Why do ETF prices sometimes diverge from NAV during stress?

The short answer

An ETF price diverges from its net asset value when the arbitrage chain that normally keeps them aligned breaks down. The chain involves authorized participants buying or selling the underlying basket against the ETF; when that becomes impossible, costly, or risky, the chain stalls.

Standard explanations frame this as the ETF being mispriced. That framing is often wrong, particularly in fixed income. When trading in the underlying bond market becomes thin or stops entirely, the NAV is computed from matrix prices that may be hours or days old. Meanwhile the ETF continues to trade in real time, absorbing fresh information about credit conditions, liquidity, and risk appetite.

In other words, in fixed income stress the ETF is often not the divergent measure — it is the measure of price discovery. The NAV is the lagging artifact.

→ New to ETFs? ETF explained for beginners

What the data shows

The March 2020 episode is the cleanest natural experiment, with both prices and trading frequency well documented.

The relevant data points (BlackRock, ICE, BIS, March 2020):

• On 12 March 2020, BlackRock disclosed in its Q1 earnings call that LQD traded approximately 90,000 times that single day while its top five underlying bonds traded an average of 37 times — the ETF was three orders of magnitude more liquid than its components
• LQD’s discount to NAV reached 5.35% in the third week of March 2020 according to ICE
• HYG (high yield) traded at peak discounts of approximately 5%, and TLT (long Treasury) reached approximately 5% premium during the same period
• The S&P 500 ETF SPY saw a peak discount of 0.80%, more than six times tighter than LQD — confirming the equity-versus-fixed-income asymmetry

The exception is that the dislocations resolved within 7-10 days following the Federal Reserve’s 23 March 2020 announcement of corporate bond ETF purchases, suggesting the divergence reflected stress in the arbitrage technology rather than fundamental valuation errors in the ETF.

→ Pillar: Passive management and ETF market structure

Why it happens — the macro mechanism

The mechanism behind price-NAV divergence is rooted in the asymmetry between the ETF’s continuous secondary market and the NAV’s matrix-based calculation.

The matrix pricing channel. Bond NAVs are not computed from observed transaction prices, because most bonds trade only a few times per day or not at all. Instead, pricing services use multi-factor matrix models that interpolate from comparable bonds. In normal markets the matrix is reasonably accurate; in stress, when bid-ask spreads widen and trading concentrates in a few liquid issues, the matrix can lag actual market levels by hours or days.

The price discovery channel. Meanwhile the ETF continues to trade in real time on equity exchanges, absorbing every piece of new information into its price. When AP arbitrage activity is high, the ETF price is roughly equal to the basket value — but in stress, AP activity drops sharply, and the ETF price decouples to reflect tradable reality. This is the angle most commentary misses: in fixed income, the ETF is often the leading price indicator, not the lagging one.

The asset class asymmetry channel. Equity ETFs rarely show large divergences because their underlying stocks trade continuously and pricing is unambiguous. Fixed-income, mortgage, and emerging-market ETFs show much wider stress divergences because their underliers are infrequently traded and rely on estimated prices.

Synthesis by regime: in normal markets all ETF asset classes track NAV within a few basis points; in equity stress (2008, 2020), divergences remain modest at 50-100 basis points and resolve within hours; in fixed income or emerging market stress, divergences can widen to 200-500 basis points and persist for days, only resolving when central bank backstops restore arbitrage capacity — the stress regime exposes the latent dependence of ETF pricing on the liquidity of the underlying.

The ETF discount to NAV is often not a glitch in the wrapper but a window onto reality the bond market was too slow to admit.

→ Framework: Market microstructure and price formation

What it means for different economic actors

ETF investors who sell into a wide discount realize a price below the NAV reported on their statement, which can feel like a loss but reflects the actual tradable value of the underlying basket at that moment.

Mutual fund investors in the same asset class redeem at NAV — but if the NAV is stale, they may be redeeming at a price that overstates the underlying’s actual market level, transferring losses to remaining holders.

Market participants and policymakers increasingly read the ETF discount as the cleanest stress signal in fixed income; the BIS has noted that ETF prices led credit spread movements in March 2020 by 24-48 hours.

A common error is to assume that the NAV is the truth and the ETF the noise; in fixed income stress, the relationship is often reversed.

Practical observation

Go deeper

Related questions

• What is an ETF authorized participant?
• How do ETFs differ mechanically from mutual funds?
• Do passive ETFs make markets more fragile?
• What are leveraged ETFs and why are they dangerous long-term?

Frequently asked questions

Is a price-NAV divergence always a sign that the ETF is mispriced?

Not in fixed income. The NAV is computed from matrix-based estimates of bond prices that often lag actual market levels in stress, while the ETF trades continuously and absorbs new information in real time. BIS research from March 2020 found that ETF prices led NAV adjustments by 24-48 hours during the most acute phase, suggesting the ETF was the leading price indicator and the NAV was the lagging artifact.

How is matrix pricing different from observed transaction prices?

Matrix pricing uses interpolation across comparable bonds — by sector, rating, maturity, and coupon — to estimate a price for bonds that did not trade today. In normal conditions the estimate is close to a hypothetical market price; in stress, when bid-ask spreads widen sharply and few bonds change hands, the matrix can be hours or days old. This is why the ICE data showed bond ETFs diverging from NAV by up to 5% in March 2020 while NAVs themselves moved more smoothly.

Why are equity ETF divergences so much smaller than fixed-income divergences?

Equity ETFs hold underliers that trade continuously on the same exchanges, with transparent prices and tight spreads even in stress. Their NAVs are computed from observed transaction prices in real time. Fixed-income ETFs hold bonds that trade infrequently, so the NAV depends on matrix estimates, and the AP arbitrage must contend with much higher transaction costs in the underlying. The asymmetry is structural, not a flaw of the ETF wrapper.

Last updated — 22 May 2026