Granger causality stock market networks: Temporal proximity and preferential attachment

Abstract

The structure of return spillovers is examined by constructing Granger causality networks using daily closing prices of 20 developed markets from 2 nd January 2006 to 31 st December 2013. The data is properly aligned to take into account non-synchronous trading effects. The study of the resulting networks of over 94 sub-samples revealed three significant findings.First, after the recent financial crisis the impact of the US stock market has declined. Second, spatial probit models confirmed the role of the temporal proximity between market closing times for return spillovers, i.e. the time distance between national stock markets matters.Third, preferential attachment between stock markets exists, i.e. spillover from market j to market i is more likely if A) market j influences other markets other than i, or when B) market i is influenced by other markets other than j.Keywords: stock market networks, Granger causality, emerging and frontier markets, nonsynchronous trading, preferential attachment, JEL classification: L14, G1Highlights: Granger causality networks are constructed among 20 developed stock markets. A detailed procedure of handling the non-synchronicity of daily data is proposed. The spatial probit model is used to study the structure of the created networks. Relationships between markets depend on a temporal proximity of closing times.

Networks of volatility spillovers among stock markets

“…Especially with respect to performing the Granger causality test, caution must be exercised because information sets must be precisely aligned with respect to time. Our return alignment procedure follows Výrost et al [19], which we briefly summarize below: 3 (1) Closing prices for two stock markets are pairwise synchronized; i.e., when there is a missing observation (non-trading day) on one market, observations corresponding to this day on the other market are deleted. (2) Consecutive returns are computed, which means that returns over non-trading days during the week are excluded.…”

“…Attention has been paid by the researchers to the equity and stock markets (Güloğlu et al [17]; Boubaker and Ali Raza [18]; Výrost et al [19]; Al Rahahleh and Bhatti [20]; Liu et al [21]) as well as to commodity markets (Ji and Guo [22]; Lahmiri [23]). Additional evidence has been put forward for spillovers between spot and futures markets (Liu and Wang [24]; Kang et al [25]) or spot and derivative markets (Kim and Ryu [26]).…”

“…Our analysis is based on the network approach that has gained currency in the econophysics literature (see for example Výrost et al [19]; Lyócsa et al [6]; Majapa and Gossel [30]; Caetano and Yoneyama [1]; Nobi et al [12]; Kuzubaş et al [14]; Majapa and Gossel [16]). Specifically, we build on Výrost et al [19] who created bi-directional 2 Granger causality networks between daily returns of developed stock markets around a world and observed that return spillovers are more probable when markets trade (in terms of trading hours) more closely to each other. Similarly, stronger return spillovers were identified also between markets which, in a given time zone, trade at similar trading hours (Al Rahahleh and Bhatti [38]).…”

Network-based asset allocation strategies

“…(2010) considered networks wherein only correlations above a threshold value were used, as topological constraints in MST and PMFG have unclear statistical and economic meaning (Výrost et al, 2015). We therefore also considered threshold networks where only edges that were statistically significant remained in the network.…”

Co-movements of NAFTA stock markets: Granger‑causality analysis

“…Rec [2009] scrutinizes the financial integration of stock markets in the former Yugoslav countries. Finally Vyrost, Lyocsa, and Baumohl [2014] analyze Granger causality networks constructed amongst 20 developed stock markets.…”