ETFS CREATED AN UNSTABLE MARKET!
There is ample evidence that ETFs have reduced liquidity, added to volatility and, possibly, exacerbated declines. The volatility of the past 4 weeks should force people to examine closely how this huge industry is affecting the stock market.
SUMMARY:
1. Stocks with high ETF Illiquidity / days of volume held in ETFs significantly underperformed the S&P500 from 2/19/20 to 3/20/20
2. Evidence from widening bid-ask spreads and widening differences between ETF prices and the value of their underlying shares
3. Studies by the EU Systematic Risk Board and the National Bureau of Economic Research (NBER) have found the following associated with ETFs and their underlying shares:
· “higher intraday and daily volatility”
· “co-movement of asset prices [that] may pose systematic stability issues”
· “increased price volatility”
---------------------------------------------------------------------------------------------------------------------------
TPA began discussing the possible problems with the rapidly growing ETF market back in 2016. There were many critics of the ETF trend in 2016 and many detractors still exist. TPA was quoted in a WSJ/Marketwatch story in entitled “Fears Grow that Popularity of ETFs is a Ticking Time Bomb”. as we were developing the ETF Illiquidity Alert. The trend in growth of ETF AUM has, however, continued unabated and as the assets in ETPs have grown from approximately $760 million in 2009 to over $4.2 trillion AUM 2019. In light of the current volatility and the omnipresent condition of ETFs in the marketplace, TPA has revisited the subject to ask if the dominance of ETFs has created an unstable market and found reasons for concern.
TPA has found 3 metrics that point to potential problems in the market that stem from the overwhelming presence of ETFs:
1. ETF stock illiquidity is trending toward finally being a problem
2. ETF average bid-ask spreads widen considerably in periods of rapid decline. Interestingly, this is only happens during market declines.
3. The difference between the ETF’s price and the intraday NAV or the value of the underlying stocks at the same time, also widens on days when the market suffers a severe decline.
In this report, TPA will explain and expand on the 3 issues above and then run through a laundry list of other potential problems that stem from the market being inundated with ETFs in particular and passive investing in general.
1. ETF stock illiquidity – From the beginning, the argument to defend stock illiquidity surrounding ETFs has always involved the Creation-Redemption process. The following video from ETF.com explains the process: https://www.etf.com/etf-education-center/etf-basics/understanding-etf-liquidity. The idea is that that because Authorized Participants stand at the ready to Create and Redeem ETFs, the market is much deeper than the available stock on any exchange.
Calculating ETF Illiquidity
In 2016, TPA created the ETF Illiquidity Reports, which is an “illiquidity” measure for the largest holdings in ETFs. TPA uses a universe of 1257 U.S. stocks. For each stock TPA calculates the Shares Held as Number of Days Volume by dividing the shares held by ETFs by the average 30 day trading volume. For example, EQR has 33.17 days volume in ETFs. That number is the percent of outstanding shares in ETFs (14.08%) multiplied by the outstanding shares (371,978,400) = shares in ETFs (52,374,559). Then dividing the shares in ETFs by the average 30 day trading volume for EQR (52,374,559 / 1,579,031 = 33.17 days volume).
The Report and Performance Measures
Once TPA had days volume in ETFs for 1257 stocks, it analyzed performance in periods when the S&P500 declined by 5% or more. TPA calculated performance for every stock in the period and then grouped the stocks according to ETF Illiquidity; 30 days or more, 20 days or more, 10 days or more, 5 days or more and the entire universe of 1257 stocks. TPA was trying to see if there was a pattern of performance for more ETF Illiquid stocks.
The table below shows results for declines of 5% or more from 2012 to the most recent decline from 3/4/20 to 3/20/20. The categories of ETF Illiquidity are color coated to highlight better and worse performance. TPA also provides benchmark indexes to add context for each period. Finally, TPA has included a chart of the growth of ETF assets below the performance table.
TPA was looking to see how stocks with higher Illiquidity Scores or more days of trading volume held by ETFs performed in times of stress or large down periods in the stocks market. Specifically, how did they perform versus the market.
