Optimization of a loan portfolio

The strategy of a large international bank in Indonesia was to increase its market share in the Small, & Medium Size Enterprise (SME) market. To realize this growth the Bank has purchased a local Indonesian bank and has now expanded its number of branches to around 100 in the Indonesian market and significantly increased its loan portfolio. The local bank focused mostly on the upper tier SME market segment. The local bank had a conservative collateral driven policy, requiring most of its clients to secure the loan up to 200% of its value. Intuitively, one would therefore expect a portfolio that carries a low risk. In this study we analyze the risk profile of the portfolio and its segments, such as regions and industries. The reason for the study is two-fold, 1] quantifying the risk profile of the portfolio meets the requirements of Basel II and 2] to align managements risk appetite with the future composition of the loan portfolio. The data from the Bank are limited and stock market information from the Indonesian Stock Exchange alone will not provide a good view on the market. Therefore, a survey was performed to obtain expert opinions from the Bank, which are then combined in a Bayesian style with stock market information and data from the Bank. The final model combines those expert views with stock market information and data from the bank. The Kronecker product is used to combine correlation matrices from the sectors and regions into one correlation matrix for the entire Indonesian market, which is essential for the final model. With this model, one is able to perform a profound analysis on the current market situation in Indonesia. The model is robust and is able to quantify the risk portfolio and can be used by the management to align the current portfolio with their risk appetite, thus it can fulfill both goals of this research.