Backtesting trading risk of commercial banks using expected shortfall

This paper uses saddlepoint technique to backtest the trading risk of commercial banks using expected shortfall. It is found that four out of six US commercial banks have excessive trading risks. Monte Carlo simulation studies show that the proposed backtest is very accurate and powerful even for small test samples. More importantly, risk managers can carry out the proposed backtest based on any number of exceptions, so that incorrect risk models can be promptly detected before any further huge losses are realized.     

The level and quality of Value-at-Risk disclosure by commercial banks

In this paper we study both the level of Value-at-Risk (VaR) disclosure and the accuracy of the disclosed VaR figures for a sample of US and international commercial banks. To measure the level of VaR disclosures, we develop a VaR Disclosure Index that captures many different facets of market risk disclosure. Using panel data over the period 1996–2005, we find an overall upward trend in the quantity of information released to the public. We also find that Historical Simulation is by far the most popular VaR method. We assess the accuracy of VaR figures by studying the number of VaR exceedances and whether actual daily VaRs contain information about the volatility of subsequent trading revenues. Unlike the level of VaR disclosure, the quality of VaR disclosure shows no sign of improvement over time. We find that VaR computed using Historical Simulation contains very little information about future volatility.     

Cross-sectional forecasts of the equity premium

If investors are myopic mean-variance optimizers, a stock's expected return is linearly related to its beta in the cross-section. The slope of the relation is the cross-sectional price of risk, which should equal the expected equity premium. We use this simple observation to forecast the equity-premium time series with the cross-sectional price of risk. We also introduce novel statistical methods for testing stock-return predictability based on endogenous variables whose shocks are potentially correlated with return shocks. Our empirical tests show that the cross-sectional price of risk (1) is strongly correlated with the market's yield measures and (2) predicts equity-premium realizations, especially in the first half of our 1927–2002 sample.     

Testing Conditional Asset Pricing Models Using a Markov Chain Monte Carlo Approach

We use Markov Chain Monte Carlo (MCMC) methods for the parameter estimation and the testing of conditional asset pricing models. In contrast to traditional approaches, it is truly conditional because the assumption that time variation in betas is driven by a set of conditioning variables is not necessary. Moreover, the approach has exact finite sample properties and accounts for errors‐in‐variables. Using S&P 500 panel data, we analyse the empirical performance of the CAPM and the Fama and French (1993) three‐factor model. We find that time‐variation of betas in the CAPM and the time variation of the coefficients for the size factor (SMB) and the distress factor (HML) in the three‐factor model improve the empirical performance. Therefore, our findings are consistent with time variation of firm‐specific exposure to market risk, systematic credit risk and systematic size effects. However, a Bayesian model comparison trading off goodness of fit and model complexity indicates that the conditional CAPM performs best, followed by the conditional three‐factor model, the unconditional CAPM, and the unconditional three‐factor model.     

A Long-Run Non-Linear Approach to the Fisher Effect

We argue that the empirical failure of the Fisher effect found in the literature may be due to the existence of non-linearities in the long-run relationship between interest rates and inflation. We present evidence that, for the U.S. during the 1960–2004 period, the Fisher relation presents important non-linearities. We model the long-run non-linear relationship and find that an ESTR model for the pre-Volcker era and an LSTR model for the post-Volcker era are able to control for non-linearities and constitute long-run co-integration vectors. Monte Carlo evidence produces support for the hypothesis that non-linearities may also be responsible for the less than proportional coefficients of inflation usually found in the linear specifications.     

Bank loan recovery rates: Measuring and nonparametric density estimation

In this paper we analyse a comprehensive database of 149,378 recovery rates on Italian bank loans. We investigate a new methodology to compute the recovery percentage that we suggest to consider as a mixed random variable. To estimate the probability density function of such a mixture, we propose the mixture of beta kernels estimator and we analyse its performance by Monte Carlo simulations. The application of these proposals to the Bank of Italy’s data shows that, even if we remove the endpoints from the support of the recovery rate, the density function estimate is far from being a beta function.     

Extending quadrature methods to value multi-asset and complex path dependent options

The exposition of the quadrature (QUAD) method (Andricopoulos, Widdicks, Duck, and Newton, 2003. Universal option valuation using quadrature methods. Journal of Financial Economics 67, 447–471 (see also Corrigendum, Journal of Financial Economics 73, 603 (2004)) is significantly extended to cover notably more complex and difficult problems in option valuations involving one or more underlyings. Trials comparing several techniques in the literature, adapted from standard lattice, grid and Monte Carlo methods to tackle particular types of problem, show that QUAD offers far greater flexibility, superior convergence, and hence, increased accuracy and considerably reduced computational times. The speed advantage of QUAD means that, even under the curse of dimensionality, it is not necessary to resort to Monte Carlo methods (certainly for options involving up to five underlying assets). Given the universality and flexibility of the method, it should be the method of choice for pricing options involving multiple underlying assets, in the presence of many features, such as early exercise or path dependency.     