TPA groups stocks into the following categories by ETF Illiquidity: 30+ days, 20+ days,10+ days, and 5+ days.
TPA’s report looks at the results in 2 ways:
1. How did each category perform versus the other categories?
2. How did the ETF Illiquid categories of 30+ days, 20+ days and 10+ days perform versus the benchmark?
TPA makes the following observations
1. The results are a mixed bag until the most recent set of declines when ETF Illiquid stocks began to underperform the market consistently. This probably coincides with ETF assets reaching a critical level.
2. Looking at the market decline since 2/19/20, the 30+, 20+, and 10% categories all underperformed the S&P500.
3. Also, these categories underperformed significantly with the 30+, 20+ and 10+ groups down 35.01%, 34.61%, and 36.31%, respectively or -3.08%, -2.67%, and -4.38% less that the S&P500 which was down 31.93% in that period.
4. Finally, the average performance of the 3 categories was down 35.31% or -3.38% worse that the S&P500 performance of -31.93%.
TPA sees this measurable underperformance as evidence of the problem ETFs create for overall market structure as the size of ETF assets reaching a critical tipping point.
![](https://static.wixstatic.com/media/eb4ead_1078c854fa1f42ab9092f1a181f7e995~mv2.jpg/v1/fill/w_980,h_381,al_c,q_85,usm_0.66_1.00_0.01,enc_auto/eb4ead_1078c854fa1f42ab9092f1a181f7e995~mv2.jpg)
Conclusion: TPA will continue to monitor ETF Illiquidity for clients, but it may be that the assets in ETFs have finally hit a critical juncture at which the liquidity paradigm is breaking down, whether or not AP’s stand ready to create and redeem.
FOR THE FULL REPORT, INCLUDING THE ETF ILLIQUIDITY FOR ALL 1257 STOCKS, SUBSCRIBE TO TPA. THE ETF ILLIQUIDITY REPORT WILL BE UPDATED REGULARLY AND AVAILABLE TO SUBSCRIBERS ONLY.
![](https://static.wixstatic.com/media/eb4ead_7c99105bc2f944839978d0fc89414062~mv2.jpg/v1/fill/w_980,h_350,al_c,q_80,usm_0.66_1.00_0.01,enc_auto/eb4ead_7c99105bc2f944839978d0fc89414062~mv2.jpg)
TPA and many other have argued that the ETF creation and redemption process is essentially “stickier” and more cumbersome than trading stocks. There are just more moving parts. In fact, an NBER study showed just that. NBER stated, “the stocks that are held within such funds experience substantially higher intraday and daily volatility than stocks without substantial ETF holdings.” (NBER Working Paper No. 20071), Itzhak Ben-David, Francesco Franzoni, and Rabih Moussawi. https://www.nber.org/digest/sep14/w20071.html.
The NBER study seems to support TPA’s ETF Illiquidity Report conclusions.
2. TPA identified 2 other metrics that indicate that ETFs are a lot less efficient in times of stress or more particularly, when stock prices decline.
· ETF bid-ask spreads widen when the market declines rapidly (not when the market rallies)
· The difference between the ETF and its intraday net asset value (INAV) widens when the market declines rapidly (not when the market rallies)
BID-ASK SPREADS - The first chart below shows the SPY price in the upper panel and the SPY average bid-ask spread in the lower panel over the past 3 years. SPY is the most liquid ETF. Note that when the SPY declines, the average bid-ask spread for SPY spikes. In fact, from 2/19/20 to 2/28/20 the SPY (S&P500) declined 12.44% and the average bid-ask spread for SPY widened by 68%. The average bid-ask spreads for the largest underlying stocks in SPY widened even more during declines. From 2/19/20 to 2/28/20, MSFT fell 15% and its average bid-ask spread increased 394%, AAPL fell 15% and its average bid-ask spread increased 296%, AMZN fell 13% and its average bid-ask spread increased 205%, FB fell 11% and its average bid-ask spread increased 265% and BRK/B fell 9% and its average bid-ask spread increased 294%. An important question is how efficient the creation-redemption process can be when the ETF and all of its constituents bid-ask spreads are becoming so inefficient?
