Attribution and Analytics: Better Together

Attribution and Analytics: Better Together

• Industry Assistant Professor of Management Science Seung Youn Cha

Abstract:

Attribution and analytics are often used for similar purposes, but they are inherently different. This paper demonstrates that using a decision-based attribution model in combination with a regression-based analytics model can improve portfolio management decisions and make the results of the analytics regressions more precise and intuitive. The reason for this result is that regressions on portfolio returns cannot accurately reflect portfolios that are managed using more than one decision process. However, running separate regressions on attribution effects can accurately capture the portfolio management risks. This research also introduces new methodologies in attribution to isolate forecast model-induced effects from portfolio effects, stock level attribution, and a simple methodology for multi-country attribution. Each of these may be used with analytics regressions to better isolate portfolio issues with risk, return, or information ratios. The resulting framework presents a path forward for better portfolio management and academic research.

All Illinois Tech faculty, students, and staff are invited to attend.

The Friday Research Presentations series showcases ongoing academic research projects conducted by Stuart School of Business faculty and students, as well as guest presentations by Illinois Tech colleagues, business professionals, and faculty from other leading business schools.