“When the world decides that there is no need for fundamental research and investors can just blindly purchase index funds and ETFs without any regard to valuation, we say the time to be fearful is now.” – FPA Capital
With increasingly more financial professionals convinced that ETFs are the scourge and real weapons of mass destruction of modern markets, serving as a indiscriminate “communist” buyers of all stocks on the way up, and a just as aggressive “baby-with-bathwater” liquidator on the way down (assuming they even work, which as August 2015 showed can be a generous assumption), here is one contrarian, and optimistic, take from DataTrek’s Nicholas Colas.
* * *
Aren’t ETFs going to blow up one day and wreck global markets? We get that question more than any other, even from heavy users of these products. The short answer is “No”. There are some badly designed products, to be sure. But the vast majority of US listed ETFs have natural shock absorbers that limit their potential for spreading market mayhem.
The most common structural bear case we hear has nothing to do with demographics, central banks or politics; it centers on US listed exchange traded funds. Here are the basics of the argument:
- While ETFs offer near-instant liquidity during trading hours, the instruments in which they invest can be much less liquid. Think high yield or emerging market bonds here…
- During periods of market stress, the sale of large blocks of ETFs could force simultaneous sales of those less-liquid underlying investments since the fund will be forced to redeem existing shares.
- This feedback loop will destabilize markets that are already under duress from some exogenous shock, potentially wrecking not just US equities but other global capital markets as well.
The closest we’ve seen to this playing out is February’s volatility shock, caused by a poorly designed inverse volatility product. We lump that into a “Design flaw” category rather than proof the macro bear case about ETFs is valid. In fact, aside from 3 hours on August 24th 2015, exchange traded funds have been remarkably invisible from most market narratives even on the choppiest days.
There is a reason why we haven’t yet seen an ETF-related market rout (that volatility product was an exchange trade note, or ETN): this structure has a built-in relief valve. Explaining it is a bit grimy, since it involves understanding some arcane market structure issues, but it is important. Three things to understand here:
#1. The vast majority of ETF trades never result in the purchase/sale of underlying assets. Want to buy 1,000 SPY or QQQ? There’s a natural seller in the crowd, and you can buy those shares. ETFs are not like mutual funds in this respect, where buys/sells net out just once a day at 4pm.
#2. Trades in ETF shares do not have to happen at net asset value. Yes, there are plenty of arbitrageurs that play the spread between NAV and the ETF’s market price to keep the two in sync. But unlike mutual funds, ETF prices are set by market forces, not strictly by the value of the portfolio.
#3. A very large sale in an ETF holding illiquid assets during periods of market duress (the nightmare scenario from our first paragraph) can run down one of two channels on its way to execution.
- Option #1: a broker authorized by the ETF sponsor can put together a sale of the underlying assets and redeem the shares.
- Option #2: the seller can mark down their asking price until it hits a natural bid.
- Option #1 can certainly pressure that illiquid underlying market. Option #2 bypasses it entirely.
That’s the theory – how well does it line up with actual US listed ETF market pricing? We pulled the discount/premium data from our friends at www.xtf.com to get the answer.
We would expect to see ETFs that track liquid assets trade closer to NAV than those with more exotic holdings in the portfolio; here is the data as of last Friday’s close:
US Treasury ETFs (44 ETFs)
Spread of discounts/premiums across these ETFs: -0.27% to +0.05%
Range of spreads: 32 basis points
US High Yield Corporate Debt (18 ETFs)
Spread of (-1.28%) to +0.42%
Range of spreads: 170 bp
Emerging Market Bonds (20 ETFs)
Spread of (1.53%) to +1.67%
Net difference: 320 bp
US Equity (605 ETFs)
Spread of (3.21%) to +3.98%
Net difference: 719 bp
Note: 95% trade between a (0.59%) discount and a +1.52% premium, making the “real” spread more like 211 basis points.
International Equities (109 ETFs)
Spread of (1.34%) to +3.45%
Net difference: 479 bp
Emerging Market Equities (69 ETFs)
Spread of (1.64%) to +1.56%
Net difference: 320 basis points
Our takeaways from this:
#1. ETFs naturally trade with an imbedded liquidity/price adjustment based on the assets they hold. The data we pulled was as of Friday’s close, so it does not include Tuesday’s volatility and is largely representative of current “normal” market conditions.
Looking at that data: US Treasury ETFs trade 90% tighter to NAV than Emerging Market Bond ETFs. EM Equity ETFs trade 50% wider than most US equity ETFs. This is exactly what you would expect to see.
#2. The bottom line here is that ETFs are more dynamic than many market watchers understand. Their pricing incorporates underlying liquidity even in calm conditions. When the next bout of volatility hits, the spreads we have cited here will certainly widen. But that, in tech-world terms, is a feature, not a bug.
If there is little/no liquidity in an underlying asset, the ETF can and will still trade, albeit at a sizable discount to notional NAV. But at this point it will be the ETF that gives the underlying market price discovery, and that’s actually a positive in a volatile market.
#3. And one caveat: the work here only included true exchange traded funds, not more exotic instruments like exchange traded notes. Those, as we saw in February, are only as good as their design. The point here was to explain the structure of ETFs and how they trade.