SFB 649 Discussion Paper 2007-064
Correlation vs. Causality in
Stock Market Comovement
* Freie Universität Berlin, Germany
This research was supported by the Deutsche
Forschungsgemeinschaft through the SFB 649 "Economic Risk".
SFB 649, Humboldt-Universität zu Berlin
Spandauer Straße 1, D-10178 Berlin
SFB 6 4 9 E C O N O M I C R I S K B E R L I N
Correlation vs. Causality in Stock Market Comovement1
Universität Mannheim and Freie Universität Berlin
Boltzmannstr. 20, 14195 Berlin, Germany
phone: +49 30 838-55792
fax: +49 30 838-54142
First version: 09/2007
This version: 10/2007
This paper seeks to disentangle the sources of correlations between high-, mid- and low-
cap stock indexes from the German prime standard. In principle, such comovement can
arise from direct spillover between the variables or due to common factors. By stan-
dard means, these different components are obviously not identifiable. As a solution,
the underlying study proposes specifying ARCH-type models for both the idiosyncratic
innovations and a common factor, so that the model structure can be identified through
heteroscedasticity. The seemingly surprising result that smaller caps have higher influence
than larger ones is explained by asymmetric information processing in financial markets.
Broad macroeconomic information is shown to enter the common factor rather than the
Keywords: Identification, Spillover, Common Factor, Structural EGARCH, DAX
JEL classification: C32, G10
1This research was supported by the Deutsche Forschungsgemeinschaft through the CRC 649 ”Eco-
nomic Risk”. I am grateful to Jürgen Wolters and Cordelia Thielitz for their help. Of course, all remaining
errors are my own.
Different stocks or portfolios often reveal a high degree of coherence in their fluctuations.
For example, from 2001 till 2007