The Research and Development Factor

The article “The Research and Development Factor” first appeared on Alpha Architect Blog.

Excerpt

Since the development of the CAPM, which explains about two-thirds of the variation of returns among diversified portfolios, academic research has attempted to find models that increase the explanatory power of the cross-section of stock returns. Models are not like cameras that provide an exact replica of the world. If models were perfectly accurate, they would be laws, like we have in physics. Instead, models are engines that advance our understanding of how markets work, and prices are set.

As new research findings were published, we moved from the single-factor CAPM (market beta) to the three-factor Fama-French model (adding size and value), to the Carhart four-factor model (adding momentum), to Lu Zhang’s q-factor model (beta, size, investment, profitability), to the Fama-French five-factor (adding value to the q-factor model) and six-factor models (adding back value and momentum to the q-factor model). There have also been versions that use different metrics for profitability and value, and Stambaugh and Yuan’s mispricing (anomaly)-based model. The empirical evidence demonstrates that stocks with high research and development (R&D) expenses have delivered a premium regardless of the model usedindicates. Intangible assets such as R&D are becoming increasingly important as the economy has moved from primarily manufacturing to service and knowledge.

For example, using a measure of R&D intensity (R&D relative to market value), Woon Sau Leung, Khelifa Mazouz, and Kevin Evans, authors of the 2020 study “The R&D Anomaly: Risk or Mispricing?,” found a statistically economically significant monotonic increase in average returns from 0.67% for Portfolio 1 (lowest decile) to 2.23% for Portfolio 10 (highest decile). They also found that the premium persisted after adjusting for size, value, and momentum effects. In addition, the zero-cost spread portfolio (Portfolio 10-1) yielded a Carhart four-factor alpha of 1.35% and a Fama French five-factor alpha of 1.52% per month, both significant at the 1% level—the R&D anomaly cannot be explained by existing pricing, including the relatively recent investment and profitability factors.

Leung, Mazouz, and Evans also found:

“The R&D premium correlates positively with innovations to the aggregate dividend yield, and negatively with shocks to the default spread and risk-free rate, demonstrating the sensitivity of R&D stocks to variables that predict future business conditions. Moreover, the loadings on these three state variable innovations are significantly priced in the cross-section of R&D stock returns and even drive out the size and book-to-market equity factors. These results demonstrate that the R&D premium represents a significant and incremental reward for bearing intertemporal risk.”

Their findings of risk-based explanations for the R&D effect are consistent with those of Jangwook Lee and Jiyoon Lee, authors of the March 2020 study “Mispricing or Risk Premium? An Explanation of the R&D-to-Market Anomaly.” They are also consistent with the risk-based theoretical prediction of Jonathan Berk, Richard Green, and Vasant Naik, authors of the 2004 study “Valuation and Return Dynamics of New Ventures,” who explained:

“The firm learns about the potential profitability of the project throughout its life, but that research and development effort itself is only resolved through additional investment by the firm.”  

The result is that the risks associated with the ultimate cash flows have a systematic component, while the purely technical risks are idiosyncratic.

In an out-of-sample test, Kewei Hou, Po-Hsuan Hsu, Shiheng Wang, Akiko Watanabe, and Yan Xu, authors of the study “Corporate R&D and Stock Returns: International Evidence” published in the June 2022 issue of the Journal of Financial and Quantitative Analysis, examined the cross-sectional return predictability of R&D in international equity markets. Their findings were consistent with prior research. For example, in global sorts the top quintile portfolio outperformed the bottom one by 1.024% (0.537%) per month in equal-weighted (value-weighted) returns. The finding that the equal-weighted spread was more significant than the value-weighted spread suggests that substantially higher subsequent returns for more intensive R&D investments are more pronounced among smaller firms. They also found that the R&D effect could not be explained by common equity factors used in asset pricing models (including market beta, size, value, and momentum).