A general procedure is proposed to identify changes in asset return interdependence
over time using entropy theory. The approach provides a decomposition of
interdependence in terms of comoments including coskewness, cokurtosis and
covolatility as well as more traditional measures based on second order moments such
as correlations. A new diagnostic test of independence is also developed which
incorporates these higher order comoments. The properties of the entropy
interdependence measure are demonstrated using a number of simulation experiments,
as well as applying the methodology to euro zone equity markets over the period 1990 to
2017.