Katarzyna Bijak

Does segmentation always improve model performance in credit scoring

Highlights? Segmentation does not always improve model performance in credit scoring. ? There is no difference in performance between the two-step and simultaneous approaches. ? Segmentation contribution to the model performance can be up to 20 percentage points. ? It is recommended to develop a single-scorecard model for comparison purposes. Credit scoring allows for the credit risk assessment of bank customers. A single scoring model (scorecard) can be developed for the entire customer population, e.g. using logistic regression. However, it is often expected that segmentation, i.e. dividing the population into several groups and building separate scorecards for them, will improve the model performance. The most common statistical methods for segmentation are the two-step approaches, where logistic regression follows Classification and Regression Trees (CART) or Chi-squared Automatic Interaction Detection (CHAID) trees etc. In this research, the two-step approaches are applied as well as a new, simultaneous method, in which both segmentation and scorecards are optimised at the same time: Logistic Trees with Unbiased Selection (LOTUS). For reference purposes, a single-scorecard model is used. The above-mentioned methods are applied to the data provided by two of the major UK banks and one of the European credit bureaus. The model performance measures are then compared to examine whether there is improvement due to the segmentation methods used. It is found that segmentation does not always improve model performance in credit scoring: for none of the analysed real-world datasets, the multi-scorecard models perform considerably better than the single-scorecard ones. Moreover, in this application, there is no difference in performance between the two-step and simultaneous approaches.

Modelling LGD for Unsecured Retail Loans Using Bayesian Methods

Loss Given Default (LGD) is the loss borne by the bank when a customer defaults on a loan. LGD for unsecured retail loans is often found difficult to model. In the frequentist (non-Bayesian) two-step approach, two separate regression models are estimated independently, which can be considered potentially problematic when trying to combine them to make predictions about LGD. The result is a point estimate of LGD for each loan. Alternatively, LGD can be modelled using Bayesian methods. In the Bayesian framework, one can build a single, hierarchical model instead of two separate ones, which makes this a more coherent approach. In this paper, Bayesian methods as well as the frequentist approach are applied to the data on personal loans provided by a large UK bank. As expected, the posterior means of parameters that have been produced using Bayesian methods are very similar to the frequentist estimates. The most important advantage of the Bayesian model is that it generates an individual predictive distribution of LGD for each loan. Potential applications of such distributions include the downturn LGD and the stressed LGD under Basel II.

Kalman filtering as a performance monitoring technique for a propensity scorecard

Propensity scorecards allow forecasting, which bank customers would like to be granted new credits in the near future, through assessing their willingness to apply for new loans. Kalman filtering can help to monitor scorecard performance. Data from successive months are used to update the baseline model. The updated scorecard is the output of the Kalman filter. There is no assumption concerning the scoring model specification and no specific estimation method is presupposed. Thus, the estimator covariance is derived from the bootstrap. The focus is on a relationship between the score and the natural logarithm of the odds for that score, which is used to determine a customers propensity level. The propensity levels corresponding to the baseline and updated scores are compared. That comparison allows for monitoring whether the scorecard is still up-to-date in terms of assigning the odds. The presented technique is illustrated with an example of a propensity scorecard developed on the basis of credit bureau data.

Operational research techniques applied throughout cancer care services: a review

Abstract Cancer is a disease affecting increasing numbers of people. In the UK, the proportion of people affected by cancer is projected to increase from 1 in 3 in 1992, to nearly 1 in 2 by 2020. Health services to tackle cancer can be grouped broadly into prevention, diagnosis, staging, and treatment. We review examples of Operational Research (OR) papers addressing decisions encountered in each of these areas. In conclusion, we find many examples of OR research on screening strategies, as well as on treatment planning and scheduling. On the other hand, our search strategy uncovered comparatively few examples of OR models applied to reducing cancer risks, optimising diagnostic procedures, and staging. Improvements to cancer care services have been made as a result of successful OR modelling. There is potential for closer working with clinicians to enable the impact of other OR studies to be of greater benefit to cancer sufferers.

Dynamic affordability assessment: predicting an applicant’s ability to repay over the life of the loan

In the credit decision-making process, both an applicant’s creditworthiness and their affordability should be assessed. While credit scoring focuses on creditworthiness, affordability is often checked on the basis of current income and estimated current consumption as well as existing debts stated in a credit report. Contrary to that static approach, a theoretical framework for dynamic affordability assessment is proposed in this paper. In this approach, both income and consumption are allowed to vary over time and their changes are described with random effects models for panel data. The models are derived from the economic literature, including the Euler equation of consumption. A simulation is run on their basis and predicted time series are generated for a given applicant. For each pair of the predicted income and consumption time series, the applicant’s ability to repay is checked over the life of the loan, for all possible installment amounts. As a result, a probability of default is assigned to each amount, which can help find the maximum affordable installment. This is illustrated with an example based on artificial data. Assessing affordability over the loan repayment period as well as taking into account variability of income and expenditure over time are in line with recommendations of the UK Office of Fair Trading and the Financial Services Authority. In practice, the suggested approach could contribute to responsible lending.