Intra-Industry and Inter-Country Effects of European Mergers
Jorg Bley
American University of Sharjah
And
Jeff Madura
Florida Atlantic University
Please direct all correspondence to:
Dr. Jorg Bley
American University of Sharjah
School of Business and Management
P.O. Box 26666
Sharjah, United Arab Emirates
Phone: 971-6-5055345
Fax: 971-6-5585065
email: jbley@aus.ac.ae
Abstract
Our main objective of the empirical tests is to determine the valuation effects of
European acquisition announcements on rivals within the target’s country (intra-industry
effects) and on other European rivals (inter-country effects). In addition, we conduct
cross-sectional analyses to isolate potential sources of intra-industry effects and intercountry effects. In essence, the magnitude of inter-country spillover effects offers a
means to gauge the degree of economic integration and market harmonization among
European countries. An inter-country spillover effect implies that information transmits
intra-industry signals beyond country borders.
We not only find evidence of intra-industry effects, but of inter-country effects as
well. This is a major finding, as it suggests that signals about rivals can transcend
national borders. M&A-induced inter-country effects are stronger if target and acquirer
are not located in the same country. Companies that share the target’s operational and
financial attributes face an increasing likelihood to be taken over, as the number of
potential domestic acquirers is raised by a much larger number of potential non-domestic
acquirers. There is no evidence that markets distinguish between EU and non-EU
acquirer. Over the four-year period, inter-country effects are larger for larger mergers.
Since the introduction of the single currency on January 1, 1999, inter-country
effects have become more pronounced. The Euro coefficient was significantly positive
for the event day t0 and the four event windows assessed in response to M&As during
1999. This offers strong evidence that the more transparent markets caused by the Euro
facilitate the transmission of private information across country boundaries in Europe.
1. INTRODUCTION
In anticipation of vanishing currency fluctuations within the 11-country euro zone,
companies are repositioning themselves in Europe’s market with about 360 million
consumers. As the uncertainty surrounding exchange rates between European countries is
removed, companies are subject to more intense cross-border competition. The euro
imposes more transparent prices across European countries. As a result, European firms
are divesting non-core businesses and expanding across country borders within their
industry sector to generate economies of scale, and increase efficiencies.
The inception of the euro also allows for greater valuation transparency, as firm
valuations can now be easily compared across countries within a given industry. There
has been a dramatic increase in the number of corporate control transactions within
Europe over the last few years. Merger and acquisition activities in Europe, accounting
for an overall transaction volume of about $220 billion in 1990, have surged to more than
$1.5 trillion worth of deals in 1999. Even more compelling is that the share of crossborder M&A transactions has risen from less than 8% to about 45%. According to
KMPG, cross-border purchases of European targets totaled $421 billion in the first half of
2000, as the European economy continues to converge and consolidate.
Numerous studies have shown that stocks of target companies experience positive
abnormal returns upon the announcement of a M&A transaction1. To the degree that a
merger may signal favorable prospects within an industry, or may signal the likelihood of
additional takeovers within the industry, rivals could experience favorable valuation
effects. Yet, as Europe becomes a single market, the intra-industry effects may even spill
over into other European countries. Thus, the concept of intra-industry effects would
take on a new meaning when applied to European firms, as they may extend across
European borders (inter-country effects). That is, intra-industry effects may convert to
inter-country effects if transmitted to firms in the same industry but located in other
European countries.
Our objective is to measure the valuation effects of merger and acquisition
announcements in Europe on rivals within the same country (intra-industry effect) and
among rivals in other countries (inter-country effect). We also attempt to identify
characteristics that explain the cross-sectional variation in intra-industry effects and in
inter-country effects. The magnitude of inter-country effects indicates the degree to
which European markets have become homogeneous.
Results of our analysis suggest that during the gradual elimination of cross-border
barriers to entry, rivals within the same country experienced favorable valuation effects
in response to M&A announcements, which were attributed to economies of scale and the
access to new markets. Moreover, our analysis finds evidence of inter-country effects,
which have become more pronounced since the introduction of the Euro.
2. RELATED LITERATURE
A. Sources of Wealth Gain in Corporate Control Transactions
For about four decades, academic studies have argued about the phenomenon of
corporate control transactions in terms of motivation and outcomes. Mergers can be
categorized as either operating or financial in nature (Lewellen 1971). Mergers of
operations are motivated by several potential sources of wealth gains. The first major
source is the opportunity to create economies of scale or other direct efficiencies in
1
For example see Dodd (1980), Dennis and McConnell (1986), Scanlon et al. (1989) and Schwert (1996).
manufacturing. The larger the scale of potential efficiency gains through a streamlined
combination of related operations, the greater is the potential wealth effect for the bidding
