Journal of Risk

Risk.net

Decomposition of portfolio risk into independent factors using an inductive causal search algorithm

Brian D. Deaton

  • This paper presents a method to decompose a portfolio's risk along independent factors using a causal search algorithm paired with a vector autoregressive model.
  • A portfolio manager can use the independent factors to understand and budget risk or to create portfolios that have their risk contributions from independent factors constrained during portfolio optimization.
  • A small application using the six most widely traded currencies shows that portfolios created by using constraints on the independent risk factor contributions compare favorably to those created by using constraints on the risk contributions of the original assets.

ABSTRACT

A method is presented to estimate and decompose a portfolio's risk along independent factors. This decomposition is based upon a market's underlying independent risk factors, which are found empirically by using an inductive causal search algorithm that is based on independent component analysis. Since independent risk factors can be understood to always add risk to a portfolio, a portfolio manager can use them to better understand and budget risk. In contrast, portfolio management using the classic marginal analysis is confusing because adding a risky security to a portfolio might actually reduce the portfolio's risk. In a small application using the six most widely traded currencies (the Australian dollar, Canadian dollar, euro, sterling, Japanese yen and US dollar), independent-factor risk contributions are constrained during portfolio optimizations, and the internal risk characteristics of the resulting portfolios are found to compare favorably with those created by using constraints on the risk contributions of the original assets.

 

To continue reading...

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here: