Fama French Three Factor Model Challenged…

I’ll be blunt: beginners may want to skip this post.

I realize that many people in the investment industry look at the capital markets in a CAPM framework. In 1993, Eugene Fama and Kenneth French released their landmark paper The Cross Section of Expected Stock Returns which showed that, historically, a model that incorporated sensitivity to what are known as the “size” and “value” factors in addition to sensitivity to the “market” factor did a better job explaining portfolio returns across portfolios of different characteristics. CAPM, for reference, really only seemed to work for large, growthy portfolios. I know “growthy” is not a real word, but humour me. :)

I’ve written extensively on this “Fama French Three Factor” model (FFTFM), but should point out that recently a new model has been proposed that seems to do a better job explaining some anomalies that the FFTFM can’t. One of these factors is “momentum”: this is the tendency for stocks that have been falling to continue falling, or vice versa for stocks that have been rising to continue rising.

Other researchers are beginning to find data to support the notion that value stocks (represented by high book-to-market metrics) do not indeed compensate investors given the higher volatilities they exhibit. This would be a problem for the FFTFM as well which assumes the opposite.

So to that end, I’m including a link to a paper that proposes a newer model, the Neoclassical Three Factor Model which suggests a solution to the momentum and distressed firm anomalies. Click here to access the paper.

NOTE: CAPM, FFTFM and the Neoclassical Three Factor Models are just that: models. They are just attempts to explain stock market behaviour but realistically even if anyone could perfectly explain past stock market returns it would hold decreasingly little predictive behaviour on a go-forward basis as this new information would then be incorporated by the market, thus changing the very nature of market valuation techniques.

Preet Banerjee
Preet Banerjee
...is an independent consultant to the financial services industry and a personal finance commentator. You can learn more about Preet at his personal website and you can click here to follow him on Twitter.
Related Posts
Showing 13 comments
  • Michael James

    These things are always tough to judge. It isn’t hard to find a model that better fits old data. The real test of various models is how they perform in the future. But, we have to wait a long time to collect enough new data. By that time some more new models show up and the cycle repeats.

  • Preet

    @MJ: Agreed, I’m just happy to believe in capitalism and leave it at that! :)

  • Richard

    A great model will explain everything except what happens tomorrow :)

    That said, doesn’t the traditional 3-factor model indicate higher returns for value stocks, which would be a way of compensating investors for volatility? Or to put it another way, investors anticipate increased volatility in the future and demand higher returns/lower prices today, making the company a value stock

  • Preet

    @Richard – you are correct, this is one of the positions of the original three factor model. However, this relation is being called into question by some other researchers.

    It’s also interesting to note that some academics have found data to support the notion that the market does a really good job determining the merits of publicly traded companies – but they tend to overpay for growth companies and underpay for value companies – in other words, they get the overall thesis right, but miss on the magnitude of price discrepancy. Perhaps I will highlight this paper in the next few weeks.


  • dj

    For some reason i can’t zoom the chart…is it firefox?

    • Preet

      @dj – don’t worry about the chart as it is unrelated to the post. Just a pretty picture in the background – sorry for the confusion! :)

  • Richard

    Don’t the growth companies end up being the ones that are overpriced and the value companies the ones that are overpriced? Another good definition for growth companies would be that people are paying (too much) for expected future earnings based on speculation, but I can’t think of a similar definition for value companies. Do people underpay for stable or declining earnings? That would make sense as a psychological bias.

    • Preet

      @Richard – there was an interesting paper which I’ll write about soon which essentially tries to show that the market is actually pretty good at predicting the future prospects of companies but overpays for growth companies and underpays for distressed companies. Stay tuned…

  • Richard

    It makes sense that people would have trouble thinking that a company with declining revenues is a good investment, turning it into a good value stock. It’s always more fun to be a contrarian :)

    • Preet

      @Richard – being a contrarian IS more fun isn’t it? :)

  • Jacob Caldwell

    Where do I find the rest of the articles on DFA? I got to XIII and can’t find the rest…the suspense is killing me!

pingbacks / trackbacks