
on Risk Management 
By:  Takaaki Koike; Marius Hofert 
Abstract:  We propose a novel framework of estimating systemic risk measures and risk allocations based on a Markov chain Monte Carlo (MCMC) method. We consider a class of allocations whose jth component can be written as some risk measure of the jth conditional marginal loss distribution given the socalled crisis event. By considering a crisis event as an intersection of linear constraints, this class of allocations covers, for example, conditional ValueatRisk (CoVaR), conditional expected shortfall (CoES), VaR contributions, and range VaR (RVaR) contributions as special cases. For this class of allocations, analytical calculations are rarely available, and numerical computations based on Monte Carlo methods often provide inefficient estimates due to the rareevent character of crisis events. We propose an MCMC estimator constructed from a sample path of a Markov chain whose stationary distribution is the conditional distribution given the crisis event. Efficient constructions of Markov chains, such as Hamiltonian Monte Carlo and Gibbs sampler, are suggested and studied depending on the crisis event and the underlying loss distribution. Efficiency of the MCMC estimators are demonstrated in a series of numerical experiments. 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:arx:papers:1909.11794&r=all 
By:  Álvaro Chamizo (BBVA.); Alfonso Novales (Instituto Complutense de Análisis Económico (ICAE), and Department of Economic Analysis, Facultad de Ciencias Económicas y Empresariales, Universidad Complutense, 28223 Madrid, Spain.) 
Abstract:  We provide a methodology to estimate a Global Credit Risk Factor (GCRF) from CDS spreads using the information provided by the defaultrelated component of observed spreads. These are previ ously estimated using Pan and Singleton (2008) methodology. The estimated factor contains higher explanatory power on CDS spread fluctuations across sectors than standard credit indices like iTraxx or CDX. We find a positive association between GCRF and implied volatility variables, and a negative association with MSCI stock market sector indices as well as with interest rates and with the slope and the curvature of the term structure. Such correlations provide useful insights for risk management as well as for the hedging of credit portfolios. Indeed, we present a synthetic factor regression model for GCRF that we apply in a stress testing methodology for credit portfolios as well as to evaluate future credit risk scenarios. Finally, we show evidence suggesting that the exposure to systemic credit risk was priced in the market during the 20062015 period. 
Keywords:  Credit Risk; Systemic Risk; Idiosyncratic Risk; Stress Tests; Factor Models; Market Pricing. 
JEL:  E44 F34 G01 G11 G23 G32 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:ucm:doicae:1927&r=all 
By:  Mo Zheng; HanSuck Song; Fredrik Armerin 
Abstract:  The Swedish real estate sector index services as a guideline and general indicator of the performance of the real estate industry, events such global financial crisis, the European debt crisis cause unprecedentedly high uncertainty and volatility. There is a need to understand and search for an appropriate method to deal with Valueatrisk and expected shortfall which comprehensively covers unconditional and conditional risk models with various estimation window and significance level. In this paper, we apply nonparametric, parametric and semiparametric methods such as historical simulation and filtered historical simulation; RiskMetrics; GARCHtype models and GARCH together with extreme value theory. The result is still in process, the draft paper will be submitted within the deadline. 
Keywords:  Expected Shortfall; Extreme Value Theory; Risk Management; Swedish Real Estate Sector; ValueatRisk 
JEL:  R3 
Date:  2019–01–01 
URL:  http://d.repec.org/n?u=RePEc:arz:wpaper:eres2019_366&r=all 
By:  Alfonso Novales (Instituto Complutense de Análisis Económico (ICAE), and Department of Economic Analysis, Facultad de Ciencias Económicas y Empresariales, Universidad Complutense, 28223 Madrid, Spain.); Laura GarciaJorcano (Department of Economic Analysis and Finance (Area of Financial Economics), Facultad de Ciencias Jurídicas y Sociales Universidad de CastillaLa Mancha, Toledo, Spain.) 
Abstract:  We use stock market data to analyze the quality of alternative models and procedures for fore casting expected shortfall (ES) at different significance levels. We compute ES forecasts from conditional models applied to the full distribution of returns as well as from models that focus on tail events using extreme value theory (EVT). We also apply the semiparametric filtered historical simulation (FHS) approach to ES forecasting to obtain 10day ES forecasts. At the 10day hori zon we also combine FHS with EVT. The performance of the different models is assessed using six different ES backtests recently proposed in the literature. Our results suggest that conditional EVTbased models produce more accurate 1day and 10day ES forecasts than do nonEVT based models. Under either approach, asymmetric probability distributions for return innovations tend to produce better forecasts. Incorporating EVT in parametric or semiparametric approaches also improves ES forecasting performance. These qualitative results are also valid for the recent crisis period, even though all models then underestimate the level of risk. FHS narrows the range of numerical forecasts obtained from alternative models, thereby reducing model risk. Combining EVT and FHS seems to be best approach for obtaining accurate ES forecasts. 
