Making sense of markets
Yann Couellan, AXA Investment Managers
Asset managers in France are increasingly using sophisticated data and aggregation tools to trade fixed income. Yann Couellan, head of fixed income execution at AXA Investment Managers is leading a push to use these tools to provide a powerful trading proposition at the fixed income trading desk.
"Actually, we had developed a suite of tools for fixed income, starting in 2006 from (i) pricing tools enabling traders to follow up each expertise from Rates to HY,(ii) Axes tools in late 2008 to manage liquidity challenges that we were facing, as well as TCA Benchmarking tool which enable to demonstrate to clients the trading performances of the Global Centralized Trading Platform of AXA Investment Managers, Trading & Securities Financing Dpt. Axes tool is an in-house application which aggregates tailored fixed income interests sent by our counterparties to Axa IM Fixed Income Trading team, to address liquidity challenge and price discovery. This application enables traders to promptly identify which are the banks which have obvious and firm interest on a specific bond for a specific size in order to reduce market impact. It helps to find the others side for a given order."
AXA first launched its market data axis tool in 2008 for internal use by traders and fund managers. Back then, however, it was only updated twice a day. Today, it is updated at 30-minute intervals. But just as importantly, it has been integrated with other related tools – most notably market data aggregation tools.
Normally, when a trader is seeking to buy or sell, that trader will look at pre-trade information. AXA has a pricing tool that interacts with the pre-trade screen, displaying counterparties in real time by slippage and the spreads that would be incurred by trading with that particular broker mapped against the market average as a benchmark.
In addition, the company uses post-trade data to directly influence the pre-trade summary. “We made the choice to start analysing Market data some years ago, as we quickly understood that market data are so useful and that we could enrich our knowledge and then to have a better understanding of our counterparties as well as our trading behaviours. We learnt a lot from market data and we still learning. All the market data that we either received from our counterparties, market vendors or simply our own market data, enable our traders to achieve more effective trading services. As soon as that we set up our trading tools with all the metrics stored in our database, we were able to run any analysis that we need to make our business more efficient. So first, we use market data internally at trading level, to define which trading strategies would be the best for a specific trades or programs; Will it make more sense to use Etrading or old school Voice trading ?, and additionally, we used also market data to provide detailed business reports to our counterparties, which also enable them to really know who we are, in which specific sectors we need liquidity access. Finally we obviously use market data for our TCA."
The firm’s TCA was envisioned as a way to measure trading and make sure the firm is not trading at the worst level. However, one of the immediate challenges was how to deal with the data the firm has stored over the last five years. The firm analyses the bid-offer spread, and the results of that analysis indicate that in bad market conditions, spreads typically widen. This information was then fed back into the trading desk as a warning sign, a possible trigger that should be observed carefully and potentially acted upon.
The second aspect of AXA’s data strategy is to aggregate the data and create a way to compare the firm’s own execution against a reference price to evaluate AXA’s performance. “We can use TCA for a wide range of benefits. As a Head of FI Trading, Tca ( in reality I prefer to speak about Trading Performances Analysis (TPA) rather than Transaction Costs Analysis (TCA) which implies a ‘cost’ and not suggest any savings from buy side FI traders) enable to demonstrate to clients the added value of a global centralized trading platform and so measure the trading performances of the FI Trading Team, as a whole. So trading performance analysis is a straight continuation of the market data output. It helps to define trading strategies, to understand where traders should pay more attention to reduce market impacts for clients. Moreover, doing a retrospective analysis our of market data enables us to profile our counterparties, and identify which one will be the best placed to offer to our traders the liquidity they need with the minimum market impact, on a specific instrument, from global players to niche players, depending the calibration of filtering criteria. It’s a virtuous circle where the post trade TCAs and data bring to powerful prices discovery by running the pre-trade. Finally I use it to demonstrate that even if the liquidity is challenging, Axa IM traders found it and still find it, but not for free, by showing our trading reality vs a perception."
The impact of a more data-led approach to trading has been significant. For example, Couellan notes that the data may prompt the trader to make an effort in an area that the trader has never considered before. It may prompt the trading desk to make use of electronic execution in some areas where it is more economic to do so than to use voice trading. The data can also be used to demonstrate best execution to clients.
