Innovative originality, profitability, and stock returns

[The financial value of innovation 5/5]

Should innovations be focused to be valued by markets?

David Hirshleifer, Po-Hsuan Hsu, and Dongmei Li investigate how innovative originality affects firm profitability and stock returns in their paper "Innovative originality, profitability, and stock returns" (2018).

They build a measure of innovative originality from the average number of unique technology classes cited across a firm's recently granted patents, and test its predictive power on US-listed firms from 1981 to 2006.

Their main conclusions include:

• Firms with higher originality deliver more persistent profitability: a one-standard-deviation increase slows the mean reversion of return on equity by around 14.5% compared to firms with no originality.
• Profitability is also more stable, as a one-standard-deviation increase in innovative originality reduces the volatility of return on equity over the next five years by roughly 11%.
• A value-weighted portfolio long in high-originality and short in low-originality firms generates monthly abnormal returns of 0.20% to 0.35%, with an annualised Sharpe ratio of 0.50 over the 1982-2008 period.
• The effect concentrates in firms that are harder to value: monthly alphas on the high-minus-low portfolio reach 0.82% to 1.08% among high valuation uncertainty firms and are negligible for high certainty ones.
• Return predictability is strongest where fewer investors pay attention, consistent with a limited attention explanation rather than a conventional risk-based story.
• Roughly a quarter of the abnormal return materialises in the three days around earnings announcements, which suggests the market underweights the cash flow implications before gradually recognising them.

This study challenges the assumption that intangible drivers of competitive advantage like innovation and originality are quickly and correctly priced by sophisticated investors.

The low correlation with size, book-to-market, and momentum, along with a Sharpe ratio comparable to the market factor, suggests an originality signal from USPTO patent data offers genuine diversification within a crowded factor landscape.

The sample however ends before ETF flows, passive ownership, and machine-readable patent data substantially reshaped how quickly cross-sectional anomalies are priced in public markets.

Extending the analysis to international markets and to the post-2010 regime would clarify whether limited attention remains a durable source of mispricing or whether it has materially eroded.