Keywords:  Extreme value theory; Skewed distributions; Expected shortfall; Backtesting; Filtered historical simulation. 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:ucm:doicae:1924&r=all 
By:  Samudra Dasgupta; Arnab Banerjee 
Abstract:  The 2008 mortgage crisis is an example of an extreme event. Extreme value theory tries to estimate such tail risks. Modern finance practitioners prefer Expected Shortfall based risk metrics (which capture tail risk) over traditional approaches like volatility or even ValueatRisk. This paper provides a quantum annealing algorithm in QUBO form for a dynamic asset allocation problem using expected shortfall constraint. It was motivated by the need to refine the current quantum algorithms for Markowitz type problems which are academically interesting but not useful for practitioners. The algorithm is dynamic and the risk target emerges naturally from the market volatility. Moreover, it avoids complicated statistics like generalized pareto distribution. It translates the problem into qubit form suitable for implementation by a quantum annealer like DWave. Such QUBO algorithms are expected to be solved faster using quantum annealing systems than any classical algorithm using classical computer (but yet to be demonstrated at scale). 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:arx:papers:1909.12904&r=all 
By:  Álvaro Chamizo (BBVA.); Alexandre Fonollosa (BBVA.); Alfonso Novales (Instituto Complutense de Análisis Económico (ICAE), and Department of Economic Analysis, Facultad de Ciencias Económicas y Empresariales, Universidad Complutense, 28223 Madrid, Spain.) 
Abstract:  We analyze whether the credit market anticipated the financial crisis before the regulators using a methodology that combines the Merton model for the determination of economic capital with Vasicek’s factor model for asset correlation. Contrary to standard practice, we estimate the credit value at risk (VaR) and expected shortfall (ES) of a global loan portfolio using CDS spreads because credit derivat ives incorporate forwardlooking information on future systemic shocks that might be essential in the estimation of economic capital. We find that onefactor model can generally be a good representation of correlations in the credit market because of the high intersector correlations, although an appro priately chosen second factor can provide additional information for risk estimation in stressed times. We show that there were, indeed, signs of stress in the credit market that were not incorporated in the determination of economic capital during the crisis and that some financial institutions did not con sider properly. The overall impression is that it is not so much that risk models were oversimplified to anticipate the financial crisis but rather, that they were backwardlooking. A potential implication of our research is that the level of regulatory capital should react to events in the credit market. 
Keywords:  Forwardlooking Asset Correlation; Economic Capital; Asset Allocation; Systemic Risk. 
JEL:  E47 G01 G28 G32 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:ucm:doicae:1925&r=all 
By:  Álvaro Chamizo (BBVA.); Alfonso Novales (Instituto Complutense de Análisis Económico (ICAE), and Department of Economic Analysis, Facultad de Ciencias Económicas y Empresariales, Universidad Complutense, 28223 Madrid, Spain.) 
Abstract:  Hedging a credit portfolio using single name CDS is affected by high spread volatility that induces continuous changes in a portfolio mark to market, which is a nuisance. Often, the problem is that CDS on firms in the portfolio are not being traded. To get around that, a derivative portfolio can be hedged by taking a contrary position in a credit index, and we examine in this paper the efficiency of such an imperfect hedge. We find over the 20072012 period an 80% hedging efficiency for a European portfolio, 60% for North American and Japanese portfolios, and around 70% for a global portfolio, as measured by the reduction in marktomarket variance. We also consider sectorial credit portfolios for Europe and North America, for which hedging efficiency is not as high, due to their more import ant idiosyncratic component. Taking into account the quality of the credit counterpart improves the effectiveness of the hedge, although it requires using less liquid credit indices, with higher transaction costs. Standard conditional volatility models provide similar results to the least squares hedge, except for extreme market movements. An efficient hedge for a credit portfolio made up of the most idiosyn cratic firms would seem to require more than 50 firms, while the hedge for portfolios made up of the less idiosyncratic firms achieves high efficiency even for a small number of firms. The efficiency of the hedge is higher when portfolio volatility is high and also when short term interest rates or exchange rate volatility are high. Increases in VIX, in the 10year swap rate or in liquidity risk tend to decrease hedging efficiency. Credit indices offer a moderately efficient hedge for corporate bond portfolios, which we have examined with a reduced sample of firms over 20062018. This analysis also shows that the current efficiency of a credit index hedge has recovered at precrisis levels. 
