The relationship between the concentration of various gas components (CO,NO,HC) at the output of a three-way catalytic converter and the input and output air-fuel ratios (AFRs) is examined. A simple linear-in-the-parameters model is developed and it is assumed that the model parameters in the lean and rich regions are different. The model is fitted to some experimental step response data obtained from fast gas response analysers and UEGO sensors, using a recursive linear least-squares estimation method. A reasonably good fit to the data is obtained, particularly for NO and CO. Results from step tests for different AFR ranges are combined to obtain an overall picture of the dependency of the gas components on measured AFR values. The proposed model provides the possibility of predicting the dynamic performance of catalytic converters from a knowledge of the input and output AFR values.