Noisy Factors in Law
For years, academic experts have championed the widespread adoption of the “Fama-French” factors in legal settings. Factor models are commonly used to perform valuations, performance evaluations, and event studies across a wide variety of contexts, many of which rely on data provided by Professor Kenneth French. Yet these data are beset by a problem that the experts themselves did not understand: In a companion article, we documented widespread retroactive changes to French’s factor data. These changes are the result of discretionary changes to the construction of the factors, and they materially affect a broad range of estimates.
In this Article, we show how these retroactive changes can have enormous impacts in precisely the settings in which experts have pressed for their use. We provide examples of valuations, performance analyses, and event studies in which the retroactive changes have a large—and even dispositive—effect on an expert’s conclusions. Our analysis has several implications. First, it demonstrates that these data are not sufficiently reliable to be used by experts. Second, it demonstrates a phenomenon we call the law of conservation of judgment: Methodologies that appear objective still rely on judgment of one kind or another. Rather than eliminating judgment, they simply move it around. Finally, our analysis points to the problems that arise from the commingling of academic and commercial interests.