Bond ETFs around the world have blown out on discounts, thanks to coronavirus-induced mega-volatility.

ETF providers – and analysts who share their perspective – have justified these discounts, claiming they are “price discovery”.

By this, they mean that bonds these days are illiquid, thanks to banking regulations introduced after 2008. ETFs, however, are highly liquid and have the added benefit of trading on exchange. This can mean its easier to trade a bond ETF than bonds themselves. And as such, bond ETFs can be a better guide as to what bonds are worth than bonds themselves.

ETF providers State Street, Vanguard, Invesco and BlackRock have all made such arguments when asked about their bond ETF discounts.

But are they right?

Panic sell-off shows limitations of junk bond ETFs

From where I sit there are a few problems. My argument below is one-sided. Others have done a good job telling the other side.

Problem 1. Bond ETFs are meant to track indexes

The first problem is the obvious one: bond ETFs are meant to track indexes. When discounts emerge, investors aren’t getting the performance that was advertised. There’s nothing complicated about this point, in my view.

You could argue – some have – that bond ETF discounts are fine, and it’s really the NAV and index prices that are wrong, as they rely on data that are triangulated and stale.

But this leads to an obvious question: “what is the point of an index tracking bond ETF if its index is stale and its NAV is wrong?”

Problem 2. ETFs hypocrisy on closed-ended funds

Closed-ended funds (CEFs) trade on discounts larger than ETFs every day. Yet, having criticised CEFs for trading on discounts for decades, the ETF industry has been hit by the same problem.

CEFs say discounts are great because they are a buying opportunity. ETF providers say discounts are great because they are price discovery.

Vanguard's biggest bond ETF discount widens 6%

Investors have been sold ETFs on the grounds that they’re safe from discounts thanks to creation redemption. Yet as the past fortnight has shown quite clearly, they’re not. This will almost inevitably lead to charges of hypocrisy.

Problem 3. Heads I win, Tails you lose

Another problem, related to the first, it feels quite like a heads I win, tails you lose sort of situation. When ETFs track their index and trade at NAV – they’re functioning as they should, we are told. They are providing the index-like performance they were always meant to.

When they fail to do this – like now – we are told they’re doing the right thing too: providing “price discovery”.

Regardless of the outcome, the ETFs are doing the right thing. We have an unfalsifiable argument.

Problem 4. What does “price discovery” mean?

The final point is: what is “price discovery"? Does the term mean anything?

Transactions either occur or they don’t. If they occur then – tautologously – they occur at a “price”. And – again tautologously – that price has been “discovered”. Transactions – tautologously – involve “prices” being “discovered”.

Yet the point being discussed is whether the price is fair and whether the discovery is valuable. Pointing out that a price has been discovered – i.e. that a transaction has occurred – adds nothing of substance.


Rather than arguing about whether discounts are price discovery, investors should be discussing whether discounts are fair.

Some have made the plausible argument that discounts are justified in today’s market, as they represent the best price traders can offer, given how illiquid many bonds are. And hey, if you’re selling out in the middle of a storm, you should expect to pay a premium. Or so they say.

Others have argued that the fairer approach is taken by open-ended funds, where the fund provides a subsidy – in effect – to investors selling out. This allows them to get out at NAV at all times, including times like this.

This, to me, seems like the better way to proceed.