Impulse-response analysis of the market share attraction model

We propose a simulation-based technique to calculate impulse-response functions and their confidence intervals in a market share attraction model [MCI]. As an MCI model implies a reduced form model for the logs of relative market shares, simulation techniques have to be used to obtain the impulse-responses for the levels of the market shares. We apply the technique to an MCI model for a five-brand detergent market. We illustrate how impulse-response functions can help to interpret the estimated model. In particular, the competitive and dynamic structure of the model can be analyzed.