Feature Stories | Nov 10 2010
By Greg Peel
On May 6, 2010, Wall Street was trading in fear as the European sovereign debt crisis accelerated and by mid afternoon the Dow was down 350 points. Then suddenly, within a blink, the Dow was down 1000 points. Almost as quickly, the tape returned to being only 350 points down. It was not a publication error. It was real.
The event, which has since become known as the Flash Crash, has been blamed as a major reason why volumes on Wall Street dropped so significantly from historical averages in the ensuing months. Investors, and particularly retail investors, were scared away by what they perceived was a stock market now under the control of mindless computers as impotent regulators looked on helplessly.
The Flash Crash was initially blamed on a “fat finger” error – one in which a human incorrectly enters a sell price (perhaps misplacing the decimal point to be wrong by a factor of ten) into an electronic trading platform and in so doing “cleans up” every buy order below the market down to a certain level. Such errors have a long history and in days gone by have immediately assumed to be just that – an error – such that humans have been able to make a judgment call and not act as if the trade were “real”. But computer programs do not necessarily have such discretionary powers.
Thus it was assumed that the Flash Crash actually came about as an exacerbation of an error which triggered vast numbers of “high frequency” computerised trades (HFT) which were following pre-programmed trading algorithms that did not have “watch out for fat fingers” code implanted. But the question was then asked, what happened to the exchange circuit breakers?
Circuit breakers come into play when sudden large moves occur – say 10% – in order to cool down markets and avoid significant volatility. Trading is halted for a short period. It was soon revealed that HFT moves faster than exchange computerised circuit breakers, and that off-market electronic quasi-exchanges did not have circuit breakers. In the latter case, when the circuit breakers were activated the selling continued from outside the exchange and into a vacuum. Maybe this was the real problem.
The truth is that no one has been able to determine in the interim, with absolute certainty, what actually caused the Flash Crash. This failure has only deepened investor reluctance to return to the market.
HFT has its detractors, but at the end of the day HFT is really only trades moving a micro-second speeds in minimal volumes individually which in essence are little different to large volume trades moving slowly. The only real problem, as suggested, is the inability of computers to “think”. Most of the visceral attacks on HFT trading over past years have been born of ignorance and more akin to a witch hunt than a genuine debate.
More recently, debate has shifted to questions over the growing preponderance of exchange-traded funds, to the point where, in retrospect, ETFs are being blamed in some quarters for the Flash Crash and questioned in general.
ETFs first appeared on the scene in 1993 but in the noughties began growing in number exponentially. The introduction of a gold ETF a few years ago has been credited with a chunk of gold's meteoric rise in price given ETFs open up a simple investment opportunity in a commodity that was once difficult for smaller players to invest in.
While a gold ETF is relatively straight forward – the sponsor of the ETF holds physical gold on the buyer's behalf – there are also now hundreds of ETFs on combinations of stocks. Indeed, stock ETFs began to boom long before commodity ETFs hit the radar.
In a gold ETF, the price of the ETF is linked to the price of gold such that an investor can invest in gold without having to buy actual bullion himself. In a stock ETF, the investor buys an instrument which contains some basket of stocks such that by investing in the single ETF, the investors' return is linked to the price of those stocks.
For example, one of the most popular stock ETFs in the US replicates the S&P financial sector index. If you want to go “long financials” for a general reason rather than long just one bank stock for a specific reason, an ETF saves you the trouble of having to buy all the stocks in the index and pay all that brokerage and clearing fees and hold scrip.
There are now EFTs on all sectors and sub-sectors and on “defensives” or “cyclicals” or “growth” stocks or “value” stocks or gold stocks or just about any combination one can dream up. A recent addition, for example, is an ETF on only rare earth metal stocks, which are themselves rare.
If one buys a gold EFT, the issuer in theory buys some actual gold and holds it on your behalf, but you don't actually own that gold. You can't “cash in” for bullion, you can only cash in by selling or redeeming your ETF. ETFs are not “over the counter”, they are listed on the exchange just as stocks are traded and yes, there are now ETFs on collections of ETFs.
Similarly, while stock ETFs will contain a basket of stocks, you do not own those stocks and, again, you can only cash in by either selling or redeeming the ETF. You simply own an instrument which has a price linked to the value of those stocks. This means an ETF is a little like a futures contract, albeit open-ended and not subject to swings through premiums and discounts to fair value.
There is nevertheless no doubting an ETF is a “derivative”.
To create an ETF, a sponsor borrows the requisite stock via agreement with a lender, be that a market-maker or mutual fund etc. The stock is bundled into one instrument for which there is one price which moves dependent on the weighted movements of the prices of the stocks in the bundle. The sponsor then issues ETF “shares” to willing buyers who now own something that for all intents and purposes acts like a share. The difference is that while ETFs can be on-sold to another party like a share, they can also be redeemed like units in an investment fund at which point that “share” in the ETF ceases to exist.
Furthermore, the only way more shares of a company can come onto the market is by specific decision by the company to issue more shares (by whatever means including capital raisings, rights issues etc), but a sponsor of an ETF can issue more shares in that ETF at any time if the demand is there.
It is important to note that sponsors of ETFs do not themselves buy the stocks that make up the ETF, they merely borrow the stocks from someone who already owns them and continues to own them.
And then it gets more complicated. There are also a growing number of “inverse” ETFs on offer which are better known as “short” ETFs for which the price of ETF shares rise when the price of the shorted shares in the bundle fall. And one can also “short” ETF shares, whether it be a long or short ETF, as long as one can borrow the ETF shares from someone else. This is just the same as borrowing stock to short it.
All the while, even if demand for EFTs grows and grows an ETF is still only a price-based instrument over a collection of existing stocks. New shares in a company are not created if a new ETF is created and sold.