![](https://static.wixstatic.com/media/eb4ead_265972b1595e42fb8875e8cbaed6f9dd~mv2.jpg/v1/fill/w_816,h_530,al_c,q_85,enc_auto/eb4ead_265972b1595e42fb8875e8cbaed6f9dd~mv2.jpg)
![](https://static.wixstatic.com/media/eb4ead_54f064411f1c43fdbe7829a03f435308~mv2.jpg/v1/fill/w_816,h_528,al_c,q_85,enc_auto/eb4ead_54f064411f1c43fdbe7829a03f435308~mv2.jpg)
![](https://static.wixstatic.com/media/eb4ead_f9638b1fe43e4d139ce52adbe555a9ab~mv2.jpg/v1/fill/w_819,h_528,al_c,q_85,enc_auto/eb4ead_f9638b1fe43e4d139ce52adbe555a9ab~mv2.jpg)
![](https://static.wixstatic.com/media/eb4ead_d19e9e3ea3004eaca26d5b3e32e0131d~mv2.jpg/v1/fill/w_821,h_531,al_c,q_85,enc_auto/eb4ead_d19e9e3ea3004eaca26d5b3e32e0131d~mv2.jpg)
![](https://static.wixstatic.com/media/eb4ead_7c4f142b841d494da69b191f66ba08c5~mv2.jpg/v1/fill/w_809,h_525,al_c,q_85,enc_auto/eb4ead_7c4f142b841d494da69b191f66ba08c5~mv2.jpg)
![](https://static.wixstatic.com/media/eb4ead_ef02897e5c0646e39c7a8591e4b74807~mv2.jpg/v1/fill/w_820,h_533,al_c,q_85,enc_auto/eb4ead_ef02897e5c0646e39c7a8591e4b74807~mv2.jpg)
![](https://static.wixstatic.com/media/eb4ead_f14e5709e0e348a0b0f668292f26797e~mv2.jpg/v1/fill/w_820,h_525,al_c,q_85,enc_auto/eb4ead_f14e5709e0e348a0b0f668292f26797e~mv2.jpg)
ETF Price – INAV
The difference between the SPY and SPY’s INAV (ETF and the value of the underlying stocks) also increases and becomes more volatile when the market falls. The table below shows the prices and INAV of SPY from 2/3/20 to 3/6/20. The average SPY-INAV difference mean, median, and standard deviation is shown below for the stable period from 2/3 to 2/19. TPA then calculated the man, median and standard deviation for declines 0n 2/24, 2/25, 2/27, 3/3, and 3/5 or -2.23%, -3.03%, -4.49%, -2.86%, and -3.32%, respectively. Note that the increase in the mean, median and standard deviation from normal to big down days was a substantial; +605%, +632%, and +477%, respectively. Note that this divergence in ETF price from underlying value occurred as the volume increase by 416%, which should have aided liquidity. Once again, how efficient the creation-redemption process can be when the ETF and its INAV begin to divert by this much?
![](https://static.wixstatic.com/media/eb4ead_f71e4a28ca6349e09e416855a163d90b~mv2.jpg/v1/fill/w_980,h_494,al_c,q_85,usm_0.66_1.00_0.01,enc_auto/eb4ead_f71e4a28ca6349e09e416855a163d90b~mv2.jpg)
The preceding was TPA’s analysis to formulate the question as to how market structure is effected by ETFs and how do ETFs perform given the size of the ETF sector and the amount of liquidity available during stressful periods (declines) in the market.
3. An EU Systematic Risk Board report (June 2019) brought up some of TPA’s concerns and also problems with co-movement of prices. The 3 situations below speak to the possibility of contagion, which may explain some of the recent volatility.
· “First, ETFs are associated with greater co-movement of asset prices: stocks tend to co-move more with their respective indices once they are included in ETF portfolios. This increase in the co-movement of asset prices may pose systemic stability issues, as it makes it more likely that many investors face losses simultaneously, therefore potentially leading to waves of insolvencies and synchronised sales.”
