Why some ETFs may blow up

Oaktree Capital’s Howard Marks warned in a recent memo that liquidity can be elusive. Financial instruments that trade with a narrow bid-ask spread today may be virtually impossible to sell tomorrow.

Particularly troublesome are structures that are meant to enhance the liquidity of a non-standardized and complex asset. In the decades leading up to the global financial crisis, mortgage-backed securities helped channel savings from European and Japanese institutional investors by creating liquid products with juicy yields. When many of the mortgages from which these MBS and CDOs went sour, liquidity disappeared in an instant.

There are reasons to believe that exchange-traded funds (ETFs) are not as bad as mortgage-backed securities. The good thing about ETFs is that we know exactly what assets they own, and for the most part these assets are liquid. So called authorized participants (“AP”) are allowed to create and redeem ETF units, which guarantees that the price of the ETF is close to the neet asset value of its underlying assets. If an ETF is trading at a discount to the market value of its assets, APs can buy the ETF and sell short the underlying assets. It can hand over this to the fund get a redemption basket of securities at NAV. It will then pocket the difference between the ETF price and NAV for a tidy profit.

The ETF liquidity crash of 24 August 2015

The market behaviour we saw yesterday (dubbed “The ETF liquidity crash” by @BluegrassCap) reminds us however that there are limits to such arbitrage. Heavy selling in ETFs over the weekend caused the S&P 500 to gap down almost 5%. Individual shares such as Baidu and HCA Corporation fell 26% and 49% before recovering most of the fall. This would never happen it is wasn’t for ETFs.

So what happened to Baidu and HCA? It appears as if owners of ETFs became worried about weak economic data from China and decided to sell into the open on Monday. Major ETFs such as SPDR’s technology stock ETF “QQQ” fell 13% from Friday’s close to bottom due to this heavy selling pressure. APs then bought the ETF and sold the underlying shares in order to make arbitrage profits.

But – the issue is that underlying shares may not be as liquid as the ETF itself. The total outstanding value of ETFs have grown from $750 billion to almost $3.0 trillion since the global financial crisis. And actively managed funds have seen significant outflows.


And so they have become very large compared to the total market capitalizations of underlying assets. And some shares may be less liquid than the average constituent, perhaps due to low free float or low institutional sponsorship. There is another issue: when you sell short shares, you need to borrow the shares from someone. Your broker actually has to find those shares and lend them to you.

So if this arbitrage mechanism is put out of play, ETF prices may fall more than is warranted by the movement in the underlying shares. And underlying shares themselves can move downwards dramatically if they are relatively illiquid due to selling pressure by APs.

There are profits to be made if you are able to spot such short-term inefficiencies and front-run authorized participants in their quest to lock in ETF price-NAV differentials. Given low holding periods and hedged position, the risk-reward can be highly favourable. Hats off to the guys who bought shares cheaply on NASDAQ and NYSE yesterday.

Yesterday’s episode also shows that the limits to arbitrage may be bigger when ETF prices are falling than vice versa. This is because the short sales of stock are more difficult to carry out than short sales of ETFs (you need a broker to find shares that you can borrow). This means ETF prices can fall very rapidly in a crash since APs will find it hard to engage in arbitrage transactions.

Reflexivity in ETFs

When retail investors buy ETFs, they typically have no idea what they are getting into. They buy certain “themes” that they have heard about from their friends or from media. Even sophisticated investors may buy ETFs without regard for values under the guise of wanting to have a certain “exposure” or “tactical asset allocation”. Certain wealth managers in Singapore are recommending their clients to invest in themes such as Biotech and Technology via low-cost ETFs as they have low management fees and are liquid.

Buying ETFs is appealing from a psychological perspective. We respond to exciting stories, and we want to be part of something that gives us hope. Buying ETFs don’t incur much psychological costs either. ETFs are perceived to be cheap as management fees are low and bid-ask spreads are tight. Worry is kept low by the fact that they are liquid and can be sold at any time through online trading platform.

When a sophisticated investor owns a stock that rises to his target price, he will sell. When a retail investor buys an ETF to get “exposure” or participate in a trend, he may not sell when the price reaches astronomical levels. Instead he will feel superior in his judgment, and see the price increase as evidence of his foresight. If ETF prices crash, he will grasp for any story that explains the price movement. If CNBC headlines tell him that this is just a hick-up and that the bull market is still intact, he may keep his ETFs but stay a bit worried. If ETF prices crash, he will hear convincing arguments of why we are entering into a new Great Depression and he will most likely sell with no regard to underlying value. The selling pressure will escalate because of “information cascades”: the more his friends are selling, the more often he will be reminded that it is the appropriate action to take. Therein lies the reflexivity.

