This post is Part 2 in a series about smart beta ETFs. See below for links to other posts in the series.
Smart beta is a relatively new term, but its roots stretch back several decades. Let’s look at the history of how the idea developed.
We’ll start with a simple question that has long been asked by students of the financial markets: what explains the difference in returns among stocks?
Back in the 1960s, economists and finance professors developed the capital asset pricing model (CAPM), which suggested that returns were directly related to risk. The formula began with a risk-free rate of return—for example, Treasury bills—plus an additional return for the stock market as a whole, called the equity risk premium. Then the model considered how sensitive a given stock is to the volatility of the overall market using a measure called beta. If a stock had a low beta—making it relatively less risky than the market—it should have a lower expected return. Stocks with higher beta should theoretically reward investors with a greater return.
CAPM is an elegant formula, one of its creators went on to win a Nobel prize in 1990, and it’s still taught in finance classes today. The problem is, it simply doesn’t hold up well in the real world. As it was tested during subsequent decades, many individual stocks and portfolios performed quite differently from what CAPM predicted. As the academic papers piled up, it became clear the model was broken. If you wanted to explain differences in returns, you had look beyond a stock’s sensitivity to the market’s volatility.
Building on reams of earlier research, Eugene Fama and Kenneth French of the University of Chicago took CAPM two steps further in the early 1990s. In a now-famous paper they created a new model that included two additional factors. It had already been observed that value stocks (those with a low price relative to their fundamentals) outperformed growth stocks, and small companies outperformed large ones. The researchers incorporated these two findings with CAPM into what would come to be called Fama-French Three Factor Model. They revealed that a stock’s sensitivity to the value premium and size premium (not just beta) can increase its expected return. The new formula did a much better job than CAPM at explaining equity returns: by some accounts, beta alone told just 70% of the story, while the three-factor model got closer to 95%.
In recent years, researchers have found evidence of other “factor premiums.” That is, they have identified groups of stocks with specific characteristics that help explain returns and risk persistently over time. These include momentum (stocks that have recently gone up in value tend to keep rising, and vice versa), quality (companies exhibiting certain characteristics related to financial health tend to outperform) and low-volatility (stocks with low beta tend to deliver higher risk-adjusted returns, which is clearly inconsistent with CAPM). Fama and French themselves published a 2014 paper that unveiled an updated five-factor model.
From theory to practice
Index investing appeared long before Fama and French, and in the beginning it was pretty straightforward. The first mutual fund tracking the S&P 500 was launched in 1975, and with a few exceptions, its successors simply mimicked the broad markets using indexes weighted by market cap: the bigger the company, the greater its influence in the index. When ETFs appeared in the early 1990s, virtually all of them used this same methodology, too.
But as the academic research accumulated, traditional ETF providers looked for ways to capture factor premiums by building indexes designed to outperform the broad market. Since traditional index funds simply capture beta, these new strategies came to be called “smart beta.”
Today, smart beta is the most significant trend in ETFs. Perhaps that’s not surprising: the market for plain-vanilla, cap-weighted funds is now saturated, and it’s impossible for new ETFs to compete on price. As fund companies design and launch new products, they’re looking to tweak traditional indexes and offer something new and better to investors.
This raises some obvious questions. Is smart beta just the latest gimmick that will ultimately disappoint investors? And why are we even discussing it on a blog devoted to indexing when it sounds like active management in disguise?
Is it really that smart?
The term “smart beta” may be clever marketing, but factor investing can’t be dismissed as a gimmick. There are decades of academic evidence behind these ideas. And there’s no secret sauce here: as we’ll discuss later in this series, the premiums may simply come from additional risk. If you believe value or small-cap stocks will deliver higher returns because they’re inherently riskier, then the decision to target them isn’t much different from choosing stocks over bonds (or corporate bonds over governments) for the same reason.
Other factor premiums have been explained by behavioural biases, such as our tendency to overpay for glamorous growth stocks and chase past performance. Even if you believe markets are mostly efficient, I think we can all agree that humans are, to use Dan Ariely’s phrase, “predictably irrational.”
While a purist would argue that anything other than a traditional cap-weighted index fund is, by definition, a form of active management, that’s going too far. There’s a lot of grey between the extremes of purely passive and highly active. Designing a transparent, rules-based index that weights hundreds of stocks according to certain characteristics is not the same as hiring a manager to pick a small number of companies according to his or her own idiosyncratic style.
