Our new paper, Deep Value, looks at episodes where the valuation spread between cheap and expensive securities is wide relative to its history. One innovation is that we do this not just for individual equities, as is most common in related studies, but also across equity index futures, currencies, and global bonds.
You might have noticed that, for the last few years, I’ve been rather cynical about the ability to time factors using only their valuations. Reasons for my cynicism include 1) the inherent difficulty, theoretically and empirically, in market/factor timing (not a lot of breadth, very low Sharpe), 2) that the timing of non-value factors mostly just ups your weight on value, and, 3) to those focusing on just today, the fact that value spreads ain’t that abnormal right now. This new paper replicates the general failure of pure contrarian value-based factor timing using a single factor (value itself). That is, we confirm all these points (so I’m still right).1 But then we take it up a notch.
We next explicitly increase the breadth of value comparisons (e.g., we compare more crosses than simple factor timing by examining deep value situations within industries, across countries, bond markets, and currencies) and find that this really helps. Suddenly, with this breadth, looking for extreme mispricings2 vs. history (contrarian value timing) is a far more viable strategy than the coarse factor timing I debate with others elsewhere (and debate and debate…). Unlike simple factor timing, deep value with lots of breadth seems to have alpha to the traditional factors. Further, we use this approach to provide new evidence on competing theories for the value premium. For instance, our approach helps to evaluate whether value investing’s success comes from rational compensation for risk, an expression of behavioral over-reaction, or arises from simple random noise in prices.
In summary, we study all the interesting things that happen when some stocks, or other assets, get deeply cheap while others get deeply expensive and, in the process, learn more about the economics of value investing, and how timing works when you sin a little but now in lots of places at once (a better form of sinning than traditional market or factor timing!).
[1] Recall I recommend “sinning a little,” not necessarily zero. [2] I say mis-pricings to avoid more awkward phraseology but, as always, these could be times of elevated rational risk premia, though the paper offers evidence and opinion on this, you’ll have to read it!
Article by AQR