Mutual funds are quite similar. Especially in the late 1990s, when retail participation in technology sector funds was large and driven by new era theories reinforced by continuous price increases. The managers of these Internet funds had no choice but to keep buying the shares that fit into their tightly defined investment mandates. Plus, the shares were going up, so why not keep riding the bull? It provides a certain excitement, after all.

The assets that ETFs invest in are broader and retail participation is lower. But ETFs are worse in one respect: they are almost all index funds. Mutual fund managers may not outperform their benchmarks, but most would not buy Shake Shack at PE 1000x. As an example: IPO ETFs whose mandates are to buy any IPO that meet liquidity conditions 5 trading days after IPO and sold 2 years after, *will* buy the stock regardless of its price/value ratio.

This causes extreme correlation between assets, as Horizon Kinetics pointed out in this quarterly letter. True value investors may short-sell the expensive stocks and buy the cheap, but when the AUM of ETFs is large and share counts are rising – why stand in front of a runaway train? It is much more profitable to front-run ill-suspecting retail investors.

There is another type of reflexivity. Asset prices are determined by supply and demand, and when global liquidity is overflowing, some of that liquidity is likely to find its way into stock markets and thus push up prices. Money supply is hard to define though. Some people view gold as the only form of acceptable means of transaction and store of value. Most people believe that liquid types of credit such as bank deposits also serve the purposes of money and can then be used as such. But reality is not as clear-cut. Professor Robert Mundell has pointed out that certain assets such as Unites States Treasuries may have most of the properties of money and can therefore be seen as “money-like”. His point is that increased liquidity of assets increase global liquidity conditions. When investment banks create liquid tranches of RMBS, the underlying mortgage goes from being stuck on an S&Ls balance sheet to something that is marginable and that can be exchanged for other assets. The velocity of money then increases and global liquidity conditions improve. The same can be said for ETFs, which can be used as collateral (which increases creditworthiness) and are often more liquid than underlying assets. ETF inflows hence improve global liquidity, which enables more transactions and more leverage and hence higher asset prices.


A more serious issue with ETFs is that APs have the authority to create new ETF shares that don’t need to be backed by anything. Right now, Market Vectors Semiconductor ETF (ticker: SMH) for example has a short interest ratio of 301%. It means that there are 3x more investors that have borrowed the shares and sold them than there are underlying stocks. This becomes especially troublesome when the valuation of the underlying assets is high and ETF AUM is large in relation to underlying market caps. Most of the buyers of these ETFs don’t know that they actually bought them from short-sellers, who in turn got them from Authorized Participants, who created the shares ex nihilo.

What happens then if 1/3 of the owners of SMH sell at the same time? The ETF price will fall. APs would buy the ETF, short-sell the underlying shares and hand over the ETF to the fund for a redemption basket of shares. After 1/3 of owners have done this there would be no shares left in the fund and the fund sponsor (firms such as Blackrock and WisdomTree) would be left with an unfunded liability to meet the redemptions. How will they meet this unfunded liability? Or will the broker-dealers that sold the non-existent shares in the first place take the hit?

What is interesting of the above scenario is that more short-selling of ETFs creates more shares, that makes the ETF even more undercollateralised. So just like in a run on a currency in a fixed-rate currency regime where market participants actively target currencies with low reserves in relation to money supply, market participants can target ETFs with high short interest ratios and sell them short even more heavily. Sooner or later, the negative price action in the underlying shares transmitted through arbitrage will scare retail investors to sell off their ETF holdings and cause the ETF to go bust. It remains to be seen if this is possible in reality, or if fund providers and regulators are clever enough to avoid such a situation from happening.


  • Don’t own any ETFs unless you know that underlying assets offer good value and/or ETF inflows are likely to continue.
  • Short selling of ETFs may provide excellent risk-reward, especially when the underlying shares are illiquid and ETF share counts are on a downtrend (which is admittedly, hard to prove). Examples may include bank debt ETFs or frontier market equity ETFs.
  • Analyzing the constituents of hot ETFs whose share counts have gone ballistic and then retreated may provide good short ideas. IBB may be a good example right now. If the individual shares are overvalued, we know full well why that is the case.
  • If you believe (as I do) that providers of ETF may be left holding the bag in a blowup, it may be worth shorting their stock. There are some risks of doing so however: inflows into ETFs may continue for another few years.


This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s