The Mutual Fund Show: The Pros And Cons Of Choosing A Quant Fund
Is it safe to choose a quant fund? Here are the pros and cons...
Mutual fund investors can either choose a passive fund that follows an index or pay higher fee for an actively managed scheme. There is one more option—a quant fund. It’s a hybrid of active and passive investing as such funds make investments based on algorithms/ set rules with minimal human intervention.
Understanding the process behind the rule-based investing and mapping the track record would be a right approach for selecting quant funds, also called smart-beta or multi-factor funds, said Kalpen Parekh, president of DSP Investment Managers Pvt., which has a quant fund. On BloombergQuint’s Mutual Fund Show, Parekh said investing in quant funds could be like buying bitcoins, which did well in the last 10 years but if investors didn’t understand the product, it would have been like taking a speculative bet.
Aparna Karnik, senior vice president and head, risk and quantitative analysis, DSP Investment Managers, said while quant funds may not eliminate the need of human interference, this discretion is definitely minimised. Once rules are set, she said, these funds update on their own.
DSP Quant Fund weathered the volatility and beat the benchmark in the last three months.
Still, Avinash Luthria, founder of Fiduciaries, advises caution. Rules-based investing may not be enough to beat the benchmark and there would be some element of risk involved to generate extra alpha, he said. Investors pay five to 10 times higher fee compared with a passive fund the hope of getting more, he said.
Quoting Nobel-winning economist Eugene Fama, whose work led to creation of passive index funds, Luthria said, “You hope to get any extra return from smart beta factor investing because you're taking some extra risk. For example, small caps which will get destroyed or hammered in a crash. That’s essentially the crux of smart beta of factor investing.”
Watch the full video here:
Here are the edited excerpts from the interview:
Kalpen, as I learned from the conversation that we just did five minutes back, in the strictest of sense, it’s probably not even a true blue quant fund of sorts.
Parekh: So, the whole concept of quant fund starts with a very fundamental reality that there are multiple routes to heaven in the world of investing. There are different styles that create good investing results. If let’s say none of us existed, as investors you could go and buy Nifty and get the return of Nifty or any basic benchmark which represents the economy of the country. Now, to do better than Nifty is why over the last 20 years, this whole range of active funds in India have been around for more than two decades, and they’ve created enough value and wealth and excess returns over the basic Nifty. So you start with Nifty as a passive product and then you have all the active funds which are through multiple slices- we can slice and dice them as large-cap, mid-cap, small-cap, growth, value, all of that - that’s the entire range of active.
When we thought about our quant fund, and actually Aparna and another colleague Pratik, they are the architects of this fund, I’m only the messenger on their behalf - our whole thought was that there are certain traits or characteristics of fund managers within DSP, or even otherwise who demonstrate a certain style of investing which creates excess return over the benchmark. They are able to generate better returns than the benchmark over long periods of time. So we studied for those styles and then we said let’s code these styles using a quant framework and design a fund keeping the art part of it, the art is the understanding of the style, and the factors that really create excess returns over the benchmark coded in a very transparent manner which we can upfront communicate to you and to your viewers and to our investors and advisors, and then strictly follow those rules. So, it’s a combination of art and science. The art that the investing principles of good fund managers truly comes together and that art is coded into rules which you showed couple of seconds back on the slide. So, there are basically three steps. First, you eliminate companies which have rules or characteristics which never created value. So companies which have high leverage, high volatility in the stock price overtime, inefficient capital allocation, poor quality of earnings, accounting, frauds--just minimise those companies from our basic index. And then the next step is select companies on the basis of good rules, the rules of high quality businesses, rules of good valuation and the rules of good future growth. The third step then is to assign weights to these companies. So we said we’ll just blend that combination of art and science and put it together in a rule-based format.
The DSP quant fund is nothing but a rules-based fund. It eliminates weak companies, and selects good companies and stays patient and then allows you to run over a long period of time.
Today we are here at the end of one year. Our whole idea here is to replicate good investing principles in a rule based-format, and because it’s not passive, because it’s not completely active, you mentioned about fees, the fee structure is somewhere in between. The fees are half of what active funds would charge but of course, they are slightly more than what completely passive funds would charge because this is not a passive fund completely. This is not an active fund either. It’s a combination of the two. So globally, they call it smart beta fund or multi-factor fund, and that’s what the DSP quant fund is all about.
Aparna, you want to add to that? Do you envisage that products in such a category could actually mushroom or grow in size and in number because of the experience that let’s say, you’ve had? Are there other peers who are doing this? Or are you as a house looking to do more of such products as well with whatever little success you had in one year? I must say one year is too short a time frame anyway.
