On 13 November 2018, Mint carried an article by Nilanjana Chakraborty that explained the basics of algorithmic trading and its current state in India. Algorithmic or automated trading refers to using a computer program that automatically submits trades to an exchange without any human intervention. This ensures that human beings’ emotions are not triggered while making the investment or trading calls and, instead, the computer program is able to place trades with a level of speed and precision that a human could not hope to achieve.
Algorithmic trading is popular among institutional investors in India and covers 35-40% of the turnover on our exchanges. Early in 2018, the Securities and Exchange Board of India (Sebi) had probed the algorithmic trading platform available on the NSE and had found that some investors had unfair access to market information and trading systems. Since then, Sebi has instituted a stronger regulatory framework for such trading.
Chakraborty also dissected what such algorithmic trading meant for the retail investor. She concluded that the technology to do this is not easily accessible to small-time investors, who would find the costs of setting up a separate set of servers to perform such trading too high. Though there are no guidelines for retail investors, some brokerage firms do offer algorithmic trading as a product. Last year, it was reported that Sebi is likely to put out guidelines for algorithmic trading for retail investors; these are still awaited. According to Chakraborty, it is important to keep in mind that algorithmic trading is a market (trading) strategy and is not meant for long-term investors.
However, globally, traditional asset management could well be an industry ripe for technological disruption by using smart algorithms. Algorithms armed with computing power can extract value from big data by recognizing and unearthing patterns from the ever-changing construct of financial markets, thereby making the human “investment officer” obsolete.
I spoke recently with Mazhar Khan, founder of a startup, still in “stealth” mode, who intends to apply the principles of algorithmic modelling to building a fund portfolio for long-term investors. Khan has more than two decades of experience and has managed advisory assets of $2 billion while working for multinational private banks. He says that the inconsistency in the performance of equity mutual funds available in the market today means it is difficult to construct investment portfolios for clients.
To support this, he cites a study by Spiva that shows that 85-90% of US equity mutual fund schemes have underperformed their benchmarks across 3-, 5-, and 10-year horizons. Spiva or “S&P Indices Versus Active” is a scorecard published by S&P Dow Jones Indices that compares the performance of market indices versus actively managed mutual funds. Khan says that this underperformance is not restricted to the US and is also prevalent in Europe and in emerging markets. As markets become more volatile and unpredictable with a high velocity of information flow, performance consistency seems to be no more within the realm of traditional mutual fund managers. Over the last four years, more than $4 trillion has moved out of the $18 trillion US mutual fund industry into passively managed index funds and exchange-traded funds or ETFs.
Khan has been on a course of exploration in behavioural science, financial theories and quantitative models, the result of which, he claims, is an algorithm-based investment model suited to the global long-term investor. According to him, the model is grounded in the fundamental performance of companies and in the behavioural constructs of stock price movements. These behavioural constructs are not based on technical analysis, but on the long-term statistical data behind stock performance. By contrast, pure algorithmic trading focuses on short-term market anomalies or arbitrage opportunities.
Considering the speed of development in Artificial Intelligence, Khan believes that large-scale analysis of historical fundamental performance and behavioural constructs is now technologically feasible. Moreover, the absence of asset managers and their minions from such “algorithmically managed” investments will mean that fund costs are kept low for the investor. The “secret sauce” will lie in the algorithmic model’s construct. Khan says that industry leaders such as Larry Fink of BlackRock, Inc. and Paul Tudor Jones of Tudor Investment Corp. have seen the light and are opening up more of their portfolios to algorithmic long-term investment. The Indian asset management industry on the other hand, he says, has been extremely slow and hesitant to adopt to the changing trends shaping global markets.
For his startup, the technology roadmap he plans is to scale his algorithms to provide a robust mechanism to innovate and roll out multiple thematic “quant” and “smart Beta” strategies that can be offered as funds, ETFs and index funds, depending on the maturity of specific buyers.
He intends to go to the market approaching both institutional investors, such as family offices and pension funds, as well as small to medium-size funds, including existing registered investment companies and “long-only” hedge funds.
The startup’s idea seems a fine one, but the actual technological and operational execution of it is where success will lie. This will be Khan’s real challenge.
Sidhharth Pai is founder of Siana Capital, a venture fund management company focused on deep science and tech in India