This study on overinvestment differs from the existing literature in that investment in machinery and equipment is modelled as a structural vector autoregression with identification achieved by imposing long-run restrictions, as in Blanchard and Quah (1989). The transitory components obtained from the model suggest the existence of a minor overinvestment in machinery and equipment of less than 5 per cent in 2000, despite a significant overevaluation of the stock market.