Nonlinear spillovers between euro area sovereign bond markets

Abstract

Summary: This paper examines nonlinear spillover effects between sovereign bond markets of six euro area countries (France, Ireland, Italy, Germany, Portugal, and Spain), four of which were among the hardest hit by the sovereign debt crisis, by applying a nonlinear Granger causality test of Diks and Panchenko (2006). The test is applied on the sovereign bond yield dynamics (i.e. yield changes) time series for the time period from 3 January 2000 - 31 August 2011. We also test for pure spillovers between sovereign bond yield dynamics, i.e. the spillovers after controlling for common and regional factors that impact the sovereign bond yield changes of all countries simultaneously. To verify if the nature of spillovers has changed after the start of the euro are sovereign debt crisis, we test for the nonlinear spillovers for the whole observed period and separately for the period before and after the start of the euro area sovereign debt crisis (period from the start of April 2010 until the end of our sample, i.e. 31 August 2011). The results of our study show that strong bi-directional Granger causality exists between the investigated sovereign bond markets. Very similar results are obtained whether the regional and world factors are or are not controlled for. We find strong bi-directional non-linear Granger causality for the investigated euro area countries prior the euro area sovereign debt crisis. After the start of the euro area sovereign debt crisis the interdependence between the markets has reduced. We can no longer detect non-linear spillovers running from Germany and France to the periphery euro area countries. The findings of this study have important implications for the policymakers as they show that shocks spill-over quickly across the sovereign bond markets and the intensity and nature of spillovers can change throughout time. The sovereign bond markets of the core euro area decoupled from the periphery euro area sovereign bond markets after the start of euro are debt crisis. The findings are also of relevance for individual investors in the sovereign bond markets for the purpose of portfolio diversification.Keywords: sovereign bond markets, spillovers, non-linear Granger causalityJEL classification: F21, F36, G15, H63(ProQuest: ... denotes formulae omitted.)IntroductionSince the early 2010 euro area sovereign debt crisis has been on the top of the international, especially euro area economic policy agenda. The euro area sovereign debt crisis, triggered by mounting concerns about the public debt sustainability of Mediterranean countries and Ireland quickly spread across (i.e. spilled-over to) the euro area sovereign bond markets, thereby raising the question of public debt sustainability and management and the macroeconomic effects of the sovereign debt crisis. Prompted by financial market pressures, large-scale fiscal austerity measures have been announced in practically all monetary union member states.The knowledge of the size and nature of exposure of sovereign bond to market spillovers can help policymakers gain insight into public financing constraints and the external risks faced by national economy and their economic agents (Metiu, 2011). This knowledge is relevant also for private financial market participant. Since the works of Markowitz (1952) and empirical evidence of Grubel (1968) financial practitioners adhere to international diversification in order to reduces total risk of their international portfolio(s). When spillovers occur, the dependence between the returns of assets increases and the advantages of international diversification of investment portfolio reduces (Ling and Dhesi, 2010). Rational investors shall respond to changing patterns in dependence by adjusting their portfolios (Savva and Aslanidis, 2010).The most frequently applied methods in the literature on financial markets interdependence include the vector autoregressive (VAR) models, Granger causality tests (Malliaris and Urrutia, 1992, Gilmore and McManus, 2002), GARCH models (Tse and Tsui, 2002; Bae et al. …