Sri Lankan trader, Raj Rajaratnam of the Galleon hedge fund is remembered today for the landmark insider trading case where he, along with Rajat Gupta and sundry others, got jail sentences. It should not be forgotten, however, that Rajaratnam was a very smart man who worked out a logical way to make money.
Rajaratnam once described his preferred methods as "arbitraging consensus". Listed companies have to declare results every quarter. Analysts project what the likely numbers will be. Their forecasts can be averaged out to produce a consensus estimate. If that consensus is matched by the actual results, the share will not see too much volatility. If the consensus is beaten, the share may jump. If the consensus is missed, the share may tank.
Galleon wasn't interested in the actual results or the long-term implications of the results. It was interested in knowing the consensus expectations, and it would take large positions when it suspected that the consensus was wrong. The only issue with that of course, is that Rajaratnam started to look for insider information to enable those "arbitrage" trades.
He also had an ability to crunch balance sheets quickly and to manage complicated trading positions including derivatives. As a result, right from his early days, he had a good track record in figuring out when consensus was likely to be wrong in tech stocks. Unfortunately, the information-gathering soon turned into actively looking for insider information, rather than just picking up on trends and understanding their implications.
The trading strategy is interesting, however, and there's no reason why a trader cannot implement the idea systematically through entirely legal means. In fact, many traders do arbitrage consensus, but they tend to do it in a hazy, unfocussed way.
Arbitraging consensus works best when one is crowd-sourcing lots of information. Ideally, a target stock should possess large and varied institutional interest and there should be many analysts tracking it. That way, the average of the estimates is more likely to be close to the reality and when it isn't, the responses are more likely to be violent.
Averaging the consensus by simply deriving an arithmetic mean or a median is possible. But it is also possible to weigh the consensus in various ways. Which institutions own the stock? Of those, which ones are likely to be aggressive traders? Which analysts track the stock? How seriously are they taken and who has links with which institution? What sort of declared insider trading has occurred? This is all either public, or semi-public data, since research reports are easily available and of course, ownership is known.
What sort of recent history does the stock have in terms of behaviour when consensus is exceeded or missed? Again, the data is public and easy to get. Does there tend to be a typical build up in price volatility and high-volume transactions prior to the quarterly releases or spikes immediately after? Do any peer companies declare their results earlier and if so, is there correlation in price trends? If there are such patterns, decoding trends may be possible.
This may seem like a lot of research and hard work. However, there is plenty of time to research and prepare if you wish to implement this strategy. There will be a narrow window of, say, five sessions per stock every three months.
Almost every stock has higher than normal volatility in the five sessions or so sandwiching its results. What is more, this isn't just random volatility with prices moving up and down with greater amplitude. The chance of a serious trend developing is much higher in the window around results.
If you can pick up the trend, you may get a long ride. So, taking focused concentrated bets at that point of time can be very rewarding. Of course, it may also lead to losses. Even somebody who does good research is going to be wrong often, if only because of the reliance on behavioural patterns.
The trader needs excess margin in hand, wider stop losses and a good sense of how investors in a given stock behave. Now is a good time to start putting together a consensus arbitrage strategy for Q3.
Rajaratnam once described his preferred methods as "arbitraging consensus". Listed companies have to declare results every quarter. Analysts project what the likely numbers will be. Their forecasts can be averaged out to produce a consensus estimate. If that consensus is matched by the actual results, the share will not see too much volatility. If the consensus is beaten, the share may jump. If the consensus is missed, the share may tank.
Galleon wasn't interested in the actual results or the long-term implications of the results. It was interested in knowing the consensus expectations, and it would take large positions when it suspected that the consensus was wrong. The only issue with that of course, is that Rajaratnam started to look for insider information to enable those "arbitrage" trades.
Also Read
Rajaratnam's early career and training (electronics engineer with a B-School degree) made him ideally suited for understanding both the intricacies of technology and financial markets. He "got" the implications of new hardware and software. Plus, he could leverage his South Asian ethnicity, which helped him to develop contacts with the influential desi techie community in Silicon Valley.
He also had an ability to crunch balance sheets quickly and to manage complicated trading positions including derivatives. As a result, right from his early days, he had a good track record in figuring out when consensus was likely to be wrong in tech stocks. Unfortunately, the information-gathering soon turned into actively looking for insider information, rather than just picking up on trends and understanding their implications.
The trading strategy is interesting, however, and there's no reason why a trader cannot implement the idea systematically through entirely legal means. In fact, many traders do arbitrage consensus, but they tend to do it in a hazy, unfocussed way.
Arbitraging consensus works best when one is crowd-sourcing lots of information. Ideally, a target stock should possess large and varied institutional interest and there should be many analysts tracking it. That way, the average of the estimates is more likely to be close to the reality and when it isn't, the responses are more likely to be violent.
Averaging the consensus by simply deriving an arithmetic mean or a median is possible. But it is also possible to weigh the consensus in various ways. Which institutions own the stock? Of those, which ones are likely to be aggressive traders? Which analysts track the stock? How seriously are they taken and who has links with which institution? What sort of declared insider trading has occurred? This is all either public, or semi-public data, since research reports are easily available and of course, ownership is known.
What sort of recent history does the stock have in terms of behaviour when consensus is exceeded or missed? Again, the data is public and easy to get. Does there tend to be a typical build up in price volatility and high-volume transactions prior to the quarterly releases or spikes immediately after? Do any peer companies declare their results earlier and if so, is there correlation in price trends? If there are such patterns, decoding trends may be possible.
This may seem like a lot of research and hard work. However, there is plenty of time to research and prepare if you wish to implement this strategy. There will be a narrow window of, say, five sessions per stock every three months.
Almost every stock has higher than normal volatility in the five sessions or so sandwiching its results. What is more, this isn't just random volatility with prices moving up and down with greater amplitude. The chance of a serious trend developing is much higher in the window around results.
If you can pick up the trend, you may get a long ride. So, taking focused concentrated bets at that point of time can be very rewarding. Of course, it may also lead to losses. Even somebody who does good research is going to be wrong often, if only because of the reliance on behavioural patterns.
The trader needs excess margin in hand, wider stop losses and a good sense of how investors in a given stock behave. Now is a good time to start putting together a consensus arbitrage strategy for Q3.