Portfolio risk management in a data-rich environment Articles uri icon


  • Bouaddi, Mohammed

publication date

  • December 2012

start page

  • 469

end page

  • 494


  • 4


  • 26

International Standard Serial Number (ISSN)

  • 1555-4961

Electronic International Standard Serial Number (EISSN)

  • 1555-497X


  • We study risk assessment using an optimal portfolio in which the weights are functions of latent factors and firm-specific characteristics (hereafter, diffusion index portfolio). The factors are used to summarize the information contained in a large set of economic data and thus reflect the state of the economy. First, we evaluate the performance of the diffusion index portfolio and compare it to both that of a portfolio in which the weights depend only on firm-specific characteristics and an equally weighted portfolio. We then use value-at-risk, expected shortfall, and downside probability to investigate whether the weights-modeling approach, which is based on factor analysis, helps reduce market risk. Our empirical results clearly indicate that using economic factors together with firm-specific characteristics helps protect investors against market risk.


  • portfolio weights modeling; factor analysis; principal components; portfolio performance; value-at-risk; expected shortfall; downside probability