Bayesian Networks for Portfolio Risk and Return Modeling

Summary

The document asks whether Bayesian networks have been studied as a way to estimate a security's exposure, or beta, to one or more risk factors. It points to a paper by Shenoy and Shenoy on Bayesian network models of portfolio risk and return, published in 1999, as a relevant starting point.

The response is only a bibliographic lead. It does not explain the paper's network structure, estimation procedure, treatment of factor betas, or empirical findings, and it does not establish whether Bayesian networks outperform conventional beta estimation. Readers would need to consult the cited research to assess how directly it addresses factor exposures and whether the approach is promising.

Key ideas

• The question concerns estimating security exposure to risk factors with Bayesian networks.
• A cited starting point is research on Bayesian network models of portfolio risk and return.
• The document provides no methodological details or empirical evaluation of beta estimation.
• The cited paper's relevance to factor betas requires examination of the original research.

Tags

# Has any research used Bayesian networks to estimate risk factor betas?


# Has any research used Bayesian networks to estimate risk factor betas?












Is there any published research on estimating the beta of a security with respect to one or more risk factors via Bayesian networks?

I'd like to see if this is a promising angle of research.

## Answer by cyborg (score 7, accepted)

https://quant.stackexchange.com/a/2214

One relevant paper is:

> Shenoy, C. and Shenoy, P.P., Bayesian network models of portfolio risk and return, 1999. PDF

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.