The Modelling of Non-Maturity Deposits

George Soulellis

Within the banking industry, accurate liabilities or deposit-based expected life modelling is widely considered a prerequisite to sound asset–liability management. Its importance in mitigating interest rate risk is undisputed. However, the techniques associated with this are still evolving. This chapter focuses on establishing a concrete analytic methodology to estimating future expected deposit balance trajectories and their associated expected remaining life. We outline a comprehensive approach to guide the reader on how to define the event variable, establish a robust segmentation scheme, introduce key parameters in a time-series multivariate regression and validate and monitor the model on an ongoing basis to ensure its appropriateness.

THE IMPORTANCE OF MODELLING DEPOSIT BALANCES

Many banks worldwide are funded with non-maturity deposits, and the way in which their average lives are modelled has significant implications when estimating their value and their effectiveness in the management of interest rate risk. For example, modelling non-maturity deposits with too short an average life will subject the bank to rising interest rate risk exposure. Forecasting the expected