Karnik: We now have more than 30-40 years of data globally on these type of quantitative techniques to investing have mushroomed with the incremental computing power which is available and the evolution of finance as a field in academics as well. So, like Kalpen said there are factor funds globally. This is one of the fastest growing categories globally with more than 3 trillion of AUM managed on smart-beta funds across different segments like hedge funds, mutual funds, prop desk. So, it’s not something very new, it’s something which is already happened globally. So, it is something which has a similar experience in India as well. The reasons for this is just, again, as information becomes more easily available and computing power increases, those things or those investment principles, which many investment managers globally, many active managers apply like somebody has a value oriented style, so then that can be converted into rules. That that can be converted to mean companies which maybe have a price to book which is cheaper or companies for instance, which have a higher dividend yield or when they say quality companies you can code that into companies which have more stable earnings or better ROEs. So, these investing principles, which are universal, which can be converted into measures and therefore applied to portfolio construction in a very seamless way with the new tools and with the kind of computing power we have. It is something which investors like because they’re able to get that additional excess return of that style at a very low cost in a very transparent manner. And that’s where factor funds are really done very well, globally. So I do believe that these funds are going to grow exponentially in India. Even if somebody doesn’t label it as a quant fund, but the application of quantitative techniques on investment management is something which is going to grow exponentially. It’s going to grow exponentially in India.
Aparna, how much of human intervention versus the quant application would be there in this fund? And is there a possibility that the human thought process to changing the components can overwrite the larger proportion of the quantitative testing or choosing?
Karnik: No, absolutely not. Once the model is created, the model is running the portfolio on a day-to-day basis. So I think where the human element comes into the picture, and this doesn’t take away from the fact that it’s a fund which is quant. Quant is nothing but quantitative and the fund is run purely on a quantitative manner. There is no human discretion, but where the human element does come in is- we do a lot of research. The design of the model or the design of the rules, there is a lot of research and that’s where our team comes into the picture. So, we do a lot of research on globally which factors are working, why are they working? What are the behavioural reasons? Or the reasons why they deserve that risk premia? What are the measures to code for these factors? So, we evaluate these on an on-going basis and then we keep on testing whether they have a potential to generate alpha in a very consistent and robust manner. And if we feel for instance that yes this rule or this factor or this principle has the ability to generate meaningful alpha and it is tying in very well with fundamental principles, then we will go to our investment committee to apply it into the model, but once the model is running, then there is no human overwrite on that.
Quant is a very generic term. It includes all kinds of funds — factor funds on which DSP quant fund is largely built on, macro-based funds, algo-trading funds. Quant is short for quantitative. Our fund is a rules-based multi-factor fund, which is coded based on fundamental principles.
Kalpen, you want to add something? Do you reckon that such funds for either lower expense factor or less human intervention factor or a combination of both could actually find greater acceptance?
Parekh: My take on this is why would I invest in this fund or why would you invest in this fund? To start with, the starting point should be that do you believe in the principles of the fund rather than the expenses alone or rather than just the fact that there is no human intervention, because then we are making an assumption that wherever there is human intervention, that’s bad. But that’s not the case.
What we are saying is, first of all, understand what are the rules. So whoever we talk to we say that don’t look at our performance first, don’t invest because you like the word quant, understand what are the rules. If you’re a believer that these are the rules, which can create better returns than the alternative benchmark, if you believe in these rules, then you should invest in this fund. Number two, once you believe in the rules, then look back that has the portfolio effectively reflected these rules and how good is our execution. So for example, our model has been tested for almost 16 years now. And it is now live for the last one year. So one year back when we launched it, there was a very valid observation that your model is very solid, looks very good, but it’s not seen real life. So let us see how you really deploy money once you raise money and then let’s see the track record. So now there is a track record over the last one year as well. So if you like the rules, and if the track record is in line with those rules, that is a good reason to take a decision whether to invest or not. Let’s just say some someone who comes in tells me that this is a quant fund, it has done very well in the last five years, please invest in this. That is akin to just buying past performance of any product. It’s like Bitcoin- done very well in the last 10 years but if you don’t know what is it all about, then you’re taking a speculative bet. So that’s how I would position it.
Avinash, can you tell us your views on this category?
Luthria: Sure, so you can you can see the active universe in simple terms as two schools of thought within it. One is the fund manager is a genius who can beat the market because it takes an investment genius to beat the market- it is that hard. The other school of thought is that you have a rules based system which may give you something more. No top world class fund manager says will give you more than the market on a risk adjusted basis, just a rule based system. What they would say is that by taking a little more risks with small-cap, you may get a little more return, hopefully. So, those are the two schools of thought and you could say that I mean, there are some pure quant funds in India and I would call the pure rules based fund in India such as the Nifty equal weight that DSP has. And then you have this combination in which is a little bit of human a little bit of rules, saying okay let’s take both. I mean, in a sense that there’s no proof that the human approach works. There is no proof that the pure rules-based works but let’s try with experimenting with both. That’s how I would put it.
