Most countries, except the OPEC, were rejoicing last week on two counts. First, the killing of Osama bin Laden. Second, the nose-diving of oil prices. The price of crude oil plunged 15% last week, its steepest drop in two and a half years. Never before had crude oil plummeted so deeply during the course of a week. In commodities markets, oil is a superstar, and its daily contract turnover, typically around $200 billion, is usually able to absorb even large inflows or outflows of investment, without much price change.
Extraordinary price changes, like that of a 15% move last week, are usually set off by dramatic events. The last time such a large change in oil price happened, albeit an increase, was during the outbreak of the first Gulf war in 1991. Of course, the news of the killing of bin Laden was pretty dramatic and like an antithesis of the Gulf war. But, irrespective of what Obama says, it is too far-fetched to believe that the world has become so safe that oil prices have started falling.
So, what caused this sudden decline in oil price? The culprit seems to be algorithmic trading. First, the evidence. Data from commodity exchanges shows that the total number of open positions in the oil market?a number that should typically fall in a selloff?instead increased big time. Normally, panicky investors selling oil en-masse would cause total open interest numbers to shrink, as exited investors close out contracts. However, in this case, it has been the opposite. The long positions have indeed decreased but there has been such a large volume of short positions that the net open positions have swelled rather than shrunk. The reason seems to be programme trading.
Computer programmes are used by large banks and funds to trade. For instance, if I am a trader in a bank and a client gives me instructions to close out a trade if a particular price point is reached before the next morning, I don?t have to keep watching the screen all night. Instead, I can input the information into a software and the trade would get automatically executed if the desired price condition is met. Similarly, if I have a long position on a WTI crude futures contract and I think that it will fall even further if the price goes below $100, I can instruct my computer to square my long position, the moment the price goes below $100. In this case, $100 is my stop loss limit, meaning my loss threshold is reached when the price touches $100. It is only fair to assume that similar machines at other firms, from New York and London to Geneva, Tokyo and Singapore, would be automatically selling in much the same manner, causing prices to decrease even more. I can instruct my machine to not only dump long positions but also build short positions if prices plummet, because when prices go down, short positions make money.
From a trading perspective, one of the good things about algorithmic trading is that as a trader I may not be comfortable changing my view so quickly. For instance, if I had a long position in crude oil, I was probably expecting prices to increase rather than decrease. It may take some time for me to change my view diametrically. However, with algorithmic trading, instead of breeding hesitation, abrupt price drops can quickly and unemotionally prompt these machines to unload a bullish long position in oil, and build a bearish short one instead. Also, potential human error involved in actively trading a volatile market can be avoided. Such algorithmic and high frequency trading accounts for about half of all the volume in oil markets. What stands out in the tumble of oil price last week was the way computers turned readjustment of positions in a huge and deep market, into a stunningly large jolt of a 15% change.
That said, some of the seeds for the drop in oil prices were sown earlier. Last month, Goldman Sachs issued two notes to clients in rapid succession, recommending they pare back positions. In one, the bank called for a nearly $20 dollar near-term correction in Brent crude. The main reason being that the recent rise in oil prices had been fuelled, in part, by the Fed pumping cash into the markets by purchasing $600 billion in bonds. This programme had pushed interest rates extraordinarily low, making borrowing essentially free once adjusted for inflation. Investors have been using the super-cheap money to invest in oil markets. But the Fed?s programme is slated to end in June. It is natural to expect that investors would likely take profits before June and were eyeballing each other to see who would take profits first. This, in turn, led to a closing out of long positions, while the algorithms ensured a build up of large short positions on the back of falling prices.
In the space of just hours, the drop in the price of crude oil had shaved nearly $1 billion off the cost of supplying the world?s daily oil needs. That is terrific news for all countries except OPEC. For a change, traders and algorithms are being given a left-handed complement for doing something that is beneficial for the real economy.
The author, formerly with JP MorganChase, is CEO, Quantum Phinance