Risk and Return Revisited
They're correlated, but not in the way many investors think.
To generate better returns, investors should take on more risk, right? The answer may seem obvious, but historically, it's depended on what the meaning of the word "risk" is.
In an absolute sense, it's of course true that the more money investors put at risk, the more they'll gain if their investments increase in value. However, with the kind of risk that's become synonymous with volatility in investment argot, the opposite has been true. On average, high-beta investments--those whose prices have swung wider than an index over a given period of time--have historically generated worse returns than less-volatile alternatives.
The pattern has been persistent. This study, which appeared last year in the CFA Institute's Financial Analysts Journal, found that between 1968 and 2008, a portfolio comprising the least-volatile quintile of the market's 1000 largest stocks swamped the most-volatile quintile over the course of 40 years. And in this explanation of why boring can be beautiful, Morningstar ETF analyst Samuel Lee cites the work of Lasse Pedersen and Andrea Frazzini. In this 2011 paper, the duo find better risk-adjusted returns resulting from "betting against beta" across a broad range of asset types and geographic boundaries over a 50-year time frame.
History doesn't always repeat. Over a lengthy stretch of time, though, investors have fared better by taking on less risk, not more.
Beta and standard deviation are widely adopted gauges of volatility, but for most fund investors, Morningstar Risk and the Sortino ratio are arguably more relevant. Few investors fret over oxymoronic "upside risk," after all, yet beta and standard deviation are agnostic when it comes to assessing upside versus downside volatility. Not so Morningstar Risk and the Sortino ratio, each of which (although in different ways) penalizes downside performance gyrations. The wilder the ride has been amid losses, the poorer a fund's Morningstar Risk Rating and Sortino ratio will be.
For this article, I examined a subset of domestic-equity funds to see if, after substituting Morningstar Risk for beta, the historical data would continue to show a low-volatility premium. To be included in the subset, a fund needed a track record of at least 10 years as of January 2012, a requirement that resulted in a group of roughly 1,700 funds. In addition to these funds' annualized 10-year trailing returns, I reviewed their percentile category rank for both total return and Morningstar Risk. Sorting the funds according to the latter data point controlled for style, allowing for analysis across risk groups comprising the least to most risky funds in each of Morningstar's domestic-equity categories.
After segmenting the funds into quartiles, here's what I found.
The top-quartile (least risky) funds in the subset returned an annualized 5.34% in the trailing 10-year period through January 2012, surpassing the bottom-quartile (most risky) funds by an annualized 0.84%. The typical top-quartile fund placed in the 41st percentile of its category; the average bottom-quartile fund ranked in the 50th percentile.
Given that two bear markets have occurred during the past decade, it stands to reason that the historical data would reflect a low-volatility premium during the measurement period. But while the results outlined above show that, at least for the subset of funds I examined, the premium persisted when Morningstar Risk was used to measure volatility, results were mixed in the second and third quartiles.
Although these segments' average annualized total returns did fall between those of the top and bottom quartiles, the third quartile's 5.03% gain edged past the second quartile's 4.83% showing. The third quartile slipped ahead of the second in terms of category placement, too, with the typical third-quartile fund ranking in the 45th percentile of its category. The average second-quartile fund placed in the 49th.
The study also reinforced the significance of style as a factor in mutual fund performance.
In the initial pass at the data, I controlled for style by sorting the data so that each category's least risky funds appeared in the top quartile while the most risky appeared in the bottom. In a second pass, I segmented the funds into quartiles based not on their percentile category rank for Morningstar Risk but on the style-agnostic Morningstar Risk score.
Irrespective of category, the higher that score is, the riskier a fund has been. Rank ordering the funds based on the score produced results that seemed dramatically different from those of the style-controlled test. Indeed, the data showed that the bottom-quartile funds outperformed the top quartile by almost 2 annualized percentage points.
That isn't a surprise, however.
During the measurement period, small-cap funds held sway over larger rivals; in the subset I reviewed, the typical small-cap fund gained 6.5% annualized versus a large-cap norm of 3.6%. Given that the vast majority of funds in the style-agnostic grouping's top quartile were large-cap vehicles and that small caps dominated the bottom, style appears to have been a larger performance contributor than the low-volatility premium the style-controlled test identified.