Review of Quantitative Finance and Accounting, 19: 399–416, 2002
2002 Kluwer Academic Publishers. Manufactured in The Netherlands.
A Generalized Method for Detecting Abnormal Returns
and Changes in Systematic Risk
Department of Finance, 907 Rawls College of Business Admin., Texas Tech University, Lubbock, TX 79409-2101
Tel.: (806) 742-3196
RAMON P. DEGENNARO
SunTrust Professor of Finance, 423 Stokely Management Center, The University of Tennessee, Knoxville,
TN 37996-0540 Tel.: (865) 974-1726
Abstract. We generalize traditional event-study techniques to allow for event-induced parameter shifts, shifting
variances, and ﬁrm-speciﬁc event periods. Our method, which nests traditional methods, also permits systematic
risk to change gradually during the event period and exit the period at higher or lower levels. We use our approach
to study 123 banks that acquired other institutions between 1989 and 1995. We ﬁnd a signiﬁcant change in the
systematic risk of the acquiring ﬁrms, signiﬁcant ARCH effects, and an event period that ends before the date of
the announcement. None of these results is detectable using conventional methods.
Keywords: event studies, abnormal returns, econometric methods
JEL Classiﬁcation: G14, G19, G21
Traditional event-study methods use a three-step procedure. First, the researcher selects a
model of returns. Second, he computes abnormal returns during some event interval as the
difference between realized returns and the expected returns conditioned on this model.
Finally, he evaluates the statistical signiﬁcance of these abnormal returns in any of several
The value of this method is indisputable. Although many researchers have challenged
traditional approaches on various grounds, empirical evidence in several studies (e.g., Brown
and Warner, 1980, 1985; Malatesta, 1986; and Henderson, 1990) has concluded that the
event-study approach is quite robust for detecting abnormal mean returns.
We argue, though, that this perspective of event studies is limited. That is, it focuses
only on mean returns during the event, while ignoring a wide range of other interesting
ﬁnancial events. When researchers have studied changes in volatility during the event period,
they have done so primarily to improve the third step of the event-study process, that
Address correspondence to: Ramon P. DeGennaro, SunTrust Professor of Finance, 423 Stokely Management
Center, The University of Tennessee, Knoxville, TN 37996-0540. Tel.: 865-974-1726.