Stress-testing US bank holding companies: A dynamic panel quantile regression approach

We propose an econometric framework for estimating capital shortfalls of bank holding companies (BHCs) under pre-specified macroeconomic scenarios. To capture the nonlinear dynamics of bank losses and revenues during periods of financial stress, we use a fixed effects quantile autoregressive (FE-QAR) model with exogenous macroeconomic covariates, an approach that delivers a superior out-of-sample forecasting performance relative to the standard linear framework. According to the out-of-sample forecasts, the realized net charge-offs during the 2007–09 crisis fall within the multi-step-ahead density forecasts implied by the FE-QAR model, but are frequently outside the density forecasts generated using the corresponding linear model. This difference reflects the fact that the linear specification substantially underestimates loan losses, especially for real estate loan portfolios. Employing the macroeconomic stress scenario used in CCAR 2012, we use the density forecasts generated by the FE-QAR model to simulate capital shortfalls for a panel of large BHCs. For almost all institutions in the sample, the FE-QAR model generates capital shortfalls that are considerably higher than those implied by its linear counterpart, which suggests that our approach has the potential to detect emerging vulnerabilities in the financial system.     

Instrumental variable quantile regression: A robust inference approach

In this paper, we develop robust inference procedures for an instrumental variables model defined by Y=D^′α(U)$Y=D^{\prime }\alpha (U)$ where D^′α(U)$D^{\prime }\alpha (U)$ is strictly increasing in U and U is a uniform variable that may depend on D but is independent of a set of instrumental variables Z. The proposed inferential procedures are computationally convenient in typical applications and can be carried out using software available for ordinary quantile regression. Our inferential procedure arises naturally from an estimation algorithm and has the important feature of being robust to weak and partial identification and remains valid even in cases where identification fails completely. The use of the proposed procedures is illustrated through two empirical examples.     

Robust penalized quantile regression estimation for panel data

This paper investigates a class of penalized quantile regression estimators for panel data. The penalty serves to shrink a vector of individual specific effects toward a common value. The degree of this shrinkage is controlled by a tuning parameter λ$\lambda$. It is shown that the class of estimators is asymptotically unbiased and Gaussian, when the individual effects are drawn from a class of zero-median distribution functions. The tuning parameter, λ$\lambda$, can thus be selected to minimize estimated asymptotic variance. Monte Carlo evidence reveals that the estimator can significantly reduce the variability of the fixed-effect version of the estimator without introducing bias.     

Endogeneity in quantile regression models: A control function approach

This paper considers a linear triangular simultaneous equations model with conditional quantile restrictions. The paper adjusts for endogeneity by adopting a control function approach and presents a simple two-step estimator that exploits the partially linear structure of the model. The first step consists of estimation of the residuals of the reduced-form equation for the endogenous explanatory variable. The second step is series estimation of the primary equation with the reduced-form residual included nonparametrically as an additional explanatory variable. This paper imposes no functional form restrictions on the stochastic relationship between the reduced-form residual and the disturbance term in the primary equation conditional on observable explanatory variables. The paper presents regularity conditions for consistency and asymptotic normality of the two-step estimator. In addition, the paper provides some discussions on related estimation methods in the literature.     

Asymptotics for panel quantile regression models with individual effects

This paper studies panel quantile regression models with individual fixed effects. We formally establish sufficient conditions for consistency and asymptotic normality of the quantile regression estimator when the number of individuals, n$n$, and the number of time periods, T$T$, jointly go to infinity. The estimator is shown to be consistent under similar conditions to those found in the nonlinear panel data literature. Nevertheless, due to the non-smoothness of the objective function, we had to impose a more restrictive condition on T$T$ to prove asymptotic normality than that usually found in the literature. The finite sample performance of the estimator is evaluated by Monte Carlo simulations.     

Productivity and efficiency at bank holding companies in the U.S.: a time-varying heterogeneity approach

This paper investigates the productivity and efficiency of large bank holding companies (BHCs) in the United States over the period 2004–2013, by estimating a translog stochastic distance frontier (SDF) model with time-varying heterogeneity. The main feature of this model is that a multi-factor structure is used to disentangle time-varying unobserved heterogeneity from inefficiency. Our empirical results strongly suggest that unobserved heterogeneity is not only present in the U.S. banking industry, but also varies over time. Our results from the translog SDF model with time-varying heterogeneity show that the majority of large BHCs in the U.S. exhibit increasing returns to scale, a small percentage exhibit constant returns to scale, and an even smaller percentage exhibit decreasing returns to scale. Our results also show that on average the BHCs have experienced small positive or even negative technical change and productivity growth.     

