This morning the ETF industry was shaken when news broke that Kweku Adoboli, an ETF trader and co-director at UBS' Delta One desk, had lost $2 billion due to fraudulent trades. My initial reaction was annoyance that another idiot had brought a negative spotlight to the ETF industry. Then I examined the facts.
The key to understanding this story and its implications for both UBS and the ETF industry is to understand delta-one trading. Despite what some may have you believe, in the case of ETFs, I can assure you it's not rocket science "¦ and it shouldn't be all that risky.
In simplest terms, delta-one trading involves the trading of products that have a delta of 1 relative to their underlying assets. In other words, there's no optionality involved. The connection between security A and security B is direct.
It's arbitrage trading in the purest sense of the word, and it fits well with ETFs. If the SPDR S&P 500 ETF (NYSEArca: SPY) were to move up 2 percent, for instance, you would expect the aggregate movement of the underlying assets (the S&P 500) to move up 2 percent as well.
On a Delta 1 ETF trading desk, ideally, a trader would be long shares of the ETF and short the underlying basket, or vice versa. The arbitrage comes in by taking advantage of moments when the two prices run out of line. The art of delta-one trading is to capture these micro-inefficiencies in market pricing, and translate them into low-risk profits.
A Simple Process Unhinged
That's what makes the UBS story so astonishing. It's hard to believe that you can just "discover" billion-dollar losses from delta-one trading. It takes time to rack up those kinds of losses, and you have to be particularly bad at it.
In the U.S., for instance, a typical cost of a creation unit"”the minimum size at which you can create new shares of ETFs"”is about $2.5 million. A delta-one trader looking to arb price inefficiencies would be long $2.5 million in ETF shares, and short $2.5 million worth of underlying assets. In this example, let's assume that the trader is perfectly hedged. When the ETF shares lose value, ideally, the trader's short in the underlying assets should gain in value with a net exposure of zero.
Of course, there are days when even what may be perceived as a perfect hedge can go out of line and you may lose in both of your positions. If you're properly hedged, however, there should be a reversion to mean"”meaning that things should come back in line soon enough.
I worked in ETF arbitrage for some time, and the way it works for the most part is simple.
As the ETF shares rise above their net asset value"”presumably because of significant buying interest"”a delta-one trader starting with a perfect hedge will begin to sell his shares of the ETF. At the same time, he covers the short position he established in the underlying assets by buying those underlying assets at net asset value. The difference is the profit that he gets to take home.
The delta-one strategy works very well for ETFs whose underlying assets can be hedged during the same market hours that the ETF is trading. This is why the difference between SPY and its indicative net asset value is always hovering around 1 cent"”it's easily hedgeable.
It gets more difficult"”and riskier"”in the case of international and fixed-income ETFs, because not all of the underlying assets are accessible due to closed markets or a lack of inventory. Traders use futures contracts or correlated assets to serve as a proxy hedge, introducing varying degrees of risk into the equation.
Read Full Article »
