A recent Bloomberg View article, published in Canada in the National Post, muses on the success of fundamental analysis, and the challenge this constitutes to the efficient market hypothesis

The author, Noah Smith, sets out the background like this:

• “Fundamental analysis succeeds if two things are true. First, the market has to have overlooked important things about a company’s value — things that can be observed by carefully scrutinizing publicly available information. Second, the market has to eventually realize the company’s true value. This second step is important because if the market never catches on, you would have to hold the stock for a very long time in order to make a profit from dividends and share buybacks…”

“In principle,” says Smith, “any technique that consistently extracts valuable but unrecognized information from balance sheets, income statements and the like counts as a big violation of the efficient markets principle.” He then points to a recent paper by the financial economists Sohnke Bartram and Mark Grinblatt, titled Fundamental Analysis Works. Here’s the abstract of that paper:

• “Stock prices cannot be the outcome of a rational efficient market if fundamental analysis based on public information is profitable. Our approach to fundamental analysis estimates the intrinsic fair values of stocks from the most common quarterly balance sheet and income statement items that were last reported in Compustat. Taking the view of a statistician with little knowledge of finance theory, we show that the most basic form of fundamental analysis yields trades with risk-adjusted returns of up to 9% per year. The trading strategy relies on the convergence of market prices to their fair values. The greatest rate of convergence occurs in the month after the mispricing signal and subsequently decays to zero over the subsequent 28 months. Profits from trading are present for both large and small firms in economically significant magnitudes.”

The authors of the paper acknowledge that their exercise employs a “particularly simple and agnostic view of how to compute fair value,” but point out: “More precise ways of obtaining fair values certainly exist, but our goal is to be conservative at assessing whether a crude form of fundamental analysis works. The fair value approach used here is unlikely to be a superior mousetrap for capturing the intrinsic values of securities. However, if the crude statistician’s approach to fundamental analysis works, then more accurate ways of measuring mispricing should work even better.”

Smith comments: “If this result holds — that is, if Bartram and Grinblatt haven’t made a big, critical mistake in their statistical procedure — then it’s a pretty serious blow to the idea that markets take advantage of all publicly available information. It means that there are some pieces of information sitting right there on corporate balance sheets and income statements, free for all the world to see, that aren’t being fully incorporated into market prices for almost two years after they appear. Whatever that information might be, these calculations suggest it’s available for the taking.”

I won’t pretend to be capable of grappling with all the complexities of the paper. But if it provides a boost to the cause of fundamental analysis, then it must also do so for the financial statements on which such analysis would be based. I’ve noted here several times that it’s common for preparers and auditors to wonder whether anyone truly engages in much depth with the statements, that even the more sophisticated and diligent users place their primary emphasis on things like non-GAAP measures, spending little time on all those detailed, scrupulously-compiled disclosures. However, when Smith wonders what the information might be that isn’t being fully incorporated into market prices, it’s reasonable to think that the answer might be, in broad terms: any of the things in financial statements that people don’t pay enough attention to.

In their conclusion, the authors note: “Because we focus indiscriminately on the most available accounting items…we have not successfully identified which accounting variables are best for determining fair value. (It) would not surprise us if only a handful of accounting variables could do as well, or improve upon, the strategies derived here. We leave that, as well as improvements in the fair value estimation approach, to future research.” No doubt such research may already be under way (not to mention Smith’s prediction that “money managers will read this paper and write code to do more sophisticated versions of what the professors did”.)  But, of course, even if such super-charged accounting variables were to be identified, the assessment of their likely effectiveness in any particular situation could only be enhanced by understanding the inherent estimation uncertainty attaching to them, or how underlying accounting policy choices might affect their comparability from one entity to the next, or any of the other complexities that might be identified in the related disclosures. For those involved in leading the credibility and visibility of financial reporting, there’s surely much here to mull on…

The opinions expressed are solely those of the author