· “Second, there is evidence that ETFs are associated with increased price volatility of the constituent securities: the high liquidity and continuous trading of ETFs enable investors, including noise traders, to take large short-term directional positions on entire asset baskets. The unwinding of these positions in the ETF market can eventually result in crashes.”
· “Third, the arbitrage mechanism between ETFs and their constituent securities may operate imperfectly: ETF prices can deviate significantly from those of the constituent securities, especially at high frequencies, for illiquid assets and during periods of financial stress.”
4. ETF’S LARGE PERCENTAGE OF TOTAL VOLUME - To buttress this study’s argument about the impact of ETFs on the market structure, the chart below shows that in the past 15 years the percent of ETFs has risen from 2.1% to about 20% of all trading. Given that during the same period the number of listed companies has basically been cut in half (chart 2 below), we have a less diverse and less liquid market.
![](https://static.wixstatic.com/media/eb4ead_b1eb762dbee44aa8a77955a379529a31~mv2.jpg/v1/fill/w_810,h_398,al_c,q_80,enc_auto/eb4ead_b1eb762dbee44aa8a77955a379529a31~mv2.jpg)
![](https://static.wixstatic.com/media/eb4ead_39aee5f66b90468387a2c0f974c36a20~mv2.jpg/v1/fill/w_759,h_503,al_c,q_85,enc_auto/eb4ead_39aee5f66b90468387a2c0f974c36a20~mv2.jpg)
5. WHO IS DOING STOCK ANALYSIS? - TPA raises the question, with all of this passive investing. Who is looking under the hood of stocks? Besides the ideas that less people really understand what they are buying, there is the idea that when the market sells off, are investors as confident in an ETF they own for exposure to a sector as they would be with a stock they have analyzed and know a lot about? Managers may feel safer owning a basket of stocks, but is diversity the priority if all stocks are heading lower?
6. VOLUMES INCREASING AT THE CLOSE - Finally, the trend is that trading volumes continue to skew more heavily to the close. The chart below shows that the closing market share of volume from 2013 to 2018 increased from approximately 3.5% to 10%. There are many possible reasons for this, but one of the most obvious is that with so many benchmark products, it has become increasingly important that buy and sell order execution prices do not stray too far from the close. Of course, this ironically is moving away from one of the initial benefits of ETFs; that, unlike mutual funds, they could be traded at times other than the close.
![](https://static.wixstatic.com/media/eb4ead_232af1e9491e4670959bd83a27f5d8a2~mv2.jpg/v1/fill/w_980,h_423,al_c,q_85,usm_0.66_1.00_0.01,enc_auto/eb4ead_232af1e9491e4670959bd83a27f5d8a2~mv2.jpg)
CONCLUSION: ETFs have many valuable characteristics that can be very beneficial for investors. This report is not meant to detract from those benefits. This report looks at the effect of an ETF industry on market structure and behavior; especially during periods of market stress. The volatility of the past 3 weeks should force people to examine closely how this huge industry is affecting the stock market. There are implications that ETFs have reduced liquidity, added to volatility and, possibly, exacerbated declines.
Below are people who also worry about the growth of ETFs:
Mohamed El-Erian, Allianz chief economic advisor, and former co-CEO of PIMCO :
"Risk has become embedded into the system as to the proliferation of instruments that look extremely attractive, because they are very cheap — ETFs — but are overpromising liquidity…It is a huge risk of contagion… " 11/28/17
Marko Kolanovic, PhD, Global Head Macro Quantitative and Derivatives Strategy, JP Morgan:
“We think the main attribute of the next crisis will be severe liquidity disruptions resulting from market developments since the last crisis.” 10/14/17
Peter Dixon, the Commerzbank economist:
"My concern would have to be that they give the impression that there's lots of liquidity out there. But if everybody is trying to get out at the same time, the question has to be, do ETFs exacerbate or magnify the impact of the decline?" 11/28/17
Jim Reid, the Deutsche Bank Strategist:
"ETFs and ETPs have not yet been fully tested in a sustained bear market. So, the real test could be when we see the next downturn and these products are faced with heavy redemptions.” 11/28/17
Sources:
Comments