Toward More Accurate Portfolio Modeling, with Equity and Credit Risk Applications

22 Pages Posted: 8 Nov 2011 Last revised: 8 Mar 2012

See all articles by Wenbo Cao

Wenbo Cao

Standard & Poor's - Quantitative Analytics

Craig A. Friedman

State++

Date Written: November 7, 2011

Abstract

We describe a numerical method to estimate joint physical probability distributions on large portfolios. This method allows for conditioning on available firm-specific side information and can accommodate portfolios with "new'' instruments (for example, from IPO's) that lack historical data. We show that for the data that we examine, our models are more accurate than certain benchmark models, in the sense that our models, on average, better place probability mass where it belongs, as measured by out-of-sample likelihood.

Keywords: Large portfolio models, copula models, factor models, correlation, dependence, equity, credit, fat tails

Suggested Citation

Cao, Wenbo and Friedman, Craig A., Toward More Accurate Portfolio Modeling, with Equity and Credit Risk Applications (November 7, 2011). Available at SSRN: https://ssrn.com/abstract=1956209 or http://dx.doi.org/10.2139/ssrn.1956209

Wenbo Cao

Standard & Poor's - Quantitative Analytics ( email )

55 Water Street
New York, NY 10041
United States

Craig A. Friedman (Contact Author)

State++ ( email )

New York, NY
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
134
Abstract Views
872
Rank
385,913
PlumX Metrics