Chapter three
Bond substitution is a common challenge faced by many sell-side fixed income traders. When a client asks the trader for a bond inquiry which is not included in the inventory, but there is one very similar, it is hard for a trader use data to find out if the similar bond will be a good fit. The Bloomberg Terminal can perform a function to suggest a substitution based on pricing data, reference data, comparables and spreads to curves. There is a huge demand from clients to use Bloomberg data and build automation tools around it.
For those sell-side banks that want to remain profitable, having their own in-house data management system is crucial.
In order to have a centralized data model, sell-side firms will need to optimize how they are purchasing their data. It is also important to have a holistic view of risk in order to view exposures across different business groups and respond appropriately to regulatory requirements. By creating an in-house data management warehouse, this holistic data can be built in at the very beginning.
While larger sell-side banks will have the budget and capacity available for these large-scale projects, smaller sell-side firms might find that they are better off outsourcing AI technology to a third-party provider.
RFQ Auto-Quoting and Price Generation Models were some of the areas where AI can be expected to deliver the greatest benefits to market-making workflows
“When you have to price a single bond with no time constraint, then a human can probably do it better than a machine today. But in practice, RFQs have timers, and large trading volumes mean many would expire before a human could look at them. Bond substitution is very interesting. Having access to all of the equivalents that are priced well can unlock a lot of liquidity.” Dan Tsou, Head of ETOMS, Bloomberg
“At an early stage of innovation, there’s no proven solution design on the market so it’s beneficial to have an in-house data infrastructure to learn and experiment with data. The challenges for this approach will be sourcing high-quality data sets and time to recruit and develop expertise to build models. Bloomberg TOMS has recognized this challenge and is investing in a new analytics platform to deliver data-driven products to support these emerging needs from our clients.” Dan Tsou, Head of ETOMS, Bloomberg
“Data is the most vital ingredient to developing analytics. It’s not that surprising firms are looking to keep it in-house despite the obvious challenges that come with that.” Ravi Sawhney, Head of Trading Automation & Analytics, Bloomberg
“At this early stage of innovation, there’s a lot of discovery and learning required. Having investment to do it on your own has many, many benefits. It also has some shortcomings, but if the investment is affordable, this should be the first choice because it will help a large organization develop the right culture and strategy through the innovation process.” Dan Tsou, Head of ETOMS, Bloomberg
“No, but I suspect this will change over time. More of them will be using outsourcing or working with a partners. For now, firms are still trying to figure out what to do with the data they have and what models can they build which truly add value back their business. The best way to do that is probably with an in-house team that’s sitting next to the traders.” Ravi Sawhney, Head of Trading Automation & Analytics, Bloomberg
“Industry security pricing composites have historically been generated based on relatively basic models. AI and advanced data science models can absolutely improve on those basic models and deliver much better composites for market reference. Data providers and trading venues already have access to the required pricing and trading data to create a better model.” Dan Tsou, Head of ETOMS, Bloomberg
“Each firm can build their own model but ultimately a third-party might have the benefit of being able to use peer-group data to build these models, which in turn makes it exponentially more valuable to each individual firm.” Ravi Sawhney, Head of Trading Automation & Analytics, Bloomberg
“AI talent with knowledge of financial markets is a difficult combination to find. If a sell-side firm wants to be a leader in the short term, the best option is to partnership with a financial technology firm to jumpstart the strategy whilst building the internal team. The work required to deliver a full solution is substantial. The solution only delivers business benefit if the right talent develops the right model with the right data integrated in the right trading tool. So when the trader is pricing the bond, deciding on the trade, deciding on the hedge, they have the right prediction model to improve their performance.” Dan Tsou, Head of ETOMS, Bloomberg
“There is an industry challenge of trying to get the best of breed in this space given the huge demand. As a result of this, you’ll start to see these firms shift the other way and embrace outsourcing more. The challenge will then be trying to combine those external capabilities with in-house domain experts to deliver value to the business.” Ravi Sawhney, Head of Trading Automation & Analytics, Bloomberg
