Changes in the parameters of a recursively identified oil market model are examined through an iterative algorithm that tests for possible breaks in coefficients and variances. The analysis detects breaks in the coefficients of the oil production and price equations, together with volatility shifts in all three equations of the model. Coefficient changes imply an enhanced response of production to aggregate demand shocks after 1980; and that the price response to supply shocks is more persistent from the mid-1990s. All variables evidence changes in the relative contributions of individual shocks to their forecast error variances, with coefficient and volatility breaks in the first half of the 1990s being particularly important in this respect. The results show that analysts of this market should eschew constant parameter models estimated over an extended period.