Construction of factor indices and allocation

Scientific project

The research themes of this program are part of a problem of understanding the existing relationships between the financial assets in which the portfolio managers invest. One way to understand this problem is to try to extract the information contained in the financial markets through the evolution of the prices of these same assets, analyzed jointly. There are many explanatory (or risk) factors, so the idea is to synthesize most of this multivariate and complex information into a limited number of factors, which we call factor indices. Once these factors have been identified and analyzed dynamically, they can be used as a vector of investment, either individually or collectively so as to best diversify the risks identified and detected on the financial markets. Two lines of research were favored over the period March 2016-December 2017:

 
  • The first is to limit as much as possible the false information (or noise) observed on the financial markets. To do this it is necessary to take into account the possible presence of extreme events in the evolution of asset prices, and then to use robust estimation methods so that these events are not considered as normal events and do not disturb the quality of the estimates made
  • The second is to estimate the number of risk factors present in the universe of the assets studied. To do this we model the set of assets as a multivariate factor model and we seek to determine the number of predominant factors in the explanation of the joint developments of all assets. Associated with the first line of research, it is therefore possible to use the remarkable results of the theory of random matrices to determine the number of preponderant factors in the system.
 

 

Scientific officer

Emmanuelle Jay
Emmanuelle Jay
See CV

Economic Partner