company (Houston and Ryngaert 1997).
A second major motive of mergers is the enhancement of competitive sales
positions through augmented monopoly power or the appeals of a more complete product
line. Examples are the $10.35 billion bid for Assurance Generales de France (AGF) by
Allianz Holding AG of Germany, Volkswagen’s takeover of British luxury car
manufacturer Rolls Royce, the mega-merger of Daimler and Chrysler, and UK’s
Vondafone $148 billion bid for Mannesmann of Germany. Berkovitch and Narayanan
(1993) concluded that synergy is the primary motive for takeovers with positive total
gains. Empirical results of a study based on the 50 largest U.S. mergers during 1979-84
revealed that merging firms experienced asset improvements relative to their industries,
leading to higher operating cash flows (particularly for firms with overlapping
businesses). Yet, many studies suggest that efficiency gains from mergers are
disappointing, which may indicate that management’s cost-cutting measures are not
implemented prudently or are not thorough enough to reap all of the benefits of the
merger-induced efficiency potential.
Mergers can also be motivated by complementary research and technology, and
complementary managerial skills, which may improve efficiency. They may also be
motivated by diversification. The conglomerate merger may utilize unused debt capacity
of an acquired firm subsequent to a merger, and reduce the variability of total corporate
earnings through the portfolio diversification.2 Most of the conglomerate merger
transactions during the 1960s and 1970s elicited a strong negative market response
Berger and Ofek (1995). Furthermore, Kaplan and Weisbach (1992) found in a sample of
large acquisitions occurring between 1971-82, that unrelated targets were 4 times more
likely to be divested by the acquirer than related targets.
Mergers are also motivated when a target appears to be undervalued. However,
information asymmetry does not apply in this context since private information, by
definition, is shared among insiders and not with outsiders, preventing them from
becoming corporate bidders. Second, successful takeovers typically require a substantial
2
The foundation of this line of reasoning can be found, e.g., in Mead (1970).
premium to be paid above the target’s current market value, diminishing the possibility of
“bargain” takeovers.
B. Research on Intra-Industry Effects
Intra-industry effects have been analyzed for a variety of corporate decisions. Pyo
and Lustgarten (1990) proposed that the amount of information about one firm conveyed
by management forecasts of another firm in the same industry depends on the earnings
covariance between the two firms and the forecast firm’s earnings variance. The
covariance reflects the competitive relation between the two firms and the variance
represents noise contained in forecasts.
Lang and Stulz (1992) found that bankruptcy announcements have both a
contagion effect and a competitive effect on other firms in the same industry. The
contagion effect is the wealth loss experienced by firms with cash flow characteristics
similar to those of the bankrupt firm because the bankruptcy announcement conveys
information about the present value of the cash flows for these firms. The competitive
effect is the wealth gain experienced by competitors because the bankruptcy
announcement conveys information about the present and future competitive position of
firms in the bankrupt firm’s industry. The authors showed that the competitive effect is
dominant in industries where leverage and the degree of competition is low. As
liquidation announcements can signal a change in industry conditions and in the market
structure of the industry, rivals may experience valuation effects.
Cheng and McDonald (1996), however, found that the direction of the intraindustry effect of bankruptcy announcements varies across industries. The study reports
positive abnormal returns in an airline industry sample but negative abnormal returns in a
railroad sample. The findings provide support for the market structure hypothesis but cast
doubt on the contagion effect hypothesis. Along these lines Slovin, Sushka, and
Polonchek (1992) had found that intra-industry effects of seasoned equity offerings
depend on the type of industry. While rivals generally experienced negative abnormal
returns following seasoned equity offerings, no such effects could be found among banks.
Analyzing intra-industry effects of downsizing announcements, Sun and Yang
(1998) found that contagion effects on rival returns prevailed over the competition
effects, leading to an overall decrease in stock prices of rivals following downsizing
announcements.
C. Intra-Industry Effects of Corporate Control Transactions
Eckbo (1983) documented the announcement effect of horizontal mergers on the
stock performance of rival companies. To account for any contemporaneous crosscorrelation of returns, he pooled the rivals of each merger into one equally weighted
industry portfolio. Over a seven-day event window, significant cumulative abnormal
returns (CAR) were found for the portfolio of rival companies. These positive intraindustry effects were explained by the productive efficiency hypothesis, which states that
merger announcements signal new information indicating that the resources of the rivals
can be more efficiently employed. Assuming the market for corporate control is efficient,
any positive CAR of rivals can therefore be the result of an increase in the probability of
a rival becoming the target for a subsequent corporate control transaction.
Mitchell and Mulherin (1996) studied industry-level patterns in takeover and
restructuring activity during the 1982 - 1989 period. They found significant differences in
both the rate and time-series clustering of these activities across industries. The intraindustry pattern in the rate of takeovers and restructurings are directly related to the