Finally, let’s remember that the fundamental problem with active management is cost. A fund charging 2% has set itself an almost insurmountable barrier over the long term. But an index-tracking ETF charging 0.35% or 0.40% to target known premiums has at least a fighting chance. Moreover, because it tracks an index it’s not subject to the style drift that plagues human fund managers.
So, it’s time to make a deep dive into the world of smart beta ETFs to understand how they work and whether it’s worth considering them as part of your own portfolio. In the next post, we’ll start with a look at the value factor.
Other posts in this series:
Smart Beta ETFs: Your Complete Guide
Great article and really looking forward to your analysis, especially from the Canadian perspective, as most of what’s out there today is US focused.
I agree with your comments 100%. Almost since day one of my index investing career, I’ve been interested in factor investing. It’s logical and makes sense. But never done it primarily due to lack of products that I had faith in (actually worked, enough to overcome their costs, etc)
The closest I’ve come is actually in US equity. Replacing some or all of a VTI allocation with VBR gets you small cap value exposure with basically no additional cost, while still being diverse, etc.
‘Finally, let’s remember that the fundamental problem with active management is cost’. Well, I was under the impression another problem with 95% of active managers was that they didn’t get as good returns as simply investing money into a broad index fund. Part of this poor return is undoubtedly fees, but isn’t some of the problem the inability of active investors to pick the Next Big Thing at exactly the right time?
@Willy: Thanks for the comment. Yes, one of the ideas we’ll look at is how difficult it is to get exposure to the various factors. A fund like VBR (US small value) is quite straightforward, and it’s large, well diversified and cheap. You can’t really build a fund like that if your target is, for example, momentum stocks in Canada. Just because a factor is real in the academic literature doesn’t mean it’s easy to access with a real fund.
@Philip: Certainly the failure of some active management is poor decision making. But overall actively managed funds tend to trail their benchmarks by an amount roughly equal to their fees. Given the choice between a well diversified and prudently managed active fund with a fee of 0% and an index fund charging 1.5% which would you choose?
Historical stock returns show greater reversion to the mean than we would expect from random chance. A side effect of this is that measured beta is different when you measure it with annual returns than it is when you measure it with decade returns. I wonder whether the low-beta stocks still have higher risk-adjusted returns when you measure beta with return samples over time periods longer than one year.
My two concerns with this are taxes for taxable investors and the fact that its getting so popular now, potentially negating any potential additional gains.
Please correct me if I’m wrong or this comment is naive, but I can’t help but think that these models become increasingly complicated because with additional research and discoveries the market adapts and becomes more efficient. My understanding of (more or less) efficient markets is that one can expect to beat the market only if they’ve found an area where the market is “inefficient”. This comes with the assumption though that the act of exploiting this inefficiency makes the market more efficient. When you increase the demand, you shorten the supply. These funds might find that their algorithm works well for a while, but I would expect their annual return to eventually trend toward market average, because by their actions they are setting a new average. Is this not the case?
@BDR: This is a great question, and I’ll touch on it repeatedly in the series.
As you point out, if a factor is expected to persist after it has been identified, there should be an explanation. The explanations fall into two broad categories. The first are risk-based: e.g. small-cap stocks deliver higher returns simply because they are more risky. If that is the case, then the premium should persist because it’s simply rewarding investors for accepting that risk.
The second category is behavioral: e.g. value stocks outperform because investors tend to overpay for glamour growth stocks and chase past performance. If you believe that markets are perfectly efficient, such a premium should not persist because it could be arbitraged away.
The interesting point is that it’s almost impossible to prove any of these explanations. They have been the subject of debate for decades.
@CCP: Thanks for the reply. I didn’t consider the risk aspect; it does make sense, intuitively, that higher returns will always be available for those willing to take the risk. I look forward to the rest of this series.
Great post CCP!
I would suggest an important decision-making factor would be the trade-off between a sure fee saving (market cap index ETF) and an uncertain risk premium (smart beta ETF).
Interesting reading as always. I’ve stuck to vanilla index funds to avoid the “maybe there is something better” rabbit hole. After moving to something you think may be better, what is to stop the next move, and the next? This smells like chasing performance and market timing.
You indicate that active managers charge 2%.
This is a little misleading as most active managers charge 1% or less these days. It’s the 1% tacked
on in embedded commissions that bring the price tag up to 2%. It’s probably time you start making this distinction.