Cointegration networks in stock markets

ABSTRACT

We use a novel approach based on a combination of network and cointegration analysis to examine linkages between stock markets across market cycles. Our results show that long-run linkages are likely to be global rather than regional and that market turbulence increases linkages. However, we find no widespread common stochastic trends between markets and neither are we able to draw a conclusion that major financial markets display influences network linkages.

KEYWORDS:

• Cointegration network
• equity markets
• linkages
• market turbulence

JEL CLASSIFICATION:

• F3
• G1

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Average coefficients are used to calculate weights in cases where two-way directed linkage exist between market pairs.

² Using Engle–Granger test for small sample periods in the presence of nonnormality can be problematic (Siklos and Ng 2001).

³ Source: http://atlas.media.mit.edu/en/, retrieved on 21 June 2016.