economic shocks borne by the sample industries. Linking economic shocks with takeover
activity suggests that the announcement of a takeover of one firm in an industry should
be accompanied by a positive revaluation of other industry members. They conclude that
takeovers and restructurings are driven by industry shocks, the positive response to
takeover announcements by industry members could represent the anticipation of
ongoing restructuring throughout the industry. Akhigbe and Madura (1999a) show
additional evidence of intra-industry effects in response to acquisition announcements,
which they attribute to signals of an increased probability of additional takeovers in the
industry.
3. INTRA-INDUSTRY AND INTER-COUNTRY EFFECTS
A. Hypotheses
To the degree that European M&As signal information about the target’s
corresponding rivals or the corresponding industry prospects or restructuring trends, they
may lead to positive abnormal returns on rival stocks. Those intra-industry effects should
be particularly pronounced for stocks of European companies that are perceived to
exhibit similar characteristics as the firms engaged in the merger. That is, European rival
firms may have a higher probability of being acquired (Akhigbe and Madura, 1999b)
and/or being subjected to other forms of restructuring (Slovin, Sushka, and Bendeck
1991).
The ongoing harmonization of Europe’s political, legal, and economic
environment has triggered a process of corporate restructuring throughout Europe. In the
wake of the ongoing formation of the EU, spillover effects of corporate control activities
are expected to occur beyond country borders, as inter-country effects. Therefore, a
signal emitted by a European M&A announcement may be transmitted to rivals in the
same industry but based in other European countries.
B. Methodology For Measuring Intra-Industry and Inter-Country Effects
The event study methodology is used to measure the average daily abnormal
returns of rivals in response to the announcement of merger transactions. The event date,
t0, is defined as the first day the plans for the corporate control transaction were made
public, according to the Securities Data Company’s databank. The mergers and
acquisitions data set used in this study covers the period from 1996 to 1999. For each
year, the 100 M&As with the highest transaction values that have a publicly traded
European company as target are selected, regardless of the country of origin of the
acquirer. Only majority ownership transfers are included in the sample. For each target, a
portfolio of European rivals is formed based on the Dow Jones Industry Indexes
classification methodology. Intra-industry effects are measured using return time series of
industry indices corresponding to the target company’s industry classification and
geographic location. Inter-country effects are measured using return time series of the
corresponding European industry indices.3 Limited availability of DJGI industry index
data reduced the M&A sample to n = 250 for the intra-industry effect analysis and to n =
334 for the inter-country analysis.
3
All DJGI indices are denominated in US dollar.
The abnormal industry index return to which the target company belongs, is
measured over a thirty-day window (t0 to t+29) following the event and calculated as the
difference between actual return and expected return:
ARit  Rit  E Rit ,
(1)
where:
ARit
=
abnormal return on the national industry index i for event
Rit
=
actual return on the national industry index i for event day t
E(Rit)
=
expected return on the national industry index i for event
day t
day t
By the same token, the abnormal return of the European industry index to which the
target company belongs is calculated as:
AR jt  R jt  E R jt ,
(2)
where:
AR jt
=
abnormal return on the European industry index j for event
Rjt
=
actual return on the European industry index j for event day
E(Rjt)
=
expected return on the European industry index j for event
day t
t
day t
For a sample of n indices, an average abnormal return on the national industry
index i (AARit) and the European industry index j (AARjt) for each day t is computed as:
(3)
1 n
AARit    ARit
 n  i 1
1 n
AAR jt    AR jt
 n  j 1
(4)
respectively.
The average cumulative abnormal return on the national industry index i (CARit)
and the European industry index j (CARjt) from event day t0 to event day t+29 is therefore
computed as:
29
CARit   ARit .
(5)
t 0
29
CAR jt   AR jt .
(6)
t 0
The expected industry return within one country is computed as a linear function of
the return on the entire national market. Applying the market model, the expected daily
return is calculated as:
E Rit    i   i Rkt
(7)
where:
E(Rit)
=
expected return on the target company’s corresponding
national
industry index i, at time t
Rkt
=
actual return on country market index k, of national
industry
index i, at time t
i
=
the intercept term
i
=
the slope coefficient associated with the market return of
country
k at time t
The expected European industry return is computed as a linear function of the
return on the entire European market. According to the market model, the expected daily
return is calculated as:
E R jt    j   j REUt
(8)
where:
E(Rjt)
=
expected return on the target company’s corresponding
European
industry index j, at time t
REUt
=
actual return on the European market at time t
j
=
the intercept term
j
=
the slope coefficient associated with the European market
return
at time t
Expected returns are generated from the market model parameters, estimated with
daily returns from the period t-140 to t-21 relative to the announcement date. Abnormal
returns are generated for the index-pooled national and European rival firms. Following
the methodology of Mikkelson and Partch (1988), a z-statistic is computed to test for
statistical significance of standardized abnormal returns and cumulative standardized
abnormal returns.
t2