Kalpen, Avinash has a point that there is no definitive proof of which model would work better. How would you respond?
Parekh: So, I think I would also spend some time later on to explain to Avinash a more detailed approach of what has gone into the design. See, everything is about the design of the fund and the rules. So, if your rules are bad or if your rules are stale or they are not reflective of changing market conditions then obviously, that will not work. It’s like I’m a fund manager been around for 20 years, I did very well 10 years back, but in today’s world, I’m not able to catch up with changing trends of how profit pools are changing, what type of industries are emerging, then even I will go wrong. So even a human being goes wrong when he’s not updated with changing trends.
What I’m coming back to here is that I would not want to generalise all quant funds as same. All quant funds are not the same. For example, if I were to say that let’s say there are 20 indices in India, 20 benchmarks- the Nifty, the Sensex, the BSE 100, BSE 200, Nifty next, and so on and so forth. They all don’t perform at the same time. There are times they do and there are times they don’t do well. Active funds, they do well at times, at times, they don’t do well. I am coming back to one common theme here that focus on the rules which are defining the portfolio, which are defining the construction to start with. When we started this fund, we also recognised that there is a lot of scepticism about rule-based investing or quant investing, also because of perceptions. So, we said let’s look for why do rules fail? A lot of our work in the first eight-nine months before we launched it was on writing down what are the failures for quant funds? What makes them fail? When do they not do well? Why do rules not work? And when we looked at those parameters, we realised that sometimes the modelling principles are wrong. Sometimes the costs taken into modelling is not realistic. On an excel sheet, it’s very easy to budget for a 0.5% or 1% fee structure and then put out the returns. We have done battle testing, not even back testing. So we have applied a 3% impact cost when we did the modelling over the last 15 years. So, we said that, okay if the portfolio has given X return, but let’s assume there will be a 3% impact cost in terms of management fees, buying, selling and stuff like that over the whole period of time. So we’ve tried to be very prudent, very conservative in the old design. And coming back, we are urging everyone to focus on what are the rules, are these the sensible rules? So what are our five rules?
Our five rules are companies with high earning stability, companies with high ROCE, companies with low price to free cash flow, companies with high dividend yields, and companies with high growth rates.
So that solves for quality of businesses and valuation. But before doing that we are doing something which probably most funds do not do in the quant framework is, massive elimination. So we have around 30 factors on forensic dimensions where we look for how do we eliminate companies.
We eliminate a lot of companies which have high leverage, we eliminate companies where governance is a question mark, where companies misreport their earnings. That is the soul of the product. And we have seen that when you eliminate such companies, that saves a lot of alpha.
Having said that, let me give you an example, in the last one month for example, since March 23, which was the low of the market; such companies have done much better than the quant fund itself. So, there are phases when eliminated companies also do well. Once in few years such companies would also do well. But we are saying if you like these rules, if you are a believer of these rules and these principles for your hard-earned money, then understand them and take it.
Lastly, what is another big USP for a category or a product like this the way we are defining it is. Today, you ask everyone does process matter or personality matter? What would everyone fashionably say? Process matters a lot. Now ask the next question, show me the process. I would challenge how many people are willing to put down a document where till the last line the process is well defined. So I think, I want to promote not the performance we might do well, we might not do well but what are those steps that we take and write it down, give it to you every day, it’s there on my website and zero deviation. So we made the last portfolio change in March, now till September, there will be no change. Only twice in a year, will there be a manual intervention otherwise come what may, we will not make any changes. So that’s the way we approach the whole category.
Avinash, is there merit in looking at an offering which is laying down the fine rules of investing and coming at a lower cost as well? Kalpen and Aparna are saying don’t look at the cost, but I mean, over a period of 15-20 years if you are indeed invested, let’s say, then cost does matter. Also, if these options are available, is it a bad category to look at or are there better options to look at if we’re talking about quant or factor funds as a category?
Luthria: I would take it one step higher which is, I have option of investing in the Nifty 50 Index fund or ETF at 0.05-0.1% fees per annum. So I already have that option. The only question for me is do I pay a fee of 5 or 10 times that in the hope of getting something more? I see that as the onus of proof is on any fund that says pay me a fee by 5 or 10 times more, for me to be willing to pay a 5 or 10 times higher fee, I need to have sufficient proof that this makes sense for me. So far, I do not see any convincing proof on that, either globally or in India. For example, we can go into the specifics, but there is no such proof. So I take the point that this is a lower fee than a pure active fund.
