The Covid epidemic had a large impact on economic activity. In contrast, the dramatic decline in mortality from infectious diseases over the past 120 years had a small economic impact. We argue that people's response to successive Covid waves helps reconcile these two findings. Our analysis uses a unique administrative data set with anonymized monthly expenditures at the individual level that covers the first three Covid waves. Consumer expenditures fell by about the same amount in the first and third waves, even though the risk of getting infected was larger in the third wave. We find that people had pessimistic prior beliefs about the case-fatality rates that converged over time to the true case-fatality rates. Using a model where Covid is endemic, we show that the impact of Covid is small when people know the true case-fatality rate but large when people have empirically-plausible pessimistic prior beliefs about the case-fatality rate. These results reconcile the large economic impact of Covid with the small effect of the secular decline in mortality from infectious diseases estimated in the literature.

We analyze how skill transferability and the local industry mix affect the adjustment costs of workers hit by a trade shock. Using German administrative data and novel measures of economic distance we construct an index of labor market absorptiveness that captures the degree to which workers from a particular industry are able to reallocate into other jobs. Among manufacturing workers, we find that the earnings loss associated with increased import exposure is much higher for those who live in the least absorptive regions. We conclude that the local industry composition plays an important role in the adjustment processes of workers.

Can staying minorities who evade ethnic cleansing affect political outcomes in resettled communities? After World War Two, three million ethnic Germans were expelled from Czechoslovakia's Sudetenland, but some were allowed to stay, many of them left-leaning antifascists. We study quasi-experimental local variation in expulsion policies, a result of the surprising presence of the U.S. Army, which indirectly helped antifascist Germans stay. We find a long-lasting footprint: Communist party support, party cells, and far-left values are stronger today where antifascist Germans stayed in larger numbers. Postwar German Communist elites appear to be behind this effect along with the intergenerational transmission of values among active party members.

We study the granular wage and employment effects of a German place-based policy using a research design that leverages conditionally exogenous EU-wide rules governing program parameters at the regional level. The place-based program subsidizes investments to create jobs with a subsidy rate that varies across labor market regions. The analysis uses matched data on the universe of establishments and their employees, establishment-level panel data on program participation, and regional scores that generate spatial discontinuities in program eligibility and generosity. Spatial spillovers of the program linked to changing commuting patterns can be assessed using information on place of work and place of residence, a unique feature of the data. These rich data enable us to study the incidence of the place-based program on different groups of individuals. We find that the program helps establishments create jobs that disproportionately benefit younger and less-educated workers. Funded establishments increase their wages but, unlike employment, wage gains do not persist in the long run. Employment effects estimated at the local area level are slightly larger than establishment- level estimates, suggesting limited economic spillover effects. On the other hand, spatial spillovers are large as over half of the employment increase comes from commuters. Using subsidy rates as an instrumental variable for actual subsidies indicates that it costs approximately EUR 25,000 to create a new job in the economically disadvantaged areas targeted by the program.

In this study, we analyzed the forecasting and nowcasting performance of a generalized regression neural network (GRNN). We provide evidence from Monte Carlo simulations for the relative forecast performance of GRNN depending on the data-generating process. We show that GRNN outperforms an autoregressive benchmark model in many practically relevant cases. Then, we applied GRNN to forecast quarterly German GDP growth by extending univariate GRNN to multivariate and mixed-frequency settings. We could distinguish between “normal” times and situations where the time-series behavior is very different from “normal” times such as during the COVID-19 recession and recovery. GRNN was superior in terms of root mean forecast errors compared to an autoregressive model and to more sophisticated approaches such as dynamic factor models if applied appropriately.