One aspect of ETF investing that is often listed as a "downside" when compared to investing in individual companies is the "expense ratio" (ER) - that usually small percentage (most frequently less than 1%) that represents the costs incurred by the fund's issuer in the course of managing the fund.
There is no "expense ratio" involved in owning stock in a company.1
While a lot of ETF investors pay close attention to a fund's ER, I tend to put very little weight on it when deciding whether to invest in a fund. In fact, for many funds, I do not look at it at all. Why? Because on the whole I do not think that spending a lot of time worrying about ERs is needed, and is not really useful. I intend to explain why I feel this way in what follows.
What is the Expense Ratio?
In its most basic concept, the ER reflects the costs involved in managing a fund.2 However, while the basic concept behind the ER may seem fairly straightforward, it is important to pinpoint as precisely as possible what it represents and how it is reached.
The following items are typically included as components of the ER:
- Manager/advisor fees
- Custodial services
- Legal expenses
- Accounting/auditing fees
- 12b-1/marketing costs
What is not included in the ER:
- Costs associated with stock transactions (stock price, commissions, etc.).
Management fees can be found in a fund's annual report. The report should have a section titled "Statement of Operations," which will include income and expenses. The problem is that issuers, while required to report the data to investors, are not required to identify exactly how much was spent on what particular item.
Many annual reports will simply identify the item "Management Fees," with a catchall "Other Fees" entry for anything that doesn't fit under the main category; others will provide a very detailed listings of fees.
Lack of standardization limits the extent to which the investor can make meaningful comparisons between funds, and there have been recent calls for making some enhancements to reports.3 It would certainly seem to be desirable to introduce some level of transparency and uniformity to these expenses that would articulate expenditures more than just a vague, overly general category.
Of the items that are included in the category of "Expenses" many elements are of a nature that would make them vary from issuer to issuer, and even fund to fund. Amongst these are:
- Number, and salaries, of members of the management team
- Size, and cost, of administrative support staff
- Costs associated with accounting, advisors and legal counsel
- Physical facilities
There would also seem to be two meta-variables, which influence, either directly or indirectly, the other variable expenses: the size of the fund and the nature of the fund. Obviously, funds with assets running into the tens or hundreds of billions of dollars may reap significant benefit due to economies of size in some respects, but will also realize increases in categories of expenses, when compared to funds consisting of a few billion dollars or less.
Perhaps not as obvious are the effects involved with the nature of a fund, either its particular focus or its dynamics. For instance, funds that have holdings on a global, international scale will encounter currency exchange and tax issues that do not apply to funds limited to U.S. holdings.
Also, the way in which holdings are weighted within the portfolio will influence costs. Most ETFs are market-cap weighted, which is a fairly inexpensive system to maintain. Holdings tend not to change their cap status vis-à-vis other holdings in the fund, making rebalancing an easy, inexpensive, task. Equal-weighted funds, on the other hand, will need regular monitoring, and may need to be rebalanced frequently, thus increasing maintenance costs.4
Calculating the Expense Ratio
To find a fund's expense ratio, the ETF's expenses included in the management fees are totaled annually, and that total is divided by the total assets currently under management by the fund (AUM). Thus, if a fund that had AUM of $25 million incurred $150,000 in applicable expenses, its ER would be 0.60%.
The average ER for all exchange-traded products was 0.62% (as of June 2013 TTM).5
There are some things that bear consideration about the ER as a statistic: it reflects no direct relation to the gross investment income realized by the fund; and, because it has no direct relation to investment income, it reflects no direct relation to distributions made by the fund to shareholders.
Indeed, the only function that seems to be served by the ER is to compare the size of a fund's expenses to the size of its assets; this has the potential to make the ER misleading, especially when comparing the ER of one fund to that of another.
In particular, while the ER is directly proportional to actual expenses, it is inversely proportional to the fund's assets - thus, the lower a fund's assets, the higher its ER. When comparing ETFs, if one fund has significantly less AUM than those to which it is being compared, its ER may seem disproportionately larger than that of larger funds, making it seem to be a less desirable alternative.
In light of the variables that influence the expenses encountered in managing an ETF, and considering what should seem to be an ambiguous influence on the part of asset size, it seems to be misleading to present a datum that infers a relationship between expenses and asset size. Thus, I am inclined to discount severely - if not disregard completely - a fund's ER; it would seem to be relevant only when comparing funds where all other things are equal.
The ETF's "Operating" Costs
The expenses incurred by an issuer in the course of managing a fund are taken from the income generated by the holdings of the fund; that is, the fund earns dividends, capital gains and interest, etc., and uses those earnings to pay the fund's expenses. The remainder of the income ("net investment income") is used to make distributions to shareholders (some may be retained in "cash" to be used for further fund activities).6
For all intents and purposes, an ETF's expenses are its operating costs. It would seem more appropriate from that perspective to examine the relationship between the fund's expenses and the monies from which those expenses are paid: the fund's investment income.
