Does Machine Learning Help us Predict Banking Crises?
Johannes Beutel, Sophia List, Gregor von Schweinitz
Journal of Financial Stability,
December
2019
Abstract
This paper compares the out-of-sample predictive performance of different early warning models for systemic banking crises using a sample of advanced economies covering the past 45 years. We compare a benchmark logit approach to several machine learning approaches recently proposed in the literature. We find that while machine learning methods often attain a very high in-sample fit, they are outperformed by the logit approach in recursive out-of-sample evaluations. This result is robust to the choice of performance metric, crisis definition, preference parameter, and sample length, as well as to using different sets of variables and data transformations. Thus, our paper suggests that further enhancements to machine learning early warning models are needed before they are able to offer a substantial value-added for predicting systemic banking crises. Conventional logit models appear to use the available information already fairly efficiently, and would for instance have been able to predict the 2007/2008 financial crisis out-of-sample for many countries. In line with economic intuition, these models identify credit expansions, asset price booms and external imbalances as key predictors of systemic banking crises.
Read article
Private Debt, Public Debt, and Capital Misallocation
Behzod Alimov
IWH-CompNet Discussion Papers,
No. 7,
2019
Abstract
Does finance facilitate efficient allocation of resources? Our aim in this paper is to find out whether increases in private and public indebtedness affect capital misallocation, which is measured as the dispersion in the return to capital across firms in different industries. For this, we use a novel dataset containing industrylevel data for 18 European countries and control for different macroeconomic indicators as potential determinants of capital misallocation. We exploit the within-country variation across industries in such indicators as external finance dependence, technological intensity, credit constraints and competitive structure, and find that private debt accumulation disproportionately increases capital misallocation in industries with higher financial dependence, higher R&D intensity, a larger share of credit-constrained firms and a lower level of competition. On the other hand, we fail to find any significant and robust effect of public debt on capital misallocation within our country-sector pairs. We believe the distortionary effects of private debt found in our analysis needs a deeper theoretical investigation.
Read article
Spillovers of Asset Purchases Within the Real Sector: Win-Win or Joy and Sorrow?
Talina Sondershaus
IWH Discussion Papers,
No. 22,
2019
Abstract
Events which have an adverse or positive effect on some firms can disseminate through the economy to firms which are not directly affected. By exploiting the first large sovereign bond purchase programme of the ECB, this paper investigates whether more lending to some firms spill over to firms in the surroundings of direct beneficiaries. Firms operating in the same industry and region invest less and reduce employment. The paper shows the importance to consider spillover effects when assessing unconventional monetary policies: Differences between treatment and control groups can be entirely attributed to negative effects on the control group.
Read article
Thou Shalt not Bear False Witness Against Your Customers: Cultural Norms and the Volkswagen Scandal
Iftekhar Hasan, Felix Noth, Lena Tonzer
Abstract
This paper investigates whether cultural norms shaped by religion drive consumer decisions after a corporate scandal. We exploit the unexpected notice of violation by the US Environmental Protection Agency in September 2015, accusing the car producer Volkswagen (VW) to have used software to manipulate car emission values during test phases. Using a difference-in-difference model, we show that new registrations of VW (diesel) cars decline significantly in German counties with a high share of Protestants following the VW scandal. Our results suggest that the enforcement culture rooted in Protestantism affects consumer decisions and penalises corporate fraud.
Read article
Should Forecasters Use Real‐time Data to Evaluate Leading Indicator Models for GDP Prediction? German Evidence
Katja Heinisch, Rolf Scheufele
German Economic Review,
No. 4,
2019
Abstract
In this paper, we investigate whether differences exist among forecasts using real‐time or latest‐available data to predict gross domestic product (GDP). We employ mixed‐frequency models and real‐time data to reassess the role of surveys and financial data relative to industrial production and orders in Germany. Although we find evidence that forecast characteristics based on real‐time and final data releases differ, we also observe minimal impacts on the relative forecasting performance of indicator models. However, when obtaining the optimal combination of soft and hard data, the use of final release data may understate the role of survey information.
