Journal of Investment Strategies

Risk.net

Winning investment strategies based on financial crisis indicators

Antoine Kornprobst

  • A framework for financial crisis indicators with a demonstrated predictive power.
  • Systematic investment strategies based on the predictive power of the financial crisis indicators.
  • A robust method to calibrate the indicators and set the parameters of the strategies.
  • Strong performance of the systematic investment strategies, which outperform static references and random paths.

The aim of this work is to create systematic trading strategies built around several financial crisis indicators, which are based on the spectral properties of market dynamics. Using singular value decomposition techniques to compute all spectrums in an efficient  way, we build two kinds of indicator. First, there are those that, for every date considered, compare the distribution  of the eigenvalues of a covariance  or correlation matrix with a reference distribution,  representing either a calm or an agitated market reference. Second, there are those that compute, for every date considered, a chosen spectral property (be it the trace, spectral radius or Frobenius norm) of a covariance or correlation matrix. Working  with the component stocks of the French CAC 40 index, we aggregate the signals provided by all these indicators in order to minimize any false positive errors; we then build systematic trading strategies based on a discrete  set of rules governing our investment decisions. Finally, we compare our active strategies against a passive reference as well as random paths. In this way, we provethe usefulness of our approach and the value added by the out-of-sample predictive power of the financial crisis indicators on which our systematic trading strategies are built.

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: