Two New Ways of Looking at Fund Risk

One of the most notable effects of the past year's horrendous bear market has been a dramatic reduction in investors' appetite for risk. Risky investments such as emerging-markets stocks and high-yield bonds had a great run before the credit crisis hit, but over the past year badly burned investors have turned away from them in favor of relatively stable, predictable fare such as Treasury bonds and consumer-staples stocks. This effect has been evident in the differing results of various mutual fund categories. The worst-performing categories in 2008 have been Latin American stock and diversified emerging markets, in which the average fund is down more than 55% through December 12. The best-performing categories (apart from bear-market funds, which have had a field day in this market) have been the government-bond categories, led by long government with an average gain of 20%.

It has been more difficult to predict the behavior of diversified domestic-stock funds, which form the core of most people's portfolios, based solely on their category. Average returns for the nine categories in the Morningstar Style Box  range from a 37% loss for small value to a 47% loss for mid-cap growth, but each of these categories includes a variety of funds with different risk levels. In fact, there are various types of risk that funds can exhibit, and some are more relevant to the current market situation than others. Standard deviation and other backward-looking risk measures tell you how volatile a fund has been in the past, but they don't always reflect future risks accurately; some hard-hit areas, such as emerging-markets stocks, exhibited unusually low volatility in the years leading up the current crisis. Portfolio measures such as average valuations and growth rates can give a general but imperfect sense of the risks in a stock fund's most recent portfolio, where high valuations and growth rates tend to correspond to greater risk. (For more on the various types of fund risk, see this article.)

Moats and Uncertainty
All these risk measures can be supplemented with two other portfolio measures that are especially relevant right now: average moat rating and average fair-value uncertainty rating, based on two of the key ratings that Morningstar's stock analysts assign to each stock they cover. A stock's economic moat represents the strength of its competitive advantages, if any. A wide-moat company such as  Coca-Cola (KO) or  Wal-Mart (WMT) has strong advantages that help keep potential competitors at bay; a narrow-moat company has less-compelling advantages; while a company with no moat lacks such advantages, and will likely face a lot of competition if it gets too successful. (You can read more in this definition.) Fair-value uncertainty measures how confident our analysts are in the fair-value estimates they assign to a stock, based on our discounted cash-flow models and analysis of possible scenarios that might occur. Companies with low fair-value uncertainty, most of which are large consumer or health-care stocks, tend to have very stable businesses with predictable cash flows; those rated medium have cash flows that are somewhat less predictable; and so on up to stocks with extreme fair-value uncertainty, many of which face serious problems. (This article provides more information about the rating.)

On an anecdotal level, it seems clear that stocks with wide moats and/or low fair-value uncertainty have been attractive in the current volatile and uncertain market, and that funds with a heavy concentration in such stocks have held up better than their peers. We decided to test this observation by quantifying the weighted average moat and uncertainty ratings for a large group of funds. We focused on the roughly 1,400 funds in the large-value, large-blend, and large-growth categories, because the vast majority of their holdings are covered by Morningstar analysts and thus have ratings. Each stock's ratings were translated into numerical values: a wide moat = 3, narrow moat = 2, no moat =1; low fair-value uncertainty = 1, medium = 2, high = 3, very high = 4, and extreme uncertainty = 5. Then we took a weighted average of all the moat ratings in each fund's portfolio, so that larger holdings count more than smaller ones, and did the same for each fund's fair-value uncertainty ratings.

When we rank all the funds according to these averages and look at how they've done over the past year relative to their peers, the results are striking. Because there's so much information, we've included several key tables in a PDF document.

Table 1 in the document shows the 20 funds in our group with the highest average moat rating as of November 30, including each fund's percentile ranking in its category for the past 12 months as of December 8. Of the 18 funds on this list that have been around for at least a year, 17 have beaten their category, and 14 rank in their category's top decile. The one fund trailing its category is Rydex Dow 2x Strategy (RYCYX), whose use of leverage has caused it to lose much more than the typical large-cap fund. Table 2 shows the 20 funds with the lowest average moat rating, also with each fund's ranking within its category. Here the percentile rankings are almost a mirror image of those in the first table; of the 18 funds with one-year records, 16 have trailed their category, with half of those ranking in the bottom decile. These are not necessarily bad funds, either; some of them have solid long-term records, and  Schneider Value  is even a Fund Analyst Pick in the large-value category. But most of these funds are heavy in financial and/or commodity stocks, which tend not to have moats and have been hammered in this year's down market.

