Research and Market Commentary
Commodities outlook: What’s next for global markets?
May 23, 2018Commodities outlook: What’s next for global markets?
The NAV of a fund is a metric that gets a lot of attention from investors and practitioners. We dive under the hood to explain what the NAV is, and what the NAV isn’t.
One of the most frequently-cited and least-understood concepts in the ETF business is that of the fund’s net asset value, or NAV. Strictly speaking, fund NAVs are a published calculation of the value of the fund’s holdings, and are determined by the methodology described in the fund’s prospectus. The responsibility for accurate fund NAVs lies with the fund’s issuer, though in practice each fund’s custodian does the bulk of the leg work. In the mutual fund structure, investors buy and sell units of the fund at this calculated value. In the ETF structure, because the funds trade continuously and through a market maker, the only parties directly exposed to the intricacies of the NAV are market makers engaged in fund creations and redemptions.
We believe that there is some confusion around the meaning of the term “NAV,” particularly among retail investors. For example we occasionally see questions such as “Can I buy this at the current NAV?” when an investor is trying to judge the fairness of the market price. The jargon involved is also frequently confusing, with terms like Intraday NAV, Theoretical NAV, and INAV that go unexplained.
In the discussion below, “NAV” refers only to the officially published daily measurement of fund asset value, and not any sort of intraday valuation or price. The generally-accepted intraday valuation term is “fair value,” which refers to a particular participant’s estimation of what the fund would be worth based on current market prices of their underlying. Unlike a fund’s NAV, fair value is generally expressed as a bid-ask quotation, because all underlying assets also contain their own bid-ask.
How is NAV determined?
In Canada, the “net asset value” for a fund is defined in National Instrument 81-102 (Investment Funds) as “the value of the total assets of the investment fund less the value of the total liabilities, other than net assets attributable to securityholders, of the investment fund, as at a specific date” with further details offered by National Instrument 81-106, part 14. At each NAV date, typically daily, the assets and liabilities of a fund must be assessed at “fair value.” For the purposes of this calculation, fair value means a market price based valuation where it is available, or where it isn’t available a “fair and reasonable” valuation according to the manager.
In reality, fund managers observe market prices for all their holdings, add them up, subtract liabilities, and publish an aggregate value. Simple? Yes. However, in practice investment fund manager still have some choices to make around specifically what prices they could use to arrive at “NAV.”
For example, this is easy for equities where market convention is to use the official marketplace closing price. However, equities are unique in that a reliable price benchmark exists for most securities, and exchanges hold closing auctions. In most other asset classes, a valuation may be difficult to come by, or at least require some decisions to be made.
For bonds, market convention in Canada for most fund managers is to use the midpoint of the benchmark price published by FTSE Russell as part of the calculation of the FTSE TMX Canadian bond indices. Some issuers use the bid price instead of midpoints. At least one ETF product we are aware of actually uses the ask. There are issuers who use a different index provider than FTSE Russell, and a different set of prices for conceivably the same bond. Finally Canadian bond prices are typically snapshot at 4 p.m., while U.S. bond prices come from a different source and are most often struck an hour earlier, at 3 p.m.
For the valuation of international holdings, issuers get to choose which FX benchmark they use in their fund NAVs. The most common two FX benchmarks are the closing spot rates published by WM/Reuters at either the London close (calculated at 4 p.m. GMT, or 11 a.m. Eastern), or the New York close (4 p.m. Eastern). However, this is not the only approach. We are aware of some issuers who use intraday snapshots at different times (for their own good reasons), with no clear benchmark. We are also aware of one product issuer whose NAV is struck on the basis of the Bank of Canada’s daily average rate – which isn’t tied to any particular point in time at all.
There is no requirement for NAVs to reflect asset and liability values observed at the same time across the entire fund. In other words, it is entirely permissible for a fund issuer to pick an average FX rate, a morning snapshot of commodity prices, an afternoon snapshot of bond prices, and a daily VWAP of equity prices to arrive at an overall fund valuation. This could be seen as the best way to reflect “asset value.” In fact, for funds with international exposure it is typical for equity prices to be struck at the close of each market, which of course happen at different times.
This myriad of choices means that for a healthy discussion of a particular fund’s NAV, one must be familiar with the methodology used to calculate it. As a general matter, fund NAVs are not comparable across issuers, and some issuers vary the approach by product. As a result, any talk of comparing premiums and discounts to NAV across issuers is not meaningful.
Now the good news: the technical nuances are largely irrelevant to investors. Funds trade on exchange on the basis of where market makers are willing to value and hedge them. Worrying about methodology differences is the market maker’s job. The investor’s sole concern ought to be whether at the time they choose to trade, the market price fairly reflects the value of the underlying assets at the time, including the market maker’s costs of hedging.
An illustration: S&P 500 ETFs
To illustrate the various valuation techniques, let’s consider some Canadian ETFs tracking the S&P 500 index (with no FX hedges). Canada is blessed with no less than five comparable products, all of which offer different NAV methodologies and also different timing for the payment of distributions. We can confidently say that investors in each of the products below can reasonably expect to receive the S&P 500 return net of the fund’s fees.
Even an ETF tracking the S&P 500 can still leave choices to be made. Nothing about the structure of any of these products is exceptional, and none are difficult for market makers to handle.
The moral of this story is that seemingly identical products, with largely identical investor experience, will show (slightly) different performance based on what the issuer chooses to use as their NAV calculation methodology.
Does NAV methodology make a difference?
The details of NAV calculation methodologies can have impacts on the ETF market. Market makers who create and redeem units are exposed to the official NAV, and effectively pay the NAV on creations (and receive it on redemptions). The degree to which NAVs methodology matters depends on the structure of each product.