Volatility dynamics for the S&P 500: Further evidence from non-affine,multi-factor jump diffusions

We apply Markov chain Monte Carlo methods to time series data on S&P 500 index returns, and to its option prices via a term structure of VIX indices, to estimate 18 different affine and non-affine stochastic volatility models with one or two variance factors, and where jumps are allowed in both the price and the instantaneous volatility. The in-sample fit to the VIX term structure shows that the second (stochastic long-term volatility) factor is required to fit the VIX term structure. Out-of-sample tests on the fit to individual option prices, as well as in-sample tests, show that the inclusion of jumps is less important than allowing for non-affine dynamics. The estimation and testing periods together cover more than 21 years of daily data.     

Estimating New-Keynesian Phillips curves: A full information maximum likelihood approach

The New-Keynesian Phillips curve has recently become an important ingredient in monetary policy models. However, using limited information methods, the empirical support for the New-Keynesian Phillips curve appear to be mixed. This paper argues, by means of Monte Carlo simulations with a simple New-Keynesian sticky price model, that single equations methods, e.g. GMM, are likely to produce imprecise and biased estimates. Then, it is argued that estimating the model with full information maximum likelihood (FIML) is a useful way of obtaining better estimates. Finally, a version of the model used in the Monte Carlo simulations is estimated on U.S. data with FIML and although the pure forward-looking New-Keynesian Phillips curve is rejected, a version with both forward- and backward-looking components provides a reasonable approximation of U.S. inflation dynamics.     

Modelling and forecasting short-term interest rate volatility: A semiparametric approach

This paper employs a semiparametric procedure to estimate the diffusion process of short-term interest rates. The Monte Carlo study shows that the semiparametric approach produces more accurate volatility estimates than models that accommodate asymmetry, level effect and serial dependence in the conditional variance. Moreover, the semiparametric approach yields robust volatility estimates even if the short rate drift function and the underlying innovation distribution are misspecified. Empirical investigation with the U.S. three-month Treasury bill rates suggests that the semiparametric procedure produces superior in-sample and out-of-sample forecast of short rate changes volatility compared with the widely used single-factor diffusion models. This forecast improvement has implications for pricing interest rate derivatives.     

Price discovery and trade fragmentation in a multi-market environment: Evidence from the MTS system

This paper proposes new metrics for the process of price discovery on the main electronic trading platform for euro-denominated government securities. Analysing price data on daily transactions for 107 bonds over a period of 27 months, we find a greater degree of price leadership of the dominant market when our measures (as opposed to the traditional price discovery metrics) are used. We also present unambiguous evidence that a market’s contribution to price discovery is crucially affected by the level of trading activity. The implications of these empirical findings are discussed in the light of the debate about the possible restructuring of the regulatory framework for the Treasury bond market in Europe.     

The Effect of Market Regimes on Style Allocation

We analyse time-varying risk premia and the implications for portfolio choice. Using Markov Chain Monte Carlo (MCMC) methods, we estimate a multivariate regime-switching model for the Carhart (1997) four-factor model. We find two clearly separable regimes with different mean returns, volatilities, and correlations. In the High-Variance Regime, only value stocks deliver a good performance, whereas in the Low-Variance Regime, the market portfolio and momentum stocks promise high returns. Regime-switching induces investors to change their portfolio style over time depending on the investment horizon, the risk aversion, and the prevailing regime. Value investing seems to be a rational strategy in the High-Variance Regime, momentum investing in the Low-Variance Regime. An empirical out-of-sample backtest indicates that this switching strategy can be profitable, but the overall forecasting ability for the regime-switching model seems to be weak compared to the iid model.     

Unconditional return disturbances: A non-parametric simulation approach

Simulation methods are extensively used in Asset Pricing and Risk Management. The most popular of these simulation approaches, the Monte Carlo, requires model selection and parameter estimation. In addition, these approaches can be extremely computer intensive. Historical simulation has been proposed as a non-parametric alternative to Monte Carlo. This approach is limited to the historical data available.In this paper, we propose an alternative historical simulation approach. Given a historical set of data, we define a set of standardized disturbances and we generate alternative price paths by perturbing the first two moments of the original path or by reshuffling the disturbances. This approach is either totally non-parametric when constant volatility is assumed; or semi-parametric in presence of GARCH(1, 1) volatility. Without a loss in accuracy, it is shown to be much more powerful in terms of computer efficiency than the Monte Carlo approach. It is also extremely simple to implement and can be an effective tool for the valuation of financial assets.We apply this approach to simulate pay off values of options on the S&P 500 stock index for the period 1982–2003. To verify that this technique works, the common back-testing approach was used. The estimated values are insignificantly different from the actual S&P 500 options payoff values for the observed period.     

