An empirical approach to residual value risk estimation in automotive leases

Purpose – The purpose of this paper is to examine residual value risk modelling issues with a focus on automotive lease portfolios. Residual value risk is approached through a re‐sampling technique that provides the probability density function of losses and VaR measures for credit portfolios. Design/methodology/approach – The methodology is applied to a portfolio of 37,523 operating leases issued between 1989 and 2001 by a major European financial institution. Findings – The results show that residual value losses are low and sometimes non‐existent. Moreover, the major part of residual value risk is idiosyncratic and can thus be eliminated through adequate diversification. Additionally, this internal model seems to prove that capital requirements stemming from the Basel Committee's proposed new framework are somehow overestimated. Originality/value – This paper advocates determining a more accurate risk weight for residual value risk in order to better reflect this relatively low‐risk part of leasing activities.