In the case of traditional in-kind products, the market maker exchanges units of the ETF (valued at NAV) for underlying assets valued at their underlying prices. As long as the NAV and the underlying prices move in unison, the way in which the underlying prices are determined is largely arbitrary.
The risk increases when market makers are directly exposed to the NAV, with no offsetting physical delivery, or when the underlying leg doesn’t fully correspond to the complete underlying holdings. This frequently occurs when creations are done entirely via cash. In this case, every creation and redemption leaves market makers fully exposed to attempting to hedge the fluctuations in the NAV using market instruments.
In some cases, the prices built into NAV are non-tradable. The simplest example occurs in the fixed income market, where products are typically valued at the midpoint of a surveyed bid-ask. No bond dealer is willing to trade bonds at the midpoint of many dealers’ contributed average. This means that the “NAV” itself is a theoretical value only. Dealers quoting ETF trades relative to published NAV in those circumstances are making a judgement call on where benchmark prices may come in relative to where their hedges could be traded. This is a risk decision and not arbitrage.
To be clear, not all cash-created ETFs have complex and unachievable NAVs. For example, a cash creation of BMO’s S&P 500 ETF, ZSP, requires the market maker to hedge the inputs into the NAV. In this case, it is the S&P 500 equities at their closing prices, and currency exposure with a benchmark of the WM/Reuters 4PM USDCAD spot rate. The equity hedge is available via the US market-on-close, and some FX dealers are willing to trade at fixed spreads to the WM/Reuters 4 p.m. benchmark. The choice of NAV methodology determines how and when market makers establish and remove their hedges. Practically speaking, this is only possible because the U.S. market has a market-on-close auction, and the FX benchmark is already widely used in the market.
The same complications apply to institutional client trades which are benchmarked to NAV. For instance, achieving the published NAV on the iShares equivalent, XUS, requires the market maker to hedge FX at 11 a.m. on the basis of where they believe IVV (not S&P 500) closing prices will be at 4 p.m. They will also trade IVV rather than the S&P 500 basket.
The bottom line: NAV methodology choices therefore affect the practical experience of market makers, and can meaningfully influence the amount of risk involved in ETF liquidity provision. If NAV methodology and the mechanics of creations introduce undue risk and costs on market makers, the product’s bid-ask will reflect these choices.
The most contentious and most difficult to hedge NAV methodology choice is “fair valuing” equity holdings. The purpose of fair valuing is to estimate the value of securities at a point in time on the basis of where they “might” be trading if their market was open. The most common example of this is to estimate the value of international holdings at the close of the North American markets, when the underlying securities are no longer trading and would otherwise have stale prices. Another example is estimating the NAV of a U.S. equity fund on days when Canada is open, but the U.S. is not.
In practice, some funds need to be fair valued to accommodate the possibility of inflows after the close of the underlying market. This is certainly the case for retail mutual funds, which allow inflows each day and have no provision for entering orders a day ahead. Issues with stale-priced NAVs led to the establishment of various rules that effectively require fair-valuing for international funds.
Fair valuing is intended to prevent situations where investors are able to interact with a fund at a NAV that is backwards-looking. For example, if a fund’s order cutoff on a particular day is noon, and NAV reflects prices of Asian or European securities, the NAV could reflect prices that prevailed before an investor places their order. This could expose the fund to trading risk. For example, if the S&P 500 is up sharply in the North American open and through the North American trading day, a fund holding Asian shares could be expected to be up on the subsequent day – and cash investments into that fund would be at higher prices than reflected in the NAV. This is certainly a potential issue for mutual funds which allow daily inflows with a cutoff after the Asian market close.
In the ETF context, the “fair valuing” concept is accomplished at all times by forcing market makers to take the other side of the investor’s trade, and therefore price the risk. Additionally, ETFs (unlike mutual funds) impose cutoff times on subscriptions and redemptions which are early enough for the fund manager to be able to trade the underlying holdings at prices reflected in the NAV. In some cases, when certain countries have trading holidays, ETFs close subscriptions and redemptions for a short period of time. In other words, potential negative impact to the fund is managed through the creation mechanics, and ETF quotations are always fair-valued through the risk being taken by the market maker.
Then what is the problem? In practice, there’s no telling where NAV might be. Fair valuing of ETF NAVs makes them unpredictable. Each day, a foreign market will close in the morning, and the NAV will be struck on the basis of how much it could be assumed to have moved between the morning and the afternoon. For example, if the S&P 500 is up 1% between Japan close and New York close, Japanese equity holdings might be valued at 1% higher than where they actually closed, on the assumption that the market “would have” been higher. Alternatively, a different NAV fair value model might arrive at the same fund being up a different amount, or maybe unchanged? This adds risk and costs to market makers.
The degree of adjustment to NAV and the underlying factors is typically proprietary to the issuer, and ultimately cannot be hedged. Meanwhile, ETF subscription and redemption cutoff times can be set early enough to address the “late trading” concern, with investors always trading at an intraday “fair value.” The incremental risks and costs to the market maker simply end up reflected in the quoted market prices, without a tangible benefit to either the investor or the fund.
The choice of NAV methodology is deep in the weeds of the ETF business and in some circumstances can meaningfully affect the way in which market makers approach products. The moral of this story for ETF issuers is to make these choices wisely. For investors seeking trades at “NAV,” we hope the mechanics illustrated provide some context to what might be occurring behind the scenes.
For questions or more expert ETF insights, please contact:
Director, Head of ETF Trading
Director, ETF Trading
Managing Director, Head of Portfolio Trading