Standard Bank, South Africa, currently employs a methodology when developing application or behav... more Standard Bank, South Africa, currently employs a methodology when developing application or behavioural scorecards that involves logistic regression. A key aspect of building logistic regression models entails variable selection which involves dealing with multicollinearity. The objective of this study was to investigate the impact of using different variance inflation factor (VIF) thresholds on the performance of these models in a predictive and discriminatory context and to study the stability of the estimated coefficients in order to advise the bank. The impact of the choice of VIF thresholds was researched by means of an empirical and simulation study. The empirical study involved analysing two large data sets that represent the typical size encountered in a retail credit scoring context. The first analysis concentrated on fitting the various VIF models and comparing the fitted models in terms of the stability of coefficient estimates and goodness-of-fit statistics while the second analysis focused on evaluating the fitted models' predictive ability over time. The simulation study was used to study the effect of multicollinearity in a controlled setting. All the above-mentioned studies indicate that the presence of multicollinearity in large data sets is of much less concern than in small data sets and that the VIF criterion could be relaxed considerably when models are fitted to large data sets. The recommendations in this regard have been accepted and implemented by Standard Bank.
The Basel II regulatory framework significantly increased the resilience of the banking system, b... more The Basel II regulatory framework significantly increased the resilience of the banking system, but proved ineffective in preventing the 2008/9 financial crisis. The subsequent introduction of Basel III aimed, inter alia, to supplement bank capital using buffers. The countercyclical buffer boosts existing minimum capital requirements when systemic risk surges are detected. Bolstering capital in favourable economic conditions cushions losses in unfavourable conditions, thereby addressing capital requirement pro-cyclicality. This paper contains an overview of the countercyclical capital buffer and a critical discussion of its implementation as proposed in Basel III. Consequences of the buffer's introduction for South African banks are explored, and in particular, potential systemic risk indicator variables are identified that may be used by the South African Reserve Bank (SARB) as early warning indicators of imminent systemic financial distress. These indicators may be valuable to the SARB since they can use it in their decision-making process regarding the build-up and release of the countercyclical buffer for South African banks.
Many banks use the loss distribution approach in their advanced measurement models to estimate re... more Many banks use the loss distribution approach in their advanced measurement models to estimate regulatory or economic capital. This boils down to estimating the 99.9% VaR of the aggregate loss distribution and is notoriously difficult to do accurately. Also, it is well-known that the accuracy with which the tail of the loss severity distribution is estimated is the most important driver in determining a reasonable estimate of regulatory capital. To this end, banks use internal data and external data (jointly referred to as historical data) as well as scenario assessments in their endeavour to improve the accuracy with which the severity distribution is estimated. In this paper we propose a simple new method whereby the severity distribution may be estimated using historical data and experts’ scenario assessments jointly. The way in which historical data and scenario assessments are integrated incorporates measures of agreement between these data sources, which can be used to evaluate the quality of both. In particular we show that the procedure has definite advantages over traditional methods where the severity distribution is modelled and fitted separately for the body and tail parts, with the body part based only on historical data and the tail part on scenario assessments.
GARCH models are useful to estimate the volatility of financial return series. Historically the ... more GARCH models are useful to estimate the volatility of financial return series. Historically the innovation distribution of a GARCH model was assumed to be standard normal but recent research emphasizes the need for more general distributions allowing both asymmetry (skewness) and kurtosis in the innovation distribution to obtain better fitting models. A number of authors have proposed models which are special cases of the class of normal mean variance mixtures. We introduce a general framework within which this class of innovation distributions may be discussed. This entails writing the innovation term as a standardised combination of two variables, namely a normally distributed term and a mixing variable, each with its own interpretation. We list the existing models that fit into this framework and compare the corresponding innovation distributions, finding that they tend to be quite similar. This is confirmed by an empirical illustration which fits the models to the monthly excess returns series of the US stocks. The illustration finds further support for the ICAPM model of Merton, thus supporting recent results of Lanne and Saikonnen (2006).
