Is value investing particularly profitable when the price spread between cheap and expensive assets (the value spread) is extremely large (deep value)? In their November 2017 paper entitled “Deep Value”, Clifford Asness, John Liew, Lasse Pedersen and Ashwin Thapar examine how the performance of value investing changes when the value spread is in its largest fifth (quintile). They consider value spreads for seven asset classes: individual stocks within each of four global regions (U.S., UK, continental Europe and Japan); equity index futures globally; currencies globally; and, bond futures globally. Their measures for value are:
- Individual stocks – book value-to-market capitalization ratio (B/P).
- Equity index futures – index-level B/P, aggregated using index weights.
- Currencies – real exchange rate based on purchasing power parity.
- Bonds – real bond yield (nominal bond yield minus forecasted inflation).
For each of the seven broad asset classes, they each month rank assets by value. They then for each class form a hedge portfolio that is long (short) the third of assets that are cheapest (most expensive). For stocks and equity indexes, they weight portfolio assets by market capitalization. For currencies and bond futures, they weight equally. To create more deep value episodes, they construct 515 sub-classes from the seven broad asset classes. For asset sub-classes, they use hedge portfolios when there are many assets (272 strategies) and pairs trading when there are few (243 strategies). They conduct both in-sample and out-of-sample deep value tests, the latter buying value when the value spread is within its top inception-to-date quintile and selling value when the value spread reverts to its inception-to-date median. Using data as specified and as available (starting as early as January 1926 for U.S. stocks and as late as January 1988 for continental Europe stocks) through September 2015, they find that:
- Across available samples, there are about 3,000 deep value episodes. These episodes tend to cluster, notably around the technology bubble of the late 1990s and around the 2008 financial crisis. Deep value strategies tend to perform best following clusters.
- Based on in-sample analyses, around deep value episodes:
- Returns to the value strategies specified above are particularly high.
- Value strategies have lower than usual betas (in fact, negative).
- Earnings of cheap stocks are particularly weak compared to those of expensive stocks, and this weakness persists after deep value dissipates. Moreover, analysts are more likely to lower earnings forecasts for cheap than expensive stocks leading up to and during deep value episodes.
- Average news sentiment is particularly worse for cheap than expensive stocks.
- Value strategy execution is inhibited by widened bid-ask spreads, elevated shorting costs and high portfolio volatility. Nevertheless, short interest and net share issuance for expensive stocks is particularly high, and mergers of cheap stocks rises.
- Out-of-sample deep value strategies applied to the seven asset classes:
- Have positive gross returns, but these returns are similar to those of standard value strategies.
- Have on average positive gross alphas (relative to market, standard value and momentum factors), but: signs are mixed; alphas for stock selection are generally insignificant; and, alphas for other asset classes are marginal. In other words, timing value based on deep value episodes work poorly for individual asset classes.
- Have significantly positive gross alpha (3.5% annualized) when aggregated into a portfolio combining all seven strategies. Individual stocks contribute little to this alpha.
- Out-of-sample deep value strategies applied to the 515 sub-classes:
- Have significantly positive gross alpha (3.0% to 5.2% annualized) when applied to individual stocks in U.S., UK and continental Europe, but not in Japan.
- Have marginally positive gross alpha (1.9% to 2.6% annualized) when applied to other sub-classes.
- Have significantly positive gross alpha (6.4% annualized) when aggregated into a portfolio combining all 515 strategies.
- Findings of significant gross alphas for aggregated deep value portfolios are robust to a variety of deep value strategy specifications.
In summary, evidence indicates that it is difficult to boost the performance of conventional value and momentum strategies by exploiting deep value. The most promising approach to exploitation involves complex aggregations of deep value strategies.
Cautions regarding findings include:
- Findings are gross, not net. Accounting for monthly portfolio reformation frictions and continuous shorting costs/constraints would reduce returns. Moreover, as noted in the paper, these costs/constraints are elevated during deep value episodes.
- Deep value tracking and strategy execution are beyond the reach of most investors, who would bear fees for delegating these tasks to a fund manager. The most promising approach, aggregating many deep value strategies, is particularly daunting.
- Testing many strategies/strategy variations on the same data introduces snooping bias (discovery of luck), such that the best-performing examples overstate expectations.
See also:
