Table 3
Impact of bond-based news about Greece and about the bailout of Greece on bank equity and sovereign bonds.
Regression equation:Rpt ¼aþbRmtþgN
G
gtþdN
B
gtþ3 t
News about Greece News about the bailout
g tstat. d tstat.
Abnormal return on bank equity
All banks from Greece 0.343 1.74 0.158 3.09
All banks 0.016 0.23 0.124 9.45
Banks exposed to Greece 0.017 0.23 0.132 9.41
Banks not exposed to Greece 0.012 0.18 0.101 4.62
Banks exposed to GIPS 0.016 0.21 0.132 9.15
Banks not exposed to GIPS 0.015 0.25 0.072 3.63
Abnormal return on sovereign bonds
Portugal 0.214 3.51 0.282 30.40
Ireland 0.160 2.79 0.234 20.61
Spain 0.052 1.96 0.113 24.64
Reportedtstatistics are calculated from Newey–West standard errors.
M. Mink, J. de Haan / Journal of International Money and Finance 34 (2013) 102–113 109
exposure to Greece. The results do not differ qualitatively from those for the full sample of banks, as
news about the economic situation of Greece never leads to significant abnormal returns, while news
about the bailout of Greece always leads to such abnormal returns. As the table shows, even banks
without an exposure to Greece respond to news about Greece’s bailout.
The next two rows consider banks with or without an exposure to any of the GIPS-countries. The
results confirm the previous ones: banks with an exposure to any of the GIPS-countries do not respond
to news about Greece, while even banks without such an exposure respond to news about the bailout.
Ourfinding that news about Greece does not have an impact on bank stock prices while news about
a bailout does, suggests that markets consider news about the bailout to be a signal of European
governments’ willingness in general to use public funds to protect private investors against losses.
When governments indicate, for instance, that they will not rescue Greece, markets consider this to be
a disturbing signal because it might imply governments also will not engage anymore in any other
financial sector rescue operations.
The last three rows of the table examine to what extent the prices of sovereign debt of the other
three GIPS-countries, Ireland, Portugal, and Spain, respond to news about the economic situation in
Greece and news about a Greek bailout. In this case we dofind significant abnormal returns associated
with news about the economic situation in Greece, while as before, news about the bailout leads to
abnormal returns as well. Thetstatistics for news about the bailout are a factor ten larger than those
for news about Greece itself.
That news about the bailout leads to abnormal returns in other countries is not surprising, as the
willingness of euro countries to bailout Greece obviously says a lot about their willingness to bailout
other GIPS-countries as well. However, that news about the economic situation in Greece leads to
abnormal returns on other countries’ bond prices might be more surprising, as it does not lead to
abnormal returns on bank stock prices (including the sub-sample with an exposure to the GIPS). An
explanation for the impact of news about Greece on other countries is that there is a learning effect.
Others refer to this as a‘wake-up call’: a crisis initially restricted to one country may provide new
information prompting investors to reassess the vulnerability of other countries, which spreads the
crisis across borders (Goldstein et al., 2000). According to this view, domestic fundamentals are likely
to play a dominant role in the transmission of the crisis (Bekaert et al., 2011). The ability of Greece to
reduce its budget deficit and government debt, and the response of rating agencies to these attempts,
are quite informative about the likelihood that other indebted countries will be able to quickly reduce
their debt levels as well. If Greece does not succeed to credibly commit to a sustainablefiscal policy, the
probability that other GIPS-countries will manage to do so may be small as well. Our results suggest
that the abnormal returns in GIPS-countries after news about Greece’s economic situation are especially due to such learning effects.
4.2. Robustness analyses
As a robustness analysis, we use changes in Greek 10-year senior sovereign CDS-spreads to identify
event dates, instead of changes in Greek sovereign bond prices. This way we examine the robustness of
our results to using a different measure of sovereign default risk. In addition, as the days with extreme
returns in CDS-spreads differ in several cases from the days with extreme returns in sovereign bond
prices, while sometimes there seems to be no clear news driving the event, we this way examine the
robustness of our results to the identification and classification of the news events. All CDS-spreads are
obtained from Thomson Datastream.
Table 4shows that the results from regressions using CDS-spreads are similar to the ones presented
above. For Greek banks both news about Greece and news about a bailout have a significant impact, but
for all other bank portfolios news about Greece does not lead to abnormal returns while news about the
bailout does.
7
Interestingly, for these other portfolios the coefficients for news about Greece and
news about a bailout are always significantly different from each other. Abnormal returns in countries’
7