One strong argument that has thus emerged against ETFs in recent times is that their original advantages are now giving way to market disadvantage. At first, ETFs were lauded as a way of enticing more interest in stock market investment if the complication of portfolio management (ie dealing with bundles of stocks) was removed from the humble investor and handed over to a market professional. But let's not forget that while ETFs might create more interest, they do not create more buyers of stocks because the stocks in question are only being borrowed from someone who retains “beneficial”ownership. Thus an increase in demand for EFT XYZ does not actually affect increased demand for the stocks within XYZ.
The flipside thus is that if investors choose only to invest in the stock market via ETFs then there is never any demand for the individual stocks per se. Detractors thus argue that the reason why volumes of individual stock transactions have been falling on US stock markets can be directly traced to the rise in popularity of ETFs. Proponents of ETFs argue this is bunkum because, so far, ETF volumes only represent a small proportion of overall market turnover.
It has also been noted by detractors that while the US stock markets have largely been trading in a sideways range throughout late 2009 and all of 2010, intraday and shorter term volatility has increased disproportionately. It is straightforward to argue that lower volumes mean thinner markets means more scope for volatility, but ETF detractors now see a “tail wagging the dog” argument.
ETF share prices are supposed to move based on movements in the prices of the underlying stocks such that strong demand for those stocks will subsequently shift the price of the EFT shares. However, so popular have some ETFs become that it is apparent it is the ETF leading the price movement and the stock prices subsequently following, rather than the other way around. Because ETFs “lump” together basket of stocks so too have stock price movements become “lumpy” such that the usual “alpha” risk factors which mean individual stocks move in different directions in any given session are giving way to more “beta” like factors which mean markets move as a whole in the same direction.
Now if you consider that you can also “short” ETF shares, then it follows that if you really wanted to sell a whole lot of stocks in a hurry it would be easier just to short an ETF now and sort out the stock positions later. You'd be covered. It is this concept that has led to accusations that the Flash Crash actually came about, or was at least exacerbated by, ETF trading. Computers can buy and sell ETFs just as quickly as they can shares.
I believe the reason why the Flash Crash has never been fully resolved is because the real cause was a confluence of many factors – a fat finger, HFT, slow circuit breakers, non-existent circuit breakers in off-market exchanges, inflexible computer algorithms, futures trading, and, indeed, ETF trading.
While the Flash Crash was very scary at the time, and subsequently scary in its implications, there is a far more sinister fear rising now with regard to ETFs and their growing popularity. It relates to what are now the three most reviled letters in the financial market alphabet – C-D-O.
A collateralised debt obligation bundled together actual home mortgages into packages containing three tranches of risk – prime, mid-prime and sub-prime. The different tranches were then on-sold to investors, be they banks or hedge funds or even individuals who may well have borrowed money to buy them. As the US housing market boom rolled on, CDOs became increasing popular as investments. Soon there was more demand for CDOs than there were mortgages, so lending standards were significantly reduced. Eventually anyone who was homeless and jobless could own a home such that the demand for mortgage securities could be satisfied.
Soon there were those in the market who saw a bubble forming and quite possibly a crash coming. But you couldn't “short” a mortgage, and thus you couldn't “short” a CDO. And CDOs were not like shares – they were each individual and traded only over the counter rather than on any exchange.
But what you could do is buy “insurance” over CDO default, by means of a credit default swap (CDS).
CDSs were as much a factor behind the GFC as were CDOs. Even before we get to mortgages, one could buy a CDS on the debt of a company – say, General Motors – such that the holder would collect if GM defaulted on its debt. Such a thought was so preposterous that CDS issuers such as AIG sold more insurance over the debt than the debt itself. It is estimated that before the GFC, CDSs were held on GM to six times the value of the debt. Thus if GM had gone down, in theory six parties would be left to fight over who should be paid, or the issuer would have to pay out six times the value of the one lump of debt it insured.
GM did actually go down. But it was bought by the government.
So take the CDS concept another step on to mortgage CDOs instead of GM bonds and what we saw before the GFC was a similar situation. There might be multiple insurance policies in place over a single CDO. And toward the middle of 2007, the demand for CDSs on CDOs became so great and the remaining pool of available mortgages from which to create CDOs so small that investment banks and hedge funds started creating “synthetic” CDOs just so more CDSs on CDOs could be issued.
A synthetic CDO was simply an artificial CDO which contained no actual mortgages but tracked the price of a real CDO. If the real CDO suffered from complete default, the holder of a CDS on that CDO would expect a payout as would the holder (or multiple holders) of a CDS on the equivalent synthetic CDO.
Do you see where I'm going here?
It eventually came to light to the whole world (only a small few saw it coming) that a house of cards had been created that was based on multiple owners of the same underlying asset and multiple owners of insurance over that asset to the point where any concept of “beneficial” ownership had long been lost. On the assumption US house prices could never go down, a multiplicity of leveraged positions was created around that assumption. The house of cards was fine as long as all holders of positions didn't try to call in their markers all at once. But when the mortgage defaults began and US house price started falling, that's exactly what they did.
The argument which is now only really beginning to brew in the US is that for all intents and purposes, the nature and machinations of the stock ETF market are not too far removed from that of the CDO/CDS market. It's all about the ability of a lot of different people to profit from, or lose from, something they don't actually own.
One of the counter-arguments from ETF issuers is that were there, for example, to be a “short squeeze” on ETFs which in theory could have all sorts of scary flow-through ramifications, then a sponsor could simply issue more ETFs (over the same finite amount of borrowed stock) and thus quell the storm.
In other words, problems of too many owners could be overcome by creating more owners.
As it is often said with tongue in cheek, what could possibly go wrong?