'Loose Lips Sinking Markets?': The Impact of Political Communication on Sovereign Bond Spreads

“Whatever it Takes!” How Tonality of Tv-News Affects Government Bond Yield Spreads During Crises

“…Another branch of literature focuses on the impact of news on fixed-income markets, most notably the effect of media coverage on government bond yields during the sovereign debt crisis in the EMU (see Table A.1 in the Appendix). In this context, Büchel (2013), Mohl and Sondermann (2013) and Gade et al (2013) analyze the impact of news on 10-year government bonds of euro area countries. Beetsma et al (2013) and Büchel (2013) focus on the GIIPS countries.…”

“…Mohl and Sondermann (2013), conducting a study of news data between May 2010 and June 2011, find a positive effect of the number of Eurozone government statements on government bond spreads in EMU when statements are related to 'restructuring' or 'bailout.' Based on their empirical study of news data between January 2009 and October 2011, Gade et al (2013) conclude that positive communication can lead to a compression of spreads, whereas negative communication dedicated to fiscal policy can cause a widening of spreads. Falagiarda and Gregori (2015) find a significant difference in the impact of the distinct Italian administrations.…”

“…Despite the fact that the aforementioned contributions provide interesting insights in the connection between media coverage and communication with bond yields, a second view might be fruitful as the existing studies face two sources of potential bias. First, most existing studies obtain their news data from news releases of agencies like Bloomberg, Reuters, Dow Jones Newswire, and Market News International (Conrad and Zumbach, 2016;Falagiarda and Gregori, 2015;Gade et al, 2013;Mohl and Sondermann, 2013). Beetsma et al (2013) use Eurointelligence, Apergis et al (2016) and Büchel (2013) obtain the news data from Factiva, an online database of newspapers that categorizes its articles by subject and provides a code that identifies articles discussing sovereign debt issues.…”

“…Other studies focus on simple Google search queries or social media like Twitter as a medium (e.g., Dergiades et al, 2015). Hence, most existing studies apply simple word counts or computer linguistic approaches (e.g., Apergis et al, 2016;Gade et al, 2013). This is especially critical if only a keyword is used to inform the algorithm on whether a report is relevant or not.…”

“…They further assume that financial markets immediately react to an event, i.e., a public statement, and that events can be determined precisely (on a daily basis) such that confounding effects are minimized. Falagiarda and Gregori (2015), Gade et al (2013), as well as Lopez and Weber (2017) have a similar strategy and additionally conduct Granger causality tests in order to determine in which direction the effect runs. However, although these contributions provide interesting insights in the possible link between the two variables, this approach may not resolve the endogeneity problems.…”