Quantitative investing simply takes the emotions out of buy or sell decisions. You may use fundamental or technical factors. You may trade for the short or long-term, but the trades are based on fixed, logical criteria, explains Sandeep Tyagi, founder of Estee Advisors. Making its debut in August 2008, Estee Advisors was one of the earliest portfolio managers in India to offer quantitative trading and algo trading to investors. Excerpts from an interview:

How do quantitative investing models work? Can you explain with respect to your own products? And can you share their track record?

Our firm is four-years-old. In India, quantitative strategies came into vogue after SEBI’s announcement in April 2008 allowing machine based and algo trading.

Take our first strategy - I-Alpha. This is a market neutral fund that identifies arbitrage opportunities. A market usually has many instruments that represent the same underlying investment opportunity. The best way to take advantage of arbitrage is through technology. The basic equations can be fed into the system as formulae. Last year this product produced a 17 per cent return with low volatility. The risk reward ratio was favourable even compared to some debt products.

A variant of this strategy – I-Future is specifically focussed just on cash-futures arbitrage. This is for people who do not want to earn interest on their margin accounts or cannot earn it, like Shariah investors or FIIs.

The third strategy is I-Systematic, a long-short fund, where we take statistical bets on stocks that are over- or undervalued. Say we have six PSU banks in the country. Now those six bank stocks will react to some bank-specific information and some sector-specific information. We take a statistical view that the prices of these six stocks should be linked. If prices move away from these relationships, we take counter-positions. If we think a stock is undervalued based on our factors we take long positions and if we think it is overvalued we take short positions.

This strategy is expected to make money whether markets go up or down as a whole. This product has fetched a return of 4 per cent since inception, while the market has fallen about 5.5 per cent in the same period. It has less than half the volatility of the index too.

Do your quant models use fundamental factors or technical indicators?

We use both. The fundamental models or factor models track over 40 different fundamental factors including earnings, revenue growth, margins, cash flow ratios and so on. Then there are models which are purely based on price and volume too. Then we have behavioural models where we have ways of gauging whether a stock is oversold or overbought and capitalising on that.

Are quant-based models usually for only short-term trading or can they be used for longer term investments as well?

Quant investing is nothing magical. It is simply saying that every decision that you make has to be systematic. We look at the same factors that a fundamental analyst does. It is only that we are very precise about what and how we will look at it. In that sense quant based investing is very transparent. You often don’t know how a fundamental analyst is reaching his conclusion. But quant-based strategies are quite the opposite. Today we run strategies with six month holding periods too. It is also possible to run strategies for multi-year periods. When you run long term strategies you need to take a call on the stability in pricing, volatility and return relationships that the model tracks.

Quant-based strategies seem to depend on markets being rational. But as the saying goes the markets can remain irrational longer than you can remain solvent….What if your trades don’t work?

That is absolutely true. In fact, all quant based trades have a specific time frame within which they must work. On relative value type of strategies, stocks can go in the opposite direction from what you expect. That is why risk control is important. You run a portfolio of trades. For your investment to deliver, all your trades need not be profitable. Indeed they can’t be. Only your overall portfolio needs to be profitable. You need to make sure your aggregate portfolio is profitable.

And the models have to be quite dynamic too. We have a team of 50 analysts who are constantly trying to find patterns from historical stock market data.

Can you share a few of those findings?

In bull markets, momentum plays a big role. Stocks which are good value, may underperform for extended periods of time. In such periods, we have to be careful about investing too much based on purely the underlying value of the stock. Such a portfolio will underperform the momentum stocks. Also we use different factors for each major sector. What works in identifying relative value in a bank will work the same for an energy company.

Finally, we are very particular about execution slippages in the markets. Many a time, analysts find good theoretical returns but can’t implement their strategies profitably in the markets due to execution or trading slippages. We analyse the historical data as well as build strong execution algorithms to get good prices in the market thus minimising slippages.

Typically what kind of clients do quant-based trading?

Corporate treasuries, family offices, high net worth individuals have invested in our strategies. I know there is a feeling that such strategies are only available to big institutions and that they marginalise the small investor. But these strategies do help markets and in turn all types of investors.

The regulation on algo trading is still evolving with SEBI recently issuing a circular on the subject. What is your view on this?

There is generally a ‘fear of unknown’ on the part of investors and the regulator about quant-based trading. Whenever there is lack of knowledge people generally assume the worst. I think it is necessary to understand that computer based trading is simply a means of interacting with the market using technology. It is no different from any advance in technology. Yes, people can misuse it. The more powerful the technology the more careful you need to be on how it is used. But in the in the hands of people with the right risk management measures, it advances the market, reduces costs and so on. So we need to make sure that it is not abused.

We do welcome SEBI’s recent regulations on algo trading. It puts in place boundary walls on algo trading, restricting unnecessary risk, placing limits on open orders etc. We were already in full compliance with them when the order came and our algos were approved with the exchanges.

Related Topics
comment COMMENT NOW