Research
Labor / Public Economics
Training in late careers - A structural approach
This study investigates the role of on-the-job training in the employment outcomes of less educated men in their late careers. Using survey data from the German National Education Panel Study adult cohort, I estimate a structural dynamic discrete-choice model reflecting the trade-offs of the employees’ training participation decision. The data set enables me to distinguish whether non-participation is due to lack of availability of training or due to individual cost-benefit considerations. As a consequence, I can investigate whether future policy interventions should target the provision of training or the individual participation incentives. I find that on-the-job training has a positive impact on the employees’ employment prospects. Counterfactual simulations show that a reduction of the individual training costs would increase training participation and positively affect the employment rate near retirement. In contrast, an increase in the general availability of training would not be effective.
Revise and Resubmit at Labour Economics. IZA Discussion Paper No. 15875.
Causal Misperceptions of the Part-Time Pay Gap with Clara Schäper and Annekatrin Schrenker
This paper studies if workers infer from correlation about causal effects in the context of the part-time wage penalty. Differences in hourly pay between full-time and part-time workers are strongly driven by worker selection and systematic sorting. Ignoring these selection effects can lead to biased expectations about the consequences of working part-time on wages ('selection neglect bias'). Based on representative survey data from Germany, we document substantial misperceptions of the part-time wage gap. Workers strongly overestimate how much part-time workers in their occupation earn per hour, whereas they are approximately informed of mean full-time wage rates. Consistent with selection neglect, those who perceive large hourly pay differences between full-time and part-time workers also predict large changes in hourly wages when a given worker switches between full-time and part-time employment. Causal analyses using a survey experiment reveal that providing information about the raw part-time pay gap increases expectations about the full-time wage premium by factor 1.7, suggesting that individuals draw causal conclusions from observed correlations. De-biasing respondents by informing them about the influence of worker characteristics on observed pay gaps mitigates selection neglect. Subjective beliefs about the part-time/full-time wage gap are predictive of planned and actual transitions between full-time and part-time employment, necessitating the prevention of causal misperceptions.
Labour Economics (2023), 83,102396.
Can the German minimum wage alleviate poverty and income inequality? with Kai-Uwe Müller
This study provides ex-post evidence on the redistributive impact of the minimum wage on disposable household incomes in Germany. Although the reduction of income inequality and poverty were emphasized as policy goals of the of minimum wage introduction, the distributional literature for Germany has focused on hourly wages and earnings or relied on ex-ante simulations. Since `natural' control groups are not available due to a nationwide introduction, this paper accumulates descriptive evidence. Based on data from the German Socio-Economic Panel we analyze changes in wage and income inequality, poverty, and different mechanisms that affect transmission from individual gross wage-rates to disposable household incomes. The minimum wage is found to be an ineffective tool for income redistribution. Increases at the bottom of the wage distribution do not translate into increases in disposable incomes of poor households because individuals affected by the minimum wage do not live primarily in households at the bottom, but are spread across the income distribution. Thus, welfare dependence is only marginally reduced. Additional simulations show that neither full compliance, nor substantial increases of the level would make the minimum wage more effective in reducing inequality and poverty.
Journal of European Social Policy, 33(2), 216–232, 2023.
Early stage:
Retirement Behavior under Cognitive Uncertainty with Marvin Immesberger and Hans-Martin von Gaudecker
The incentives induced by public pension systems are complex. At the same time we observe many people retiring at distinct retirement ages that are unlikely to reflect the pure trade-off between leisure and consumption preferences. To what extent can a neoclassical model with a comprehensive reflection of the financial incentives by the tax-and-transfer system explain behavior? And to what extent can cognitive uncertainty driven by the complexity of the decision explain excessive bunching?
Individual lifetime employment dynamics revisited: The role of worker and job characteristics with Mattis Beckmannshagen
Household decision making in retirement behavior with Hans-Martin von Gaudecker, and Christian Zimpelmann
In this project, we design a comprehensive structural model of household decision-making concerning employment and retirement behavior. The aim is to provide a model that allows to reliably simulate the consequences of German pension reforms under couples decision making.
Experimental / Behavioral Economics
Individual Strategy Choice in the Repeated Prisoner’s Dilemma with Yves Breitmoser
Reanalyzing 12 experiments on the repeated prisoner’s dilemma (PD), we find strong evidence for players’ use of behavior strategies. Starting with unrestricted memory-1 strategies, the most parsimonious non-rejected representation of behavior distinguishes three subject types: defectors, cautious cooperators and strong cooperators. The defectors defect with a high probability in every round. Both cooperating types play semi-grim behavior strategies with different cooperation rates in round 1. This simple three-type mixture fits significantly better than 1046 combinations of (generalized) pure strategies from the literature, which we fitted at the treatment level. Semi-grim behavior strategies fit better than all 1046 mixtures of (generalized) pure strategies even when we use a constant and pre-defined specification, without using free parameters or any kind of post-hoc econometric magic. Furthermore, the resulting type shares correlate with the treatment parameters in a predictable manner, and the strategies themselves are largely predictable thanks to their approximate invariance, but the strategies cannot be rationalized as responses to expected payoffs.
Inequity Aversion and Limited Foresight in the Repeated Prisoner’s Dilemma with Yves Breitmoser (extended Version)
In a meta-study we reanalyze 12 experiments on the repeated prisoner’s dilemma (PD) and identify three distinct types of players: defectors, cautious cooperators and strong cooperators. The defectors defect with a high probability in every round. Both cooperating types play semi-grim behavior strategies. This simple three-type mixture fits significantly better than any model consisting of combinations of (generalized) pure strategies from the literature, which we fitted at the treatment level (considering 10^51 pure-strategy mixtures), even when we use constant specifications of the three types across all experiments. The three best fitting strategies vary slightly across experiments, however. Structurally analyzing these strategies, we find that subjects have limited foresight and subjectively assign utility values to the four states (cc,cd,dc,dd) of the supergame, which relate to the original stage-game payoffs in a manner compatible with inequity aversion. This subjectively transforms the prisoners dilemma game into a coordination game and can reliably explain the strategies used across all 32 treatments.
Learning through period and physical time with Steffen Huck, Johannes Leutgeb, and Ryan Oprea
Subjects in a laboratory experiment play a two-player Cournot-Tullock game over hundreds of timed periods with and without information of the payoff function. By varying the length of periods across treatments, we are able to separately measure the effect of ``period time'' to that of ``physical time'' (clock time). In high information games, learning is driven purely by period time, while in low information games, physical time has a substantial impact on learning. We conclude that learning models should account for both, adaptive learning that occurs in period time and for contemplative learning that occurs in physical time.
Games and Economic Behavior (2023), 141, 305-322.
Early stage:
Repeated games with asymmetric payoffs with Yves Breitmoser
Older Projects
Evolution and Determinants of Rent Burdens in Germany with Kathrin Gebers and Carsten Schröder (2015)
The affordability of housing has become a major topic of discussion in Germany among both social scientists and the public at large. Using data from the German Socio-Economic Panel (SOEP), we provide rent-income ratios over more than two decades and show how they change with households’ disposable needs-adjusted income. We find a substantial increase in the ratios over the 1990s. In the decade that followed, they remained relatively constant. Moreover we find that rent-income ratios decrease in income at a decreasing rate, suggesting that rising square-meter prices put particular financial pressure on low-income households. Our analysis also indicates economies of scale from shared living space for multi-member households.
SOEP Paper 806 (2015).