standardized returns are considered3. Because these results are contrary to the results presented in
table 1, we have examined them in more detail. An important finding is that Dutch companies
generally surround the news of an issuance of CBs or WBs with other (good) news. A number of
announcements are made together with the publication of the provisional or definite annual results
(11 CBs and 2 WBs). In the other cases the announcement is accompanied by the publication of
the issuance prospectuses and/or the presentation of the contents of the issuance prospectuses.
Besides technical details about the issue, these prospectuses include a great deal of other informa-
tion about the firm and its future prospects.
An interesting example is the case of the issuance of a CB of 300 million Dutch guilders
in February 1994 by the Dutch transport concern Nedlloyd. This issue was first announced in the
Dutch financial newspaper Het Financieele Dagblad on February 8, 1994 (day 0). The abnormal
returns on respectively days -1, 0 and 1, with t-values between brackets, are -0.52% (-0.26),
+6.98% (3.41) and +2.41% (1.10). This leads to cumulative abnormal returns on day 0 and day 1
of respectively +6.45% (2.23) and +8.86% (2.50). On February 14, 1994 an editorial comment in
Het Financieele Dagblad makes the following statement about these announcement effects
(literally translated):
"Meijer (the chairman of the Board of Nedlloyd) has packaged his message in good news. Much
earlier than usual, he has brought out his provisional results for 1993 as well as his global
prognoses for 1994".
Two weeks after the issue announcement, on February 22, 1994, Nedlloyd announced its
intention to increase the size of the issue from 300 to 400 million Dutch guilders. At this time
there was no more good news to surround the announcement with. The market reacted accordingly.
If we take February 22, 1994 as day 0, the abnormal returns on day -1, 0 and 1 are respectively -
3.83% (-1.87), -2.33% (-1.14) and +0.16% (0.08). The cumulative abnormal returns on day 0 and 1
are respectively -6.12% (-2.13) and -6.00% (-1.70). Unfortunately our sample only includes a few
issue size increases. Therefore a more profound study of this effect is not possible.
There is almost no issue in our sample that is not surrounded with announcements of good
news. Therefore it is not possible to correct for this by splitting the sample or by using a dummy
variable for good news in a regression analysis. The effect of surrounding announcements of CBs
and WBs with other (good) news seems to be less relevant for the United States. For example,
Brennan and Her (1993) study a sample of 155 CBs. This sample is drawn from an initial sample
3 Table 2 also presents sign-tests to investigate whether there are more positive than negative
returns on an event day. Under the null-hypothesis that the event does not have any effect, there
should be as many positive as negative returns, i.e. the percentage of positive returns should be
equal to 50%. This hypothesis is never rejected.