Join our reading group! https://hudsonthames.org/reading-group/

A problem of empirical correlation matrices is that they are purely observation-driven, and do not impose a structural view of the investment universe, supported by economic theory. The TIC algorithm is aiming to estimate a forward-looking correlation matrix based on economic theory. The method is using a theoretical classification of assets (hierarchical structure) and fits the empirical correlation matrix to the theoretical structure.

TIC algorithm is described in Estimation of Theory-Implied Correlation Matrices by Marcos Lopez de Prado.