Contributions in Merlin

Covariance matrix estimation and principal component analysis (PCA) are two cornerstones of multivariate analysis. Classic textbook solutions perform poorly when the dimension of the data is of a magnitude similar to the sample size, or even larger. In such settings, there is a common remedy for both statistical problems: nonlinear shrinkage of the eigenvalues of the sample covariance matrix. The optimal nonlinear shrinkage formula depends on unknown population quantities and is thus not available. It is, however, possible to consistently estimate an oracle nonlinear shrinkage, which is motivated on asymptotic grounds. A key tool to this end is consistent estimation of the set of eigenvalues of the population covariance matrix (also known as the spectrum), an interesting and challenging problem in its own right. Extensive Monte Carlo simulations demonstrate that our methods have desirable finite-sample properties and outperform previous proposals.

In this paper we analyze R&D collaboration networks in industries where firms are competitors in the product market. Firms’ benefits from collaborations arise by sharing knowledge about a cost-reducing technology. By forming collaborations, however, firms also change their own competitive position in the market as well as the overall market structure. We analyze incentives of firms to form R&D collaborations with other firms and the implications of these alliance decisions for the overall network structure. We provide a general characterization of both equilibrium networks and endogenous production choices, and compare it to the efficient network architecture. We also allow for firms to differ in their technological characteristics, investigate how this affects their propensity to collaborate and study the resulting network architecture.

Many postulated relations in finance imply that expected asset returns strictly increase in an underlying characteristic. To examine the validity of such a claim, one needs to take the entire range of the characteristic into account, as is done in the recent proposal of Patton and Timmermann (2010). But their test is only a test for the direction of monotonicity, since it requires the relation to be monotonic from the outset: either weakly decreasing under the null or strictly increasing under the alternative. When the relation is non-monotonic or weakly increasing, the test can break down and falsely ‘establish’ a strictly increasing relation with high probability. We offer some alternative tests that do not share this problem. The behavior of the various tests is illustrated via Monte Carlo studies. We also present empirical applications to real data.

Many governments subsidize regional rail service as an alternative to road traffic. This paper assesses whether increases in service frequency reduce road traffic externalities. We exploit differences in service frequency growth by procurement mode following a railway reform in Germany to address endogeneity of service growth. Increases in service frequency reduce the number of severe road traffic accidents, carbon monoxide, nitrogen monoxide, nitrogen dioxide pollution and infant mortality. Placebo regressions with sulfur dioxide and ozone yield no effect. Service frequency growth between 1994 and 2004 improves environmental quality by an amount that is worth approximately 28-40 % of total subsidies. An analysis of household behavior shows that the effects of railway services on outcome variables are driven by substitution from road to rail.

This paper describes and analyzes research on the dynamics of long-term care and suggests directions for the literature to make progress. We discuss sources and causes of dynamics including inertia/state dependence (confounded by unobserved heterogeneity); match-specific effects; and costs of changing caregivers. We comment on causes of dynamics including learning/human capital accumulation; burnout; and game-playing. We suggest how to deal with endogenous geography; dynamics in discrete and continuous choices; and equilibrium issues (multiple equilibria, dynamic equilibria). Next, we evaluate the advantages of different potential data sources (NLTCS, PSID, AHEAD/HRS, SHARE, ELSA) and identify first order data problems including noisy measures of wealth and family structure. We suggest some methods to handle econometric problems such as endogeneity (work, geography) and measurement error. Finally, we discuss potential policy implications of dynamics including the effect of dynamics on parameter estimates and direct policy implications of inertia (implications for family welfare, parent welfare, child welfare, and cost of government programs).

This paper investigates empirically how similarity of demand structures - approximated by similarity of income distributions - affects trade patterns along both the extensive and intensive margin. The idea that similarity of demand structures intensifies trade goes back to the well-known Linder hypothesis. Based on a sample of 102 countries, I find that bilateral trade volumes are increasing in the overlap of two countries income distributions. This effect is driven by both the extensive and intensive margin. I establish two novel measures of income similarity - the average income level of the overlap area and the range of incomes for which two distributions overlap - and document that both are important determinants of bilateral trade margins. My analysis shows that the positive relationship between similarity of income distributions and bilateral trade margins is present at the aggregate and disaggregate level of trade flows.

We assess the role of financial linkages in the transmission of sovereign risk in the Euro Crisis. Building on the narrative approach by Romer and Romer (1989), we use financial news to identify structural shocks in a vector autoregressive model of daily sovereign CDS premia for eleven European countries. To estimate how these shocks transmit across borders, we use data on cross-country bank exposures to sovereign debt. Our results indicate that exposure to Greek sovereign debt and the debt of Greek banks constitute important transmission channels. All else being equal, the transmission rate to the country with the greatest exposure to Greece (1.22 percent of GDP) has been roughly 46 percent higher than the rate to the country with the least exposure (0.08 percent of GDP).

This paper explores how extended unemployment insurance (UI) benefits targeted to older workers affect early retirement and social welfare. The trade-off of optimal UI between consumption smoothing and moral hazard requires accounting for the entire early retirement system, which often includes extended UI and relaxed access to disability insurance (DI). We argue that extended UI generates program complementarity (increased take-up of UI followed by DI and/or regular retirement benefits) and program substitution (increased take-up of UI instead of DI). Exploiting Austria's regional extended benefit program, which extended regular UI benefits to up to 4 years, we find: (i) program complementarity is quantitatively important for workers aged 50+; and (ii) program substitution is quantitatively relevant for workers aged 55+. We derive a simple rule for optimal UI that accounts for program complementarity and program substitution. Using the sufficient statistics approach, we conclude that UI for older workers was too generous and the regional extended benefit program was a suboptimal policy.

This paper analyzes whether hosting the most prestigious European cultural event, the European Capital of Culture, has an impact on regional economic development or the life satisfaction of the local population. Concerning the economic impact, we show that European Capitals are hosted in regions with above average GDP per capita, but do not causally affect the economic development in a significant way. Even a positive impact on GDP per capita would not imply a positive impact on individual utility or social welfare of the regional population. Surprisingly, using difference-in-difference estimations, a negative effect on the well-being of the regional population is found during the event. Since no effect is found before the event, reverse causality and positive anticipation can be ruled out. The negative effect during the event might result from dissatisfaction with the high levels of public expenditure, transport disruptions, general overcrowding or an increase in housing prices.

During industrialization, Protestants were more literate than Catholics. This paper investigates whether this fact may be led back to the intrinsic motivation of Protestants to read the bible and to what extent other education motives might have been involved as well. We employ a historical data set from Switzerland which allows us to differentiate between different cognitive skills: reading, numeracy, essay writing and Swiss history. We develop an estimation strategy to examine whether the impact of religious denomination was particularly large with respect to reading capabilities. We find support for this hypothesis. However, we also find evidence which is consistent with the view that Protestants’ education otives went beyond acquiring reading skills.