Price, NAV and market makers

Two numbers answering different questions

An ETF can show a NAV of €100 and an exchange price of €100.20. Which is its “true” value? The question already contains a misunderstanding: those two numbers are not born the same way and do not answer exactly the same question.

The NAV per unit describes which portion of the fund’s net assets corresponds, in accounting terms, to one unit — at a defined moment and under the fund’s valuation rules.

The market price describes instead the level at which buyers and sellers are willing to exchange the unit on a given trading venue. It can move all through the session and embeds the information, costs and risks operators are weighing at that moment.

The NAV is not necessarily a price the small investor can trade at. The exchange price is not necessarily a definitive valuation of the assets. They are two different observations, connected by the mechanism explained in the previous chapter.

How the NAV is calculated

In simplified form:

NAV per unit = (fund assets − fund liabilities) ÷ units outstanding

Assets can include securities, cash, receivables, accrued dividends and the positive value of derivatives. Liabilities can include accrued costs, payments due and the negative value of derivative contracts. The real calculation follows the prospectus, the accounting rules and the fund’s valuation policies.

If a fund holds assets worth 1,002 million, has liabilities of 2 million and 10 million units outstanding, net assets are 1,000 million and the NAV is 100 per unit.

The delicate point is not the division: it is the value assigned to the assets. A very liquid share with a recent closing price is relatively simple to value. A thinly traded bond, a closed market, a suspended security or a complex derivative require evaluated prices, models or fair-value procedures. The NAV is a disciplined calculation, not a perfect, simultaneous photograph of every possible sale price.

The official NAV has a time

The NAV is normally determined at the valuation point — the moment of the day when the fund officially fixes its values. Often it happens once a day; the timing and the price sources depend on the fund’s documentation.

This creates a practical problem: an ETF can keep trading while part of its portfolio is closed, or while the most recent NAV is still the previous day’s.

A European ETF on US equities, for example, can trade in the European morning before Wall Street opens. The unit’s price can react to futures, currencies, news and related instruments, while the official prices of the underlying shares have not yet been refreshed by a new session. Comparing that price with the previous evening’s NAV and calling the difference an “error” would be misleading.

The same difficulty applies to non-coinciding holidays, time zones, bonds that do not trade continuously and emerging markets with limits or suspensions.

iNAV: an estimate, not a second official NAV

To offer a reference during the session, some exchanges, issuers or providers publish an indicative intraday value, often called iNAV, IIV or IOPV. The label and the methodology can vary.

The iNAV tries to update the basket’s value using available prices and exchange rates. It can be useful, but it has important limits:

• it may update at a preset frequency, not continuously;
• it may use stale prices for closed markets or illiquid securities;
• it may be based on the previously published portfolio or basket;
• it may not reflect trading, hedging or funding costs;
• it is not necessarily a price at which anyone commits to buy or sell.

An iNAV with many decimals does not become more accurate because of them. Displayed precision must not be confused with the quality of the information feeding the calculation.

Market price: there is no single number

“The ETF’s price” can mean: the last traded price; the best bid; the best ask; the midpoint between bid and ask; the official closing price of a given exchange; a consolidated price built from several venues.

They are not equivalent. The last trade may date back several minutes; the closing price may come from an auction; bid and ask are proposals, not completed transactions. To measure a premium or discount during the session, the midpoint can be more informative than a stale last price — but it is still not a guaranteed execution price.

The same ETF class can also be listed on several exchanges and in several trading currencies. The ISIN can be the same, but each venue has its own book — the list of buy and sell proposals —, its own hours, its own last trade and venue-specific liquidity. Prices should remain economically linked after currency conversion, without being digit-for-digit identical at every instant.

Finally, the trading currency does not automatically change the fund’s currency exposure. Buying in euro a quotation of the same ETF whose assets are denominated in dollars may spare you a broker conversion, but it does not create a currency hedge. Hedging is a separate feature of the class or the strategy.

Premium and discount: the simple formula

When price and NAV are genuinely comparable:

Premium or discount (%) = [(market price − NAV per unit) ÷ NAV per unit] × 100

With a NAV of 100 and a price of 100.20, the premium is 0.20%. With a price of 99.70, the discount is 0.30%.

Before interpreting the formula, ask: do price and NAV refer to the same instant? Are they in the same currency, and at which rate? Were the underlyings open and tradable? Does the NAV use observed or evaluated prices? Is the chosen price recent and representative? Have distributions, auctions or corporate events intervened?

Without these checks, an apparent premium may just be an old NAV; an apparent discount may be the market pricing in new information ahead of the official calculation.

Who sets the price

No single party normally “fixes” an ETF’s price on the secondary market. Investors’ orders meet the proposals on the book. Market makers publish prices at which they are willing to buy or sell, within set quantities and conditions. Other intermediaries and investors can compete with those quotes.