Keywords:  Market Risk; CDS; Credit Indices; Credit Hedge; Asset Allocation; Systemic Risk. 
JEL:  G01 G12 G13 G14 G15 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:ucm:doicae:1928&r=all 
By:  Laura GarciaJorcano (Department of Economic Analysis and Finance (Area of Financial Economics), Facultad de Ciencias Jurídicas y Sociales Universidad de CastillaLa Mancha, Toledo, Spain.); Alfonso Novales (Instituto Complutense de Análisis Económico (ICAE), and Department of Economic Analysis, Facultad de Ciencias Económicas y Empresariales, Universidad Complutense, 28223 Madrid, Spain.) 
Abstract:  We introduce three dominance criteria to compare the performance of alternative VaR forecasting models. The three criteria use the information provided by a battery of VaR validation tests based on the frequency and size of exceedances, offering the possibility of efficiently summarizing a large amount of statistical information. They do not require the use of any loss function defined on the difference between VaR forecasts and observed returns, and two of the criteria are not conditioned on any significance level for the VaR tests. We use them to explore the potential for 1day ahead VaR forecasting of some recently proposed asymmetric probability distributions for return innovations, as well as to compare the APARCH and FGARCH volatility specifications with more standard alternatives. Using 19 assets of different nature, the three criteria lead to similar conclusions, suggesting that the unbounded Johnson SU, the skewed Studentt and the skewed Generalizedt distributions seem to produce the best VaR forecasts. The added flexibility of a free power parameter in the conditional volatility in the APARCH and FGARCH models leads to a better fit to return data, but it does not improve upon the VaR forecasts provided by GARCH and GJRGARCH volatilities. 
Keywords:  Value at risk; Backtesting; Forecast evaluation; Dominance; Conditional volatility models; Asymmetric distributions. 
JEL:  C52 C58 G17 G32 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:ucm:doicae:1923&r=all 
By:  Marcel Bräutigam (LabEx MMEDII  UCP  Université de Cergy Pontoise  Université ParisSeine, LPSM UMR 8001  Laboratoire de Probabilités, Statistique et Modélisation  UPD7  Université Paris Diderot  Paris 7  SU  Sorbonne Université  CNRS  Centre National de la Recherche Scientifique, CREAR  Center of Research in Econofinance and Actuarial sciences on Risk / Centre de Recherche Econofinancière et Actuarielle sur le Risque  Essec Business School); Marie Kratz (SID  Information Systems, Decision Sciences and Statistics Department  Essec Business School, LabEx MMEDII  UCP  Université de Cergy Pontoise  Université ParisSeine, CREAR  Center of Research in Econofinance and Actuarial sciences on Risk / Centre de Recherche Econofinancière et Actuarielle sur le Risque  Essec Business School) 
Abstract:  In this study, we derive the joint asymptotic distributions of functionals of quantile estimators (the nonparametric sample quantile and the parametric locationscale quantile) and functionals of measure of dispersion estimators (the sample standard deviation, sample mean absolute deviation, sample median absolute deviation)  assuming an underlying identically and independently distributed sample. Additionally, for locationscale distributions, we show that asymptotic correlations of such functionals do not depend on the mean and variance parameter of the distribution. Further, we compare the impact of the choice of the quantile estimator (sample quantile vs. parametric locationscale quantile) in terms of speed of convergence of the asymptotic covariance and correlations respectively. As application, we show in simulations a good finite sample performance of the asymptotics. Further, we show how the theoretical dependence results can be applied to the most wellknown risk measures (ValueatRisk, Expected Shortfall, expectile). Finally, we relate the theoretical results to empirical findings in the literature of the dependence between risk measure prediction (on historical samples) and the estimated volatility. 
Keywords:  asymptotic distribution,sample quantile,measure of dispersion,nonlinear dependence,VaR,ES,correlation 
Date:  2018–12 
URL:  http://d.repec.org/n?u=RePEc:hal:wpaper:hal02296832&r=all 
By:  Laura GarciaJorcano (Department of Economic Analysis and Finance (Area of Financial Economics), Facultad de Ciencias Jurídicas y Sociales, Universidad de CastillaLa Mancha, Toledo, Spain.); Alfonso Novales (Instituto Complutense de Análisis Económico (ICAE), and Department of Economic Analysis, Facultad de Ciencias Económicas y Empresariales, Universidad Complutense, 28223 Madrid, Spain.) 
