with Moritz Kuhn and Philip Jung; Revision conditionally accepted at the Review of Economic Studies
available as CEPR Discussion Paper No. 13328 , IZA Discussion Paper No. 12001
Media mentionings of our work at Frankfurter Allgemeine Zeitung, Die Zeit, Makronom, IZA Newsroom, Berliner Zeitung, General Anzeiger, Rheinische Post, Sueddeutsche Zeitung; radio interview with my co-author Moritz Kuhn at Deutschlandfunk "Laenderzeit"
Abstract: A key question in labor market research is how the unemployment insurance system affects unemployment rates and labor market dynamics. We revisit this old question studying the German Hartz reforms. On average, lower separation rates explain 76% of declining unemployment after the reform, a fact unexplained by existing research focusing on job finding rates. The reduction in separation rates is heterogeneous, with long-term employed, high-wage workers being most affected. We causally link our empirical findings to the reduction in long-term unemployment benefits using a heterogeneous-agent labor market search model. Absent the reform, unemployment rates would be 50% higher today.
ECB Working Paper Series, No. 2933
Abstract: Banks in the euro area can generate high-quality liquid assets (HQLA) by borrowing central bank reserves from the Eurosystem against non-HQLA collateral. This paper quantifies the extent of this liquidity transformation and finds that on average EUR 0.92 of net HQLA are generated for each euro of credit provided by the Eurosystem. The paper then identifies intentional liquidity transformation using two novel approaches: The first approach compares the liquidity profile of already pledged vs new collateral, and the second approach compares the liquidity profile of the pool of pledged securities with banks’ total eligible securities holdings. Both approaches show that banks use their least liquid assets as collateral first and pledge more liquid assets only at the margin. This intentional liquidity transformation is sizable and accounts for 30-60% of generated HQLA. These results are relevant for calibrating the collateral framework as well as the optimal size and composition of the Eurosystem balance sheet.
with Moritz Kuhn and Philip Jung
available as IZA Discussion Paper No. 10341
Abstract: We provide new estimates on worker flow rates in and out of unemployment for Germany covering the last six decades. In the 1980s, Germany emerged as the sick man of Europe with a labor market characterized by persistently high unemployment rates. We attribute a substantial fraction of the rise in unemployment to a dramatic increase in inflow rates compared to the 1960s. Germany’s recovery started in the mid-2000s after the Hartz reforms, when inflow rates persistently decreased. Comparing the German and U.S. labor market during recessions uncovers a striking similarity between the recent financial crisis in the U.S. and the German recession in the 1980s. We relate these findings to existing theories on labor market differences between the U.S. and Germany.
Abstract: The U.S. labor market has seen a steady decline in the rates at which workers and jobs are reallocated between firms since the 1980s. Existing explanations have mainly focused on structural changes that increased the costs of hiring or firing employees, e.g. through higher employment protection or occupational licensing, which have negative implications for allocative efficiency. This paper explores a new channel that puts the negative reasons for the declining dynamism into question: Since the 1980s profits and price markups in the professional service industries have fallen which increased final-good firms' demand for these services. As firms started to procure more labor inputs from external providers, they increasingly reacted to idiosyncratic shocks by cutting or expanding externally procured services rather than hiring or firing inhouse staff. The availability of a cheaper outsourcing option therefore shielded inhouse employees from firm-level shocks which reduced the aggregate reallocation rate. In support of that mechanism, this paper uses disaggregated data on worker flows, job flows and the input-output structure of the US economy to establish three new stylized facts: First, occupations which experienced larger shifts towards the professional and business services sector (PBS) had substantially larger declines in worker reallocation rates. Second, job creation and destruction rates decreased more in those industries which experienced a larger rise of PBS-inputs as a share of gross output. Third, employment changes became less responsive to business cycle fluctuations in industries that increasingly relied on PBS-services as intermediate inputs. The paper then builds a structural model in which firms face a trade-off between employing workers inhouse subject to adjustment costs or procuring external services at a price markup. Reducing the cost of procurement from external providers exogenously to match the increase in the share of outsourced services in the data, the model can explain 29% of the fall in aggregate worker flow rates since the 1980s which is not accounted for by shifts in the composition of the work force. The cheaper access to external services enables firms to circumvent the labor adjustment friction which increases allocative efficiency and raises output by 0.3 percentage points.
Abstract: There is an open debate on whether the European Monetary Union (EMU) should be supplemented by a fiscal stabilizer that allows member states to share the risk of asymmetric shocks across countries. One suggested instrument is a common European unemployment insurance (EUI). While the risk-sharing benefits of an EUI scheme are evident, imposing a one-size-fits-all scheme across heterogeneous labor markets might cause first-order distortions. This paper computes the welfare effect of these distortions in a structural labor market model where the social planner chooses UI replacement rates optimally to balance the gains from insuring individual labor market risk and the losses from reduced search incentives. Calibrating the model separately for each country to match the current policies as well as the level and cyclicality of worker flow rates in the data, allows to back out policy-invariant country-specific parameters. The model features duration dependence in both benefits and job-finding rates using a novel measure of high-frequency worker flow rates. In stark contrast to existing policies, the optimal benefit profile is increasing in unemployment duration, both for country-specific UI schemes as well as for the federal UI policy. The welfare effects of replacing the sub-optimal decreasing UI profiles currently in place with optimal country-specific policies are large (2.96% of consumption). The distortions of moving from country-specific optimal policies to a single union-wide optimal policy are comparatively small (-0.22%). Duration dependence is a key driver of these results: An alternative calibration without duration dependence in job finding rates and only one benefit level yields smaller welfare gains of optimal national policies (0.72%). These welfare gains are more than offset by the welfare distortions of having only one common UI policy across all countries (-1.48%).