Risk management in banking is clearly based on criteria of prudence and selectivity, involving several financial variables but also the knowledge of customer’s preferences and behaviours. In this communication, thus, we try to suggest how to manage and reduce bank risk through the identification of risky customers than more virtuous ones. This study investigates, as usual, the risk profile of clients through the use of the profiling questionnaire: the MiFID questionnaire; indeed, the goal of this questionnaire is to identify the experience, the competence and the client's risk profile, but not always its questions can reach the desired results. First we tried to identify the most significant and discriminating items in the MiFID questionnaire, by means classic methods (Markowitz, 1987) as well as innovative analyses (by example, segmentation analysis: Breiman et al., 1984); after that, such items were treated by using weights in order to improve results. Then we determined the level of the most virtuous clients loyalty through statistical methodologies, as Survival Analysis (Collett, 2003; Miller, 1997), also by using mixed techniques “survival trees” (Therneau and Atkinson, 1997). In this way the different types of clients (more loyal and less loyal) are distinguished, in order to find the causes of their differences in loyalty.
Risk management in banking and customer loyalty
Francesco Domenico d'Ovidio
;Najada Firza;Dante Mazzitelli
2018-01-01
Abstract
Risk management in banking is clearly based on criteria of prudence and selectivity, involving several financial variables but also the knowledge of customer’s preferences and behaviours. In this communication, thus, we try to suggest how to manage and reduce bank risk through the identification of risky customers than more virtuous ones. This study investigates, as usual, the risk profile of clients through the use of the profiling questionnaire: the MiFID questionnaire; indeed, the goal of this questionnaire is to identify the experience, the competence and the client's risk profile, but not always its questions can reach the desired results. First we tried to identify the most significant and discriminating items in the MiFID questionnaire, by means classic methods (Markowitz, 1987) as well as innovative analyses (by example, segmentation analysis: Breiman et al., 1984); after that, such items were treated by using weights in order to improve results. Then we determined the level of the most virtuous clients loyalty through statistical methodologies, as Survival Analysis (Collett, 2003; Miller, 1997), also by using mixed techniques “survival trees” (Therneau and Atkinson, 1997). In this way the different types of clients (more loyal and less loyal) are distinguished, in order to find the causes of their differences in loyalty.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.