The Expense Paradox

The term paradox has a number of meanings.  When the subject of fund expenses is raised a couple of its meanings might apply.

In recent years, more often than not, the impact of fund expenses on fund returns has been included in discussions of the relative merits of two or more funds.  Retail investors are informed, primarily by means of the media, to avoid “expensive” funds because of an expected adverse impact of fund expenses on future returns. Fiduciaries are counseled to avoid funds (and share classes) that have “above average” expense structures.  If any of this is to make sense, we should be able to find a clear relationship between fund expenses and returns.

It’s easy to construct a case where the admonitions make sense. We can do so by comparing the returns, excess over bench in this example, of all of the share classes of the same fund, as shown below for the Allianz NFJ Small Cap Value, and by relating them to associated expenses.  (Keep in mind that mutual fund returns are reported net of expenses.)

Upon doing so, a clear indirect adverse relationship between expenses and returns is apparent.  But this connection is not surprising because the underlying portfolio return of each of the share classes of the fund is the same, so return differences among the share classes correctly may  be attributed to their respective expense structures.  (Let’s consider loads to be trade commissions and exclude them from this discussion.)

Assuming that you have decided to invest in the Allianz fund and that you have complete freedom to choose one share class from among all of the share classes, common sense  seems to dictate the selection the institutional (Instl) share class.  It has the highest return, which necessarily means that it has the lowest expense ratio.

It’s reasonable to expect to see essentially the same return/expense pattern when examining other multi-share class fund products.  So let’s broaden our focus by including another fund, Van Kampen Small Cap Value.

The return/expense pattern for the Van Kampen fund also is as one might expect.  With the exception of the B shares, all of its share classes have both a higher return and a higher expense ratio than the competing Allianz share class. All other factors being equal (of course they are not), wouldn’t it be prudent to favor the Van Kampen fund even though faced with higher expenses?  The answer clearly is yes, that is, before considering other important factors, such as risk-adjusted return.

When faced with the need to evaluate a large number of competing products, the task gets messy. The next chart displays all share classes of all small cap value funds. Right off it’s apparent that we are looking at data and not information.  A clear relationship between expense and return is not discernable by the eye, nor is it evident from statistical analysis.

What if the focus is narrowed to a single share class to remove the influence of the multi-share class expense structure, as shown in the next visual? Will it help to clarify our perception of the relationship between return and expense?

We certainly do get a reduction in “noise”, and our attention is directed to the funds appearing in the upper left quadrant formed by the intersection of the median return and median expense values of the plot. But do you see a meaningful relationship between expense and return?

When examining other share classes of the two funds, the result is the same, as is the case when inquiring into most other sectors of the market.  However, as you might expect, the returns of index funds tracking the same benchmark definitely are affected by the level of expenses.  The same holds for certain bond fund categories where gross excess return is so small that expenses have a prominent negative effect on them.

What can we take away from this brief exposition? First, it may be unwise when evaluating funds to place undue emphasis on a single factor, such as the expense ratio. Second, advisors ought to arm themselves with the information needed to counter the misdirection by the media and by some marketers of “low cost” funds on the subject of expenses. It might be useful to have on hand visuals like the ones shown above. A comparison of risk-adjusted return versus expenses could be very effective.  And third, while most vendors of mutual funds do not seem to feel that they have the ability to charge higher fees for superior results, many vendors of poorer performing funds have no such reluctance. Why allow this practice to continue?

(Edited 10/22/09)