Did Exchange Traded Funds drive the recent market rollercoaster?
The supply of ETF shares moved opposite the market. The S&P 500 fell about 16% in December and rose around 19% from Dec 24 to March 5. In December, says the Investment Company Institute, US ETFs created, or introduced, $260 billion of ETF shares, and redeemed, or retired, $211 billion.
So as the market tumbled, the number of ETF shares increased by $49 billion.
We saw the reverse in January as the market soared, with $208 billion of ETF shares created, $212 billion redeemed, the supply shrinking a little.
If ETFs track indexes, shouldn’t available shares shrink when the market declines and increase when the market rises? Why did it instead do the opposite?
One might point to the $46 billion investors poured into equity ETFs in December at the same time they were yanking $32 billion from Active funds, says Morningstar.
Again a contradiction. If more money flowed to equities than left, why did the S&P 500 fall? Don’t stocks rise when there are more buyers than sellers, and vice versa?
The fact that data and market behavior are at loggerheads should cause consternation for both investors and public companies. It means we don’t understand supply and demand.
One explanation, the folks from the ETF business say, is that inflows to ETFs may have been short. That is, when ETF shares increase while stocks are falling, ETF creators are borrowing stocks and trading them to Blackrock and Vanguard to create ETF shares for investors, who borrow and sell them.
These people explain it in a tone of voice that sounds like “aren’t we geniuses?”
But if true, the unique characteristics of ETFs that permit them limitless supply and demand elasticity contributed to the market correction.
We cannot manufacture shares of GE to short. But ETF market-makers can manufacture ETF shares to short. How is that helpful to long-only investors and public companies? The behavior of stocks separates from fundamentals purely on arbitrage then.
Here’s another statistical oddity: The net shrinkage in January this year marks only the third time since the 2008 Financial Crisis that the monthly spread between ETF creations and redemptions was negative. The other two times were in February and June last year, periods of market tumult.
And still the ETF supply is $45 billion larger than it was when the market corrected (near $55 billion if one adds back market-appreciation).
We conducted an experiment, tracking week-over-week gains and losses for stocks comprising the eleven General Industry Classification System (GICS) sectors and comparing changes to gains and losses for corresponding sector ETFs from State Street, called SPDRs (pronounced “spiders”) from Dec 14 to present.
Startlingly, when we added up the nominal spread – the real difference between composite stocks and ETFs rolled up across all eleven sectors – it was 18%, almost exactly the amount the market has risen.
What’s more, on a percentage basis the spreads were not a penny like you see between typical best bids to buy and offers to sell for stocks. They averaged 5% — 500 basis points – every week. The widest spread, 2,000 basis points, came in late December as stocks roared.
Now the spread has shrunk to 150 basis points and markets have stopped rallying. Might it be that big spreads cause traders to chase markets up and down, and small spreads prompt them to quit?
Now, maybe a half-dozen correlated data points are purely coincidental. False correlations as the statistics crowd likes to say.
What if they’re not? Tell me what fundamental data explains the market’s plunge and recovery, both breath-taking and gravity-defying in their garishness? The economic data are fine. It was the market that wasn’t. What if it was ETF market-making?
The mere possibility that chasing spreads might have destroyed vast sums of wealth and magically remanufactured it by toying with the supply of ETF shares and spreads versus stocks should give everyone pause.
Investors, you should start thinking about these market-structure factors as you wax and wane your exposure to equities. If fundamentals are not setting prices, find the data most correlated to why prices change, and use it. We think it’s market structure. Data abound.
And public companies, boards and executives need a baseline grasp on the wholesale and retail markets for ETFs, the vast scope of the money behind it — $4.5 TRILLION in 2018, or more than ten times flows to passive investors last year – and what “arbitrage mechanism” means. So we’re not fooled again (as The Who would say).
What do data say comes next? Sentiment data are the weakest since January 7 – and still positive, or above 5.0 on our ten-point Sentiment scale. That’s a record since we’ve been tracking it.
So. The market likely stops rising. No doom. But doom may be forming in the far distance.