Zaks Investment Advisory Service

OK, But Could We at Least Pick Up a Good Mutual Fund?

We know that despite all the stock analysts’ work, it is practically impossible to consistently identify stocks that will outperform the market on a risk-adjusted basis. Which, by the way, doesn’t mean that what the stock analysis do is meaningless: it is partly thanks to their diligence in studying financial statements, analyzing industries and making intelligent assessments of companies’ future incomes that market is efficient, making it impossible to stock-pick. (Of course there are stocks that perform much better than others, and sometimes stock analysts identify those – but they also pick the duds and on average perform no better than the market).

Nevertheless, there is a notion of legendary portfolio managers who consistently outperformed the market, Peter Lynch of Magellan fund being one of them. Lynch ran Magellan from 1970 to 1990 and indeed was tremendously successful, achieving the average return of 29%. His successors were not as lucky (or talented) and during the last 10 years, for example, the fund was generally under-performing S&P 500. It’s interesting to note that if one takes the data from March 1987 – the only readily available long-term data we could find – and compares Magellan with the Vanguard’s S&P 500 index, it turns out that while Magellan (FMAGX) outperformed Vanguard (VFINX) on the pre-tax basis (10.81 vs. 10.48 average annual return for the period), on the after-tax basis Vanguard performed better: Magellan’s average dividend yield during this period is 6% as opposed to Vanguard’s 1.8% (this reflects a much higher turnover ratio at Magellan than Vanguard). If we use an average tax rate of 30%, the after-tax rate of return on Vanguard would be 10.38% and on Magellan – 9.01%.

This leads us to the following question: is it possible to pick up an actively managed fund that would outperform the market with at least some consistency? We decided to perform the following experiment: Charles Schwab, using the data from Morningstar, lists 625 funds that were around long enough to accumulate at least a 10-year history of annual returns. Morningstar also provides the average returns for the last five, three and one year. From the original list we eliminated all bond funds and index funds. What remained were 459 funds (we left specialty funds, such as REIT, gold and other commodity funds, and high-yield bond funds on the list). We should note that we are interested in a correlation, if any, between the fund’s performance during different periods; we are not interested in the actual returns per se. Therefore, we do not attempt to calculated average returns, which would requite us to use market weights of each fund.

From the data provided, we calculated the average annual returns for the last 5 year and for the previous 5-year period and subtracted average market (S&P 500) returns to arrive at the relative performance numbers (March 1996 through March 2001 and March 2001 through March 2006). We also calculated two subsequent 2-year periods (March 2001 through March 2003 and March 2003 through March 2005).

If there was consistency in the mutual funds’ performance, the ones that out-performed the market during one period would out-perform it in the subsequent period, and the ones that under-performed the market would do the same in the following period. So, if that were the case, we would see a chart that would look something like this:

And indeed, if we were to compare the average returns during two consecutive 2-year periods we would find that there seems to be a certain correlation between returns: there are more funds that performed either better than the market during both periods or worse than the market during those periods than the funds that over-perform during one period and under-perform during another one (to be precise, out of 459 funds, 254 over- or under-perform in both 2-year periods and 205 over-perform in one periods and under-perform in another one. Statistically, this is not a significant difference but one could say that there is a better than 50% chance that if one were to pick a fund based on the previous 2-year experience, the fund would outperform the market in the subsequent 2-year period. (Again, we should note that these calculations are performed ex-post, and, in theory would have to be confirmed ex-ante).

The predictive power of the two-year period returns seems to be weak. Let’s now turn to the five-year periods. Results are presented on a chart below:

As we can see, this chart looks very differently. There seems to be very little correlation between the two periods. Again, let us stress that we are displaying here the results adjusted for the average market performance. For these two periods, of the 459 funds, 304, or more than half, under-performed the market in one period and over-performed it in another. It is almost as if the rule could be: “take a look at a laggard and pick it up for the future.” We realize that this should not be a rule, but only that the past performance of a fund seems to be a very bad predictor of its future performance.

So what advice we have to offer? One way of avoiding the problems discussed above is investing in index funds. These funds, almost by definition, mirror certain markets. Picking a winning mutual fund seems to be almost as difficult as picking a winning stock. It is wiser to diversify one’s holding across the appropriate asset classes and not worry about the short-term market fluctuations.