ETF wisdom from the dugout

Yogi was my favourite Yankee. Not only is he the greatest catcher ever, but also, as a kid, I liked his friendly smile. (It wasn’t until years later that I discovered his propensity for malapropism.)

I’ve often thought the investment business would be a richer place if Yogi went to Wall Street. The investment business is at once simple and complex—a perfect backdrop for Yogi. Prices go up when there are more buyers than sellers and vice versa. Millions of dollars are spent trying to determine what’s going to happen before it does. Millions also go into explaining why that research got it wrong.

Risk equals return

For investors, the risk-return relationship is very important. We’re taught more return comes only with a commensurate amount of risk. Of course, the definitions for risk vary depending on the audience. For most clients, risk is about losing money—absolute loss.

Professionals use standard deviation to describe risk, partly because it can be measured and partly because it’s persistent—and therefore somewhat predictable.

Don’t believe me? Daily standard deviation of the S&P 500 has an auto-correlation of 0.67, while the correlation of returns is 0.02. This means today’s volatility tells you something about tomorrow’s volatility, but today’s returns tell you virtually nothing about tomorrow’s returns. Hence Yogi’s conundrum: if risk is predictable but return is not, how can they be equal?

Low volatility strategies: U.S.

In their publication Low-Risk Stocks Outperform Within All Observable Markets of the World, Nardin Baker and Robert A. Haugen measured the 24-month volatility of U.S. stocks from 1990 to 2011.

They then put them into deciles from low-risk to high-risk, and plotted their annualized returns (see “Low volatility: United States,” left). They repeated this calculation monthly during the testing period. A relatively uniform relationship between low volatility and higher returns was the result. This is a surprising result if you believe risk equals return.

Baker and Haugen also measured the 24-month volatility of emerging-market stocks from 2001 to 2011 (see “Low volatility: emerging markets,” left). The relationship between low volatility and higher returns was not as uniform as in the U.S., but does exist.

The frequency of rebalancing (monthly) influences results, but the main point of interest is that an investor loses less money during volatile periods.

Why? My colleague, Ioulia Tretiakova, director of Quantitative Strategies, has an as-yet-untested hypothesis that this is a statistical phenomenon related to volatility drag. In other words, any combination of securities that reduces volatility will show a return benefit, particularly during periods of higher volatility.

Manufacturers are introducing low-volatility ETFs to exploit the “Yogi conundrum” described above. If you compare the BMO Low Volatility ETF (ZLB) with the iShares S&P/TSX Composite ETF (XIC) from the end of October 2011 to the end of April 2012, ZLB appears to have avoided two spikes in market volatility, one in November-December 2011 and another in April 2012 (see “Low volatility: Canada,” below, left).

While far too short a period from which to draw any solid inferences, the price history is instructive. ZLB is actually a low-beta ETF that measures sensitivity to market movement. It could benefit from relabelling.

In Canada, Powershares has also issued low-volatility ETFs using different methodologies. Powershares S&P 500 Low Volatility (ULV) ETF is based on a more traditional calculation of volatility, standard deviation. The Powershares S&P 500 High Beta (UHB) uses beta as the volatility measure, similar to BMO’s ZLB.

There will be new ETFs that use different aspects of volatility and other factors that will benefit investors. We need to continually question how we manage money. As far as investment research goes, “It ain’t over till it’s over.” And for portfolio managers, “It ain’t the heat, it’s the humility.”