“Understanding beta and alpha” is eligible for CE credits, see Accreditation details for more information


For most investors, beta is the number they want to beat—or buy. It’s provided by an index, but beta is much richer and more complex.

In his 1955 doctoral dissertation, Harry Markowitz established the notion that a diversified portfolio of stocks could smooth fluctuations among individual stocks, providing those stocks weren’t closely correlated. Markowitz’s three principal achievements were correlation analysis, diversification as a risk-dampening measure, and the definition of risk as volatility.

Those achievements are summarized in the notion that there’s an efficient frontier beyond which diversification adds no boost to returns.

That was step one in modern portfolio theory. Step two was the articulation of beta. That goes back to the work of William Sharpe, among others. His Capital Asset Pricing Model established the market portfolio as the sum of all risky assets. And instead of analyzing the correlation of individual stocks to each other, his model looked at their correlation to the market portfolio.

Beta, said Sharpe, is the variability of an individual security’s returns against the market return—usually a benchmark such as the S&P 500. But in popular parlance, it’s come to stand in for the market return itself. CAPM involves more than that.

Securities have to be evaluated against a risk-free asset like a treasury bill. The result is the capital markets line, which tracks a portfolio starting with a risk-free asset and slopes upward as riskier assets are introduced until it reaches a portfolio of risky assets only. The higher the risk, the higher the reward.

Simple theory; the reality is less so. “There are a few different ways to define beta, but really, it’s a description of the risk/return characteristics of a particular asset class,” says Michael Cooke, head of distribution, PowerShares Canada.

“It captures the performance characteristics of a particular market or asset class, for example equities or fixed income, and it is representative of the total invested capital in that particular market. So, in theory, it’s reflective of the beliefs and expectations of every market participant about the future prospects of individual securities, and in aggregate it reflects the market portfolio.”

Adds Rotman School of Finance professor Eric Kirzner, “A high-beta stock is expected to be more volatile than the market. When the market is strong, a high-beta stock [tends to] go up more than the market; and when the market is weak, a high-beta stock [tends to] go down more than the market.”

But, he notes, in practice, betas are not the best way to pick individual stocks.

“Beta is not a terrific measure for individual stocks. Although the Capital Asset Pricing Model shows that there is a strong relationship between return and volatility, betas are quite unstable for individual stocks. So using beta for individual stocks is probably not a very powerful tool.”

Kirzner says it works much better on a portfolio level when the beta of that portfolio can be benchmarked against a relevant index.

Built-in unpredictability

All the same, Tyler Mordy, director of research and co-CIO at HAHN Investment Stewards, comments, “I love what Andrew Lo at MIT says—that finance suffers from physics envy. We would like our models to be as predictive as they are in physics. You labour in this business for a while and you realize that humans run financial markets. So things aren’t as predictive.”

He finds price variability too restrictive a definition of risk. “If you think about short-term variability and you look back, even if you took a broad-based volatility measure five years ago […] volatility was heading lower right into the financial crisis.”

Beta may not be predictive on the risk front. But there are also problems on the return front, particularly when stock market beta is compared to the beta of other asset classes. Here, stocks are assumed to be riskier, and therefore should fetch a risk premium.

But, “if we’re going to capture that positive risk premium,” says Cooke, “we have to buy assets when they’re cheap and sell them when they are expensive. In fact, the opposite holds in the market-capitalization-based portfolio,” because it forces you to do the opposite.

The performance of emerging market bonds and stocks since 1994 offers a good example. Emerging-market stocks return 5.6% annualized, making it look like the additional risk was compensated.

But that wasn’t the case, he notes. “Money-market instruments in emerging markets annualized 7.3% a year. In other words, I underperformed cash by 170 basis points per annum—a very stark example of a negative risk premium.”

The disappointing returns, Cooke says, stem from investors’ perception of beta. “A lot of foreign investors view emerging markets [through the lens of] a local bank, utility, cement manufacturer, pharmaceutical company, or energy company.”

So much so, Cooke says, that “The top 10 stocks in the Russian stock market count for 81% of the total market capitalization of the index. That wouldn’t be a problem if equity markets in general, and emerging markets in particular, were efficient. But history has shown they’re not.

“During the same period, not once did the top 10 stocks in any emerging market index collectively outperform the rest of the market over a subsequent five-year period. And yet that’s where the cap-weighted index is putting most of investors’ capital.”

So index construction may capture a stock’s market capitalization, but not its economic footprint or weight in the real economy.

There are two issues here. CAPM refers to all risky assets. But not all are represented on an investable index. One example would be privately held companies.

The other problem is finding a better representation of the economic weight of tradable companies that are components of an investable index.

“There’s something practitioners call the beta puzzle,” explains Mordy. “The Capital Asset Pricing Model suggests that higher-beta stocks should have higher returns. Empirical evidence suggests that it’s actually the opposite.”

Like Cooke, he argues capitalization-weighting is driven by high valuations, and valuations need more attention as a risk factor if investors are using beta as a proxy for risk.