Research

Most juvenile offenders are given some form of community-based sanction, but little evidence exists on whether these can improve the lifetime trajectories of youth. I evaluate the effects of a youth justice reform in Sweden that sharply increased the use of court-ordered community service --- i.e. unpaid, low-skilled work. The new sanction replaced fines and rehabilitation. On average, the reform did not affect recidivism or labor market outcomes, but these average effects mask considerable heterogeneity depending on the most likely alternative sanction. In particular, post-reform recidivism and adult incarceration rates are lower for individuals for whom community service replaces fines, while the opposite is true for rehabilitation. I then evaluate the net financial effect of the policy and analyze how the program could be targeted for improved efficiency. Overall, the results show that community service is a good alternative to fines for all youth offenders, but repeat offenders and youth from disadvantaged areas might benefit more from a pure rehabilitative sanction.

Using harmonized administrative data from Scandinavia, we find that intergenerational rank associations in income have increased uniformly across Sweden, Denmark, and Norway for cohorts born between 1951 and 1979. Splitting these trends by gender, we find that fatherson mobility has been stable, while family correlations for mothers and daughters trend upwards. Similar patterns appear in US survey data, albeit with slightly different timing. Finally, based on evidence from records on occupations and educational attainments, we argue that the observed decline in intergenerational mobility is consistent with female skills becoming increasingly valued in the labor market. 

We explore the role of identity along multiple dimensions in high-stakes decision-making. Our data contain information about demographic and socioeconomic indicators for randomly assigned jurors and defendants in a Swedish court. Our results show that defendants are 15 percent less likely to get a prison sentence if they and the jurors belong to the same identity-forming groups. Socioeconomic background and demographic attributes are at least as important, and combining several identities produces stronger effects. 

We estimate the change in the gender wage gap between 1968 and 2010 in Sweden accounting for (1) changes in the intensive margin of labour supply; (2) changes in the overall wage inequality; (3) changes in selection into the labor market using parametric and non-parametric selection corrections. Our results show that between 1968 and 1991, about half of the changes in the gender wage gap can be attributed to changes in the overall wage distribution. Conversely, changes in the wage distribution in 1991-2010 masks a larger closure of the gender wage gap. Our corrections for selection into the labor force suggest that uncorrected estimates miss about half of the around 20 percentage points decrease in the gender wage gap over the 1968-2010 period.