AR
jt



1 n 
t t 1



 t2

n j 1 
 VAR   ARjt  
 t t 1


(9)
where:
ARjt
=
abnormal return as previously defined,
t1
=
first day of the time interval,
t2
=
last day of the time interval,
n
=
number of observations, and
t2
2 

~
R mt  T  R m 

2

t2


T
t1

Var   ARjt   Vj 2 T 
 t ED
2 
 t t 1

ED
~

R mi  R m 


i 1


(10)


where:
Vj2
=
residual variance for firm j from the market model over the
estimation period,
ED
=
number of days in the estimation period used to estimate
the
market model,
T
~
R mi
=
number of days in the interval (t2 - t1 + 1)
=
market portfolio return for the ith day of the estimation
~
R mt
=
market portfolio return for day t
Rm
=
market portfolio average return over the estimation period
m
=
the market portfolio is represented by the country market
period
index k
corresponding to the national industry index i, and the European
market index corresponding to the European industry index j
This approach differs from the standard formula for the variance of an individual
abnormal return because it adjusts for the dependence created by cumulating individual
abnormal returns calculated using a single set of parameter estimates derived from the
market model. Thus, the variance of the sum of individual abnormal returns is measured.
C. Results From Measuring Intra-Industry Effects
Using the Dow Jones Industry Index classification methodology, average
abnormal and cumulative average abnormal returns are measured on the target
company’s corresponding DJGI national industry index. Table 1 presents the breakdown
of the M&A sample by target company’s country and industry classification. A total of
250 M&A transactions are assesed over the period from 1996 to 1999.
The M&As with the highest transaction volume in each of the years 1996 to 1999
were not equally distributed across countries or industries. In each year, the United
Kingdom was the single country with the largest number of corporate targets, indicating
the highest level of large scale corporate restructuring. The second most active national
market for corporate control was France, with M&A activities occurring across a wide
spectrum of industries. The industry focus changed over the four-year observation period.
In 1996, the electric utilities industry experienced the highest number of top M&A
transactions, most of which took place in the UK. From 1997 through 1999, however, the
focus shifted towards the financial sector. Most target companies were found in the
banking industry, followed by the insurance industry, and, in 1998, in the real estate
sector. The fact that industries seem to experience several transactions in particular
periods suggests that an initial merger in an industry may signal impending mergers as
rivals respond to maintain their competitive position.
M&A announcement-induced intra-industry effects are analyzed based on a fouryear sample, n = 250, with the number of annual transactions ranging from 59 to 70.
Average abnormal returns on rivals are measured on the event day, t0, average cumulative
abnormal returns are measured over four event windows: t0 – t4, t0 – t9, t0 – t19, and t0 –
t29. Table 2 presents the results of the intra-industry effect analysis. Over the four-year
period, the abnormal mean return is 0.43% on the day of the M&A announcement, t0,
indicating a positive valuation adjustment of target company’s industry rivals by the
market. The z-statistic indicates significance at the 1%-level. The cumulative abnormal
mean return is 0.78% (significant at the 1%-level) over the 5-day event window and
0.82% (significant at the 1%-level), 0.67% (significant at the 5%-level), and 0.74%
(significant at the 10%-level) over the 10-day, 20-day, and 30-day event window,
respectively.4 These results offer strong support for the hypothesis that M&A
announcements emit favorable signals about the target’s corresponding rivals in the same
4
Year-to-year results of the intra-industry effect analysis are also available.
industry.
D. Results From Measuring Inter-Country Effects
Using a target company’s corresponding DJGI European industry index, M&A
announcement-induced average abnormal and cumulative average abnormal returns are
measured. Inter-country spillover effects were analyzed based on a four-year sample of
size n = 334, and annual samples of between 77 and 91 transactions. Table 3 presents the
annual breakdown of the M&A sample by target company’s country and industry
classification. Between 36% and 42% of each year’s top M&As had a target located in
the United Kingdom, followed by France with a range of between 11% and 16% and
Sweden with 6% to 10%.
Tables 4 shows the results of the inter-country spillover analysis. Over the 1996 –
1999 period, the mean abnormal return on the event day, t0, was 0.28%. The z-statistic
revealed significance at the 1%-level. The mean cumulative abnormal returns measured
