Past performance is no guarantee of future results. We all know the disclaimer. Nonetheless, we use past performance to shed some light on anticipated performance. We're not looking for a guarantee. Rather, we want a sense of how a particular equity mutual fund (for example) might behave in relation to the general equity market.

The problem is in the word "guarantee." So let's back away from that particular word and focus on correlation of performance over time, understanding that correlation does not imply a guarantee. Historical correlation simply conveys information about how funds have performed in relation to one another in the past. We'll also take a look at risk correlation, to see if standard deviation correlates to future performance.

The data set for this analysis was the largest 200 U.S. equity mutual funds as of Nov. 30, 2009 (using Morningstar data). All the mutual funds in the study had to have at least 10 years of performance as of Dec. 31, 2008. Only distinct share classes were included.

PERFORMANCE CORRELATION

As shown in "Using the Past" on page 93, there are dramatic swings in the correlation of performance. The first bar in the chart represents the correlation of the performance of the 200 funds in 1999 with the performance of the same funds in 1998. The correlation was just over 0.40, which is fairly low. What this correlation suggests is that in 40% of the cases, a fund that had a positive return in 1998 had a positive return in 1999.

Recall that the range for correlation coefficients is -1.00 to +1.00. A negative correlation indicates that as one variable increases, the other variable decreases. A positive correlation indicates that as one variable goes up, the other does also. Correlation coefficients closer to -1.00 or +1.00 indicate stronger relationships, whereas correlation coefficients closer to zero indicate a very weak relationship.

In 2000, the correlation was -0.52. This indicates that in about 50% of the cases, a fund that had a positive return in 1999 had a negative return in 2000.

The correlation between 2001 and 2002's return was over 60%. This high positive correlation indicates that funds with a positive return in 2000 tended (over 60% of the time) to have a positive return in 2001, and likewise between 2001 and 2002.

Despite the good showing in those two years, the correlation between this year's and last year's performance has generally been weak. In eight of the 10 years, the correlation in performance from year to year was below positive 60% and below negative 60% (meaning between zero and -60%). The closer to zero (positive or negative), the lower the correlation. Prior performance is not a reliable indicator of future performance. The standard disclaimer is vindicated.

RISK CORRELATION

How about using a risk measure to gauge performance going forward? If there is a correlation between risk and return, let's use it to see more clearly into the fog of future performance.

Standard deviation is the typical risk measure. A fund with a high standard deviation tends to have annual returns that vary significantly from (above and below) the average return of the fund. The larger the standard deviation of return, the higher the risk of the fund. This is the typical interpretation of standard deviation as a measure of performance risk.

Standard deviation is not without its drawbacks. The main problem is obvious. It assigns equal value to return variance above and below the mean return. But, of course, investors prefer upside variance and want to avoid downside variance. Having stated the obligatory concerns about standard deviation, let's examine its usefulness as a correlate (not a guarantee) of future performance.

The 200 funds were placed into their respective Morningstar categories for this part of the analysis. Only those funds in large value, large blend and large growth categories were used. Remember that this sample of funds represents the largest funds (by net assets) within these three categories.

As "Risk and Return" on page 94 shows, the correlation between volatility of past performance (as measured by standard deviation over the previous 36-month period) to subsequent one-year performance was a mixed bag. In 2001, the one-year performance of the 30 largest large-cap value funds had a -0.02 correlation with the prior 36-month standard deviation (over the three-year period from 1998 to 2000). This indicates no correlation between the two.