The two components did not work—the human genius can’t work and the formula can’t work, how do I know that combining them can work? So why should I pay a fee of 5 or 10 times more?
Avinash, a follow up question to you is this that, if indeed passive low cost funds would always be the preference of choice, unless there is a past performance and I think from everybody I’ve heard that don’t look at past 10 years’ performance and then go out and invest in fund, you will not be able to invest in an actively managed fund at all because the cost factor will always be in favour of the passively managed fund. Isn’t it?
Luthria: Yes, but think of it this way, the customer of the mutual fund is not saying that I want Nifty equal weight. What he’s saying is you, the fund manager telling me Nifty equal weight will give better returns. So, I come to you and I put my money with you. So it’s the job of the fund manager to say that I’ve picked this strategy which I think will beat the index. And the customers are really coming in and saying I want this strategy. So it’s really the job of the fund manager to say this strategy with a higher fee has robust proof that it beats the index on a risk adjusted basis. And globally there is no such proof.
Parekh: There are life funds globally. Over a long period of time, equal-weight strategies have a meaningful source of alpha including in India over the last 15-18 years. The Nifty equal weight has beaten its counterpart Nifty by around 200 basis points over the last 17 or 18 years. The last two three years have been a different story in India, primarily because six stocks have given all the returns in Nifty. So, we all know it will be weighted in the past also. So, there’s a huge amount of polarisation of the top six stocks which have taken up all the returns. It has happened globally also and it’s happened in recent times in India also. But in spite of that polarisation, in spite of the last two years where the market cap strategies have done better than equal weight, if you take over the last 15 years-- world over and in India, equal weight as a strategy is a source of alpha, and there’s a reason for that because stock number 11 to 50 also in the Nifty is coming across from diversified industries and sectors and market leaders from each of these sectors also contribute meaningfully. So, there are phases when the top six stocks of the index do well and then there are phases when the broader markets do well. So these are phases and one has to be aware of that, that there are phases when one style works and there are phases where the other style works. So, over a long period of time, has this style worked or not? There is enough academic evidence, there is enough actual practical evidence of equal weight as a strategy having done better than market cap weight.
Karnik: I would actually like to ask Avinash because I have myself read a huge amount of research published by many very respected people like Fama & French, etc. There are so many quant fund houses like AQR, there are hedge funds which apply quant. See the point is that not everybody is a genius. Okay? I agree with you there but to say that no active fund manager even has proved that their strategies deliver alpha, I will not agree. There are a few, statistically you can prove over 10-15 years that they have information ratio. I don’t know what that statistical evidence is which needs to be looked for, but there are people with very high sharp ratios, very high information ratios. Not everybody’s genius, but there would be some geniuses too in the active space and if we look at the strategy performance over many years, it can be statistically proven that there is some skill there using different metrics, you can use information ratios etc. Coming to factor investing, the fact is that, there are two reasons why factors work and is beta across 40 to 50 years in markets like the U.S. One, risk-adjusted returns- excess returns is because of additional risk of value investing or maybe a small-cap investing where you are earning excess returns because you’re taking excess risk. There are other reasons, however, which are equally academically sound is a behavioural alpha. So, factors also harvest behavioural alpha. So, factors like lowball, factors like momentum, factors like quality, have evidence across decades and you can read any amount of research which shows that they have got excess returns even adjusting for risk.
What we have done in case of DSP quant fund is, we have gone by universal principles. Universal principles are like quality investing, value investing, growth investing, where we know that these are principles for either behavioural alpha reasons or risk premia reasons that have delivered returns over 40-50 years across geographies, even across asset classes and not only equity.
We say that if these have worked over 40-50 years and they have delivered excess returns, we have tested them in India. In India, whatever 15-20 years data we have, we believe, we have seen them and have enough evidence to believe that yes, it can deliver excess returns in India. Hence, we have applied them and back tested them as robustly as we can. That’s a totally different field. What I’m saying is, we’ve tried to apply them very robustly and as Kalpen said, we didn’t want to just apply that without factoring in some nuances of the Indian market and eliminating these high leverage companies, companies which have ownership issues. So we eliminate those using metrics which we have designed. And we have a reason to believe that following this approach will deliver 3-4% alpha over a cycle.
Avinash, please finish your point.
Luthria: I take Aparna’s point. I think just one point that Eugene Fama who is the father of factor investing or smart beta, himself said that you hope to get any extra return from smart beta factor investing because you’re taking some extra risk. For example, small-caps which will get destroyed or hammered in a crash. So that’s essentially the crux of smart beta of factor investing.