Minimum-ism

ou-get-what-you-pay-for." It's a basic economic mantra, and it's often true. But, when it comes to investing in mutual funds, it doesn't seem to apply.

This article is centered on the following question: "All else being equal, do U.S. equity mutual funds that have a low initial investment requirement under-perform funds with higher initial investment requirements?" Or, asked differently, "is the initial investment requirement in any way related to fund performance?"

The funds included in this analysis were extracted from the January 2005 release of Morningstar Principia, and thus represent data as of year-end 2004. Only distinct U.S. equity funds with at least five years of performance were included. The term "distinct" indicates that only one share class of funds that are marketed via multiple share classes was included. The "distinct fund" is either the fund with the largest net asset base or the earliest inception date, and often those two criteria go together. Funds that were closed (as of 12/31/04) were omitted from the sample. Finally, exchange-traded funds were also excluded inasmuch as they do not have a stipulated minimum initial investment in the same way that open-end mutual funds do. The total number of funds in this study was 2,276.

The first wave of analysis was to simply arrange the funds from low to high on the basis of their minimum initial investment (into a regular, non-IRA account). As shown in Figure 1, there were a total of 31 different minimum initial investment requirements, ranging from $0 to $100,000,000. For instance, there were 120 distinct, available U.S. equity funds with a minimum initial investment requirement of $0. Pragmatically, this generally means an initial investment of at least $1 or more. (In many cases, the normal minimum initial investment requirement is waived if the investor begins an automatic investing plan, but that is a separate issue. This study only focused on the stated minimum (i.e. lump sum) initial investment into a regular, non-IRA account.

The most common minimum initial investment requirement was $1,000. In fact, $1,000 was the initial requirement for just over 34% of the 2,276 funds. The next most common initial requirement was $2,500, followed by a requirement of $500. The fourth most common requirement was $0, representing just over 5% of the funds. It is interesting to note that there were no funds with an investment requirement in between $5,000 and $10,000, or in between $50,000 and $100,000, and so on. The 31 different minimum initial investment figures shown in Figure 1 represent all the possibilities in this sample of U.S. equity funds.

The second wave of analysis consisted of grouping the funds into minimum initial investment categories, as shown in Figure 2. As can be seen, funds in the "Low Min" category (minimum initial investment of $500 or less) represented 13.8% of the sample yet possessed nearly 21% of the assets of the sample. Low Min funds attract proportionally more assets than "High Min" funds. Moreover, Low Min funds were the largest funds - with an average size of $1.6 billion.

Funds in the second category ($1,000 - $5,000 minimum initial investment) represent nearly two-thirds of the funds in the sample, both in terms of percentage of funds (64.3%) and percentage of assets (65.0%). The average size of funds in the $1k-$5k category was just over $1 billion.

The 5-year performance data in Figure 2 suggest that funds with higher initial investment requirements have superior performance, particularly when comparing load-adjusted 5-year returns. The average 5-year, load-adjusted return for the 219 High Min funds was 2.54% versus 0.79% for the 314 Low Min funds. If the analysis stopped here these findings would lead to erroneous conclusions inasmuch as the data in Figure 2 do not account for style differences (value vs. growth) or capitalization differences (large, mid, small).

The third wave of the analysis is reported in Figure 3, where funds were further categorized by "Best Fit Index" (which accounts for both capitalization and style orientation). Of the 2,276 funds in the analysis, there were 276 that had the Russell 1000 as their "best fit index" (as determined by Morningstar). Those 276 funds were then sorted on the basis of their minimum initial investment (from low to high) and then divided into quartiles (i.e. 25% of the group).

The 1st quartile consisted of the 69 funds (276 divided by 4) with the lowest minimum initial investment requirement (Low Min) and the 4th quartile consisted of the 69 funds with the highest initial investment requirements (High Min). The mean initial investment requirement of the 1st quartile was $536 and $1.4 million for the 4th quartile, which are statistically different means at the 99% confidence level. The first figure in each box in Figure 3 is the mean for the 1st quartile, or the Low Min funds. The second figure in each box is the mean for the 4th quartile, or the High Min funds.

In Figure 3, we observe that that the average front-end load for the 1st quartile (or Low Min) funds which have the Russell 1000 as their best fit index was 1.91% and 0.43% for the 4th quartile (or High Min) funds. The difference between those two means was statistically significant at the 90-99% confidence level. The 5-year return (regular and load-adjusted) between the 1st and 4th quartiles was not significantly different, suggesting that minimum initial investment had no bearing on performance over this particular 5 year period. Conversely, the expense ratio for the Low Min funds was significantly higher than the High Min funds (as noted by the green shading).