The market maker builds its own estimate of current value and adds the costs and risks of taking the position: prices and liquidity of the underlyings, futures and currencies for hedging, cost of capital and funding, availability and cost of creation/redemption, the risk of moves during the time needed to hedge, inventory limits, and the hours and state of related markets.

The result is not an altruistic declaration of “fair value”. It is a pair of tradable prices within a certain size. The more expensive or risky the hedge, the more cautious the quotes can become.

Market maker and authorized participant are not the same thing. The market maker operates in the secondary market and quotes bid and ask. The authorized participant can create or redeem blocks of units in the primary market. The same institution can play both roles, but there can also be a chain: the market maker manages the position and leans on an AP to access the fund. That is why it is not correct to simply say “the market maker brings the price back to the NAV”: alignment can involve trading the units, hedging the underlyings, transferring inventory and, when worthwhile, creation or redemption.

How arbitrage works

Suppose, in a simplified way, that a unit trades at 101 while the economically replicable value of the portfolio is 100. An eligible operator could buy the basket for 100, deliver it in the primary market, receive new units and sell them around 101. The new sales increase the supply of units; the basket purchases act on the underlying market. The process tends to narrow the gap.

In the opposite case, with the unit at 99 and the replicable portfolio at 100, the operator could buy units, redeem them and receive assets worth more, then sell them or use them to close hedges. The redeemed units are cancelled and supply shrinks.

This picture explains the direction of the incentive, but leaves out almost everything that determines whether the operation actually pays: spreads and market impact, creation or redemption fees, taxes and settlement costs, availability of the securities and borrowing costs, currency and funding, the risk that prices move before completion, the minimum size of the operation, balance-sheet and counterparty limits.

Arbitrage therefore does not force price = NAV. It creates an economic band: as long as the gap is smaller than the costs and risks of the operation, there may be no profit to capture.

When premiums and discounts widen

Gaps can become more visible when the bridge between unit and underlying costs more, or works with less reliable information.

Closed underlying markets. The ETF’s price embeds new information through futures and related instruments, while the available NAV still uses earlier closes. The measured gap may be mostly a timing gap.

Illiquid or hard-to-value securities. For some bonds there is no recent, meaningful last price. The NAV may use valuations from pricing services, while the ETF trades in real time. Under stress, the ETF’s price can become a more current signal of the level at which the market is willing to transfer risk.

During the March 2020 turbulence, several bond ETFs showed wide discounts to NAV. The ECB documented those gaps together with the difficulties met by dealers and authorized participants. BIS Bulletin No. 6 by Aramonte and Avalos highlighted another reading as well: many bond prices used in NAVs were absorbing the new information more slowly than ETF prices, which could therefore contribute to price discovery. Not all discounts necessarily had the same cause, and they were not costless for sellers — but the NAV could not automatically be treated as the more timely truth.

Volatility and risk limits. When prices change quickly, hedging requires more capital and the probability of executing at worse terms rises. Market makers and APs may reduce sizes or demand more compensation.

Fragmentation and operational problems. Suspensions, market closures, settlement difficulties or the unavailability of an intermediary can slow arbitrage down. IOSCO recommends monitoring dependence on a few APs and market makers precisely because theoretical and operational capacity do not always coincide.

A discount is not tracking difference

Premium and discount compare price and NAV at a specific moment. The tracking difference compares, over a period, the fund’s return with the benchmark’s.

An ETF can trade very close to its NAV today and still have a negative annual tracking difference because of TER, taxes, replication costs and other causes. It can also show a premium or a discount for a few hours while having tracked its index very well over the long run. These are three distinct levels:

The third level is developed in the next chapter, on liquidity, spreads and orders — coming soon.

How to read the data correctly

When a platform shows price, NAV and premium/discount, the reader should also find: the trading venue of the price; the type of price used (last, close, bid, ask or midpoint); date and time; currency and conversion rate; the NAV’s date and valuation point; whether the intraday value is indicative; source and methodology.

A number without its time and its source is much less informative than it looks.

Rebalix archives NAVs and series as declared by the issuers, with their documentary date — as you can see on every fund page. It does not present the NAV as an executable price and, until it has licensed live market data with adequate coverage, it does not turn delayed prices into fake “real time”. This absence is preferable to an apparent precision that mixes times, venues and sources. The criterion is stated in the methodology.

Six mistakes to avoid

• “The NAV is the correct price at every instant.” It is a valuation performed under set rules and timings.
• “A price above the NAV proves the ETF is overvalued.” NAV and price may refer to different moments or information.
• “The iNAV is tradable.” It is an indicative estimate, not a binding quote.
• “The market maker decides the price.” It quotes proposals in competition with other orders, embedding costs and risk.
• “Arbitrage always removes the gap.” It steps in only when expected benefit, costs, capacity and risk make it practicable.
• “Discount and tracking difference are the same thing.” One is a point-in-time price gap; the other is a return difference over time.