Abstract:  We provide evidence suggesting that the assumption on the probability distribution for return in novations is more influential for Value at Risk (VaR) performance than the conditional volatility specification. We also show that some recently proposed asymmetric probability distributions and the APARCH and FGARCH volatility specifications beat more standard alternatives for VaR fore casting, and they should be preferred when estimating tail risk. The flexibility of the free power parameter in conditional volatility in the APARCH and FGARCH models explains their better performance. Indeed, our estimates suggest that for a number of financial assets, the dynamics of volatility should be specified in terms of the conditional standard deviation. We draw our results on VaR forecasting performance from i) a variety of backtesting approaches, ii) the Model Confi dence Set approach, as well as iii) establishing a ranking among alternative VaR models using a precedence criterion that we introduce in this paper. 
Keywords:  Valueatrisk; Backtesting; Evaluating forecasts; Precedence; APARCH model; Asym metric distributions. 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:ucm:doicae:1926&r=all 
By:  John Armstrong; Cristin Buescu 
Abstract:  We study the optimal management of a collectivised pension fund, where all investors agree that the assets of deceased members are shared among the survivors. We find that for realistic parameters based on the UK pensions market, a collectivised fund achieves an approximately 20% better return than either an annuity or a personal investment fund. We introduce models of investor preferences over a stream of pension payments in the presence of mortality, incorporating a new concept of adequacy. We find that for riskaverse individuals, pension adequacy plays an important role in determining the optimal fund management strategy. A key issue in the design of collective funds is how to ensure the fund treats all investors fairly. This is a trivial problem in the case that all investors have identical preferences, wealth and mortality, but becomes challenging for heterogeneous funds. We give a strategy for the management of heterogeneous funds in complete markets and prove that it is asymptotically optimal in the absence of systematic longevity risk. 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:arx:papers:1909.12730&r=all 
By:  Riza Demirer (Department of Economics and Finance, Southern Illinois University Edwardsville, Edwardsville, IL 620261102, USA); Konstantinos Gkillas (Department of Business Administration, University of Patras – University Campus, Rio, P.O. Box 1391, 26500 Patras, Greece); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, 0002, South Africa); Christian Pierdzioch (Department of Economics, Helmut Schmidt University, Holstenhofweg 85, P.O.B. 700822, 22008 Hamburg, Germany) 
Abstract:  We analyze the predictive power of timevarying risk aversion for the realized volatility of crude oil returns based on highfrequency data. While the popular linear heterogeneous autoregressive realized volatility (HARRV) model fails to recognize the predictive power of risk aversion over crude oil volatility, we find that risk aversion indeed improves forecast accuracy at all forecast horizons when we compute forecasts by means of random forests. The predictive power of risk aversion is robust to various covariates including realized skewness and realized kurtosis, various measures of jump intensity and leverage. The findings highlight the importance of accounting for nonlinearity in the datagenerating process for forecast accuracy as well as the predictive power of noncashflow factors over commoditymarket uncertainty with significant implications for the pricing and forecasting in these markets. 
Keywords:  Oil price, Realized volatility, Risk aversion, Random forests 
JEL:  G17 Q02 Q47 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:pre:wpaper:201972&r=all 
By:  Makoto Nakajima (Federal Reserve Bank of Philadelphia) 
Abstract:  In this paper, we systematically analyze several labor income definitions by drawing on PSID data, and estimate how the volatility and skewness of income shocks moves with the aggregate activity. We find that volatility is countercyclical, with individual and joint (head+wife) labor incomes being the most volatile, and with hourly wages and postgovernment joint labor income being the least ones. This suggests that \textit{(i)} intrafamily insurance is limited, \textit{(ii)} taxes and transfers remove a large portion of fluctuations in risk, and that \textit{(iii)} hours, not wages, make individual earnings fluctuate over the cycle. By reestimating the volatility of earnings shocks on the ``young'' (ages 2339) and ``old'' (ages 4060) subsamples separately, we find that nearly all the countercyclicality of shocks comes from the young workers, the old subsample exhibits quantitatively muted fluctuations in risk. We then allow for timevarying skewness, and find that the probability of large negative events increases in recessions by more than the probability of large positive events. Taxes and transfers reduce the probability of tail events by a factor of 2 to 3 as compared to other income definitions considered. 