There are two ways one can look at the relationship between a fund's income and its expenses: the ratio between expenses and gross income (let's call this the expense/income ratio, or E/IR) and the ratio between the net income (what is left after expenses are paid) and gross income (this is strictly analogous to a company's operating margin, so let's call it the expense margin, or EM).
The E/IR would be determined by dividing the total expenses for a fund by the total investment income for that fund. The EM would be calculated by dividing the net income (gross income minus expenses) by the total investment income. The two ratios are essentially looking at the same coin from two sides; the sum of the E/IR and the EM would be 100.00%.
Let's look at an example. Invesco's PowerShares CEF Income Composite ETF (NYSE:PCEF) had total investment income of $23.63 million for the fiscal year ended October 31, 2013. Expenses for that period were $2.24 million, leaving a net income of $21.39 million. The following chart shows the two relevant ratios:
In the diagram, "Expenses" (E/IR) amount to 9% of PCEF's total income, and "Net Income" ((EM)) constitutes 91%.
I have to admit to being a little devious in choosing PCEF as an example here. PCEF has a very high expense ratio of 1.77% - nearly triple the average ER; PCEF also has a high dividend yield of 7.85%.
To my mind, PCEF's ER does not adequately speak to how the fund's expenses relate to its dividends, leaving one to look at what seems to be a very large bundle of expenses. Several readers were outraged when I announced that I was making PCEF a component of my retirement portfolio, arguing that an ER of 1.77% was far too much to pay for an ETF.7
However, if one considers that expenses consume only 9% of the fund's gross income, and that the fund is left with more than 90% of its income to distribute to shareholders, this puts a more appropriate light on the amount being paid to operate the fund, and it does not seem to be too much to pay.
Let's take a look at another example: Guggenheim S&P 500 Equal Weight ETF (NYSEARCA:RSP). This fund has AUM of $9.6 billion, and is tagged with an ER of 0.40%. The fund had income of $83.57 million in its last fiscal year. On the basis of ER alone, one would assume RSP had less in the way of expenses than PCEF; on closer scrutiny, however, that turns out not to be the case.
As illustrated above, RSP's expenses were almost eight times those of PCEF; moreover, RSP's expenses took up 20% (E/IR) of gross income, leaving $67.22 million for distributions - compared to PCEF's E/IR of 9%.
Expense ratio does not give an adequate - or even a helpful - picture of a fund's expenses, particularly when comparing the ERs of several funds.
All three funds cover the complete S&P 500, although IVV and SPY are cap-weighted funds, while RSP is equal-weighted. The relationship between expenses and distributable income as reflected in the diagrams helps to explain why RSP carries a lower dividend yield than the other two funds (1.35%, as opposed to 1.75% and 1.78% for IVV and SPY, respectively).
The above information, combined with the earlier observation that equal-weighted portfolios cost more to manage than cap-weighted portfolios, helps also to understand why RSP pays out proportionately more in expenses than its competitors; the question the investor is then in a position to ask is: is RSP worth the higher management fees?
Year to date, RSP has performed only nominally better than IVV and SPY; however, in the 11 years since its inception, RSP has significantly outperformed the other two funds, and has more than doubled the performance of the S&P 500 itself.9 The information provided in the earlier comparison indicates (to me) that yes, RSP's significantly better performance is worth the extra management fees, even if that performance is purchased at the expense of a few cents in dividends.
As I have perused the comments that have followed my recent articles, I have been dismayed at the amount of misinformation and misunderstanding that exists among otherwise informed, conscientious investors. A significant part of my efforts, as reflected in this article, has been intended to develop a way of presenting a fund's data that is more edifying and useful than the standard expense ratio.
The issue is really to find an effective way of expressing a fund's expenses - a way that provides the investor with useful information, and not simply an arcane comparison - so that the investor is in a better position to effectively evaluate the fees that are paid by the fund, and the impact those fees have on the distributable income the fund generates.
The use of the expense margin - EM - enables the potential investor to more accurately determine what the actual expenses might be, as well as assess the impact those expenses have on the shareholders' distributions.
Consider a comparison of PCEF and iShares Mortgage Real Estate Capped ETF (BATS:REM) - a comparison I made for myself in early November. 10 At the time, REM was offering a dividend yield of 12.80%, which amounted to $1.58 per year. PCEF was just shy of 8%, at approximately $1.81 per year. REM pays dividends quarterly, while PCEF pays monthly.