Read article
The Value of Smarter Teachers: International Evidence on Teacher Cognitive Skills and Student Performance
Eric A. Hanushek, Marc Piopiunik, Simon Wiederhold
Journal of Human Resources,
No. 4,
2019
Abstract
We construct country-level measures of teacher cognitive skills using unique assessment data for 31 countries. We find substantial differences in teacher cognitive skills across countries that are strongly related to student performance. Results are supported by fixed-effects estimation exploiting within-country between-subject variation in teacher skills. A series of robustness and placebo tests indicate a systematic influence of teacher skills as distinct from overall differences among countries in the level of cognitive skills. Moreover, observed country variations in teacher cognitive skills are significantly related to differences in women’s access to high-skill occupations outside teaching and to salary premiums for teachers.
Read article
Bankruptcy Spillovers
Shai B. Bernstein, Emanuele Colonnelli, Xavier Giroud, Benjamin Iverson
Journal of Financial Economics,
No. 3,
2019
Abstract
How do different bankruptcy approaches affect the local economy? Using US Census microdata, we explore the spillover effects of reorganization and liquidation on geographically proximate firms. We exploit the random assignment of bankruptcy judges as a source of exogenous variation in the probability of liquidation. We find that employment declines substantially in the immediate neighborhood of the liquidated establishments, relative to reorganized establishments. The spillover effects are highly localized and concentrate in nontradable and service sectors, consistent with a reduction in local consumer traffic and a decline in knowledge spillovers between firms. The evidence highlights the externalities that bankruptcy design can impose on nonbankrupt firms.
Read article
Drivers of Effort: Evidence from Employee Absenteeism
Morten Bennedsen, Margarita Tsoutsoura, Daniel Wolfenzon
Journal of Financial Economics,
No. 3,
2019
Abstract
We use detailed information on individual absent spells of all employees in 4140 firms in Denmark to show large differences in average absenteeism across firms. Using employees who switch firms, we decompose days absent into an individual component (e.g., motivation, work ethic) and a firm component (e.g., incentives, corporate culture). We find the firm component explains 50%–60% of the difference in absenteeism across firms, with the individual component explaining the rest. We present suggestive evidence of the mechanisms behind the firm effect with family firm status and concentrated ownership strongly correlated with decreases in absenteeism. We also analyze the firm characteristics that correlate with the individual effect and find that firms with stronger career incentives attract lower-absenteeism employees.
Read article
The Internet Effects on Sex Crime Offenses – Evidence from the Broadband Internet Expansion in Germany
André Diegmann
Journal of Economic Behavior and Organization,
September
2019
Abstract
This paper studies the effects of the introduction of a new mass medium on sex crime in Germany. I use unique data on criminal offenses and broadband internet measured at the municipal level to shed light on this issue. In order to address endogeneity in broadband internet availability, I exploit technical peculiarities at the regional level that determine the roll-out of high-speed internet. Results provide evidence of a substitution effect of internet exposure on sex crime. The substitution effect is neither driven by differences in reporting behavior, nor by matching processes at the victim and offender side. This suggests that the consumption of extreme media plays an important role in explaining the documented high-speed internet effect.
Read article
How Forecast Accuracy Depends on Conditioning Assumptions
Carola Engelke, Katja Heinisch, Christoph Schult
IWH Discussion Papers,
No. 18,
2019
Abstract
This paper examines the extent to which errors in economic forecasts are driven by initial assumptions that prove to be incorrect ex post. Therefore, we construct a new data set comprising an unbalanced panel of annual forecasts from different institutions forecasting German GDP and the underlying assumptions. We explicitly control for different forecast horizons to proxy the information available at the release date. Over 75% of squared errors of the GDP forecast comove with the squared errors in their underlying assumptions. The root mean squared forecast error for GDP in our regression sample of 1.52% could be reduced to 1.13% by setting all assumption errors to zero. This implies that the accuracy of the assumptions is of great importance and that forecasters should reveal the framework of their assumptions in order to obtain useful policy recommendations based on economic forecasts.
Read article
