Broadridge Onboards Clients to Corporate Bonds Platform Set to Launch Next Year

Broadridge is planning a full launch of its artificial intelligence (AI)-driven corporate bond trading platform, LTX, in the first quarter of next year.

The vendor, which announced in June that it had executed its first trades on LTX, is now gathering liquidity on the platform to enable the full launch.

Vijay Mayadas, who was recently appointed president of capital markets at Broadridge, says the technology is already live, but the platform needs to onboard 30 to 45 buy-side firms and five to 10 dealers by early next year before it reaches the level of liquidity necessary to drive trading activity.

“We are in that process of onboarding dealers and buy-side firms right now,” he says.

The vendor has been working on the platform for over three years, having begun discussions with sell-side and buy-side firms in 2016.

Mayadas says traders in the corporate bond market can take a lot of time to figure out who are natural counterparties for a particular trade. The firm was approached by some sell-side clients a few years ago asking for help in trading corporate bonds, particularly with larger, more illiquid bonds. These clients wanted to know if there was a way to leverage Broadridge’s dealer network and access to large quantities of data to help with the trades.  

“We spoke to about 100 buy-side firms [in 2016],” Mayadas says. “The buy side validated that it is really hard for them to move a large block of corporate bonds without taking a meaningful profit-and-loss hit because of the way the market structure had changed since the financial crisis.”

In response, the vendor built an AI engine using the open-source machine learning platform TensorFlow.  

“The corporate bond market is always this balance of inviting a bunch of your customers who have the highest probability of doing a trade but minimizing information leakage,” Mayadas says. “It’s always a delicate balance to strike. That is a complex thing to figure out, and underlying that is a lot of intrinsic complexity in the corporate bond market itself.”

Mayadas says the bond market is vast, with tens of thousands of underlying Cusips—codes used to identify securities to be sold publicly—which adds to the complexity of the trade. There is a lot of data involved in these trades and there needs to be a way to narrow down the data so as to provide actionable insight to a dealer.  

The AI used on the platform is called a convolutional neural network, a type of deep neural network used for image analysis. “It is a neural network with multiple layers of neurons,” Mayadas says. “One of the reasons we chose that model was because of the complexity of the corporate bond market. There is so much information in a set of trades that we felt a neural network was the best way to understand how to interpret all of that data in such a way that you can help a dealer come up with actionable insight.”

The AI will look at the dealers’ data and publicly available data to determine the optimal set of customers available that would most likely be the ones to invite to the trade with the greatest chance of getting the trade done. After that, the AI picks the customers from the data and puts them into a protocol, which then drives a type of auction mechanism to generate the best pricing for the buy-side customer, and also helps the customers on the other side of the trade with price discovery.

“A lot of these bonds only get traded a couple of times a month, so the pricing data that is available publicly may not be the best data for you to figure out how you want to place a bid against a certain trade. This auction mechanism helps you figure out what that is,” Mayadas says.