Multi-factor investing can help reduce drawdowns, deliver better risk-adjusted returns: Sukanya Ghosh

Sukanya Ghosh, Fund Manager at SBI Mutual Fund.

SBI Mutual Fund, India’s biggest asset management company, is launching a quant fund, which will follow a multi-factor investment approach.

This fund marks the entry of SBI Mutual Fund, which manages assets worth Rs 11 lakh crore, into the quant space.

Quant funds, also known as quantitative mutual funds, use quantitative analysis to make investment decisions.

SBI Quant Fund, specifically, will follow a multi-factor approach based on momentum, value, quality, growth and low volatility.

Moneycontrol caught up with Sukanya Ghosh, Fund Manager at SBI Mutual Fund, to understand the fund better.

Edited excerpts:

Take us through SBI Quant Fund’s investment philosophy.

We started working on this framework back in 2015, and our thought process at that point of time was that we wanted to create a broad framework that would be required for spotting or for researching stocks. That’s why we thought of a multi-factor framework, because there is no single factor that always performs well.

Over the years, we have tested a lot of factors, around 30, which have worked in global markets. We then shortlisted a bunch of factors based on their performance in Indian markets.

We then zeroed in on four broad factors in the market -- quality, value, growth and momentum -- to screen stocks for our quant fund, and we don't penalise any specific factor. We want to give each factor a chance to be part of the framework.

Also read | How abrupt new goals can wreck long-term financial plans

During our research, we found that value is far more volatile than the other four factors while quality is the least volatile. That makes sense because quality usually performs when the markets are down, so it's less volatile. On the other side, when markets are rising, value has a better chance of capturing that upside, as compared to quality.

This is broadly the crux of why we wanted to go for a multi-factor framework because it will ensure that there is diversification and we are able to kind of capture different factors based on the market cycles. This will also help us in reducing drawdowns in each of the factor. Hence, we get a better risk-adjusted return.

In momentum, we look at the near-term and medium-term stock performance. For value, we look at all parameters like trailing & forward valuations, dividend yield and enterprise value to sales (EV/sales). For quality, we looked at a bunch of fundamental ratios, such as return on equity, debt to equity, and then, sustainability of earnings.

The growth factor is slightly different from the existing listed indices. It takes trailing earnings and trailing revenue growth into consideration, but also looks at forecasted earnings as well forecasted upgrades and downgrades.

Using these four factors, we created a multifactor model.

Take us through the stock selection process. How does it differ from other schemes in the category?

Using the four broad factors, we created a multifactor model for the fund. Our universe currently consists of the top 200 companies by market capitalisation.

I think the other strategies that are there in the market have initial filters where they're excluding certain stocks based on certain quality scores. We consider each and every stock within the universe. All the 200 companies are eligible candidates to enter into the portfolio.

Also read | Nifty can fall towards 22,000-22,500 level, which is a fair value for markets: Pankaj Murarka

Then, we give each company a momentum rank, a quality rank, a value rank, and a growth rank. We then combine them together in a weighted way to give the companies a composite rank. Those ranks will be between 1 and 200, 1 being the best and 200 being the worst.

That composite quant rank is the main criterion for investing in a stock or not.

If there is a large concentration on a few stocks which are not liquid, you will not be able to kind of replicate the model portfolio when it changes. So we look at the average traded volume of all the 200 stocks. So for good liquid stocks, we take a higher active stock weight and for less liquid stock, we take a lower active stock weight.

Do you adjust for the weights of each four factors in the fund?

When it comes to the composite rank, one way to go about is to take 25 percent of each factor -- momentum, value, quality and growth. It's like equal weighted.

But that penalises the factor that is working at that moment. If quality is working, you are increasing the weight, but when you're rebalancing, you're again cutting the weight back.

Also read | Active fund managers hiked exposure to these sub-sectors. Are you invested in any of these?

That is why, every month, we look at the performance of all the four factors and then we increase or decrease the weight based on whichever factor is performing.

However, we have set extreme limits for outperformance or underperformance of a factor. When this limit, on either side, is hit by any of the four factors, the framework goes back to an equal weight.

In effect, whenever this upper band or the lower band gets hit by a factor, the framework goes back to an equal weight and then we again start looking at the monthly performance and increase each factor.

That's kind of what happened this year in April. The quality factor's underperformance was so extreme, the band got hit and the framework went back to an equal weight.

Broadly, this is what is unique in our framework and different from the other quant strategies that are existing.

Because of the dynamic allocation of the factor, our model recovers faster because it quickly picks up whichever factor is working.

Will there be any human intervention in making buy-side or sell-side decisions in your quant fund?

This is purely a rule-based fund. We didn't want to create any black box where we don't know why or what is happening. It is systematic, rule-based and is driven by a framework, instead of a fund manager's views. So, there is no fund manager intervention. The fund manager intervention happened when I created the framework. That is where I came into the picture, and now that the framework is done, it is just replicating the model portfolio.

How will your fund look to adapt to fast-changing trends in the market in your quant model?

We have a monthly rebalancing, which is decently fast. We did back-testing with different rebalancing periods – fortnightly, monthly, quarterly, semi-annually. Monthly rebalancing gave us the best risk-adjusted returns. The framework is nimble enough to pick changes in the market, but also not so fast or frequent, so that the costs kill the returns.

Also read | Profit From The Past: Lessons from history for investing in cyclical stocks

The investible stock universe is 200 top companies. Will you look to add small-cap stocks in the future?

The reason today we have the top 200 companies is liquidity. We have to be nimble in our fund. As of now, the top 200 stocks have decent liquidity where you can enter or exit a stock. Eventually, in the future when markets become even more liquid, we will expand our universe. Also, we use a lot of estimates like earnings momentum, where the coverage as of today is limited when it comes to small-cap stocks.

Does your model look at cash allocation?

We don't take any cash calls. We will just keep around 3-5 percent cash allocation to take care of inflows or outflows.