انتشار اخبار: تاثیر اخبار بر بازارهای اوراق قرضه اروپا در طول بحران

We investigate how “news” affected domestic interest spreads vis-à-vis Germany and how it propagated to other countries during the recent crisis period, thereby distinguishing between the so-called GIIPS countries and other European countries. We make original use of the Eurointelligence newsflash to construct news variables based on the amount of news that is released on a country on a given date. We find that more news on average raises the domestic interest spread of GIIPS countries since September 2009. In addition, we find that it leads to an increase in the interest spreads of other GIIPS countries. The magnitude of this effect is related to cross-border bank holdings. A split of news into bad and good news shows that the upward pressure on domestic and foreign interest spreads is driven by bad news. We also find spill-overs of bad news from GIIPS countries onto non-GIIPS countries. However, the magnitude of these spill-overs is substantially smaller than that to other GIIPS countries.

The European sovereign debt market has been in a continuous crisis since the new Greek government in the fall of 2009 announced that the deficit for that year would turn out to be much higher than originally predicted. Since the announcement yields on Greek public debt have steadily risen, in spite of repeated promises of further austerity. The turbulence in the Greek debt market subsequently spread to other countries as well. This has led to a first rescue package for Greece and the installment of a crisis mechanism with funds from the EU (the European Financial Stabilisation Mechanism, EFSM) and other euro-zone countries (European Financial Stability Facility, EFSF). However, the relief that followed these measures was only short-lived. After an initial fall, bond yields started creeping up again, while capital market access became impaired. A second rescue package was negotiated with Greece in the summer of 2011, but only ratified at the beginning of 2012. Meanwhile, Ireland and Portugal have also received rescue packages. The debt crisis also inspired a wave of new European legislation to deal with fiscal profligacy and macroeconomic imbalances. However, so far, these European measures have proven ineffective in solving the crisis. In this paper, we explore co-movements among interest spreads vis-à-vis Germany on European public debt and spill-overs in response to macroeconomic and financial news. We extract our news variables from the newsflash of Eurointelligence, which is an independent internet-based service providing daily morning news briefings of the European media for readers with an interest in euro-area news. Founded in 2007 as a simple daily platform for debate and commentary on news for the euro-area (with a focus on macroeconomics, politics and macro-finance) Eurointelligence witnessed a spectacular rise in readership, from just a few hundreds to 4000 daily visitors.4 Although Eurointelligence is widely read by the most influential policymakers and experts in the private sector, we do not expect it to be the main source of information of investors. However, we consider it as a compact and consistent form of information provision that captures the main daily economic, financial and political concerns. To the best of our knowledge, this way of using the newsflash is new. We focus in particular on the GIIPS countries (Greece, Ireland, Italy, Portugal and Spain) over the period since mid-2007. However, we also compare our analysis for these countries with that of a set of other European countries and make a comparison between sub-samples where we split the full sample at September 2009. Our results are the following. We find that more news about a country, as measured by the number of times a GIIPS country is mentioned in the newsflash, drives up the interest rate spread of the country. In addition, we find spill-overs of the news concerning the country onto other GIIPS countries related to the value of the financial claims of the banking sector of the other countries on the country under consideration.5 By contrast, interacting news with the intensity of cross-border trade linkages does not yield significant results, thereby indicating that investors view banking sector linkages among countries as particularly important. In other words, our analysis can be viewed as rationalising the spill-overs across sovereign debt markets on the basis of cross-border stakes of the banking sectors. We establish the robustness of the aforementioned news effects for variations on our baseline regressions. We also establish that, not surprisingly, the news effects are concentrated in the second half of our sample period, i.e. the period September 2009–February 2012. Further, while most of the attention during the past couple of years has focussed on the GIIPS countries, we also find spill-overs from the GIIPS to several non-GIIPS European countries (except Germany). However, while those spill-overs are in the same direction, they are smaller in size. Finally, when we split our news variable into bad and good news, we show that the domestic and cross-border effects of news are confined to bad news. This is the case both for the spill-overs from GIIPS to other GIIPS countries and the spill-overs from GIIPS to non-GIIPS countries. This paper connects to different, sometimes overlapping, strands