The Gateway to Algorithmic and Automated Trading

Asia's biggest fund takes an algorithmic passage to India

Published in Automated Trader Magazine Issue 26 Q3 2012

Jul 16th, 2012 - Eastspring Investments is the largest fund manager in Asia with some 2,000 employees. Adam Cox caught up with one of them, Sanjay Awasthi, a Singapore-based director at the firm who trades the fast-moving Indian markets. Sanjay got his start with India's financial regulator and has witnessed India's evolution from a cash-only market with three-week settlement periods to the hyper-fast, colourful market it is today. He says India has a reputation for being extremely porous, so when a fund the size of Eastspring wades in ($80 billion and counting), the use of algorithms becomes critical.

Sanjay Awasthi

Sanjay Awasthi

Adam Cox: We'll start with how you came to join Eastspring.

Sanjay Awasthi: I've been working with Eastspring Investments for close to seven years, and have been with the dealing desk for around four years now, looking at India and derivatives. India is my main market and that is what I focus on. Prior to joining Eastspring, I was an independent trader for close to two years and before that I was with the regulator in India, the Securities and Exchange Board of India (SEBI), for nine years.

Adam Cox: So what attracted you to the trading side, why did you want to move on from SEBI?

Sanjay Awasthi: I joined SEBI straight out of management school when the Indian markets were going through a huge change. SEBI's role in those times was largely developmental, and I had the opportunity to be involved in many interesting developmental initiatives, like setting up the derivatives markets in India and setting up the surveillance systems. It was great fun and a deeply enriching experience.

Nine years just flew by. I realised thereafter that I wanted to actually trade the markets. So, I just put in my papers one day and started to trade on my own. Although it was my first foray into trading, I was amongst the first group of options traders in the market. After trading on my own, I realised that I needed to experience work in an institutional environment. An opportunity with Eastspring Investments came along and I moved here to Singapore.

Adam Cox: Are there times when you feel more comfortable looking at parts of the market because of your experience at SEBI?

Sanjay Awasthi: In '95, when I began my career with SEBI, the Indian markets had a settlement cycle of 15 days. Settlements were largely paper settlements and dematerialisation processes were not in place. This meant that settlement actually occurred 21 days after the trade. Now, the Indian markets are cutting-edge in terms of technology, sophistication and processes. I was lucky to witness the entire transformation of the markets from a uniquely close perspective.

As a regulator you get a top-down view of the markets, which is very illuminating. I can relate to what previously existed and what has changed. In my day-to-day work, it does help to know a little bit about the market micro-structure and the various changes around it.

Adam Cox: How long has Eastspring traded in India?

Sanjay Awasthi: Eastspring Investments has been managing Asian equity funds and, more specifically, India equity funds successfully for a long time now.

Adam Cox: Are you the only one focused on India or are there others?

Sanjay Awasthi: We have an eight-person dealing desk covering equities, fixed income, FX and derivatives. I focus on India but as a team we cover most markets in the region. India has a very active single stock futures market. It's probably one of the most liquid single stock futures markets in the world.

Adam Cox: Now the vast majority of equity trade is concentrated at the National Stock Exchange, which has the lion's share, and the Bombay Stock Exchange. And you trade on both. Are there arbitrage opportunities between them?

Sanjay Awasthi: There's a huge local arbitrage community which trades a lot of cash futures, and inter-exchange arbitrage that's largely captured by locals.

Adam Cox: So what are some of the challenges you see that are specific to India?

Sanjay Awasthi: Purely from our point of view, there are minor issues like the lack of a block window. Everything has to happen on the floor of the exchange, even if you have a counterparty and you want to take a block of, say, $10 million. That hinders institutional investors like us because, if you're buying large positions or selling large positions, to reduce impact cost we do look for counterparties who can match us in our size, or part of the size. That is a bit of a hindrance.

Adam Cox: What do you guys do on the algorithmic front? Is it all about execution?

Sanjay Awasthi: We use algorithms for trade execution. Algorithms were largely developed to help the sell side execute their trades. For example, if a client placed an order and asked for the order to be worked through the day, or that the orders stay at 20 percent of the volume, the sell side had a system in place instead of having somebody do it manually. That is what has transferred to the client side in the last 3-4 years.

So now, instead of a person like me calling up a broker, giving him an order and telling him how to execute it, I have a window which gives me access via the broker's pipe straight to the exchange floor to execute these trades. And I can set my own parameters, depending on my understanding of the requirement, without having to communicate these orders to somebody else. This gives a trader more control over the execution of the transaction. As to its efficacy, and deciding whether it turns out better or not, it is always a matter of what happens subsequently. When you pass the order on to a broker it's called high touch, and they're also adding their own input, based on their understanding and their read of the market at that particular point in time.

The quantum of adoption of the algorithm is dependent on the effective understanding of the markets. So if one is more comfortable with the markets one probably uses it much more. India is a very unique market in the region because of its micro-structure. It has uniform tick-sizes, single share trading etcetera, which bring in a lot of noise. Further, there are various kinds of participants in India. There's large retail and institutional participation and a variety of traders. The Indian market is slightly more volatile than most markets in the region. All these factors give rise to a unique set of challenges.

Sanjay Awasthi

Adam Cox: How do you determine what parameters to set? Is that mostly based on the nature of the orders coming in, or do you try to add value?

Sanjay Awasthi: When you're executing an order, you need a framework for the execution and that stems from an appropriate benchmark. At Eastspring, we believe dealing is not standalone, but part of the investment management function. We've adopted Implementation Shortfall as our benchmark for execution. The IS philosophy benchmarks execution to the decision price, i.e. the price around which the investment decision to buy or sell a stock was taken. We also carry out detailed transaction cost analysis to measure execution performance against the benchmark. Once you have that in place, you design your execution strategy.

