Return spillovers around the globe: A network approach

Abstract

Using a rolling windows analysis of filtered and aligned stock index returns from 40 countries during the period 2006-2014, we construct Granger causality networks and investigate the ensuing structure of the relationships by studying network properties and fitting spatial probit models. We provide evidence that stock market volatility and market size increases, while foreign exchange volatility decreases the probability of return spillover from a given market. We also show that market development and returns on the foreign exchange market and stock market also matter, but they exhibit significant time-varying behaviour with alternating effects. These results suggest that higher market integration periods are alternated with periods where investors appear to be chasing returns. Despite the significance of market characteristics and market conditions, what in reality matters for information propagation is the temporal distance between closing hours, i.e. the temporal proximity effect. This implies that choosing markets which trade in similar hours bears additional costs to investors, as the probability of return spillovers increases. The same effect was observed with regard to the temporal distance to the US market. Finally, we confirm the existence of the preferential attachment effect, i.e. the probability of a given market to propagate return spillovers to a new market depends endogenously and positively on the existing number of return spillovers from that market. JEL Classification: G01, L14 40 developed, emerging, and frontier markets, we test for Granger causalities among returns while controlling for the size of multiple Granger causality tests and taking care of return alignment with respect to non-synchronous trading effects. A possibly high number of return spillovers creates a complex network of relationships, which depicts world-wide market linkages. This is described via measures used in the network theory.The remainder of the paper is organized as follows. In Section 1 we briefly introduce the reader to the networks and their use in finance. Section 2 describes the data, including the return alignment procedure used to deal with non-synchronous trading effects. Econometric and network methodology is described in Section 3. Section 4 describes the results and the last section concludes.

A Weighted and Directed Perspective of Global Stock Market Connectedness: A Variance Decomposition and GERGM Framework

“…The inner logic of financial networks is that the intricate connectedness, either physical based on bilateral exposures/flows between financial institutions, or association-based depicting return dependency among financial markets, could be captured and analyzed in complex financial systems using a network approach [1,2,6]. The network approach, which describes relationship architecture and regularities involved in complex multivariate systems, has become a powerful tool in financial crises early warning and tracking [7,8], risk spillover sources tracing [9,10], or exploitation of asset allocation [11,12]. Three research paradigms exist in the current financial network literature [1], namely, (i) mean-spillover network or Granger-causality network [13], (ii) volatility spillover network represented by variance decomposition-based network [14] and GARCH-based network [15], and (iii) risk spillover network with the main forms in tail-risk driven network [16] and extreme risk network [8].…”

A Network Approach to the Study of the Dynamics of Risk Spillover in China’s Bond Market

“…Morana and Bagliano [ 15 ] analyzed business cycle spillovers and synchronization within groups of old and new European Union countries and found out that spillovers are beneficial for the common monetary policy of the European Union. Lyocsa et al [ 16 ] studied the connectedness of a sample of 40 stock markets across five continents using daily dosing prices and return spillovers based on Granger causality by building a complex network of the global stock market. In conclusion, they found that the probability of return spillover from a given stock market to other markets increases with market volatility and market size and decreases with higher foreign exchange volatility.…”

Volatility Similarity and Spillover Effects in G20 Stock Market Comovements: An ICA-Based ARMA-APARCH-M Approach

“…However, a large number of parameters have to be estimated in these models when it comes to more than two or three financial assets. To overcome the curse of dimensionality, some network models have been proposed in recent years [1,18,20,[39][40][41][42][43][44]. Geng et al [18] construct volatility networks of energy companies using the connectedness network approach and provide a reference for risk management.…”