Last April, almost a year before the cataclysmic volatility-repricing event of February 5, which nuked vol sellers and led to countless casualties among those using inverse VIX ETFs to collect pennies in front of a central-bank primed vol steamroller, FPA Capital joined others such as Howard Marks, Eric Peters, and this website, in warning that “ETFs Are Weapons Of Mass Destruction” and that we could get an onslaught of selling once markets wake up from the passive-investing daze:
“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.”
It now appears that from the purely abstract, concerns about the role of ETFs – which at this point greatly outnumber the actual underlying securities that make up the universe of exchange traded funds and notes – are starting to spread to the broader investor community, and as BMO’s credit strategists write, “one very interesting aspect of equity market volatility that seems to come up more and more in our client meetings is the role ETFs play in the volatility, and the growing concern that the proliferation of ETFs may actually fuel the next disorderly repricing of financial assets.”
As a result, the strategist decided to investigate the ETF story to see if the huge growth in passive investment poses a risk to high quality spreads. Here’s what he found.
First, some statistics:
ETFs have grown significantly both in absolute and relative size over the past five years. To put it in numbers,assets invested in bond or equity ETFs reached a high of about $3.4 tn in late 2017 compared to under $1.4 tn at the beginning of 2013, which equates to a growth rate of 143% in just five years.
This growth is concerning to some market participants who worry that the proliferation of passive investing has inflated the price of assets as investors own the basket of securities that comprises ETF indices, regardless of fundamentals.
These fears are compounded when considering that ETFs have yet to be tested by a significantly stressed market environment, and some are concerned that this concentration of investment may exacerbate price declines during a risk off market. Said differently, if a large percentage of investors own the same securities, and all of those investors want to sell at the same time, what will happen to asset valuations?
Supporting this theory, ETF trading volume is highly correlated with market volatility (Figure 5). Volatility spiked on Monday February 5th when the Dow Jones Index fell by 4.6%. The 5 largest equity ETFs saw trading volume jump 213% from the prior Thursday, significantly outpacing the 108% and 60% increases in trading volumes of the Dow and S&P 500 indexes.
Further, on the most volatile days this year, some high yield bond funds have displayed sudden and extreme price swings amid significant outflows and high trading volumes.
Conversely, there are other market participants who argue that ETFs should actually work to counteract declining asset prices in a stressed market given the presence of arbitrageurs known as Authorized Participants. Unlike mutual funds, ETFs trade on an exchange with a price determined by the market rather than strictly the value of the underlying securities that makes up the ETF. If that’s the case, then what ensures that the price of the ETF will accurately reflect the value of the security basket underlying it? This is where Authorized Participants (APs) come in.
Authorized participants are entities, typically investment banks, that have the power to create or redeem ETF shares. To create one, the AP buys the securities that will underlie an ETF and delivers them to an ETF sponsor, who then issues ETF shares to be sold by the AP. The inverse is true of the redemption process, when an AP takes ETF shares to the sponsor and redeems them for cash or the underlying securities. It is important to note here that ordinary investors do not have the power to redeem ETF shares, only an AP can do so.
Being an AP has its advantages, as they are able to arbitrage ETF price changes. For example, if the price of the ETF increases to more than the value of the underlying securities, the AP can buy the security basket and deliver it to the ETF sponsor in exchange for an ETF share and that it can then sell into the market. The purchase of the security basket and subsequent sale of ETF shares works to increase the price of the security basket and decrease the price of the ETF, thereby restoring a fair trading relationship.
Naturally, the opposite pattern also holds where the price of the ETF falls too low. In this example, the AP would buy
ETF shares and redeem them with the ETF sponsor, driving up the price of the ETF and reducing the price of the security basket.
While in theory some claim this force should help counter a significant decline in ETF prices, the BMO team says it finds itself more swayed by the arguments that ETFs could exacerbate security price declines in a stressed market because of the essential role that liquidity plays in the supposedly countercyclical AP arbitrage.
To demonstrate why, let’s consider a scenario where ETF prices are driven below what would be considered fair value given the value of the underlying security basket as investors flee the market. To arbitrage, an AP can buy the ETF and redeem it for the security basket.