Determinants of house prices in Istanbul: a quantile regression approach

This paper uses quantile regression methods where a hedonic equation is estimated for each quantile of the conditional distribution of housing prices. The survey data are used to investigate the relationship between house prices and housing characteristics in Istanbul. This data set includes some housing characteristics of the dwellings like numbers of room, bathroom, heating system, location of house etc. In the results of this paper show some similarities and differences from earlier studies on housing prices. We find that age, cable tv, security, heating system, garage, kitchen area, increasing numbers of room and bathroom increase the house prices. Our findings also show that side variable which is a special factor for Istanbul real estate market has negative effect on the prices. It is clear that the factors of housing prices can change because of the properties of country, region or city. The results of this study may give some important interpretations for developing real estate market.     

Corporate social responsibility and corporate performance: a quantile regression approach

This paper investigates the impact of corporate social responsibility activities on corporate performance. In view of the inconsistent empirical findings in the literature, and the limitations of least squares regressions, we adopt a quantile regression method to fill this gap in the literature. An important finding is that the sensitivity of a company’s performance to its engagement in corporate social responsibility activities does not vary with the quantile location of the firm’s performance level, and the engagement in corporate social responsibility activities has a significant positive relation with corporate performance across all quantiles. This study argues that undertaking corporate social responsibility leads to greater financial returns than the related costs. Therefore, this study concludes that engaging in corporate social responsibility is beneficial for firms, and thus worth implementing.     

Local employment growth in West Germany: A dynamic panel approach

In this paper we study the dynamics of local employment growth in West Germany from 1980 to 2001. Using dynamic panel techniques, we analyse the timing of the impact of diversity and specialisation, as well as of the human capital structure of local industries. Diversity has a positive effect on employment growth in the short run, which is stronger in manufacturing than in services. Concerning specialisation we find also positive effects. They are, however, not strong enough to reinforce growth permanently, since there is mean reversion in the process. A positive effect of education is only found in manufacturing. Additionally, we look at the impact of firm size and regional wages on local employment growth.     

Semiparametric quantile regression estimation in dynamic models with partially varying coefficients

We study quantile regression estimation for dynamic models with partially varying coefficients so that the values of some coefficients may be functions of informative covariates. Estimation of both parametric and nonparametric functional coefficients are proposed. In particular, we propose a three stage semiparametric procedure. Both consistency and asymptotic normality of the proposed estimators are derived. We demonstrate that the parametric estimators are root-n$n$ consistent and the estimation of the functional coefficients is oracle. In addition, efficiency of parameter estimation is discussed and a simple efficient estimator is proposed. A simple and easily implemented test for the hypothesis of a varying-coefficient is proposed. A Monte Carlo experiment is conducted to evaluate the performance of the proposed estimators.     

Quantile regression for dynamic panel data with fixed effects

This paper studies a quantile regression dynamic panel model with fixed effects. Panel data fixed effects estimators are typically biased in the presence of lagged dependent variables as regressors. To reduce the dynamic bias, we suggest the use of the instrumental variables quantile regression method of Chernozhukov and Hansen (2006) along with lagged regressors as instruments. In addition, we describe how to employ the estimated models for prediction. Monte Carlo simulations show evidence that the instrumental variables approach sharply reduces the dynamic bias, and the empirical levels for prediction intervals are very close to nominal levels. Finally, we illustrate the procedures with an application to forecasting output growth rates for 18 OECD countries.     

The relationship between board characteristics and performance of bank holding companies: before and during the financial crisis

We examine the relationship between performance of the bank holding company and several board characteristics. We find that board size, CEO tenure and board tenure enhance bank performance. However, we find no evidence that board structure or CEO power influences bank performance. More importantly, we show that the effect of board characteristics during the crisis is quite different. During the crisis, board size has a negative effect on Tobin’s Q and the non-performing asset ratio, which supports Jensen’s (1993) argument that large boards are less likely to function effectively. Further, we report that the non-performing asset ratio decreases with board independence during the crisis.     

Nonparametric efficiency analysis: A multivariate conditional quantile approach

This paper focuses on nonparametric efficiency analysis based on robust estimation of partial frontiers in a complete multivariate setup (multiple inputs and multiple outputs). It introduces α-quantile efficiency scores. A nonparametric estimator is proposed achieving strong consistency and asymptotic normality. Then if α increases to one as a function of the sample size we recover the properties of the FDH estimator. But our estimator is more robust to the perturbations in data, since it attains a finite gross-error sensitivity. Environmental variables can be introduced to evaluate efficiencies and a consistent estimator is proposed. Numerical examples illustrate the usefulness of the approach.     

