Innovation and market value: a quantile regression analysis

We construct a new database by matching firm-level Compustat data to NBER patent data, for four 2-digit complex technology sectors. Whilst conventional regression estimators show that the stock market does recognise efforts at innovation, quantile regression analysis adds a new dimension to the literature, suggesting that the influence of innovation on market value varies dramatically across the market value distribution. For firms with a low value of Tobin's q, the stock market will barely recognize their attempts to innovate. For firms with the highest values of Tobin's q, however, their market value is particularly sensitive to innovative activity.