Optimal Peers

30 Pages Posted: 12 Apr 2024

Date Written: March 18, 2024

Abstract

This paper provides a theoretical foundation for constructing optimal benchmarks via machine learning (ML). For a broad class of models, the optimal benchmark is given by an appropriately weighted portfolio of peers. While Ordinary Least Squares (OLS) provides the theoretically optimal weights in the population, ML methods, notably the lasso, can provide a robust, implementable solution. In an application to a large sample of U.S. public firms, ML-based benchmarks strongly outperform traditional industry benchmarks in out-of-sample explanatory power. This suggests that ML-based benchmarks can substantially improve outcomes in a wide range of applications, such as incentive contracts or relative performance evaluation.

Keywords: Peer groups, benchmarking, machine learning, lasso, relative performance evaluation, industry classification. JEL Classification: D2, D4, D8, G1, G3, L1, L2

JEL Classification: D2, D4, D8, G1, G3, L1, L2

Suggested Citation

Peters, Florian S., Optimal Peers (March 18, 2024). Available at SSRN: https://ssrn.com/abstract=4763280 or http://dx.doi.org/10.2139/ssrn.4763280

Florian S. Peters (Contact Author)

University of Amsterdam ( email )

Plantage Muidergracht 12
Amsterdam, 1018 TW
Netherlands

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
35
Abstract Views
189
PlumX Metrics