Understanding Alpha and Beta

What are alpha and beta?

The return on any portfolio consists of alpha and beta return.

Beta is the part of a return that is correlated with the markets, and alpha is the uncorrelated part that comes from active management.

Exposure to beta can be obtained cheaply – through low fee indices or ETFs. Everyone who wants beta exposure can capture it.

Alpha is a zero sum game. Investing in alpha managers makes sense only if the right manager is selected. Alpha is not available to everyone.

Is it important for investors to know the difference?

Yes. Without knowing the difference, investors will be less diversified than desired and end up paying higher, active management fees for beta exposure, which could be obtained cheaply.

Managers who derive the bulk of their return from exposure to risk factors can be classified as beta grazers.^[i] Investing in beta grazers provides less diversification, because markets and hence betas become highly correlated when volatility increases. Only real alpha provides true diversification.

For example, the correlation of hedge funds to global equities is 4% when global equities produce returns greater than one standard deviation above their mean return, but it rises to 80% when equities generate returns more than one standard deviation below their mean (see Exhibit 1). Thus, diversification tends to fail exactly when it is most needed, i.e. in falling markets. Beta labeled as alpha is still beta.

Exhibit 1: Diversification Often Fails When It Is Most Needed

How can we separate alpha from beta?

The Arbitrage Pricing Theory (APT) offers a framework for modeling the expected return on any portfolio. APT posits that the expected return can be expressed as a linear function of certain macro-economic factors or market indices.

For example, Fung and Hsieh (2002) ^[ii] used an Arbitrage Pricing Theory (APT) to model hedge fund returns. They use seven hedge fund risk factors, i.e.

Return = Alpha + Beta1 + Beta 2 + Beta3 ...

This is similar to how the APT models have been used to relate equity returns to various betas such as firm size, book value to market value ratios, market price to earnings ratios, and so on.

Could you provide an example for alpha and beta separation in equity and fixed income?

Yes. The APT model provides the framework, i.e.

Equity Portfolio Return = Alpha + Market Beta + Small Size Beta + Value Beta

Bond Portfolio Return = Alpha + Duration Beta + Credit Beta + Prepayment Beta

What is the correct list of beta factors in any given asset class?

This is a matter of empirical analysis as long as two conditions are satisfied. First, there must be complete transparency in how factor returns are derived. Second, the return series must be sufficiently long to produce reliable statistical results.

The numbers of factors might vary, but the framework is unchanged. The part of the return which stems from exposure to risk factors and could be easily replicated by simple trading strategies is beta and investors shouldn't pay active fees for it. The real alpha is what is left after accounting for such exposure. Simply assuming that all return represents alpha would mean to pay for skills where no skill is present. Exhibit 2 demonstrates that real alpha is often much smaller than previously assumed after accounting for exposure to the appropriate risk factors.

Are the returns from currency management alpha or beta?

As with equities and fixed income, the returns from currency management can be separated into alpha and beta components once we identify the appropriate risk factors.

Pojarliev and Levich (2008)^[iii] propose four factors to model currency returns. The model specification takes the form.

Currency Portfolio Return = Alpha + Carry Beta + Trend Beta + Value Beta + Volatility Beta

Exhibit 2: There are a Multitude of Betas