key: cord-0931160-zqacdcep
authors: Fendel, Ralf; Neugebauer, Frederik; Zimmermann, Lilli
title: Reactions of euro area government yields to Covid-19 related policy measure announcements by the European Commission and the European Central Bank
date: 2020-12-31
journal: Financ Res Lett
DOI: 10.1016/j.frl.2020.101917
sha: 11510daa87a2fdcb56ddf90f4e4eecebbb4aef7f
doc_id: 931160
cord_uid: zqacdcep

This paper evaluates the impact of announcements of Covid-19 related monetary and fiscal policy measures by the European Central Bank and the European Commission. Applying an event study, we find that the announcements predominantly affect the government bond yields of more solvent countries such as Germany and the Netherlands. Since this finding mainly holds for the announcements of fiscal measures, we conclude that the investors are primarily concerned about the future fiscal burden that has to be shouldered by those solvent countries within the euro area.

The outbreak of the Covid-19 pandemic has severely hit economies around the world and has triggered ample policy responses. In addition to individual national policy measures, the European Central Bank (ECB) and the European Commission (EC) have undertaken unprecedented steps to support recovery of the economies in the euro area. This letter investigates the immediate impact of those monetary and fiscal policy measure announcements on the government bond market in the euro area. The reaction of government bond yields can be seen as the first element in monetary policy transmission mechanism as well as investors' perception regarding the ability of the fiscal authority to cope with the crisis by introducing fiscal policy packages. In particular, the yield responses seize investors' confidence in future fiscal space and capacity, which in turn determines the risk premium for the government bonds.

This letter is closely related to Fendel and Neugebauer (2020), who among others 1 look at short-term responses of government bond yields with respect to announcements of unconventional monetary policy measures of the ECB by employing event study techniques. In order to assure comparability with previous results from this strand of the literature, our event study employs the same methodology The paper contains original research and has not been submitted and published elsewhere.This paper represents the authors' personal opinions and does not necessarily reflect the views of the Deutsche Bundesbank or the Eurosystem. and data set (among them the control variables) as Fendel and Neugebauer (2020) for the Covid-19 pandemic period. Only the set of events changes from ECB announcements on unconventional monetary policy to Covid-19 related announcements by the ECB and the EC (including the related entities of the Eurogroup and the European Stability Mechanism). In line with Afonso et al. (2020) , we consider both fiscal and monetary policy announcements, as they are crucial for developments in financial markets and in particular for the government bond markets.

Only very few papers have examined this issue so far. Klose and Tillmann (2020) examine the reaction of stock market prices as well as bond yields in European countries in the pandemic and find significant effects. In contrast to our analysis, they look at national measures and European-wide measures in a pure panel setting and do not report results for individual countries. Further, their sample period ends on April 24, 2020 and thus excludes important events such as the economic stimulus package "Next Generation EU" announced on May 27, 2020. D'Orazio and Dirks (2020) investigate the impact of several Covid-19 related measures such as the imposed lockdowns on the euro area stock markets. They do not include the events we are examining and the bond prices only appear as explanatory variables for explaining the stock price movements. Hartley and Rebucci (2020) consider 24 Covid-19 related announcements of 21 central banks and find significant effects on local 10-year government bond yields. In comparison, our study focuses on the euro area and jointly evaluates fiscal announcements, which should be of no less importance. Andrieş et al. (2020) assess the impact of the pandemic in general on sovereign CDS spreads in Europe by looking at the cumulative abnormal changes in bond prices. They do not concentrate on monetary and fiscal measures. Delatte and Guillaume (2020) analyze the effects of a set of various policy measures in the euro area announced during the pandemic period including both fiscal and monetary measures. In contrast to our study, they only analyze sovereign bond yield spreads and find, for example, that Italy benefited the most from the policy interventions by a significant reduction in the Italian yield spread. We explicitly show below, that such a simplified analysis of only spreads (as opposed to level effects) might lead to a crucial misinterpretation.

