to players who (marginally) failed to make the cut. This success-effect is substantially larger when the subsequent (outcome) tournament involves more prize money. The results therefore suggest that past successes provide an important prerequisite when performing high-stakes tasks. Essay 2: Recent experimental evidence suggests that women in general are more discouraged than men by failures which potentially can explain why women, on average, are less likely than men to reach top-positions in firms. This paper provides the first quasi-experimental evidence from the field on this issue using data from all-female and all-male professional golf tournaments to see if this result can be replicated among competitive men and women. These top-performing men and women are active in an environment with multiple rounds of competition and the institutional set-up of the tournaments makes it possible to causally estimate the effect of the result in one tournament on the performance in the next. The results show that both male and female golfers respond negatively to a failure and that their responses are virtually identical. This finding suggests that women’s difficulties in reaching top-positions in firms are caused by external rather than internal barriers. Essay 3: Voting is a fundamental human right. Yet, individuals that are younger than 18 do typically not have this right since they are considered uninformed. However, recent evidence tentatively suggests that the political knowledge of youths is endogenous to the voting age. I test for the existence of such dynamic adjustments utilizing voting age discontinuities caused by Swedish laws. I employ a regression discontinuity strategy on Swedish register data to estimate the causal effect of early age voting right on political knowledge around age 18. The results do not support the existence of positive causal effects of early age voting right on political knowledge. Thus, we should not expect that 16-year-olds respond by acquiring more political knowledge if they are given the right to vote. This finding weakens the case for a lowering of the voting age from 18 to 16. Essay 4 (with Lena Hensvik): We postulate that firms’ production losses  from absence depend on the employees’ internal substitutability, incentivizing firms to keep absence low in positions with few substitutes. Using Swedish employer-employee data we show that absence is substantially lower in such positions even conditional on establishment and occupation fixed effects. The result reflects sorting on both entry and exit margins, with stronger separations responses when it was difficult to predict the absence of the employees beforehand. These findings highlight that internal substitution insures firms against production disruptions caused by absence and that absence costs are important aspects of firms’ hiring and separations decisions.