Thursday, September 18, 2014

Does the Value Premium Exist? Redux

A couple of months ago I wrote about John Bogle’s doubts that stocks have a persistent value premium. I’ve since persuaded myself that Bogle defines value stocks differently from other researchers. It seems that a value premium has persisted for some time. I’m not certain that it will exist into the future in a form that can be captured in returns after costs, but I’ve decided to modify my portfolio on the premise that I might be able to capture a value premium.

I had owned the Vanguard exchange-traded fund VB, which holds U.S. small cap stocks. Vanguard also offers two ETFs that split small cap stocks into value stocks (VBR) and growth stocks (VBK). I’ve sold my VB and bought VBR. This isn’t exactly a huge change in my portfolio, but it is the biggest change I’ve made in some time apart from adding new money.

I was adding a block of new money and did a currency exchange using the Norbert Gambit. By making the change from VB to VBR at the same time, I saved a few transactions.

Over the past decade, U.S. small cap growth stocks (VBK) have actually outperformed small cap value stocks (VBR) by 1.5% per year. So, I’m lucky to have split the difference by owning VB instead of VBR before now. Here’s hoping that VBR gives the higher returns in the coming decades.


  1. The value premium is elusive. So, you start getting funds that screen stocks out of the academic definition of value, etc. This does look like performance chasing.

    Anyway, the theory is clear: the return of the total market is what it is. Investors (as a whole) cannot get more than this return minus costs. Indexing the total market will give you your fair share of market return (minus minimal costs).

    If you want to bet on a small part of the market, good for you. I suspect that with all the recent media exposure of small-value funds, big-growth is likely to outperform in the near future, but I won't bet on it. I'll stay the course with my total market funds.

  2. @Anonymous: I haven't made any changes to my allocations to large stocks, Canadian stocks of any size, or international stocks of any size. The only change is selling some U.S. small growth stocks to buy more U.S. small value stocks. I'm open to arguments that this may be misguided, but performance chasing doesn't fit. U.S. small growth stocks have outperformed U.S. small value stocks over the past decade by 1.5% per year. So, I'm switching from a recent winner to a recent dog.

  3. So, you're probably smarter than the people that will sell you their shares. :-)

    1. @Anonymous: In a short-term sense, that's almost certainly not true. I assume that traders take half the spread and possibly a little more from me every time I make a trade. As for longer-term, who knows, but I've voted with my money.

  4. I think people tend to get caught up on the value premium partially because of its name.

    The three factor model says that value stocks are riskier than growth stocks, but if you were to ask most people they’d probably say growth stocks are riskier. The disconnect is because when people think of value stocks, they think of the Benjamin Graham definition of value which considers price-earnings, earnings per share, and growth. The Fama-French definition of value is a high book-market ratio which is common with stocks that have had recent drops in price due to reduced earnings and growth expectations or have uncertainties surrounding their long term viability. Many of the stocks that the Fama-French model would consider value stocks would not be considered a value stock to many investors.

    The theory of a value tilt isn’t that it’s a free lunch and you get better performance over short periods. The theory is that you can increase the risk of your portfolio by increasing your exposure to value stocks. Whether that additional risk pays off over a certain time period is unknown, but over the long term increased risk should give increased returns.

    1. @Wo: There are differences in risk levels between value and growth stocks and between small and large stocks, but Fama and French's research found that there were value and small premiums beyond what could be explained by risk levels. Not everyone believes these premiums will persist into the future, but I've decided to bet a slice of my portfolio on them. I already had a small-cap tilt, but now I've made that slice a small-value tilt.

    2. I agree with the theory. I also apply a small cap and value tilt to my portfolio.

      I’m not sure I entirely agree with you with regards to risk level, although it might just be the terminology. Fama and French’s research found that a lot of the alpha in the classic CAPM model could be explained by adding two additional risk factors: small cap and value. The excess return you get from small cap and value are the risk premiums. Increasing your exposure to the value factor increases your exposure to the risk factor their research identified.

      For example, according to the Fama-French three factor model, return = Bmkt*MKT + Bsmb*SMB + Bhml*HML + alpha. You can then calculate the Fama-French factors for VB and VBR over the past 10 years.

      VB: Bmkt=1.07, Bsmb=0.69, Bhml=0.12 and alpha was 0.05% (statistically insignificant).
      VBR: Bmkt=1.02, Bsmb=0.62, Bhml=0.41, and alpha was 0.02% (statistically insignificant).

      As you’d expect VBR has a higher Bhml value and therefore increased exposure to HM L. By putting a value tilt on your portfolio you’ve increased your exposure to the HML factor. In any given month HML may be positive or negative which would have a larger impact on VBR than it does VB and therefore increases the volatility of your portfolio. According to Fama-French’s three factor model, both the maximum possible loss and the maximum possible gain is higher with VBR than with VB. To me that means an increased risk.

    3. @Wo: I think we're just using terminology differently. I thought you meant risk in terms of volatility only. If we view value and size factors as risks as well, then it's all risk premiums.