As equity and fixed income markets face increased volatility, quantifying trading performance can become harder. Best practice in trading analysis can make a significant difference to trading desks’ performance.
On Trader TV This Week, Ash Sharma, trading and analytics manager at Aviva Investors, considers market impact models performance under different scenarios, the use and suitability of different models and inputs.
Interview:
Dan Barnes: How does the current market environment affect traders’ use of analytics?
Ash Sharma: So the current increased volatility of the market has been an interesting one with lack of clarity around Trump’s tariffs, along with US jobs data showing unemployment ticking higher and US tech declining for different reasons. So TCA in these scenarios has to be placed in context because you’ll see large positive and negative skews in the data. The appropriate metrics need to be considered, such as market condition measures, which compare actual versus expected conditions, which is extremely important. So we’re able to access these via our equities TCA vendor, who can produce percentiles for volatilities, spreads and liquidity to indicate the market conditions during trading. Market impact models become even more important to consider, to understand how they deal with these different scenarios. and the trading desk. will be in contact a lot more than usual to find out how orders are running during real time, but also post -trade as well. And that may range from larger orders through to smaller orders that are submitted down the algo wheel.
Dan Barnes: How does that impact your use of specific inputs and broker rankings?
Ash Sharma: Market impact models have to be analyzed to ensure they can adapt to volatility events in an appropriate manner. And it’s always useful to contact your vendors to discuss how they’re accounting for this to make sure it’s fully understood. When we consider broker rankings for different execution channels, we analyse the data to ensure that the skews are normalized because we understand that there are some areas that are outside the broker’s control. This makes sure that when we have those regular broker meetings, it makes a dialog more targeted and ensures that we can compare brokers fairly. In volatile environments, the trading desk will want to know how real -time performance is looking via analytics available in the EMS, as well as the standalone intraday TCA. They may want to analyze whether taking a risk price, for example, from a broker is better than spreading the order out over a longer duration. As we use a few different market impact models, it’s always interesting to compare how the methodology differs and which one is suitable within different environments.
Dan Barnes: How are you evolving your capabilities for analytics within the equity and the fixed income derivatives markets?
Ash Sharma: So recently we’ve started to delve into how performance analytics can be used for our credit flow. So we’ve created a framework where we analyze price quality and responsiveness within e -trading. Price quality includes metrics like percentage of positively or negatively slipped tickets, percentage price held and percentage inside CBBT just to name a few. And then price responsiveness includes metrics like percentage hit rate and average quote times. So we apply different weightings to each of these metrics and then that provides a broker ranking at the end where we can clearly see which dealers are doing better than others. And then we can have those targeted conversations with them, which are essential. On the equity derivative side, we’ve designed a framework where we can measure role and non -role performance separately using different benchmarks at different points in the order life cycle based on order tagging. This has highlighted some key areas internally, so it’s been a very useful exercise. I’ll be talking about this in more detail at TradeTech in one of the sessions on the equity derivatives track. So overall, I think it’s prudent to always work with the trading desk to ensure that we can produce analysis that actually matters and that the traders actually value and will utilize going forward.
Dan Barnes: Given the level of information that’s available out there, how are you taking that information efficiently and inputting it into your decision -making engines?
Ash Sharma: So last year, we started using APIs for our rates analysis, which my colleague extracts from our TCA Vendors platform. And it’s essential for us that we use APIs for this because it provided much more information that was available in the front end. But I think if vendors have a lot of detail available in the front end, then APIs maybe become a little bit less necessary. But this allowed us to create analysis focusing on mid -slippages at raring timestamps. illustrate how the bond prices move during the order intervals. We can then split the data up into different groupings such as comp versus non -comp, buy investment desk and instrument type just to name a few. We also incorporated liquidity scores so we can bucket the most liquid to the most illiquid bonds and then we can use these results to observe any patterns in the data. So hopefully we can build on that internally going forward.
Dan Barnes: Finally, in your personal opinion, do you think AI can be used for trading analytics on the buy side?
Ash Sharma: So AI is obviously the buzzword at the moment and it’s everywhere you look. Extracting data using code has been around for decades and it’s used a fair bit at Aviva Investors. One of my colleagues, as I mentioned, is tasked with extracting the data into a readable format and it’s a very meticulous task. So these areas I think AI and automation is extremely useful. And I think the important piece to consider here is the trading flow differences on the buy side, where much larger firms would yield more from AI and automation. Also, I think where the firms are positioned in the market is an important factor, as I think the sell side and exchanges will generally have more data points to analyze, and this is where AI I think becomes vital. Hedge funds and quant houses would probably fall into this category as well. So overall, I think… AI is great in relevant scenarios and shouldn’t be utilized just for the sake of it. There has to be value in targeted goals.
Catch the show each week:
To watch other Trader TV This Week shows go to – Video
Or make sure to follow us on LinkedIn – Trader TV’s LinkedIn
Available as a podcast – here
