Contributions in Merlin

We investigate the effects of constraining leverage and shrinking covariance matrix in constructing large portfolios, both theoretically and empirically. Considering a wide variety of setups that involve conditioning or not conditioning the covariance matrix estimator on the recent past (multivariate GARCH), smaller vs. larger universe of stocks, alternative portfolio formation objectives (Global Minimum Variance vs. exposure to profitable factors), and various transaction cost assumptions, we find that a judiciously-chosen shrinkage method always outperforms an arbitrarily-determined leverage constraint. By extending the mathematical connection between leverage and shrinkage from static to dynamic, we provide a new theoretical explanation for our finding from the perspective of degrees of freedom. In addition, both simulation and empirical analysis show that the DCC-NL estimator results in risk reduction and efficiency increase in large portfolios as long as a small amount of leverage is allowed, whereas tightening the leverage constraint often hurts a DCC-NL portfolio.

Small businesses (SMEs) depend on banks for credit. We show that the severity of the Eurozone crisis was worse in countries where firms borrowed more from domestic banks (“domestic bank dependence”) than in countries where firms borrowed more from international banks. Eurozone banking integration in the years 2000–2008 mainly involved cross-border lending between banks while foreign banks’ lending to the real sector stayed flat. Hence, SMEs remained dependent on domestic banks and were vulnerable to global banking shocks. We confirm, using a calibrated quantitative model, that domestic bank dependence makes sectors and countries with many SMEs vulnerable to global banking shocks.

This paper addresses the modern optimal tax progressivity literature, which clarifies the key role of the behavioral response to taxation and accounts for the incomes of the superrich being qualitatively different than others. Some may be “superstars,” for whom small differences in talent are magnified into much larger earnings differences, while others may work in winner-take-all markets, such that their effort to climb the ladder of success reduces the returns to others. We stress that pivotal tax-rate elasticities are not structural parameters, and will be smaller the broader and less plastic is the tax base and the more effective is the enforcement of tax evasion. For this reason, normative analysis of tax rates should be accompanied by attention to the tax base, with a special focus on capital gains, which comprise a large fraction of the taxable income of the superrich.

We study the intergenerational transmission of time preferences in a setting without reverse causality concerns. We find substantial transmission of patience from parents to children, which is insensitive to the inclusion of comprehensive sets of administratively reported controls and persists as children age. We further explore heterogeneity in the transmission with respect to two theoretically important but distinct dimensions of socialization through which parents can influence children’s traits: parenting values and parental involvement. Our results show that, in contrast to authoritative parents, authoritarian and permissive parents transmit patience to their offspring. Meanwhile, parental involvement is not an important moderator. These patterns replicate in an independent sample with richer measures of parental involvement.

This paper constructs a new estimator for large covariance matrices by drawing a bridge between the classic Stein (1975) estimator in finite samples and recent progress under large-dimensional asymptotics. The estimator keeps the eigenvectors of the sample covariance matrix and applies shrinkage to the inverse sample eigenvalues. The corresponding formula is quadratic: it has two shrinkage targets weighted by quadratic functions of the concentration (that is, matrix dimension divided by sample size). The first target dominates mid-level concentrations and the second one higher levels. This extra degree of freedom enables us to outperform linear shrinkage when optimal shrinkage is not linear (which is the general case). Both of our targets are based on what we term the “Stein shrinker”, a local attraction operator that pulls sample covariance matrix eigenvalues towards their nearest neighbors, but whose force diminishes with distance, like gravitation. We prove that no cubic or higher-order nonlinearities beat quadratic with respect to Frobenius loss under large-dimensional asymptotics. Non-normality and the case where the matrix dimension exceeds the sample size are accommodated. Monte Carlo simulations confirm state-of-the-art performance in terms of accuracy, speed, and scalability.

