During the reported period, I worked in several directions to achieve objective (i). I first had to gather and to clean the data for analysis. This was an important amount of work, as several administrative records – covering the entire population annually during 1986-2013 – were used. Then, I had to set up an convincing empirical strategy. The key issue in our framework was that graduation year should be exogeneous to individual characteristics. I discussed this issue with my co-authors over the whole reporting period, and especially in November 2016 during a one-week visit to the University of Uppsala. We focus on vocational students who graduated during the deep Swedish economic crisis of the early 1990s. Our empirical strategy exploits the fact that during 1987-1991, some vocational students were given the choice of attending a three- rather than an ordinary two-year track at the end of compulsory school. Using a difference-in-differences approach, we compare individuals enrolling in vocational studies the same year, but graduating a year apart (hence, different economic conditions). For each enrolment cohort, the choice to enrol in two- versus three-year track was made before 1992, before the crisis hits, so anticipation effects are unlikely.
During the reporting period, I produced a causal estimate of recession at career entry on labour-market outcomes and family formation in the life-course (Objective (i)-(b)). Our results show that individuals who graduate in a bad economy temporarily withdraw from the labour-market (up to 4 years after graduation). A one-percentage point increase in unemployment rate at graduation decreases by 4% the probability of being employed that year, and earnings by 6%. These estimates are comparable to the ones obtained for the US and Germany for low-educated workers. The catch-up process is rather quick, as both men and women have the same employment and earnings as their luckier counterparts four years after graduation. Interestingly, individuals who enter the labour market when unemployment rates are high have a higher probability of staying at their parents’ home in the two subsequent years. Initial economic conditions affect both the timing and the occurrence of marriage. Women who graduate in a recession marry more, but also earlier than their counterparts. They also yond timing effects, completed fertility seems to be impacted too, as women facing adverse initial economic conditions have more children on average at age 38. These results for women can be seen as conservative, in the sense that Sweden is an equal country, where labour participation among women is historically very high. They suggest that as the opportunity cost of childbearing and marriage decreases, women temporarily withdraw from the labour-market and accelerate their entry into marriage and childbirth. What remains for future analysis is to what extent these results are linked to health in the longer-run.