over the four event windows were 0.36% (significant at the 1%-level), 0.46% (significant
at the 1%-level), 0.48% (significant at the 10%-level), and 1.02% (significant at the 5%level) for the 5-day, 10-day, 20-day, and 30-day event window, respectively.
4. CROSS-SECTIONAL ANALYSES
A. Characteristics That May Affect Intra-Industry Effects
The magnitude of intra-industry effects of M&A announcements varies among
merger announcements, as illustrated by the quartile statistics disclosed in Table 2. The
variation may be partially attributed to identifiable characteristics that are specific to each
merger. Specifically, we hypothesize that the magnitude of intra-industry effects is
conditioned on the following characteristics:
Type of M&As Economies of scale are more likely when the operations between firms
are related. Houston and Ryngart (1997) found the degree of business overlap between
acquirer and target to be positively related with the overall valuation effect. To the extent
that a specific type of merger signals additional mergers of the same type, intra-industry
effects are expected to be more pronounced for horizontal than for vertical mergers.
Thus, a dummy variable is set equal to one if the corporate control transaction is
horizontal and zero otherwise. This variable is expected to be positively related to the
intra-industry effects among European M&A announcements.
Cross-border transaction Grullon, Michaely, and Swary (1997) found positive valuation
effects for inter-state M&A transactions to be more pronounced than for intra-state M&A
transactions. The acquisition of a European target by a firm in a different European
country may signal the potential for corresponding rivals to be part of an international
market for corporate control. That is, they are subject to a takeover by a larger pool of
potential acquirers within Europe. The prospect of becoming a potential target for local as
well as for foreign corporations would increase the likelihood of a rival’s actual takeover.
Non-European acquirer
A European target may be acquired by a non-European
acquirer because the European acquirer wants to gain a foothold in Europe rather than for
reasons related to favorable industry-specific prospects.
Therefore, these types of
acquisitions may transmit less information about industry prospects, and intra-industry
effects are expected to be less pronounced.
Transaction value The takeovers of small firms may not have a substantial effect on the
market structure of the indusrty, which means that the rivals are less likely to take action
in response to this event. In addition, some of the larger rivals may be perceived as
immune to possible takeovers because of the costs of acquiring them. Conversely, large
takeover targets indicate that firm size is not a barrier to corporate control transactions,
and that all rivals are fair game. Thus, the larger the size of the target, the stronger is the
expected signal of subsequent acquisitions in the industry.
Tax treatment differentials
National differences in accounting rules can influence
takeover premia. According to the German tax system, for example, goodwill has a taxdeductible status.5 As a result, German firms tend to pay higher takeover premia than
5
According to the German tax code, goodwill can be depreciated straight-line over 15 years. The five
other EU countries that allow for the depreciation of goodwill apply different schedules: In Denmark ,
firms that are required to amortize goodwill to earnings, as documented by Lee and Choi
(1992). Given that the market is aware of these tax treatment-induced premia
differentials, takeovers involving acquirers of German origin should result in less
pronounced inter-country effects.
Relative industry performance An industry that underperforms the market could be
associated with a higher potential for such efficiency gains as well. Yet, to the extent that
merger prospects are based on growth potential, and growth potential is valued by the
market, intra-industry effects could be stronger for high performance industries. The
industry performance variable is measured as the average daily return difference between
the target’s national industry index i and the corresponding country market index k over
the one year period prior to the M&A announcement.
B. Cross-Sectional Model Applied to Intra-Industry Effects
The following OLS regression model is applied:
Valuei   0   1 Horizont   2 Cross   3 Non  EU
(11)
  4 Trans   5 Tax   6 Perf   i
where:
Valuei =
the valuation effect, measured as the CAR at time t on national
rivals in industry i
0
=
the intercept term
1-6
=
the slope coefficients associated with each independent
=
dummy variable set equal to one for a horizontal merger,
variable
Horizont
zero
goodwill is depreciated over 10 years on a straight-line basis if acquired on or after May 19, 1993 but
