Lumesis introduces artificial intelligence for pricing, scales platform

Lumesis has incorporated artificial intelligence to its municipal pricing and scales software in an effort to address limitations of the interpolation and extrapolation methodologies in predicting curve structures that are currently widely used in the market, the firm said.

The improvements have been added to the company’s DIVER Pricing and Scales software and is currently available, Lumesis said.

The artificial intelligence software is used “to predict the shape of a deal-specific municipal yield curve based on the unique structural and credit-characteristics of the deal being priced and to use that predicted shape to fill-in portions of a yield curve that cannot be built from observable market transactions,” Lumesis said in a press statement.

Sudha Balla, Lumesis
Lumesis Managing Director for Architecture and Technology Sudha Balla said the new approach addresses proposed new issue scales for maturities that lack recent market observations.

“When particular maturities lack recent trades, traditional interpolation and extrapolation techniques to fill these gaps in a scale can fall short, particularly at ‘transitional’ points in a curve, such as where coupons change,” said Lumesis Chief Executive Officer Tim Stevens.

Managing Director of Architecture and Technology Sudha Balla said that “by utilizing Machine Learning technology and linking the same with Lumesis’ proprietary obligor database, we have created models that can predict, with very high accuracy, the difference in curve shape and steepness between individual maturities for different types of credits with different structural features.”

“Before implementing this logic and models into our Pricing and Scales platform, we back-tested the models and algorithms,” Stevens said. “For tax-exempt structures, we found the solution to be superior or equivalent to traditional approaches in 94% of the scenarios tested and, for taxable structures, the AI-based method performed superior or equivalent to traditional approaches in 98% of the scenarios tested.”

This is the first time Lumesis is using Artificial Intelligence in software offered to clients. Public finance bankers, municipal advisors, underwriters, traders and issuers use the pricing platform. Lumesis said clients have had a positive reaction to the new software.

For reprint and licensing requests for this article, click here.
Software development Artificial intelligence Machine learning Capital markets
MORE FROM BOND BUYER