Most of the existing evidence on the effectiveness of family leave policies comes from studies focusing on their impacts on affected families - that is, mothers, fathers, and their children - without a clear understanding of the costs and effects on firms and coworkers. We use data from Denmark to evaluate the effect on firms and coworkers when a worker gives birth and goes on leave. Using a dynamic difference-in-differences design, we compare small firms in which a female employee is about to give birth to an observationally equivalent sample of small firms with female employees who are not close to giving birth. Identification rests on a parallel trends assumption, which we substantiate through a set of natural validity checks. When an employee gives birth she goes on leave from her firm for 9.5 months on average. Firms respond by increasing their labor inputs along several margins such that the net effect on total work hours is close to zero. Firms' total wage bill increases in response to leave take up, but this is driven entirely by wages paid to workers on leave for which firms receive reimbursement. There are no measurable effects on firm output, profitability or survival. Finally, coworkers of the woman going on leave see temporary increases in their hours, earnings, and likelihood of being employed but experience no significant changes in well-being at work as proxied by sick days. Overall, our results suggest that employees going on parental leave impose negligible costs on their firm and coworkers.

This paper examines the labor market effects of unemployment insurance extensions. It uses administrative Social Security matched employer-employee data from Austria. Critical components of the analysis are effects on wages as well as retirement/job separation effects. The paper found that: - Older workers are very responsive to unemployment insurance benefit extensions, which in our setting may have served as a bridge into early retirement. - These separations stem from high-value employment relationships, and may constitute inefficient separations. - Wages appear unresponsive to such shifts on workers’ nonemployment outside options, implying that these workers may not be able to use such reforms to improve their bargaining power - even if they are older and on the retirement margin. - However, in a second study, we find no effects on separations of such policies even for older workers. The key difference is that one reform was potential benefit extensions (entailing early retirement) rather than the benefit levels (in increases in the generosity of which even older workers did not separate from their jobs). The policy implications of the findings are: - Unemployment insurance benefit generosity may interact with retirement policies and trigger early retirement among older workers eligible for these benefits. - These separations may be socially inefficient and may therefore have welfare costs. Subsidizing early retirement policies does not entail wage pressure.

We study experimentally when, why, and how people intervene in others' choices. Choice Architects (CAs) construct opportunity sets containing bundles of time-indexed payments for Choosers. CAs frequently prevent impatient choices despite opportunities to provide advice, believing Choosers benefit. We consider several hypotheses concerning CAs' motives. A conventional behavioral welfarist acts as a correctly informed social planner; a mistakesprojective paternalist removes options she wishes she could reject when choosing for herself; an ideals-projective paternalist seeks to align others' choices with her own aspirations. Idealsprojective paternalism provides the best explanation for interventions in the laboratory and rationalizes support for actual paternalistic policies.

Previous research shows that many people seek financial advice from non-experts, and that peer interactions influence financial decisions. We investigate whether such influences are beneficial, harmful, or simply haphazard. In our laboratory experiment, face-to-face communication with a randomly assigned peer significantly improves the quality of private decisions, measured by subjects' ability to choose as if they properly understand their opportunity sets. Subjects do not merely mimic those who know better, but also make better private decisions in novel tasks. People with low financial competence experience greater improvements when their partners also exhibit low financial competence. Hence, peer-to-peer communication transmits financial decision-making skills most effectively when peers are equally uninformed, rather than when an informed decision maker teaches an uninformed peer. Qualitative analysis of subjects' discussions supports this interpretation. The provision of effective financial education to one member of a pair influences the nature of communication but does not lead to additional improvements in the quality of the untreated partner's decisions, particularly in novel tasks.

Unraveling, the excessively early matching of future workers to employers, is a pervasive phenomenon in entry-level labor markets that leads to hiring decisions based on severely incomplete information. We provide a model of unraveling in one-to-one matching markets for prestigious positions. Its distinguishing feature is that the market operates over an extended time period during which information about potential matches arrives gradually. We find that unraveling causes potentially thick markets to spread thinly over a long time period. In equilibrium, an employers desirability is correlated neither with the time at which they hire, nor with the expected productivity of their matched worker. Unraveling thus significantly redistributes welfare among employers compared to a pairwise stable match. We study policies that manipulate the availability of information about students and show that they are effective only if they provide a sudden surge in information. Our main application is the market for U.S. federal appellate court clerks, a significant input into the efficiency of the justice system. Consistent with the model, hiring times in our dataset are spread over a period of six months and are uncorrelated with the desirability of a judge as an employer.