A statistical model for the temporal pattern of individual ATM withdrawals

Models of consumer behaviour based purely upon empirical relationships in data can perform well in the short-term, but often degrade rapidly with changing circumstances. Superior longer term performance can sometimes be attained by developing models for the deeper processes underlying the consumer behaviour. In this paper, we develop a random-effects point-process model for ATM (Automated Teller Machine) withdrawals. Estimation, prediction and computational issues are discussed. The model may be used to predict behaviour for an individual, assess when state changes in individual behaviour have occurred, and as a description of behaviour for a portfolio of accounts.