In the last 10 years, algo trading and co-location have been in the news for the right reasons and for all the wrong reasons. Algorithmic trading or Algo trading was permitted after SEBI permitted Direct Market Access (DMA) in 2008. Let us first understand what algo trading is all about. To begin with, algo trading is not the same as co-location although that is something people tend to use interchangeably. Algo trading is not even High Frequency Trading (HFT), although they normally go together. So what is algorithmic and co-location trading and what do we understand by algorithmic trading risk management. Above all, it is also essential to understand the risks of algo trading.
What is algo trading all about?
Imagine that you are a broking house with a large institutional business volume each day. You would obviously be having a large army of sales traders and dealers talking to clients and executing orders as per their requirements. Even with the best of dealers at your disposal having great motor skills, you have a few key limitations. Firstly, humans can never match up to the speed of computers and high-end servers when it comes to speedy and accurate execution. Secondly, machines can execute repetitive transactions with speed and accuracy. Thirdly, human dealing can never execute conditional orders that considers an entire matrix of conditions. That is only possible through machines.
An algorithm or an algo is nothing but a computer program that automatically executes conditional orders on the stock exchange based on pre-set parameters. At a very simple level, a slice order will help you spread the buying or selling of a stock through the day to overcome the volatility in the price. At a more complex level there are options algorithms that can execute based on conditional undervaluation and overvaluation of options or arbitrage positions that can be executed automatically based on the spread defined.
Pros and cons of algo trading..
Globally, markets operate substantially on algorithm driven trading. It makes the market broader and the large institutions act as automatic market makers. This improves the market quality. From a broker’s perspective or from a trader’s perspective, the algo can be set in such a way as to get the best possible price. Secondly, since algos are executed automatically and with great speed, there is tightening of spreads in buy and sell orders. That substantially reduces the risk for traders as the basis risk is largely reduced for buyers and the sellers. Lastly, algo trading also proffers the advantage of improving the liquidity in the market. An algo keeps throwing orders in the market and then withdrawing it. This is something that SEBI plans to regulate more closely now. The bottom-line is that this aggressive interplay of orders helps improve the liquidity in the market and facilitates the execution of transactions seamlessly.
Are there are downside risks to algorithmic trading? To begin with, algos can be unfair to small and retail investors. Most small investors do not have access to algos and these are primarily used by institutions and proprietary desks of brokers. Since these algo users manage to corner most of the volumes in the market, this can be unfair to small traders. Retail investors find that either the volumes are not available in the market or they are not available at the price of choice. Secondly, there is a major systemic problem with algo trading. Not only India but other countries too have faced this problem quite often. They are called fat finger trades. Sometimes the person feeding the algo may make a mistake or set the wrong price or set wrong conditions. This can lead to a virtual flash crash in the markets as the NSE saw in the case of Infosys around 7 years ago. Execution flaws can also happen because of flaws in the design of the algos. For example, algos may go into an endless loop or they may execute outside the condition due to software glitches. Algos are tested and approved by the NSE but obviously the stress testing is done only under normal trading room conditions.
How co-location is something entirely different..
To begin with, co-location is a facilitator of high frequency trades. That means you get price feeds a split nanosecond quicker than the rest of the market gets. For that, the broker is required to put up a co-location server at the exchange which will give them that fractional advantage. Let us understand the relationship. Co-location of servers on the exchange actually facilitates high frequency trades (HFT). This HFT, in turn, facilitates execution of complex trading programs through algorithms. It is only with co-location and HFT that algo trading actually begins to add value to traders!