Fitting duration models on an inflow sample of jobs in Germany starting in 2002 to 2010, the author investigates the impact of employers’ use of temporary agency work on regular workers’ job stability. In line with dual labor market theory, the author finds that nontemporary jobs are significantly more stable when employers use temporary agency workers. The rise in job stability stems mainly from reduced transitions into nonemployment, suggesting that nontemporary workers are safeguarded against involuntary job losses. The findings are robust to controlling for unobserved permanent employer characteristics and changes in the observational window that comprises the labor market disruption of the Great Recession.

Using a 50 % sample of all private sector establishments in Germany, we report that spinoffs are larger, initially employ more skilled and more experienced workers, and pay higher wages than other startups. We investigate whether spinoffs are more likely to survive than other startups, and whether spinoff survival depends on the quality and size of their parent companies, as suggested in some of the theoretical and empirical literature. Our estimated survival models confirm that spinoffs are generally less likely to exit than other startups. We also distinguish between pulled spinoffs, where the parent company continues after they are founded, and pushed spinoffs, where the parent company stops operations. Our results indicate that in western and eastern Germany and in all sectors investigated, pulled spinoffs have a higher probability of survival than pushed spinoffs. Concerning the parent connection, we find that intra-industry spinoffs and spinoffs emerging from better-performing or smaller parent companies are generally less likely to exit.

Using linked employer–employee panel data for Germany, we investigate whether firms implement real wage reductions in a selective manner. In line with insider–outsider and several strands of efficiency wage theory, we find strong evidence for selective wage cuts with high-productivity workers being spared even when controlling for permanent differences in firms' wage policies. In contrast to some recent contributions stressing fairness considerations, we also find that wage cuts increase wage dispersion among peers rather than narrowing it. Notably, the same selectivity pattern shows up when restricting our analysis to firms covered by collective agreements or having a works council.

We show that the major part of the plant size–wage premium in Germany is reflected in different wage growth patterns in plants of different size. This is consistent with the hypothesis that large firms ‘produce’ more skilled workers over time.

After more than 25 years of empirical investigation, the analysis of works council effects on establishment productivity is still an active field of economic research. I present first evidence for works council regression coefficients along the unconditional log labour productivity distribution in Germany. The highest (lowest) coefficients were found at the bottom (top) of the distribution, which suggests that works councils have the most beneficial effects at low levels of productivity. Importantly, coefficients are positive at each productivity level. The results shed light on a classical theoretical argument concerning moderating factors for the productivity effect of employee participation.

We examine how connections to German federal parliamentarians influence firm dynamics using a novel dataset linking politicians and election candidates to the universe of firms. To identify the causal effects of political access, we exploit (i) new appointments to the company leadership team and (ii) discontinuities around the marginal seat of party election lists. Results reveal that political connections reduce firm exits and gradually increase employment growth, with heterogeneous productivity effects depending on the political mandate. Incorporating data on credit ratings, subsidies, and procurement contracts allows us to identify the mechanisms driving the effects over the politician’s career.

We provide new evidence on the effects of economic shocks on political support, voting behaviour and political opinions over the last 25 years. We exploit a sudden, large and long-lasting shock in the form of job loss and trace out its impact on individual political outcomes for up to 10 years after the event. The availability of detailed information on households before and after the job loss event allows us to reweight a comparison group to closely mimic the job losers in terms of their observable characteristics, pre-existing political support and voting behaviour. We find consistent, long-lasting but quantitatively small effects on support and votes for the incumbent party, and short-lived effects on political engagement. We find limited impact on the support for fringe or populist parties. In the context of Brexit, opposition to the EU was much higher amongst those who lost their jobs, but this was largely due to pre-existing differences which were not exacerbated by the job loss event itself.

We provide the first systematic evidence on the effectiveness of a contested policy in Germany to help displaced workers. So-called “transfer companies” (Transfergesellschaften) employ displaced workers for a fixed period, during which time workers are provided with job-search assistance and are paid a wage which is a substantial fraction of their pre-displacement wage. Using rich and accurate data on workers’ employment patterns before and after displacement, we compare the earnings and employment outcomes of displaced workers who entered transfer companies with those that did not. Workers can choose whether or not to accept a position in a transfer company, and therefore we use the availability of a transfer company at the establishment level as an IV in a model of one-sided compliance. Using an event study, we find that workers who enter a transfer company have significantly worse post-displacement outcomes, but we show that this is likely to be the result of negative selection: workers who lack good outside opportunities are more likely to choose to enter the transfer company. In contrast, ITT and IV estimates indicate that the use of a transfer company has a positive and significant effect on employment rates five years after job loss, but no significant effect on earnings. In addition, the transfer company provides significant additional compensation to displaced workers in the first 12 months after job loss.

We conducted a field experiment in the natural setting of blood donations to test how oxytocin relates to empathy and altruism. We randomly assigned blood donors in the Croatian Institute for Transfusion Medicine to three groups with the aim to induce different levels of empathy by showing a neutral video to the donors from the control group and an emotional to the donors from the first and second treatment groups. In addition to watching the emotional video, donors from the second treatment group are given a gift which relates to the emotional story from the video. We find no effect of our treatment on induced levels of oxytocin. Null effects of our treatments could be explained by the above average baseline levels of oxytocin and inability of our treatments to provoke emotional stimuli in blood donors. Nonetheless, for our empathy measures we find the effect of gift exchange on empathic concerns, but not on perspective taking. After our experimental treatments, we followed the return of our blood donors for a whole year. We find that only variable which consistently predicts return for blood donation in stated period is the number of previous donations. From policy perspective it is an important finding. Especially for hospitals and other blood providers when faced with time and resource constraints.

We present causal evidence on the rent-sharing elasticity of German manufacturing firms. We develop a new firm-level Bartik instrument for firm rents that combines the firms’ predetermined energy input mix with national energy carrier price changes. Instrumental variable estimation yields a rent-sharing elasticity of approximately 0.20 implying that a 10 percent change in rents leads to a 2 percent change in wages. Rent-sharing induced by energy price variation is asymmetric and driven by energy price increases, such that, on average, workers do not benefit from energy price reductions but are harmed by price increases. Reduced-form evidence shows that a 10 percent increase in firm-level energy prices depresses firm-level wage growth by 0.34 percent.