PCEF, as mentioned earlier, has an expense ratio of 1.77%, but has an expense margin of 90.52%. REM's expense ratio is 0.48%; its expense margin is 96.61%. It is important to note that PCEF has significantly less AUM, with less than 12% as much in assets as REM. In light of the nature of the expense ratio, it would be fair to wonder how much of the 129bps difference between the two funds is due to that fact. 11
While both expense ratio and expense margin differ greatly between the two funds it is through the expense margin that we get a more complete picture of specifically how the two funds differ - REM gives more of its investment income in distributions than does PCEF (to the tune of 609bps).
We are also able to pin specific figures to the expenses: PCEF's dividend of $1.81/year is 90.52% of approximately $2.00 (per share), meaning that $0.19/share goes towards expenses. For REM, the $1.58/year dividend is 96.61% of approximately $1.64 (per share), with $0.06/share going towards expenses.12
Thus, by relating expenses directly to the investment income from which they are paid, we are able to quantify our data in a way that can be readily, and reliably, compared between funds. We are able to do this irrespective of the relative sizes of the funds; we are able to view the data as an organic, interdependent whole. It is not clear to me that the same can be accomplished by way of the expense ratio - at least, not as simply and straightforwardly.
To be sure, use of the expense margin does not suffice as a complete and thorough examination of a fund's worthiness; it is just one piece of the puzzle. However, as a piece of the due diligence puzzle, it is a far better fit than the expense ratio has been.
On the whole, the expense ratio is a vacuous statistic that serves only as an oblique means of informing the investor of a fund's management fees without providing any meaningful context within which to interpret the information. That and the lack of standards for articulating the manner in which those management fees are actually reported serve only to perpetuate a lack of understanding on the part of investors of the real nature and burden of those fees.
Relating the management fees directly to the funds from which they are paid provides a more meaningful datum for the investor: it provides one with a direct means for understanding the relationship between expenses and distributions; it gives one a more relevant point of comparison to use in choosing between several funds; and, given a more articulated accounting of the management fees, the investor would be in a better position to evaluate whether a fund was using its resources efficiently.
Given the two alternatives I've discussed above (E/IR and EM), I will begin referring to these figures, rather than to a fund's ER, in future articles about ETFs. My personal preference is for the EM ("expense margin"), as this is most analogous to a company's OM ("operating margin"), and would seem to fit more naturally into discussions. I will continue to note a fund's ER for the convenience of readers (at least for the time being), but I do not intend to discuss it to any length.
Funds mentioned herein are presented for illustrative purposes only; no recommendation or representation as to the fitness of those funds for investment purposes is intended.
1 Or is there? Any company has its revenue, and there are costs involved in securing that revenue - the items that make up operating costs, and which are reflected in the operating margin.
2 This is true whether one is talking about mutual funds or exchange-traded funds; I will be considering ETFs exclusively.
3 In "The Right Way to Show Fund Expenses," Morningstar's John Rekenthaler advocates breaking down management fees into five elements: portfolio management fees, administrative fees, operational fees, distribution fees and advisory fees. His schema is based on one being prepared by Morningstar's Paul Ellenbogen.
4 I show a list of various weighting schemes in my article "Guggenheim's RSP: Equal Weight or Deal Weight?" One could probably spend a great deal of time on the relative costs of these methods; the real issue, though, is when is the added cost worth it?
5 "ETF Fees Creep Higher," Rick Ferri, Forbes.com. According to Ferri, figures indicated that since the 1990s, ERs have been gradually increasing.
6 The fund must distribute 90% of net income to shareholders in order to avoid paying taxes on that income.
7 See my article "Adjusting the ETF Retirement Portfolio." I have since divested myself of PCEF, but not because of its ER or its dividends; after about 6 months in the portfolio it was the only fund that registered a negative total return - despite having the highest yield in the portfolio.
8 See note above. I also compared Vanguard S&P 500 ETF (NYSEARCA:VOO), but that is omitted here for the sake of clarity, and to avoid certain issues that arise when considering Vanguard funds, issues referred to as "the Vanguard Factor" in Morningstar's "Full Report: The Cost of Owning ETFs and Index Mutual Funds."
9 Again, see note above. Since inception, RSP has increased in value by 268.01%. Over that same period, IVV has increased by 184.32%, while SPY has increased by 182.47%.
10 Specifics for REM can be found in my article "REM and MORT: Mortgage REIT ETFs for Growth and Yield," Seeking Alpha, Nov. 5, 2014.
11 Keep in mind that the expense ratio is based in part on the size of the fund's AUM - the less the AUM, the higher the ratio.
12 In total, PCEF paid $2.24 million in expenses in its last fiscal year, while REM paid $5.08 million. Also, funds may hold some net income back for future use by the fund - up to 10% may be held back.
Disclosure: The author is long REM.
The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.