before January 1, 1998. Goodwill acquired thereafter may be amortised over a period of 7 years with 1/7 th
annually. While Italy accepts amortization of goodwill up to a maximum amount of one-tenth per year,
goodwill is generally amortised over a period of 3 to 5 years in the Netherlands. Norway and Sweden use
the declining balance method at 30% maximum depreciation rate per annum. Alternatively, Swedish
acquirer can depreciate goodwill under the straight-line method at the rate of 20% per annum.
otherwise
Cross
=
dummy variable set equal to one if the M&A transaction
occurred
across country borders, zero otherwise
Non-EU
=
dummy variable set equal to one if the European target was
acquired by a non-EU company, zero otherwise
Trans
=
the transaction value variable, measured as the natural
logarithm of the transaction volume in US dollars
Tax
=
dummy variable equal to one if goodwill is tax deductible
under
the national tax laws of the acquirer, zero otherwise
Perf
=
the industry performance variable, measured as the average
daily
return difference between the national industry index i and the
corresponding country market k over the one-year period prior to
the M&A transaction
i
=
the random error term
4
C. Results Of Cross-Sectional Analysis Of Intra-industry Effects
The intra-industry effects are measured on the day of the M&A announcement, t0.
The cumulative abnormal returns are measured over a five-day, ten-day, twenty-day, and
thirty-day event window ( t0 – t4, t0 – t9, t0 – t19, and t0 – t29, respectively). The results
from the cross-sectional anslysis of intra-industry effects are disclosed in Table 5.
Depending on the subperiod and the event window, between 1.1 percent and 5.5 percent
of the variation in the intra-industry effects is explained by the cross-sectional model.6
The results are partitioned by each characteristic that was hypothesized to affect the intraindustry effects.
Horizontal M&As
6
The Horizont variable was not statistically significant. Thus, there
The results of the year-to-year analysis are not shown in detail. However, depending on the subperiod
and the event window, adj. R2 – statistics were as high as 28 percent.
is no support for the hypothesis that the announcement of horizontal M&As induces a
more pronounced intra-industry effect then the announcement of vertical M&As.
Cross-border transaction
The cross-border transaction dummy variable, Cross, is
generally not significant. Only in 1997 did the 2-coefficient display significance at the
10%-level. There is little evidence of a relationship between the average intra-country
spillover effect and the country of origin of the acquirer relative to the target.
Non-European acquirer
Over the four-year period, none of the Non-EU
dummy coefficients was significant. Only in 1996, however, was the negative coefficient
of Non-EU of statistical significance (at the 5%-level), indicating that over the 20-day
event window from t0 to t19, the intra-industry spillover effect was lower if targets were
acquired by US and other non-European companies.
Transaction value
The coefficient associated with Trans was positive and
significant on the day of the M&A announcement (at the 1%-level) and over the four
event windows (at the 5%-level, 1%-level, 1%-level, and 5%-level for the 5-day, 10-day,
20-day, and 30-day event window, respectively). These results clearly indicate a positive
relationship between the dollar value of the M&A transaction and the magnitude of the
intra-industry valuation effect.
Tax treatment differential
The coefficient 5, associated with the dummy variable
Tax, was positive on the event day and over the four event windows. Yet, the coefficient
is only significant (at the 10%-level) for the 30-day event window.
The 6-coefficient associated with Perf was positive
Relative industry performance
for the event day but changed sign over the event windows. None of the results, however,
was statistically significant. Relative industry performance does not seem to influence the
magnitude of the intra-industry effect.
5. CROSS-SECTIONAL ANALYSIS OF INTER-COUNTRY EFFECTS
A. Inter-Country Model
The inter-country effects are cross-sectionally regressed on a set of country and
industry dummy variables:
Value j   j 
(12)
41

i 1
ji
IDi 
16

k 1
 jk CDk
where:
Value j =
the valuation effect, measured as the AAR and ACAR at time t
on the European industry index j
j
 ji
=
the intercept term
=
the slope coefficient associated with the industry dummy
variables
 jk
=
the slope coefficient associated with the country dummy
variables
IDi
=
one if
the target company belongs to industry i; zero otherwise
CDk
equal
the target company’s industry dummy variable, equal to
=
the target company’s national market dummy variable,
to one if the target is located in country k; zero otherwise
The industry dummy variables and country dummy variables, however, add up to
the unit vector across companies. