Finally, Figure 3 provides addition

Finally, Figure 3 provides additional time-series evidence on corporate investment growth by size and b/m portfolios. Over these 23 sample years, firms that researchers would classify as high b/m consistently experience lower rates of investment growth (cegth2) prior to portfolio classification than firms that would be classified as low b/m firms. The evidence from Figure 3 and the other exhibits in this section supports implications ofproposed models along the lines of Berk et al. (1999). Classification of stocks according to firm-level fundamentals such as size and b/m is implicitly conditioned on recent levels ofinvestment growth. In particular, stocks classified under Fama–French methods as low b/mstocks show increases in investment activity prior to the classification year,consistent with exercise of growth options. For these stocks, size increases and b/mdecreases prior to portfolio formation. For high b/mstocks, the opposite pattern seems to hold. In short, exercise of growth options conditions commonlyused portfolio sorting methods that are based on valuation measures.