in the literature. First, and foremost, our work relates to the literature that investigates whether news has an impact on financial markets. Examples are Andersen et al., 2003 and Andersen et al., 2007 and Fleming and Remolona (1999), who study the (almost) immediate impact of U.S. macroeconomic news releases on the bond, foreign exchange and stock market. Kaminsky and Schmukler (1999) use data from the Asian crisis to investigate the impact of news on stock markets. Other works are Baig and Goldfajn (1999) and Albuquerque and Vega (2009). However, closest in spirit to our work is Aizenman et al. (2012) who explore the spill-overs of the recent global and euro-zone debt crisis on regions of developing countries. Unlike us, they use an event study approach, while, like us, they distinguish the effects of bad and good news. However, they use different news sources, while their news measure does not measure its intensity. Second, there is a substantial literature on contagion and co-movements in financial markets. An overview is given in Pericoli and Sbracia (2003). More recent work is due to Bekaert et al. (2011). Co-movements may be caused by interdependence as a result of fundamental and financial cross-country linkages. While there exists no unique definition, contagion generally refers to some form of discontinuity in the cross-border relation among financial markets as a result of a crisis. A third strand of relevant literature explores the role of trade and financial linkages among countries in the contagion of currency crises – see, for example, Eichengreen et al. (1996), Van Rijckeghem and Weder (2001) and Albuquerque et al. (2011). The final relevant strand is the recent literature dealing with European bond markets. Examples of contributions to this literature are Beber et al. (2009), Favero et al. (2010), Bhanot et al. (2011), De Santis (2012), Kallestrup et al. (2012) and Ang and Longstaff (2012). These contributions mostly deal with the effects of credit risk and liquidity on yields or yield spreads. A notable contribution to this final strand of the literature, which is closer in spirit to our paper, is Mohl and Sondermann (2013). They construct a dataset scanning thousands of news agency reports (from Bloomberg, Dow Jones Newswire, Market News International and Reuters) for statements of European politicians about “restructuring”, “bailout” and the “EFSF” and test their effects on the European bond market. They find that the intensity of these statements impacted bond spreads of the GIIPS vis-à-vis Germany during the period between May 2010 and June 2011. The main innovations of this paper relative to the above literature are the following. We use an up-to-date sample period of daily yield data up until the end of February 2012. We have a novel news dataset that identifies the most important news on economic, financial, political and institutional developments in Europe. This news variable is not merely a dummy, but it also measures the ‘amount’ or ‘intensity’ of the news by the length (number of words) of the news items as well as by the number of times a particular word and country is mentioned on each given day.6 We also split this news variable into a “bad’ and “good” news variable (analogous to Baig and Goldfajn, 1999; who split news on the Asian crisis in bad and good news dummies) and test for asymmetric reactions. Finally, we model the interaction of the news variable with economic and financial integration variables. The remainder of this article is structured as follows. In Section 2 we set out the empirical model, while in Section 3 we describe the data that we use in this paper. Section 4 presents our baseline empirical results and some robustness checks on those results. In Section 5 we investigate the role of our news variable when it is split into bad and good news. Finally, Section 6 concludes the main body of the paper.

In this paper we have explored how ‘news’ affects domestic interest spreads in the euro-zone and how it propagates to other countries during the recent crisis period. To this end, we have distinguished between the so-called GIIPS countries and other countries. Part of the originality of this paper concerns the use of the Eurointelligence newsflash to construct “news variables” based on the amount of news that is released on a country on a given date. We have explored in detail the 5 and 10-year public debt market and found that more news on average raises the domestic interest spread of GIIPS countries since the fall of 2009. In addition, we find that more news in one GIIPS country leads to an increase in the interest spreads of other GIIPS countries. The magnitude of the spill-overs is strongly related to the size of the cross-border bank holdings. A split of news into bad and good news shows that the upward pressure on domestic and foreign interest spreads is driven by bad news and that the effects are confined to the second sub-sample period, i.e. the public debt crisis period that started in the fall of 2009. Many of the results for the GIIPS countries carry over to the non-GIIPS countries: we also find spill-overs from GIIPS to non-GIIPS countries and those spill-overs are again confined to bad news during the second sub-period. However, the spill-overs from GIIPS countries to other GIIPS countries are substantially larger than the spill-overs from GIIPS to non-GIIPS countries.