Cost of Ownership for ETFs

When deciding among various investment products, it would be nice if there were just one number with which to summarize and compare the costs. You can start with the Morningstar Quote pages, which prominently display a fund's expense ratio alongside its current price, but that is just the first step. There are other important costs to consider, which can be quite nuanced even when you are comparing similarly structured products. A true side-by-side comparison becomes even more difficult when you are comparing mutual funds with ETFs.

For mutual funds, not only do we need to pay attention to the expense ratio, but we must also compare loads and fees such as marketing fees, 12b-1 fees, and the costs the fund incurs when buying and selling shares. Mutual fund loads and marketing fees can be quite substantial (on the order of 3% to 5%), but there are a few things we can do to minimize them, such as shop for a lower-cost fund within the same category and star-rating band, or purchase the fund through a lower-cost distribution channel or from a broker that charges a flat fee instead of loads.

Part of the reason ETFs have become so popular over the last decade is because of their low  expense ratios and avoidance of backdoor shenanigans like loads and distribution kickbacks. The explicit costs of ETFs--payments you will know before you buy--are fully transparent, and in most cases, extremely low. However, because ETFs trade on an exchange, we know we need to look at the implicit costs such as bid-ask spreads and premiums or discounts--in addition to the expense ratio--to really know how much a fund will cost to own.

Commissions, Premiums, and Discounts
There are several things the investor can do to minimize ETF trading costs. Several brokerage firms offer low- or no-cost trading in certain ETFs. If you do need to pay a large commission, avoid making recurring small-dollar investments such as through a dollar-cost-averaging strategy. The trading commission is perhaps the most obvious source of trading costs, but other sources of trading costs should not be ignored, particularly if you are trading in a less-liquid asset class such as corporate bonds or emerging-market stocks.

Premiums and discounts are recorded on the Morningstar Quote page once a day and, by clicking on a longer-term chart, you can see the extent to which an ETF's share price has deviated from its net asset value (NAV). If a fund bounces around its NAV by a large margin (perhaps 0.5% to 1.0% on a daily basis), you might be able to time your purchase to avoid buying when the fund just happens to be at a large premium. Other funds trade at consistent premiums, and no amount of waiting around will reduce that cost. These sometimes relate to market access and liquidity factors as well as simple supply and demand.

Bid-Ask Spreads
Bid-ask spreads are not explicitly stated before you buy a fund, but they reduce your return by forcing you to buy at a slightly higher price and sell at a slightly lower price. The bid ask spread is something you can also monitor on the Quote page. When you are ready to place a trade to buy, take a look at the bid and the ask. The current ask will be higher than the bid and will represent the lowest price at which any seller is offering to sell shares. Likewise, the bid will be lower than the ask; this is the amount you will receive for your shares if you are selling. If you want to trade immediately, the ask is the price you will have to pay. But, if you have the patience to wait around a half hour or so, you can often achieve a purchase price closer to the midprice between the bid and the ask. Of course, by using this tactic you risk the possibility that the market will move higher and you will be forced to buy later at an even higher price. Trades that are large (say, 10% of the daily volume of a given ETF) will require even more patience. Thus, there is a balance between getting a good execution on a trade and the opportunity cost of not being invested.

Buying at a price even lower than the current ask will require the use of a limit order. A limit order is an instruction to your broker to put the word out to the market that you are willing to buy, but only at the price of your choosing. Once the market maker sees your buy order, he may reassess his offer to sell and may lower his price. If demand for the ETF is strong perhaps the market maker will go to the authorized participant to create more shares. If the prices of the underlying stocks are going up, you may regret hesitating over one or two cents per share. However, patience is often rewarded.

Expense Ratios
The expense ratio is the percentage of assets a fund management company charges. While we like to think of the expense ratio as one simple number, a peek behind the curtain reveals that it is not always black and white. The prospectus lists one expense ratio while the annual report may list another based on the actual costs incurred. Managers may grant fee waivers or discounts that are not guaranteed forever but may serve to reduce a fund's costs over a specific period of time. Morningstar's reports show the prospectus net expense ratio for each fund. That figure encompasses the management fees and operating expenses associated with running the fund, as laid out in the prospectus, less any fee waivers the fund currently has in place.

As with any business, the fund manager has some costs that are fixed and some that are variable. For example, the rent on the fund company's headquarters is fixed while certain marketing costs are variable (such as broker sales incentives or postage fees). Most of the costs a fund company incurs are predominately fixed. The costs to implement an accounting software system or the salary of the portfolio managers does not change much if a mutual fund goes from $100 million to $200 million in assets. But the revenue earned on a 1% expense ratio doubles. This is why the asset management industry is said to be scalable.

If we expect our assets to grow at 10% per year, it might seem acceptable to pay a 1% annual fee. If we expect a near-zero rate of return, such as on a bank deposit account, paying a 1% fee would seem outrageous. That's why investors in stock funds, where the range of prospective returns is very large, should be willing to tolerate higher expense ratios than they would for bond and money market funds, where the range of returns is small and gains also tend to be small in absolute terms. We don't actually cut a separate check or see a separate line item on our statement describing the fee we paid, and over the course of a year, the volatility in our investment may be high so we may not notice that the expense ratio has been deducted. But rest assured that whether the markets go up or down, the fund company will deduct its fee. The fund company just sold 1% of the stock in the fund and transferred the money into its own account. So absent any capital appreciation in your fund, a fund that is charging a 1% expense ratio will slowly shrink by 1% a year.

One fund in which the effect of the expense ratio is easy to see is  SPDR Gold Shares (GLD). When it was launched in 2004, each share of the fund was backed by one 10th of an ounce of gold. But there has been a leakage of gold and each share no longer represents a 10th of an ounce due to the relentless effect of the 0.40% expense ratio. Let's take a look at the numbers. Every business day, about .0000015 ounces of gold per share (at today's price of about $1,400 per ounce this amounts to about $0.002, or two 10ths of a cent) has been sold from the fund; now it no longer contains one 10th of an ounce of gold but rather 0.09762 ounces. That's right: Over the last six years, the constant erosion from the expense ratio has left the fund missing .00238 ounces, or about $3 per share.

That doesn't sound so bad in dollar terms. In fact, if the expense ratio were a constant proportion of assets, the fund will never run out of gold. That is a mathematical certainty according to Zeno's paradox, which states that, if we move toward a goal in fixed proportion to our distance, we will never actually reach the goal. However, as we mentioned before about the fixed and variable nature of expense ratios, State Street eventually will need to raise the fee to keep up with fixed costs.

Does this mean that GLD is not a good investment? Not when compared with the high costs of dealing in physical gold on your own. For example, the bank-vault storage fees or the bid-ask commissions charged by gold dealers far outweigh the costs in GLD. If we are unhappy about the declining amount of gold in the fund, all we need to do is buy a few more shares every year.

In summary, there are four main costs to owning an ETF: trading commissions, premiums or discounts, bid-ask spreads, and the expense ratio. A buy-and-hold investor incurs the trading costs only when they trade, but the annual expense ratio is incurred forever.