Diversification and risk-adjusted performance: A quantile regression approach

The effect of diversification on firm performance has been debated. We reexamine the effect using a sample of 44,248 observations of non-financial US firms for the 1997–2009 period employing the quantile regression approach. Our empirical results show that the effect of diversification on firm performance is not homogeneous across various quantile levels: the diversification discount (premium) shows up in firms with high (low) RoE quantiles. Further, we find that diversification affects firm risk as well. Therefore, we consider a risk-adjusted performance measure and find that both diversification discount and premium disappear, which is consistent with the risk-return trade-off principle.     

Institutional investor stability and crash risk: Monitoring versus short-termism?

This study tests two opposing views of institutional investors—monitoring versus short-termism. We present evidence that institutional investor stability is negatively associated with 1-year-ahead stock price crash risk, consistent with the monitoring theory of institutional investors but not the short-termism theory. Our findings are shown to be robust to alternative empirical specifications, estimation methods and endogeneity concerns. In addition, we find that institutional ownership by public pension funds (bank trusts, investment companies, and independent investment advisors) is significantly negatively (positively) associated with future crash risk, consistent with findings that pension funds more actively monitor management than other types of institutions.     

Block ownership and firm-specific information

This study examines the impact of block ownership on the firm’s information environment. Previous research shows that stock price efficiency depends on the cost of acquiring private information, as well as on the precision of this information. Blockholders have a clear advantage over diffuse, atomistic shareholders in terms of the precision and acquisition cost of their private information. We hypothesize that this informational advantage will manifest itself primarily in the firm-specific component of stock returns. Our empirical findings confirm that blockholders increase the probability of informed trading and idiosyncratic volatility, and decrease the firm’s stock return synchronicity. These results hold for both inside and outside blockholders, but are insignificant for blocks controlled by employee stock ownership plans (ESOPs). Overall, our findings support the contention that ownership structure plays a significant role in shaping the firm’s information environment.     

Economic value,competition and financial distress in the European banking system

In this paper we examine the impact of a large number of factors at the bank level (liquidity and credit risks, asset size, income diversification and market power), at the industry level (banking concentration) and macro-level (real GDP growth) on bank financial distress using an unbalanced panel of 308 European commercial banks between 1996 and 2009. The observations falling below a given threshold of the empirical distribution of the Shareholder Value Ratio proxy bank financial distress. We employ a panel probit regression and, given the presence of overlapping data giving rise to residual autocorrelation, we use the Bertschek and Lechner (1998) robust estimator of the covariance matrix of parameters. We show that credit risk, liquidity risk and bank market power are the most influential determinants of distressed Shareholder Value Ratio. Finally we evaluate the model out-sample forecasting performance over the 2008–2009 crisis period.     

Measuring time-varying financial market integration: An unobserved components approach

We measure the time-varying degree of world stock market integration of five developed countries (Germany, France, UK, US, and Japan) over the period 1970:1–2011:10. Time-varying financial market integration of each country is measured through the conditional variances of the country-specific and common international risk premiums in equity excess returns. The country-specific and common risk premiums and their conditional variances are estimated from a latent factor decomposition through the use of state space methods that allow for GARCH errors. Our empirical results suggest that stock market integration has increased over the period 1970:1–2011:10 in all countries but Japan. And while there is a structural increase in stock market integration in four out of five countries, all countries also exhibit several shorter periods of disintegration (reversals), i.e. periods in which country-specific shocks play a more dominant role. Hence, stock market integration is measured as a dynamic process that is fluctuating in the short run while gradually increasing in the long run.     

Correlation dynamics in equity markets: evidence from India

This study is aimed at understanding the correlation dynamics of the equity markets from a developing country perspective using daily data from July 1997 to August 2006. A simple unconditional correlation estimate and dynamic time varying correlation estimate from a DCC-MVGARCH of Engle and Sheppard (2001) are derived for S&P CNX Nifty and other 10 world indices that includes four developed and six Asian country indices. The results show low correlation across S&P CNX Nifty with both Asian and developed nations. In addition a Logistic Smooth Transition Regression (LSTR) model is implemented and finds that the S&P CNX Nifty index is moving towards a better integration with other world markets but not at a very noteworthy phase. The low correlation provides space for the global funds to diversify risk in Indian markets.     

The autumn effect of gold

This paper studies recurring annual events potentially introducing seasonality into gold prices. We analyze gold returns for each month from 1980 to 2010 and find that September and November are the only months with positive and statistically significant gold price changes. This “autumn effect” holds unconditionally and conditional on several risk factors. We argue that the anomaly can be explained with hedging demand by investors in anticipation of the “Halloween effect” in the stock market, wedding season gold jewelery demand in India and negative investor sentiment due to shorter daylight time. The autumn effect can also be characterized by a higher unconditional and conditional volatility than in other seasons.