This study examines whether firm innovation predicts valuation in the semiconductor industry after controlling for persistent firm traits and market-wide shocks. I investigate the relationship between R&D intensity, patent stock, and citation stock with Tobin’s Q, while controlling for size, age, investment ratios, capital structure, share repurchases, and M&A activities. The sample comprises listed semiconductor companies from 2015 to 2024, with up to 290 firm-year observations. Innovation output is measured as industry-year–normalized, depreciated stocks of patents and forward citations at the 4-digit SIC level, then logged. I assume a 20 percent annual depreciation as the baseline and verify robustness with a 30 percent rate. Methodologically, I use pooled OLS and two-way fixed-effects models, incorporating firm and year effects, with standard errors clustered at the firm level. I analyze recognition lags of one, two, and five years and include random effects for comparison. The fixed-effects models outperform pooled OLS, and the Hausman test favors fixed effects over random effects, supporting within-firm identification. Results are sensitive to timing. I conclude that markets tend to assign only modest value to firm-level innovation over short periods, whereas operational discipline shows a stronger and more consistent link with valuation. These findings highlight the need for longer recognition windows and more comprehensive measures of patent quality in future research. They also imply that managers should combine R&D efforts with stricter cash flow management and well-defined commercialization milestones.
Lynn Succari (Wed,) studied this question.