Taking too long? Close loading screen.

Our Work


Client Description
Broker/dealer with fixed income sales and trading operations
Background
Generate fast, accurate municipal bonds price quotes that will allow our client, a mid-sized broker-dealer, to bid on and win more trades (primarily odd lots) in the secondary market.

Approach
In the municipal bond market, there are far more Requests for Quotes (RFQs) on any given day than a small- to mid-sized broker-dealer can process; of course, if the firm doesn’t supply a quote, it can’t compete for the trade. Calculating a quote that can win the trade and generate a reasonable profit depends on a bond’s unique characteristics, its similarities with bonds that have recently traded, a dynamic rates environment, and the latest market updates.

Incorporating all of these factors fast enough to be relevant requires a non-trivial effort when responding to each individual RFQ. Relying on manual processing meant the firm could only compete for a subset of trade opportunities each day.

We combined our AI expertise with our first-hand knowledge of capital markets sales and trading to design a pricing incorporating both traditional term structure models and neural network algorithms to generate quotes based on real-time market conditions, as well as the firm’s proprietary history of trade color reports detailing all prior bids submitted and trades won and lost.
Outcome
An end-to-end solution that:

• allows the firm to respond to more RFQs more quickly, with more accurate price quotes;

• provides seamless connections via FIX protocols to multiple automated trading systems/ECNs;

• resulted in a significant increase in flow almost immediately after the system was implemented.