Do Intangible Assets Explain the Failure of the Value Factor?

Is the failure of the value factor over the last decade due to the inability of book value to incorporate so-called “intangible assets,” such as the intellectual property that has propelled companies like Amazon, Alphabet and Apple? New research provides the answer.

The value factor underperformed over the past decade, leading many to argue that a driver of value’s poor performance has been the deteriorating quality of book assets as a fundamental anchor due to the omission of internally generated intangible assets, which have become a steadily increasing percentage of corporate assets. As one example, Amitabh Dugar and Jacob Pozharny, authors of the 2020 study, “Equity Investing in the Age of Intangibles,” concluded that the relationship between financial variables and contemporaneous stock prices has weakened so much for high-intangible-intensity companies in both the U.S. and abroad that investors can no longer afford to ignore the changes in the economic environment created by intangibles.

Andrea Eisfeldt, Edward Kim and Dimitris Papanikolaou contribute to the literature with their August 2021 study, “Intangible Value.” They began by noting: “Correctly defining the fundamental anchor for the value factor is important both in the context of rational explanations of value, in which book assets capture assets in place, and for behavioral explanations, in which market to book ratios represent a measure of mispricing.” In support of this view, they cited research finding that intangibles had grown from about one-third of corporate assets in 2003 to about half – resulting in a growing mismeasurement of book assets. To address this problem, they proposed an intangible-augmented value factor (“intangible value,” HMLINT) – adding intangible assets to the book equity of each firm – and constructed it using a simple modification to the standard Fama and French value factor (HMLFF). They also performed their intangible value sort within industries.

Eisfeldt, Kim and Papanikolaou applied the perpetual inventory method to flows of 100% of selling, general and administrative (SG&A) expenses, given assumptions about depreciation and initial values to estimate the value of three main categories of intangibles: computerized information, R&D and economic competencies. Their measure of HMLINT added intangibles to book equity and subtracted goodwill. Their factor construction matched the original Fama and French data construction methodology as closely as possible. Their sample period was 1975 to 2018. Additionally, they conducted analyses for subperiods from 1995-2018 (post-internet era) and 2007-2018 (post-crisis era). Following is a summary of their findings:

  • HMLINT was highly correlated with the traditional value factor (78%).
  • HMLINT priced standard test assets with lower pricing errors than HMLFF.
  • HMLINT substantially and significantly outperformed HMLFF.
  • The average returns to a portfolio that was long HMLINT and short HMLFF were 2.1% annually, with a standard deviation of only 6.4%.
  • HMLINT had an information ratio of 0.33% with respect to HMLFF; and the long-intangible-value, short-traditional-value portfolio’s Sharpe ratio was 0.33 over the full sample and 0.59 in data since 2007. The outperformance held over the entire sample and was more pronounced in the post-crisis era in which the returns to traditional value have been particularly disappointing.
  • The alpha of intangible value in a single traditional value factor model was 3.77% and highly statistically significant (t-stat = 6.1) over the full period. Over the most recent subperiod (2007-2018), its alpha was 2.18% (t-stat = 1.86) – outperforming traditional value by 3.44%.
  • Intangible value had a positive and significant alpha of 2.96% (t-stat = 5.05) controlling for the market, size, value and momentum factors.
  • Isolating the intangible value factor by going long only stocks that were uniquely in HMLINT and shorting only stocks that were uniquely in HMLINT, they found positive alphas.
  • Intangible value goes long firms with better fundamentals (they are higher quality companies) – the long leg of intangible value contains firms with higher productivity, higher earnings to price ratios, higher profits to assets and lower debt to earnings. By contrast, traditional value contains firms with lower gross profitability to total assets, lower sales to stockholders’ equity, lower sales to book assets and higher debt to earnings.
  • Intangible value tends to be long slightly smaller firms and short slightly larger firms than tangible value.
  • The intangible value portfolio avoids “value traps” and avoids shorting low book-to-market firms whose book values do not reflect their total capital stock.
  • Value has consistently been a within-industry phenomenon for both traditional and intangible value – book-to-market’s ability to predict stock returns is almost entirely driven by within-industry variation.