When an order comes to me, I have a few choices. I can use a sales-trader. If I see a counter-flow and the possibility of a block trade, I can use that. Or I can work it quietly with an algorithm. You have various tools available to you, and algorithms are amongst them.

Adam Cox: You've talked about studies showing two-thirds of trade costs being impact cost. Do you have any sense of how much, when you're using algos, you're reducing that?

Sanjay Awasthi: We send all our data to a third party, who carry out the necessary analysis of all our trades vis-à-vis the benchmark. We use the data to not only measure the performance of our brokers, but also our dealing KPIs are linked to our execution performance.

Adam Cox: That's interesting. Are any of your competitors doing likewise?

Sanjay Awasthi: The feedback we've got is very few of the buy side in Asia do such detailed analysis. Last year, Eastspring received an award for being the most innovative buy side dealing desk in Asia.

Adam Cox: So from a performance point of view, you actually have real numbers to work on.

Sanjay Awasthi: Once you have real numbers and real benchmarks, devising strategies becomes easier. It's not just saying, 'Oh, I feel the market is going to go here!'

Adam Cox: You talked about the three options for when you get an order - is it often the case that the algo option is the best?

Sanjay Awasthi: It is very situational. If I see a block seller in the market and I'm buying something, if I can get liquidity at a price, I'll take it. Unless the portfolio managers give explicit instructions like 'Work it over the day' or 'Work it within a limit', then that is how I am bound to execute it.

In India, the proportion of trading using algorithms is much higher because algorithms tend to give you anonymity. India is notoriously porous. Especially when institutions with large transactions like Eastspring are executing trades, there have been times when the word has got out and affected the execution. Algorithms help in that sense, to keep the execution process from being affected by the porous nature of the Indian markets.

Adam Cox: How do you think it compares to other markets in that way?

Sanjay Awasthi: Honestly, I can't compare the Indian markets with others. A lot of the times people say a lot of things about India because it's very different in terms of the microstructure. That understanding, at times, eludes some people. Unique features include the bid-ask spread, the uniform tick size etcetera. So a stock which is 1,000 rupees and one which is 20 rupees have the same tick size where one can trade up to one share. On the screen, you only see maybe five or six shares and there are a lot of hidden orders on both sides of the screen. In this sort of scenario and if you really don't understand this microstructure, while executing, a trader tends to signal by his action. An action, which is perfectly normal in some other market, when used here ends up signalling, particularly where there is a large retail trading community who probably catch on. They're very smart, because they're trading with their own money. At times that is misunderstood as being porous as well.

Adam Cox: Tell us about the algorithms you use.

Sanjay Awasthi: We use broker algorithms. The brokers give us the option to customise the algorithms based on our requirements.

Adam Cox: Do you use a number of different ones?

Sanjay Awasthi: We have 10 counterparties, and we work with them on this.

Adam Cox: And do you do any customisation?

Sanjay Awasthi: From time to time, yes, we do request our brokers to customise the algorithms to our requirements.

Adam Cox: Now there are so many algorithms out there that are good at sniffing out other algorithms, if you are using broker algorithms that are not customisable in some cases, are you at all concerned that what you're trying to achieve can possibly get determined by another algorithm?

Sanjay Awasthi: You're right. I hear such algorithms are active in India. You find ways to go around it. You change parameters and avoid following a pattern. We started off largely from the point of view of anonymity and I think that goal has been achieved.

Adam Cox: Where do you see it going from here? Can you see changes on the horizon in terms of the wider use of algorithms and your own use of them?

Sanjay Awasthi: In the past quarter or so, we're seeing local institutions in India actively using algorithms. It is difficult to say at this stage whether algorithms will alter the Indian markets. Only time will tell.

Adam Cox: Can you see a scenario where you might use algorithms for alpha capture, and where you might offer investment products based on that for your clients?

Sanjay Awasthi: We do not use algorithms to capture alpha. We hear of some participants coming into Asia but I don't think it's as much as other markets in Europe or the U.S.

Adam Cox: As that kind of trading does develop in Asia, do you find it impacts your own trading?

Sanjay Awasthi: There's a lot of debate around high frequency trading. From our execution perspective, we're very happy if there's a system or structure or a participant who helps us reduce the impact cost. So if the volumes increase because of high frequency trading, the spreads come down and it helps us. I'm sure there are pros and cons to it, but a lot of it is still a debate and we're seeing very little on the ground in terms of data to show the disadvantage of HFT.

So my take is, that since it is not conclusive as yet, why throw out the baby with the bathwater?

Adam Cox: As someone who worked on the regulatory side years ago, are there any specific regulatory risks in trading in India that are different from other markets? It's such a big theme for markets in Europe and the US at the moment, but could that be for different reasons than what's going on in India?

Sanjay Awasthi: In India the exchanges have been cautious in permitting algorithms. When a broker has to introduce a new algorithm, they have to have exchange approval, where they have to disclose to the exchange exactly how the algorithm is going to behave. It's only then that they can roll it out to their clients.

What I'm happy about is that there are people - the regulator and the exchanges - who are looking into it seriously. So for a fund house like us, it basically helps strengthen the integrity of the system.

Of course it may result in minor headaches for some people from time to time, but if you have somebody looking at and evaluating a system and a process, as a user, I think it's good and in the end the client benefits, which is what it is all about.

Eastspring Investments