A very similar pattern emerges when we look at the funds with the lowest and highest average fair-value uncertainty. (Unlike moats, with which a higher number is better, a lower number is better when it comes to the uncertainty rating.) Table 3 shows the 20 funds in our group with the lowest average fair-value uncertainty ratings as of November 30, some of which are the same as in the high-moat list. In fact, the same fund, John Hancock Large Cap Select , tops both lists thanks to a concentrated portfolio full of blue chips such as Wal-Mart,  Johnson & Johnson (JNJ), and  Procter & Gamble (PG). This group has held up even better than the high-moat funds; 19 of these 20 funds rank in their category's top quartile over the past year (once again Rydex Dow 2x Strategy is the exception), and 17 of them rank in the top decile. Table 4 shows the funds with the highest fair-value uncertainty, and once again they are almost a mirror image of the previous table; an amazing nine of the top 10 funds on this list rank in their category's bottom decile over the past year, and only two of the 20 have beaten their categories over that time.

Turning Back the Clock
These figures certainly look compelling, but somebody might object to using current portfolio data with returns over the past year, because a fund's portfolio may have been significantly different a year ago. Therefore, we also looked at each fund's most recent portfolio as of Nov. 30, 2007, roughly one year ago, and calculated weighted-average moat and risk figures in the same way we did for the current portfolios. These figures allow us to see which funds had the highest and lowest average ratings a year ago, and to see how those funds performed over the subsequent 12 months.

The results are very similar to those we saw for the current portfolios, with a few interesting outliers. Table 5 shows the funds with the highest average moat ratings a year ago and their relative performance since then. Not surprisingly, some of these are the same funds we saw in Table 1, so it's also not surprising that most of this group feature excellent one-year returns, with 14 of the 20 funds ranking in their category's top decile. The major exception is  Thompson Plumb Growth (THPGX), which has been pummeled to bottom-decile returns over the past year. Its presence here becomes understandable when we realize that this fund was hurt this year by big positions in  AIG (AIG),  Fannie Mae (FNM), and  Freddie Mac (FRE)--all of which we categorized as wide-moat stocks a year ago, and all of which lost essentially all their value in this year's financial meltdown. This example shows that moat ratings are not infallible, though to be fair virtually nobody predicted the collapse of these venerable companies, and the fact that such exceptions are so rare and dramatic speaks well of the accuracy of the moat ratings overall.

The rest of the tables also depict similar results. Table 6 shows the funds with the lowest average moat ratings a year ago, and once again these funds have been overwhelmingly poor performers over the past year. Morningstar's fair-value uncertainty ratings in their current form didn't exist a year ago, but Tables 7 and 8 show the funds with the lowest and highest average risk ratings back then; this was the predecessor to the fair-value uncertainty rating--not the same but similar enough for our purposes. As we saw with the similar Tables 3 and 4, the low-risk funds as a group have done exceptionally well relative to their categories, while the high-risk funds have done poorly as a group. The most notable exception is manager Ken Heebner's CGM Mutual , which has been one of the best-performing large-growth funds of 2008 despite being among those with the lowest average moat and highest average risk a year ago. That's because a year ago, this fund's portfolio consisted almost entirely of energy and commodity stocks, most of which had no moat and high risk. These stocks went on a tear in the first half of 2008, but Heebner has since sold almost all of them amid plunging commodity prices. This fund and Heebner's other offering,  CGM Focus , are so distinctive that it's probably not a good idea to draw any conclusions from them.

What It All Means
Different market environments call for different portfolio metrics that can adequately explain varying performance among funds. Average moat and fair-value uncertainty have clearly been the metrics of choice in separating the winners and losers over the past year, given their strong correlation with fund performance. It's important to reiterate that a low average moat or high average uncertainty aren't inherently bad, as there are some good funds with these characteristics. But such funds will behave very differently from high-moat, low-uncertainty funds in many environments, and these measures highlight these differences in a tangible way. They give investors some valuable new forward-looking tools for evaluating portfolio risk, using the insights of Morningstar's stock analysts to indicate how a fund is likely to hold up in the event of an economic downturn or other disruption. They can be used for any portfolio in which a significant percentage of the holdings have moat and fair-value uncertainty ratings. Thus, they're currently most helpful for the domestic large-cap stock funds we've been looking at, but they could also be helpful for at least some domestic mid-cap and small-cap funds, and foreign stock funds. In all these cases, a fund's average moat and uncertainty ratings need to be put into context relative to a peer group to be really useful--either by ranking, as we've done here, or by comparing them with a category median or average.

Of course, the past year has been an exceptional time in the markets, and the correlation between these two measures and fund performance will almost certainly be less striking in normal markets; there may even be a negative correlation in some speculative markets, such as those of 1999 or 2003. But the fact that these measures have worked so well in this exceptional market (with a few isolated exceptions) is a testament to the validity of the moat and fair-value uncertainty ratings, and to the analysts who assign those ratings. There's plenty more work that can be done on this topic, but this is at least a starting point.