The rest of the letter is structured as follows. Section 2 introduces the data set employed. Section 3 explains the underlying regression model. Section 4 presents the results, and Section 5 concludes.

The data set includes eleven euro area countries: Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands, Portugal, and Spain. The sample period runs from January 1, 2020 until July 31, 2020. Our main variables of interest are the 10-year government bond yields, which are usually chosen for that kind of studies. Fig. 1 displays their development over the sample period. The yields show a similar pattern, for instance they all peak in March 2020 when the Covid-19 cases spread quickly in Europe.

Our event set includes 20 Covid-19 related announcement dates, which are listed and described in Table A1 in the Appendix. 2 Six announcements are from the ECB and, thus, are related to monetary policy measures ("ECB events" thereafter). The most prominent event among these six is the announcement of the Pandemic Emergency Purchase Programme (PEPP) worth 750 billion euro on March 18, 2020. The remaining 14 announcements are policy announcements from the EC respectively from the Eurogroup or the European Stability Mechanism (ESM) such as the economic stimulus package "Next Generation EU" from May 27, 2020. These 14 announcements are fiscal measures and for the sake of simplicity we subsequently refer to them as "EC events".

To ensure comparability with previous studies, we employ the identical event study set up as in Fendel and Neugebauer (2020) giving the following linear regression model:

where Δy t = y t − y t − 1 represents the daily change in the yield rates of government bonds with a maturity of 10 years, and α is the constant. The dummy variable Covid t is the variable of interest. It is set equal to one on days when the Covid-19 related announcement (s) took place, and zero otherwise. Following Urbschat and Watzka (2020) we include the first lag Δy t − 1 as control variable. Further control variables, which have proved to be of particular information content, are included in the vector X t : country-specific stock market indices, the Citigroup Economic Surprise Index for the Eurozone, the 3-month Euribor future rate, the VIX, the US-$/Euro spot exchange rate, and the 10-year US Treasury Inflation-Protected Securities yield rates. 3 Eq.

(1) is estimated using the ordinary least squares estimator for each country individually. We use Newey-West standard errors. In a variant of Eq. (1), we alternatively regress the dummy and the controls on the changes in the spread of the yields relative to the German yield (instead of the change in the yield levels). Both variants (level changes and spread changes) are additionally set up as a panel model, which we estimate using a country fixed effects estimator. Table 1 presents the results for individual countries both for the changes in the yield level (left part) as well as for the changes in the spread (i.e., yield difference to German government bonds; right part). For convenience, we have transformed the original coefficients into basis points.

From Table 1 , we see that the dummy covering all events under consideration (first column) significantly increases the yield levels in only seven out of the eleven countries: Austria, Belgium, Germany, Finland, France, Ireland and the Netherlands. Interestingly, these countries are considered as the euro area member countries with a more sustainable debt burden and a more sound fiscal position. This observation suggests the interpretation that financial markets may consider the future financing aspects of main relevance, which puts a burden on this particular set of countries that have sustainable fiscal positions. This interpretation is supported when differentiating between EC and ECB events: The 14 EC events (third column) seem to be the main driver of these results. If the dummy variable covers Notes: For the sake of simplicity, only the estimated coefficients of Covid t are displayed. They are displayed in basis point (i.e., multiplied by the factor 100) to ease legibility. *, **, and *** denote a significance at the 10, 5, and 1 percent level, respectively.

fiscal announcements only the results turn to be statistically even more significant, whereas in case of a dummy covering only the 6 monetary measures (second column), the results do not show any significance. For the complementary case of regressing the events on the change in the spread relative to the German yield level, we observe the interesting fact that for those countries with a less sustainable debt burden (Spain, Italy and Portugal) the spread decreases statistically significant. Only the Greek spread does not display a significant effect. In line with Delatte and Guillaume (2020), we find that the Italian spread observes the largest decline. However, from the previous results of the level regression, we can conclude that the reduction of the spreads is mainly driven by the significant increase in German yields and not so much by a decrease in the yield of the particular countries themselves. Consequently, the spread reduction in case of Italy should not be interpreted as a positive development in terms of more relaxed financing conditions for the Italian government resulting from the Covid-19 related measures. This particular finding shows the clear advantage of level regressions over spread regressions, since the latter could potentially lead to a misinterpretation about the underlying driver of the spread changes.

