Factor investing has grown rapidly in recent years. Although it is difficult to quantify the exact dollar amount of money following factors, it is likely that there is somewhere close to, or maybe even north of, $1 trillion following factors via ETFs, mutual funds and other investment products.
And based on the long-term historical data, there is good reason for that.
Factors like value and momentum are supported by extensive academic research showing that they have produced a positive premium over an 80+ year period. That data gives investors who use factors confidence that they will see the premiums in their own portfolios that have shown up in the historical data. But in many cases, the results haven’t been as good in practice as they have been in theory.
Identifying that factors haven’t worked as well in the real world as they have in testing is the easy part, though. Identifying why it has occurred, and what that means for the success of factors going forward is much more difficult. The question of why factors haven’t worked as well in the real world as they have in testing is being asked the most loudly by value investors right now since the value premium is now at the point where it has been negative over a decade plus period, but the same thing has happened to a lesser extent with other factors as well.
We spoke to Adam Butler of Resolve Asset Management for our Excess Returns Podcast last week, and he offered up an interesting framework for looking at the life cycle of factors and how their adoption can lead to changes in the premiums associated with them. Although I am more optimistic than Adam on the future of some of the major factors, I think this framework offers an excellent way to look at the way factors become used by investors, and the potential problems that come along with widespread adoption.
Here is a high-level look at the life cycle of factors using the value factor as an example.
Before a factor can be used by investors (or at least before they know they are using it), it first needs to be discovered. This work is typically done in the academic community. Academics typically test factors on a long/short basis and want to see that stocks with the highest exposure to the factor produce a positive excess return, and that the stocks with the lowest exposure have negative excess returns.
Although value investing has been around for a very long time, the discovery of the value factor is largely attributed to a paper published by Eugene Fama and Ken French in 1992.
 Early Adoption
Once a factor is discovered, its use by practitioners begins to increase. The competition to try to beat the market is obviously very intense, and managers are looking for any edge possible. So if academic research identifies a way to achieve a significant excess return over the market, you can bet that managers will find ways to use it. In the early part of the adoption phase, more intrepid managers use the factor, but as time goes by, its use becomes more widespread and more and more money is deployed into it.
 Widespread Adoption
This is where potential problems can come into the picture.
Factors typically work because there is some reason that investors avoid the types of stocks selected by the factor. For example, stocks selected using the value factor are typically riskier than the market and have significant short-term problems with their businesses. The fact that investors avoid these stocks allows investors who are willing to hold them to generate a premium. But as money flows into these types of stocks, the premium can be reduced, eliminated, or even become negative.
In visual form, it’s very similar to the lifecycle of technology and how technology is adapted over time. You have the innovators and early adopters, which would be similar to the early days of factor investing when premiums are high. As you get more and more of the market understanding and investing in the factor, i.e. the early majority and late majority phase, the premium offered by the factor falls and eventually can even become negative.
The Relationship Between Flows and Returns
Let me give you a theoretical example based on the framework Adam gave us in the interview.
Let’s say that $1 billion is avoiding value stocks for the reasons I listed above. But now a researcher comes along and finds that investors who are willing to hold value stocks can take advantage of this and generate a premium. If $500 million flows into that factor, the premium ends up being reduced, but it still exists. If $1 billion flows in, it could be eliminated altogether. If $2 billion flows in, it could become negative because the capital that has come in now exceeds the capital that was avoiding value stocks in the first place. But the process can also work in reverse. If investors capitulate when value stocks struggle and withdraw their capital, the premium can become positive again. This process can continue until an equilibrium is reached.
Under this framework, a factor that has a positive long-term premium can have long periods where it has a negative premium if it attracts too much capital.
This is obviously just a theoretical framework, and many other factors play into the equation in the real world, but it shows how widespread adoption of a factor can lead to a reduction in its effectiveness. And the problem can be the biggest for the factors with the strongest long-term data to support them and most sensible explanations behind them because they will attract the most capital.
The Death of Factors?
So does this mean that we all should go and liquidate our factor-based portfolios?
I don’t think so. Although this framework can help explain how largescale adoption can reduce or eliminate the premium associated with a factor, the fact that investors will withdraw capital when that happens also explains how it can come back to life. If over time more investors avoid the factor than follow it, a long-term premium will still exist.
But this life cycle does support the evidence we have seen in the real world that factors likely won’t work as well once they achieve widespread adoption as they did in testing, and they will likely have long periods of time where they don’t work. If there is a long-term premium associated with any factor, the ability to stay the course during these periods (while hoping that others who follow the factor will not and will withdraw their capital) is essential.
Like all things in investing, there is no excess return without risk. With factors, that risk comes in the form of prolonged periods of underperformance. It also comes in the form of the chance that any factor that has worked in the past will not work in the future. But regardless of where you think any given factor is in its life cycle, understanding how the process plays out can be an important tool in analyzing its effectiveness.
Jack Forehand is Co-Founder and President at Validea Capital. He is also a partner at Validea.com and co-authored “The Guru Investor: How to Beat the Market Using History’s Best Investment Strategies”. Jack holds the Chartered Financial Analyst designation from the CFA Institute. Follow him on Twitter at @practicalquant.