Standard Bank, South Africa, currently employs a methodology when developing application or behav... more Standard Bank, South Africa, currently employs a methodology when developing application or behavioural scorecards that involves logistic regression. A key aspect of building logistic regression models entails variable selection which involves dealing with multicollinearity. The objective of this study was to investigate the impact of using different variance inflation factor (VIF) thresholds on the performance of these models in a predictive and discriminatory context and to study the stability of the estimated coefficients in order to advise the bank. The impact of the choice of VIF thresholds was researched by means of an empirical and simulation study. The empirical study involved analysing two large data sets that represent the typical size encountered in a retail credit scoring context. The first analysis concentrated on fitting the various VIF models and comparing the fitted models in terms of the stability of coefficient estimates and goodness-of-fit statistics while the second analysis focused on evaluating the fitted models' predictive ability over time. The simulation study was used to study the effect of multicollinearity in a controlled setting. All the above-mentioned studies indicate that the presence of multicollinearity in large data sets is of much less concern than in small data sets and that the VIF criterion could be relaxed considerably when models are fitted to large data sets. The recommendations in this regard have been accepted and implemented by Standard Bank.
The Basel II regulatory framework significantly increased the resilience of the banking system, b... more The Basel II regulatory framework significantly increased the resilience of the banking system, but proved ineffective in preventing the 2008/9 financial crisis. The subsequent introduction of Basel III aimed, inter alia, to supplement bank capital using buffers. The countercyclical buffer boosts existing minimum capital requirements when systemic risk surges are detected. Bolstering capital in favourable economic conditions cushions losses in unfavourable conditions, thereby addressing capital requirement pro-cyclicality. This paper contains an overview of the countercyclical capital buffer and a critical discussion of its implementation as proposed in Basel III. Consequences of the buffer's introduction for South African banks are explored, and in particular, potential systemic risk indicator variables are identified that may be used by the South African Reserve Bank (SARB) as early warning indicators of imminent systemic financial distress. These indicators may be valuable to the SARB since they can use it in their decision-making process regarding the build-up and release of the countercyclical buffer for South African banks.
Many banks use the loss distribution approach in their advanced measurement models to estimate re... more Many banks use the loss distribution approach in their advanced measurement models to estimate regulatory or economic capital. This boils down to estimating the 99.9% VaR of the aggregate loss distribution and is notoriously difficult to do accurately. Also, it is well-known that the accuracy with which the tail of the loss severity distribution is estimated is the most important driver in determining a reasonable estimate of regulatory capital. To this end, banks use internal data and external data (jointly referred to as historical data) as well as scenario assessments in their endeavour to improve the accuracy with which the severity distribution is estimated. In this paper we propose a simple new method whereby the severity distribution may be estimated using historical data and experts’ scenario assessments jointly. The way in which historical data and scenario assessments are integrated incorporates measures of agreement between these data sources, which can be used to evaluate the quality of both. In particular we show that the procedure has definite advantages over traditional methods where the severity distribution is modelled and fitted separately for the body and tail parts, with the body part based only on historical data and the tail part on scenario assessments.
GARCH models are useful to estimate the volatility of financial return series. Historically the ... more GARCH models are useful to estimate the volatility of financial return series. Historically the innovation distribution of a GARCH model was assumed to be standard normal but recent research emphasizes the need for more general distributions allowing both asymmetry (skewness) and kurtosis in the innovation distribution to obtain better fitting models. A number of authors have proposed models which are special cases of the class of normal mean variance mixtures. We introduce a general framework within which this class of innovation distributions may be discussed. This entails writing the innovation term as a standardised combination of two variables, namely a normally distributed term and a mixing variable, each with its own interpretation. We list the existing models that fit into this framework and compare the corresponding innovation distributions, finding that they tend to be quite similar. This is confirmed by an empirical illustration which fits the models to the monthly excess returns series of the US stocks. The illustration finds further support for the ICAPM model of Merton, thus supporting recent results of Lanne and Saikonnen (2006).
Uploads
Papers by Riaan De Jongh