Pooling economies into a panel confirms the former findings (see Table 2 ). We also allocate the eleven countries into two subgroups: "south" (Greece, Italy, Portugal and Spain) and "rest" (Austria, Belgium, Germany, Finland, France, Ireland and the Netherlands). 4 For the level regression with all 20 events, we see that only the latter subgroup (again) displays a significant increase in yields. Interestingly, in the panel setup, the monetary announcements are also significant even for the complete panel of eleven 28 -5.12 Notes: For the sake of simplicity, only the estimated coefficients of Covid t are displayed. They are displayed in basis point (i.e., multiplied by the factor 100) to ease legibility. *, **, and *** denote a significance at the 10, 5, and 1 percent level, respectively.

Crucial policy announcements by the ECB and the EC.

Notes: More events than event dates exist since multiple announcements happened on March 13, April 2, and May 15. ECB event dates are highlighted in grey; the other events refer to EC policy measures. Source: Press releases by the ECB, EC, and ESM.

countries (see row "all"). In the panel setting of the spread regression, a contradiction to the previous result seems to emerge. For all 20 events as well as the "EC events", the spread declines significantly for the complete set of countries (row "all") and the monetary announcements are also spread reducing for the more stable countries (row "rest"). However, the potential explanation is that in the spread regression of the panel setup Germany is not present in the latter subgroup anymore, since its government bond yields serve as the reference level for the yield differences (the spread). Also those countries whose yields increased (see results of the level regressions in Table 1 ) experience a spread reduction relative to Germany, because Germany experienced an even higher increase in its yields. Markets seem to place the biggest future financing burden on Germany.

Markets are presumably very sensitive to the national developments of the corona crisis. However, the results still hold when including the daily number of new Covid-19 cases per 100k inhabitants for each country as an additional control variable.

Our event study analysis has revealed that euro area government bond yields reacted significantly following the Covid-19 related announcements by the ECB and the EC. The spreads of high-debt countries declined. However, not because of a decrease in their yields themselves, but due to an increase in the yields of the more stable economies such as Austria, Germany, and the Netherlands. This is particularly true for the fiscal policy measure announcements. This result suggests that especially the financially stable countries are expected to carry most of the financing burden in the future and therefore their government bonds are traded with a higher premium.

The findings hint at an imbalance of yield reactions among euro area members to the common European-wide fiscal programs. A transition to a more integrated fiscal policy in the EU would mitigate such differences in the future. Complementary measures and reforms could support this process. Acharya et al., 2018 emphasize that sovereign bonds become riskier when their concentration in domestic banks increases. Finnerty and Gonzalez (2017) propose reforms of the European bank regulation to reduce the risk of undercapitalized banks and thereby also the fiscal burden expected by euro area members.

Lender of last resort versus buyer of last resort-evidence from the European sovereign debt crisis

The effects of macroeconomic, fiscal and monetary policy announcements on sovereign bond spreads

Measuring Euro area monetary policy

The effectiveness of unconventional monetary policy announcements in the euro area: an event and econometric study

The Covid-19 pandamic and sovereign bond risk

Covid-19: a new challenge for the EMU

Country-specific euro area government bond yield reactions to ECB's non-standard monetary policy program announcements

Why European Banks are undercapitalized and what should be done about it

An event study of Covid-19 central bank quantitative easing in advanced and emerging economies

Quantitative easing in the Euro area-an event study approach

 Table A1