Health and political regimes: Evidence from quantile regression

This study investigates how political regimes affect health conditions such as infant and child mortality rates and life expectancy using data from 180 countries observed between 1960 and 2013. Panel quantile regression is used to examine the effects at different intervals throughout the distribution of health outcomes. The estimation results indicate that democracy has significant positive effects on health outcomes and that its impacts are greater when health outcomes are worse. These results are robust to different democracy and health indices. The effects of different types of democracies and dictatorships are also considered, that is parliamentary, mixed (semi-presidential) and presidential democracies, and civilian, military and royal dictatorships. The parliamentary form of democracy has the largest positive impact on health outcomes at the worst quantile of health outcomes, although the difference in the impacts of the three types of democracies is not necessarily large. Furthermore, all types of dictatorships have a negative impact on health outcomes, with military dictatorship having the worst outcome when health outcomes are worse. Finally, the effects of democratization on health outcomes are significantly positive when the health outcomes are worse.     

Private school vouchers and student achievement: A fixed effects quantile regression evaluation

Fundamental to the recent debate over school choice is the issue of whether voucher programs actually improve students' academic achievement. Using newly developed quantile regression approaches, this paper investigates the distribution of achievement gains in the first school voucher program implemented in the US. We find that while high-performing students selected for the Milwaukee Parental Choice program had a positive, convexly increasing gain in mathematics, low-performing students had a nearly linear loss. However, the program seems to prevent low-performing students from having an even bigger loss experienced by students in the public schools.     

Forecasting bank failures and stress testing: A machine learning approach

This paper presents a forecasting model of bank failures based on machine-learning. The proposed methodology defines a linear decision boundary that separates the solvent banks from those that failed. This setup generates a novel alternative stress-testing tool. Our sample of 1443 U.S. banks includes all 481 banks that failed during the period 2007–2013. The set of explanatory variables is selected using a two-step feature selection procedure. The selected variables were then fed to a support vector machines forecasting model, through a training–testing learning process. The model exhibits a 99.22% overall forecasting accuracy and outperforms the well-established Ohlson’s score.     

Dynamic bank runs: an agent-based approach

Journal of Economic Interaction and Coordination - This paper simulates bank runs by using an agent-based approach to assess the depositors’ behavior under various scenarios in a...     

A nonlinear approach to US GNP

A univariate nonlinear model is estimated for US GNP that on many criteria outperforms standard linear models. The estimated model is of the threshold autoregressive type and contains evidence of asymmetric effects of shocks over the business cycle. In particular the nonlinear model suggests that the post-1945 US economy is significantly more stable than the pre-1945 US economy.     

Incentive effects of social assistance: A regression discontinuity approach

Before 1989, childless social assistance recipients in Quebec under age 30 received much lower benefits than recipients over age 30. We use this sharp discontinuity in policy to estimate the effects of social assistance on various labour market outcomes using a regression discontinuity approach. We find strong evidence that more generous social assistance benefits reduce employment. The estimates exhibit little sensitivity to the degree of flexibility in the specification, and perform very well when we control for unobserved heterogeneity using a first difference specification. Finally, we show that commonly used difference-in-differences estimators may perform poorly with inappropriately chosen control groups.     

Income convergence within the MENA countries: A panel unit root approach

This article aims at testing the convergence hypothesis in MENA region using new tests of a unit root in panel data. Evans and Karras [Evans P., & Karras G. (1996). Convergence revisited. Journal of Monetary Economics, 37, 249–265] and Bernard and Jones [Bernard A., & Jones C. I. (1996). Productivity across industries and countries: Time series theory and evidence. The Review of Economics and Statistics, 135–146] recommend this technique to evaluate the income convergence hypothesis. According to them it avoids econometric problems of the cross-countries growth regressions testing convergence and sample bias of the multivariate cointegration techniques. We test for both absolute and the conditional convergence with panel unit root tests using the Summers and Heston's data 5.6 and 6.1 on the periods of 1960 to 1990 and from 1960 to 2000. The absolute convergence hypothesis use panel unit roots test with no fixed individual effects. The catching-up hypothesis is not rejected for most groups of countries of the region during both periods. If we allow a break in the unit root tests, the hypothesis is not rejected for more groups. The conditional convergence requires panel unit root tests with fixed individual effects. Again, during the whole periods, the conditional convergence is not rejected for the major part of the remaining groups of MENA countries.