However, in order to make the arbitrage profit, the AP must then be able sell the security basket into the market. If it cannot, the AP is stuck holding securities, which clearly makes this a decidedly non-arbitrage opportunity. As a result, APs may disappear at times of illiquidity, which is precisely when the market would need it the most. The growth in synthetic and leveraged ETFs makes this an even larger concern.
* * *
In light of the growing size of the ETF markets and persistent equity market volatility, BMO next set out to analyze the potential for ETFs to exacerbate an environment of market stress or volatility, and if that would have an impact on the high quality spreads.
How Do Equity ETFs Compare to Bond ETFs in Size?
We analyze the size of the largest 250 bond and equity ETFs as a proportion of the underlying indices they each track (Figure 6). Equity ETFs make up a significantly larger share of the equity market at around 12% compared to general bond funds, where aggregate bond ETFs comprise just 2% of the underlying index.
Further, equity ETFs are substantially larger in dollar terms. For instance, the 250 largest equity ETFs hold over $2 tn in total assets; the 250 top bond ETFs hold less than $450 bn.
Breaking bond ETFs out into categories based on their strategies reveals that there is significant heterogeneity among bond ETFs. We separate bond ETFs into the following classes: aggregate, corporate, high yield, government, corporate≥5yr, corporate<5yr, government≥3yr, and government<3yr (Figure 7).
The next chart shows the shares of the underlying indices held by ETFs over time. While equity ETFs are responsible for over 10% of the indices they track, bond ETFs hold a relatively small proportion of their markets. Most of these ETFs hold less than 3% of their indices.
However, high yield bond ETFs are an obvious outlier to these trends, and present an interesting case for a few reasons, one of which being that they make up a significant portion of their underlying market (Figure 8). In this sense they have more in common with equity ETFs than government or investment grade bond ETFs.
How Do Equity ETFs Compare to Bond ETFs in Stability of Flows?
As BMO notes, for ETFs to present a destabilizing influence on bond markets, they must also exhibit volatile price action on certain days. As a result, the Canadian bank’s team analyzed the volatility of ETF price action through daily creation and redemptions among the 5 largest ETFs in their respective classes. Specifically, it compared equity, high yield, and aggregate bond ETFs.
Figure 9 shows the average daily flows, normalized by assets under management, for the 5 largest equity and bond ETFs. As expected, equity ETF flows tend to be more substantial than general bond ETFs. This is not surprising considering bond ETFs are generally less risky and more likely to be held by investors looking for stable long-term returns. On the other hand, high yield ETFs display more volatile inflows and outflows than equity ETFs (Figure 10). This result is somewhat unexpected as BMO had expected to see roughly equal rates of inflows and outflows between equity and high yield funds.
BMO also finds that the relative volume of high yield ETF trading to high yield bond trading by FINRA also spikes on volatile days. That is, on volatile days, trading becomes more concentrated in ETFs than in individual bonds. Additionally, this trend has become more pronounced over time; over the past several years, high yield ETF volume traded relative to individual bonds has increased, especially on volatile trading days.
BMO’s troubling conclusion: “we view this as supportive of the notion that the growing prevalence of ETFs in risk assets can lead to destabilizing flows that can exacerbate market moves.“
What are the Implications?
While it will not come as a surprise to ETF critics, BMO finds direct evidence that that there is significant threat of herding in equity and high yield ETFs and significant volatility in these markets could be exacerbated through excessive ETF selling and redemptions.
Surprisingly, bond ETFs generally are less exposed to such disorderly moves because they make up a smaller percentage of the market, exhibit less volatile flows adjusting for their size, and are less prone to outflows given a sudden risk-off move. However, within this universe, high yield ETFs behave more like equity ETFs than investment grade or government bond ETFs. They make up a fairly significant portion of the underlying market, are risk assets and so underperform on a risk-off spike in volatility, and exhibit significant flows.
Therefore, in the event of a sustained stressed market, BMO warns that ETFs in equity and high yield debt markets have the potential to exacerbate selloffs. The good news is that unlike stocks and junk bonds, the study paints a fairly bullish picture for high quality spreads in an environment of market stress. Of course, even that optimistic assumption will be severely tested during the next major IG bond selloff, which – on top of everything – will come at a time when yields are rapidly rising as the Fed continues to tighten monetary policy every quarter.