Partial Natural Beta: An Argument for a ‘Better’ Beta

Abstract

Beta is the most widely used measure of risk exposure for most investment managers when reporting risk exposure to their investors.

We propose a conceptually simple method for estimating Beta, and its associated results, that has the benefit of using exactly the same calculations as traditional Beta but using a brute-force approach comparable to Value-at-Risk methods based on historical simulation. The Beta is calculated for the returns of a user-defined (PARTIAL) sample of the data.

The result is a Beta exposure value that is more stable and more representative of the expected risk but using the standard (NATURAL) calculations that have made Beta so familiar and easily available.

The Partial Natural Beta is being used successfully by both investment managers and institutional investors.

Keywords: Beta, Regression Analysis, Investment Risk Measurement

JEL Classification: C,G