The two key questions which motivate our work are: do bubbles exist (in the sense that stock market prices do not always correspond to the present value of expected future profitability) and, if bubbles exist, do they have an effect on business fixed investment? The case of Japan is particularly interesting because of the dramatic movements in the Japanese stock market and the wide perception that these were associated with a bubble. We use a variety of techniques to analyze these questions. Fist, we examine financing and investment patterns to gauge firms' reactions to the 1980s stock market run-up. Second, we test subsets of the orthogonality conditions associated with the empirical first-order conditions for fixed investment. Third, we use a linear projection to decompose stock market prices into fundamental and bubble components, allowing us to carry out parametric estimates of the effect of the bubble component on fixed investment. The data strongly suggest that there was a bubble that had an economically important statistically significant effect on business fixed investment in Japan.