Date:  2019 
URL:  http://d.repec.org/n?u=RePEc:red:sed019:1233&r=all 
By:  Peter Carr; Liuren Wu; Zhibai Zhang 
Abstract:  In this paper we formulate a regression problem to predict realized volatility by using option price data and enhance VIXstyled volatility indices' predictability and liquidity. We test algorithms including regularized regression and machine learning methods such as Feedforward Neural Networks (FNN) on S&P 500 Index and its option data. By conducting a time series validation we find that both Ridge regression and FNN can improve volatility indexing with higher prediction performance and fewer options required. The best approach found is to predict the difference between the realized volatility and the VIXstyled index's prediction rather than to predict the realized volatility directly, representing a successful combination of human learning and machine learning. We also discuss suitability of different regression algorithms for volatility indexing and applications of our findings. 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:arx:papers:1909.10035&r=all 
By:  J. Beleza Sousa; Manuel L. Esquível; Raquel M. Gaspar 
Abstract:  Due to bond prices pulltopar, zero coupon bonds historical returnsare not stationary, as they tend to zero as time to maturity approaches. Given that the historical simulation method for computing Value at Risk(VaR) requires a stationary sequence of historical returns, zero couponbonds historical returns can not be used to compute VaR by historical simulation. Their use would systematically overestimate VaR, resultingin invalid VaR sequences. In this paper we propose an adjustment of zero coupon bonds historical returns. We call the adjusted returns “pulledtopar” returns. We prove that when the zero coupon bonds continuously compounded yields to maturity are stationary the adjusted pulledtoparreturns allow VaR computation by historical simulation. We first illustrate the VaR computation in a simulation scenario,then we apply it to realdata on euro zone STRIPS. 
Date:  2019–09 
URL:  http://d.repec.org/n?u=RePEc:ise:remwps:wp0932019&r=all 
By:  Gopalakrishnan, Balagopal; Mohapatra, Sanket 
Abstract:  One of the arguments often advanced for implementing a stronger insolvency and bankruptcy framework is that it enhances credit discipline among firms. Using a large crosscountry firmlevel dataset, we empirically test whether a stronger insolvency regime reduces firms' likelihood of defaulting on their debt. In particular, we examine whether it reduces default risk during increased economic uncertainty and various external shocks. Our results confirm that a stronger insolvency regime moderates the adverse effects of economic shocks on firms' default risk. The effects are more pronounced for firms in the top half of the size distribution. We also explore channels through which improved creditor rights influence firms’ default risk, including dependence on external finance, corporate leverage, and managerial ethics. Our main results are robust to an alternative measure of default risk, inclusion of currency and sovereign debt crisis episodes, and alternative estimations. 
Keywords:  Insolvency; Bankruptcy; Default risk; Economic policy uncertainty; Sovereign debt crisis; Currency crisis 
JEL:  G30 G32 G33 
Date:  2019–10–01 
URL:  http://d.repec.org/n?u=RePEc:pra:mprapa:96283&r=all 
By:  Andreas Kindt 
Abstract:  With the increasing digitization and big data, automated valuation models (AVMs) are becoming increasingly important within the real estate industry (internationally). The further potential for the use of AVM seems enormous. However, the mainstream AVM research is hitherto largely onedimensional and requires a wider and focus. Especially stakeholders with situationalrecurrent (e.g. mortgage lending, transaction, etc.) or regular valuation tasks (e.g., risk management, performance analysis, etc.), have the need for individual AVM solutions. An efficient access to the subject is still difficult because of the high complexity. Thus, the respective stakeholders have an increased interest in systematic and integrated decision support. Working on this point, the dissertation tries to give guidance for the optimal development and application of AVMs. 
Keywords:  Automated Valuation Models; AVM; Big data; Digitalization; Property Valuation 
JEL:  R3 
Date:  2019–01–01 
URL:  http://d.repec.org/n?u=RePEc:arz:wpaper:eres2019_240&r=all 
By:  Kyle Dempsey (The Ohio State University); Felicia Ionescu (Federal Reserve Board) 
Abstract:  How much do changes in credit supply affect consumers’ ability to insure against income risk over the business cycle and what is the valuation of such insurance? Using loanlevel data from the Senior Loan Officer Opinion Survey (SLOOS), we construct measures of key credit supply variables, such as lending standards and terms for consumer credit in the U.S. and build a heterogeneous model of unsecured credit and default that accounts for credit supply dynamics as estimated from these data. Our economy is quantitatively consistent with key features of the unsecured credit market, earnings dynamics, and measures of consumption volatility in the U.S. We find that variability in standards and terms for credit is welfare improving despite the loss in consumption insurance that such an environment may induce. The key mechanism behind this result is the asymmetric effect that changes in standards induce for loan pricing in good and bad states of the economy. 
Date:  2019 
URL:  http://d.repec.org/n?u=RePEc:red:sed019:1428&r=all 