“There are three reasons that suggest caution when evaluating the evidence on the investment or asset growth factor.”
“First, in the post-1963 period, the relationship between value or profitability and returns is monotonic. For example, if one constructs five portfolios with increasing profitability, their returns increase uniformly from one portfolio to the next. That is not true for investment. In fact, most of the return differences come from the extreme underperformance of high-investment firms.”
“Second, the results seem to be sensitive to how one measures investment (e.g. Cooper et al., 2020). This suggests that the effects of investment are not clear cut as suggested by earlier studies. For a phenomenon to be robust, it should not be overly sensitive to measurement choices.”
“Finally, the fact that investment generates no variation in returns prior to 1963 suggests that it is not nearly as stable as value and profitability. To be clear, the economic mechanism for investment/asset growth to matter is sound – it is the evidence that appears to be shaky.”
“The data are what they are: the research says that the investment factor subsumes the ‘standard’ value factor after 1963 in the US. The more important question is what this means both for the underlying economics, as well as for investors. My view is that this is a very sample-specific result and one should not make too much of it. As I mentioned earlier, investment does not subsume value prior to 1963, or outside the US.”
‘Pure value and pure profitability strategies commit the sin of ignoring information relevant for each other’
“Economic primitives link value and profitability. The evidence linking the two is pervasive in three different data sources: pre-1963 in US data, post-1963 in US data and post-1990 in international data. The implication is that considering these jointly makes a lot of sense for investors.”
“Suppose one thinks of the long-only value strategy as buying cheap assets. When you just buy cheap assets (i.e. value only), some fraction of them could be cheap for a reason – some value firms will inevitably go bankrupt. In other words, value-only portfolios ignore the information in profitability. Similarly, when one implements a long-only profitability strategy, one ignores the information in prices – you are simply buying productive assets but perhaps overpaying for them.”
“Therefore, in isolation, pure value and pure profitability strategies commit the sin of ignoring information relevant for each other. Of course, that does not prevent investment managers from delivering such products, nor should it. Caveat emptor applies. But Eduardo Repetto and I point out in our work that this is false choice – there really is no reason for investors to ignore useful and implementable information.”
“All investors have been subject to a cold and dark winter. But investors who have focused on factors, particularly arbitrary factors with limited economic content, have had a tougher winter. Some of the drawdowns have been quite large, but it is best to view them in the context of historic volatility.”
“There are two things that investors would be well advised to remember. First, realized portfolio volatility (and drawdown) is likely to be higher than factor volatility because it includes liquidity premiums. In other words, realizations are likely to be worse than expectations. Second, how one develops expectations of volatility and drawdowns matters. Relying on normal distributions is not a good idea. In my work with Eduardo Repetto(3), we use bootstrapping to get a sense for what the true distribution of portfolio returns might look like.”
“As to the question of if and/or when the winter will end, that is very much strategy dependent. For investment strategies with valid economic underpinnings, spring has to come simply because economic fundamentals require it. But when spring comes remains an open question.”
‘Whether factors are arbitraged away or not depends on whether the factors are genuine and they reflect mispricing or risk’
“Whether factors are arbitraged away or not depends on whether the factors are (a) genuine, and (b) they reflect mispricing or risk. If they are not genuine (i.e. if they are generated via willful or inadvertent data mining), they will disappear. If they represent mispricing, they can be arbitraged away so long as the cost of arbitrage drops sufficiently. But if they reflect economic fundamentals and risk, then arbitrage cannot change the expected return series. Key factors in academia (like value and profitability) can represent mispricing and/or risk and it is very hard to tell the difference.”
“There are two topics that might be of particular interest to your readers. The first has to do with opportunity costs in trading – in other words, measuring the cost of the ability or unwillingness to trade. This is a cost that is particularly relevant for large institutional investors (such as sovereign wealth funds) as their portfolios move prices. The second is ongoing work that looks at the role of intangible assets. We already know from existing work that goodwill adjustments are important to value portfolios. But preliminary evidence suggests that there is other information that is also informative. Stay tuned.”
1 Wahal, S., February 2019, “The Profitability and Investment Premium: Pre-1963 Evidence”, Journal of Financial Economics.
2 Wahal, S., Repetto, E., June 2020, “On the Conjoint Nature of Value and Profitability”, working paper.
3 Wahal, S., Repetto, E., November 2020, “The Joint Distribution of Value and Profitability: International
Evidence”, working paper.