Model risk and capital reserves

Kerkhof, Jeroen; Melenberg, Bertrand; Schumacher, Johannes

doi:10.1016/j.jbankfin.2009.07.025

Cited by 94 publications

(79 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Therefore, if the assumption of stationarity is not met, the model of Vasicek may offer misleading capital estimates even if the hazard rate series are assumed to follow a simple linear dynamic of constant coefficients and normally distributed innovations. It can be concluded that the non-stationarity generates model risk, Kerkhof et al (2010). In comparing the different time windows using the proposed approach, Figure 4 and Figure 5 show a displacement to the right of the loss densities generated by Π D and Π F with respect to those generated by Π C and Π I , respectively.…”

Section: Risk Unitmentioning

confidence: 90%

A New Approach to the Unconditional Measurement of Default Risk

2014

View full text Add to dashboard Cite

This paper analyzes the unconditional measurement of default risk and proposes an alternative modeling approach. We begin the analysis by showing that when conducted under non-stationarity, the objective of the unconditional measurement changes and that some relevant problems appear as a consequence of the sample dependence. Based on this result, we introduce our approach and discuss its consistency, practical advantages, and the main differences from the conventional static framework. An empirical analysis is also conducted. Under nonstationarity, the regulatory model for the unconditional probability of default distribution performs badly when compared to our approach. Results also show that the capital figure presents a determinant and nontrivial dependence on the homogeneity and severity of the economic scenario represented in the sample. We begin the analysis by showing that when conducted under non-stationarity, the objective of the unconditional measurement changes and that some relevant problems appear as a consequence of the sample dependence. Based on this result, we introduce our approach and discuss its consistency, practical advantages, and the main differences from the conventional static framework. An empirical analysis is also conducted. Under non-stationarity, the regulatory model for the unconditional probability of default distribution performs badly when compared to our approach. Results also show that the capital figure presents a determinant and non-trivial dependence on the homogeneity and severity of the economic scenario represented in the sample.

show abstract

Section: Risk Unitmentioning

confidence: 90%

A New Approach to the Unconditional Measurement of Default Risk

2014

View full text Add to dashboard Cite

show abstract

“…For reference, the mentioned rolling-window approach for a symmetric GARCH model would produce an average autoregressive coecient of 0.869 with a standard deviation of 0.073 and a spike at 0.346, whereas the growing sample approach delivered an average coecient of 0.920 with a standard deviation of 0.002. We believe that is the main reason for the poor performance of the GARCH models in, e.g., Kerkhof et al (2009) andBao et al (2006). Hence, in this study we do not include rolling window versions of the GARCH or CAViaR models, relying on the results of similar models employed by other authors for reference.…”

Section: Backtesting and Multiple Hypothesis Testingmentioning

confidence: 92%

“…Bao et al (2006), Kerkhof et al (2009), using the sample measures of the empiril overge proility (ECP) ρ :=…”

Section: Backtesting and Multiple Hypothesis Testingmentioning

confidence: 99%

Multiple Hypothesis Testing of Market Risk Forecasting Models

Esposito

Cummins

2015

SSRN Journal

View full text Add to dashboard Cite

Extending previous risk model backtesting literature, we construct multiple hypothesis testing (MHT) with the stationary bootstrap. We conduct multiple tests which control for the generalized condence level and employ the bootstrap MHT to design multiple comparison testing. We consider absolute and relative predictive ability to test a range of competing risk models, focusing on Value-atRisk (VaR) and Expected Shortfall (ExS). In devising the test for the absolute predictive ability, we take the route of recent literature and construct balanced simultaneous condence sets that control for the generalized family-wise error rate, which is the joint probability of rejecting true hypotheses. We implement a step-down method which increases the power of the MHT in isolating false discoveries.In testing for the ExS model predictive ability, we design a new simple test to draw inference about recursive model forecasting capability. In the second suite of statistical testing, we develop a novel device for measuring the relative predictive ability in the bootstrap MHT framework. The device, we coin multiple comparison mapping, provides a statistically robust instrument designed to answer the question: which model is the best model?.

show abstract

“…However, it has been argued that parametric methods often entail unavoidable misspecification risk and estimation risk (Kerkhof et al, 2010). The model risk stems from the inconsistency between the assumptions underlying the probability model of asset returns and the realized asset returns (Tsay, 2010), while estimation risk comes from the uncertainty of parameter estimation (Talay and Zheng, 2002).…”

Section: Datastreammentioning

confidence: 99%

Downside Risk and Stock Returns: An Empirical Analysis of the Long-Run and Short-Run Dynamics from the G-7 Countries

2016

View full text Add to dashboard Cite

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractThis paper presents a fractionally cointegrated vector autoregression (FCVAR) model to examine various relations between stock returns and downside risk. Evidence from major advanced markets supports the notion that downside risk measured by value-at-risk (VaR) has significant information content that reflects lagged long-run variance and higher moments of risk for predicting stock returns. The evidence supports the positive tradeoff hypothesis and the leverage effect in the long run and for some markets in the short run. We find that US downside risk accounts for 54.36% of price discovery, whereas the own effect from the country itself contributes only 27.06%.JEL classification: G11, G12, G15, C24, F30

show abstract

Model risk and capital reserves

Cited by 94 publications

References 30 publications

A New Approach to the Unconditional Measurement of Default Risk

A New Approach to the Unconditional Measurement of Default Risk

Multiple Hypothesis Testing of Market Risk Forecasting Models

Downside Risk and Stock Returns: An Empirical Analysis of the Long-Run and Short-Run Dynamics from the G-7 Countries

Contact Info

Product

Resources

About