ETF costs: separate the bid-ask spread, NAV premium and expense ratio

Three percentages that should not be added blindly

An ETF can have a low annual expense ratio and still be costly to trade. To understand the difference, separate three comparisons: the gap between the bid and ask, the market price relative to net asset value, and the fund's annual operating expenses. They use different reference points and time horizons. Adding them without checking those definitions can count the same effect twice.

This worksheet concerns an ordinary, unlevered equity ETF. It does not evaluate a particular fund or recommend a trade. FINRA distinguishes ETFs, which hold pooled investments, from exchange-traded notes, which are issuer debt obligations. Confirm the product structure before applying an ETF example to something with a similar ticker description.

The objective is practical: reconstruct the dollars paid and received, then explain which differences arise from execution prices and which arise inside the fund. Start with a blank worksheet rather than a ranking of expense ratios. A cheap annual fee is one input, not a complete statement of the investor's cost.

Write the quote and valuation timestamp beside every number

The SEC's ETF bulletin explains that retail investors transact at market prices, while a fund calculates NAV per share each business day. NAV reflects assets minus liabilities, divided by shares. A market price above NAV is a premium; one below it is a discount. Bid and ask refer to buying and selling interest in the market, with their difference forming the spread.

For a research note, record the quote time, bid, ask, displayed size and the date of the NAV being used. Label whether the price is an actual execution, a closing price or the quote midpoint. A premium calculated from yesterday's NAV and today's quote should not be described as a contemporaneous valuation gap without further evidence.

Consider a hypothetical quote of $49.90 bid and $50.10 ask. Its midpoint is $50.00. If a reference NAV is also $50.00, the midpoint matches that reference, but buying at the ask still costs more than selling at the bid. The reference price does not remove the difference between the two available sides of the quote.

These numbers are invented. Assume the quote remains unchanged, sufficient shares are available, and commissions, taxes and market impact are zero. Those assumptions make the arithmetic visible; they are not promises about execution in a real market.

Calculate a round trip from cash paid and cash received

Suppose the hypothetical investor buys 100 shares at $50.10 and immediately sells all 100 at $49.90 under those unchanged conditions. The purchase costs $5,010; the sale returns $4,990. The difference is a $20 loss before any other costs. That is approximately 0.3992% of the purchase amount.

The quoted spread is $0.20 per share. Dividing it by the $50.00 midpoint produces 0.40%. These two percentages differ slightly because their denominators differ: one describes the realized loss relative to the cash invested, while the other describes the quoted spread relative to the midpoint. Keep the label next to the percentage instead of treating the difference as an error.

The ask is 0.20% above the assumed NAV and the bid is 0.20% below it. In this deliberately symmetric example, those deviations describe the same $20 round-trip gap already calculated. Adding a separate $20 spread charge to another $20 premium-and-discount charge would invent a $40 loss. The actual cash flows still show only $20.

This is an accounting check, not a prediction that every trade loses a full quoted spread. A different execution price, changing quote or insufficient displayed size would change the result. Use actual fills when reviewing a completed transaction; do not overwrite them with an average spread from a fund factsheet.

Separate NAV performance from the prices an investor obtains

Now use a second, independent hypothetical holding period. The starting NAV is $50.00, and the investor pays $50.10 per share, a 0.20% premium. At the end, NAV is $52.00 and the assumed sale execution is at a 0.30% discount to that ending NAV. Assume there are no distributions, changes in share count, commissions or taxes during this example.

The ending sale price is $52.00 times 0.997, or $51.844. NAV increased by 4.00%, but the investor's price return is $51.844 divided by $50.10, minus one: approximately 3.4810%. For 100 shares, sale proceeds are $5,184.40 and the gain over the $5,010 purchase cost is $174.40.

A useful reconciliation is one plus the investor return equals one plus the NAV return, multiplied by one plus the ending premium, divided by one plus the starting premium. Discounts enter as negative premiums. Here that is 1.04 times 0.997 divided by 1.002. The formula makes the starting purchase price part of the denominator instead of simply subtracting percentage points.

The gap between 4.00% and 3.4810% is about 0.5190 percentage points. It is not a newly discovered expense ratio. It arises from the assumed purchase and sale prices relative to NAV. Those execution prices already contain their trading effects, so subtracting another assumed spread would double count them.

A counterexample clarifies the interpretation. If the starting and ending execution prices were both at the same 0.20% premium, with all other assumptions unchanged, the price return would match NAV's 4.00% rise. A premium's existence alone does not determine the holding-period drag; its change matters too.

Put annual expenses on their own time horizon

The SEC's fees bulletin describes the expense ratio as annual operating expenses relative to average net assets. Operating expenses are paid from fund assets and reduce share value. The prospectus fee table does not capture every cost an investor may incur, including some brokerage and trading costs. A fee waiver or reimbursement may also affect the disclosed figure, so read its conditions and duration.

For scale only, imagine a constant $10,000 balance and a 0.20% annual expense ratio. Simple multiplication gives $20 over a year. Prorating on an assumed 365-day basis gives approximately $1.64 for 30 days. This is a simplified comparison, not a statement about a specific fund's accrual method or the amount a broker will deduct from an account.

Compare that annualized estimate with the time horizon of the trade. An immediate $20 execution loss and a $20 illustrative annual operating cost happen on different clocks. Equal dollar figures do not make the mechanisms identical. Conversely, an investor who trades repeatedly should not assume that one round-trip example describes the cost of an entire year's activity.

When starting from a fund's observed NAV performance, remember that expenses paid from fund assets have already affected that value. Subtracting the full expense ratio again would duplicate that effect. When starting from a hypothetical gross return instead, state how expenses enter the model. Those are different starting points, and a reproducible calculation names which one it uses.

Treat historical disclosures as evidence with limits

FINRA notes that exchange-traded products can have liquidity risk, including wide spreads or substantial premiums and discounts. The SEC describes authorized-participant arbitrage as a mechanism that generally brings ETF prices toward NAV. That mechanism does not give a retail investor a guaranteed fill at NAV or eliminate the need to examine the actual quote.

Historical information can help frame a question without answering today's execution question. Record the period covered by a disclosed median spread and the dates of premium or discount observations. A median describes its sample; it does not guarantee the spread available for the next order. If the current quote differs materially, investigate the current conditions instead of replacing it with the historical summary.

A useful countercheck is to ask what would make the comparison invalid. The NAV might refer to an earlier valuation time; the quote might cover fewer shares than the proposed order; or the product might use a structure outside this worksheet's scope. Each issue calls for additional evidence, not an automatic conclusion that the fund is mispriced.

Avoid turning a discount into a promised bargain. The worksheet establishes a relationship between two prices or values under specified assumptions. It does not establish why the relationship exists, when it will change, or whether the underlying portfolio will rise.

Keep a short cost reconciliation with the research note

Save four groups of inputs: the product and disclosure dates; the bid, ask, size and quote time; the NAV and its valuation date; and the expense ratio with any waiver terms. For a completed investment, add executed prices, share quantities, distributions and separately charged fees. Keep hypothetical assumptions visibly separate from observed records.

Then run three checks. First, can the cash paid and received reproduce the investor return? Second, do the premium and spread calculations describe overlapping parts of those same prices? Third, has a fund-level expense already affected the NAV series being used? Resolve these questions before presenting an all-in percentage.

Our related guide on portfolio deposits and returns addresses how external cash flows change account-level performance calculations. This article focuses on the prices and costs of an ETF transaction. Both worksheets work best when the numerator, denominator and measurement period remain explicit, and when an unknown input stays marked as unknown rather than being replaced with a convenient assumption.