Throughout Figure 3, every box shaded green indicates that the Low Min funds had a statistically higher mean than the High Min funds. Purple shading indicates the opposite, that the High Min funds had a statistically higher mean than the Low Min funds.

Finally, for the Russell 1000 "best fit index" funds, the mean 5-year tax cost ratio for the Low Min funds was lower than the mean for the High Min funds (where the lower the Tax Cost Ratio the better). However, the box is not shaded purple because the difference between the two means was not statistically significant at the 90-99% confidence level. The figures 0.72 and 0.87 are in bold however, which indicates that the difference in the 1st and 4th quartile means was significant at the 80-89% confidence level. Every other non-shaded box with figures in bold indicates an 80-89% confidence level difference in means.

The striking findings in Figure 3 are as follows:

The difference in minimum initial investment between the 1st quartile (Low Min) and 4th quartile (High Min) is statistically significant at the 90-99% confidence level in 12 out of the 13 categories (four large cap categories, five mid cap categories, and 4 small cap categories). In the remaining category the difference was significant at the 80-89% confidence level. The "Minimum Initial Investment Requirement" column in Figure 3 was intentionally not shaded purple so as to focus the attention on the differences among the other variables in this study (front load, return, expense ratio, and tax cost ratio).

Low Min funds have front-end loads (a commission on the purchase of fund shares) and in most cases High Min funds do not. In fact, in only 4 categories (Russell 1000, Russell 1000 Value, Russell 1000 Growth, Russell 2000 Value) did the High Min funds have a non-zero, front-load mean. In all 13 categories, the mean front-load for Low Min funds was statistically higher than the mean front-load for High Mean funds (in 12 cases at the 90-99% confidence level).

In no case was the mean 5-year annualized return (not load-adjusted) for High Min funds statistically higher than the mean return for Low Min funds. Rather, in eight cases, the Low Min mean 5-year return was higher (in two cases it was statistically higher, once at the 90-99% confidence level and once at the 80-89% confidence level).

When comparing 5-year load-adjusted returns the findings were mixed. In seven categories High Min funds had a higher mean return, but only one difference (Wilshire REIT) was significant at the 90-99% confidence level. In the Russell 2000 Growth category, the mean return of -0.6% for the High Min funds was higher than the -2.9% mean return of the Low Min funds, but only at the 80-89% confidence level. In six cases, Low Min funds had a higher 5-year load-adjusted return, but in only one case (S&P 500) was the difference statistically significant.

In all but one box, the expense ratio column is entirely shaded green. This indicates that Low Min funds have higher expense ratios. This is somewhat surprising considering that as shown in Figure 2, Low Min funds tend to be the largest in terms of net assets. In theory, as funds grow they should be able to reduce the expense ratio. Remember also that reported performance is "net" of the expense ratio, meaning that the expense ratio has already been subtracted out. Therefore, the comparable (or slightly superior) performance manifested by Low Min funds compared to High Min funds is all that much more impressive.

Finally, in 10 of 13 categories, Low Min funds had a mean tax cost ratio that was lower than High Min funds (where the lower the tax cost ratio the better). In four cases (noted by purple shading) the mean tax cost ratio for High Min funds was significantly higher than the mean tax cost ratio for Low Min funds.

In summary, we suggest that as it pertains to mutual funds, "What you pay (initial investment requirement) doesn't have much to do with the performance you get." Clearly, in the short run (i.e. 1 year) High Min funds will likely outperform Low Min funds, but only because of the impact of the front-end load. Over time, however, the front-end load drag on performance (front-end loads being more common among Low Min funds) is diluted.

When it comes to the minimum initial investment, small is beautiful…particularly when there's no load attached.

Figure 1. From "Low Min" to "High Min"

Figure 2. Distinct, Available U.S. Equity Funds with a 5 Year Performance History

Figure 3. Low Min vs. High Min Performance by Specific Cap & Style Categories

____________________________________________________________________________________
Craig L. Israelsen, Ph.D. is an associate professor at Brigham Young University. He teaches family finance in the Department of Home and Family Living. His research interests include mutual fund analysis. He writes monthly for Financial Planning magazine.

Erin Lindsley is an undergraduate student at Brigham Young University.