One potential area for improvement is the availability of benchmarks. For example, fixed income traders may benefit from the ability to compare themselves verses their peers with a similar profile. Couellan believes it would be quite possible for market price vendors to provide such a benchmark, although currently he is not aware of any such commercially available solution.
While effective use of data may be desirable for the buy-side, it isn’t always cheap or easy to achieve. Couellan warns that the technology cost is often underestimated. AXA has groups of three to five staff dedicated to the technology. Nevertheless, he believes the end result justifies the costs.
“The main limitations to run Fixed Income TCAs are data quality, transparency and aggregation, I certainly forgot some points, but to me those are the most important. I mainly refer here to assets like emerging markets or high-yield where it’s often more challenging, but not impossible, to obtain a reference price at the time you did trade to compare your execution against. Some data that we need today are missing, as not yet fully calibrated. Fixed income TCA is related to the consumption of accurate and sensitive market data, which could be misused by ruthless users, and then, at some points, could impact negatively the market."
The impact of regulation could be significant on the fixed income trading desk, not just at AXA but at all buy-side firms active in the market. One of the main impacts is that the buy-side firm will need to know whether each bond is considered liquid or not under the European Commission’s MiFID II legislation. The distinction is important, because if it is sufficient liquid, the buy-side firm carries an obligation to trade the bond in question on an electronic platform.
However, Couellan says the immediate impact on AXA is likely to be limited, due to the caveats in the legislation. "Our average trading sizes are generally over the threshold required to have to engage pre-trade transparency. This is something that is likely to hit the sell-side more.
As a buy side, I am really concerned about the post trade transparency impact, especially for sensitive trades , programs trades, larges blocks or issuers which could be under the limelight which will be made public after 48 hours, unless they will benefit from the deferred publication regime extended to 4 weeks by the NCAs. This post trades publication as defined and proposed today (i)could and will lead the banks to take less risk as principal as 48 hours later all the street will know, (ii) and also increase the snowball effect of such publications transparency, causing higher market impact. Moreover, it’s unclear to me how NCA from different countries will coordinate as well as what will be the criteria which lead to such decisions. I don’t know yet if this will change trading behavior but buyside will certainly have to adjust their trading strategies.
That’s not to say the rules won’t have an impact, however. One of the concerns for AXA, like many other asset management firms, is that the ability to trade large blocks could be negatively impacted by the impact of additional transparency mandated by MiFID II.
While technology can clearly impact on a buy-side firm’s trading and investment strategies, ultimately it is still human beings that have to make the big decisions (at least for the time being). For this reason, many firms are implementing Trade Management Oversight Committees. At AXA, the firm has a best execution committee, which meets on a quarterly basis. These meetings are attended by a compliance officer, as well as internal quality control specialists. Couellan believes clients should also be a part of these meetings, along with the trader and the fund manager.
“We have regular Best Execution Committee (what you call TMOC) with fund manager, heads of FI Trading and Business quality team members which ensure that Fi executions are done in respect of internal procedures and regulation.. We generally review our internal KPI and provide rationale regarding , as an example, to what extend we used more electronic trading on specific areas. These are the kinds of issues it could help to address.”
He adds that while external vendors offer TCA, they could improve the flexibility of the reports they produce, which is currently somewhat lacking. This is an important consideration for the buy-side, which places a high value on customisation. Even if an external provider ticks the box on paper, Couellan estimates that the buy-side will mainly use TCA to enrich their own data – and one provider won’t be enough. “The data is there but we don’t know how to capture it. We just need the IT capabilities to do so, more electronification with market data standardisation, and make the connectivity as easy as plug and play. “
In terms of the rest of the buy-side community, Couellan is relatively open to the idea of sharing information with other similar buy-side firms and working with vendors, if that will help to make TCA more effective. “First, I am convinced that the buy side industry will benefit to increase the collaboration which is essential to avoid diversification in the FI projects. It will be successful for leading projects like TCA, transparency, etc… as well as for the future of the FI Markets. In addition, Partnership with banks will remain essential too and should remain on top of the pile. The FI primary market needs to be improved as today , there is a lack of transparency, we are missing data quality, and the sell side must continue to push forward on this topic. The industry should develop sharing tools to maximise the benefit of post-trade transparency under MiFID II.”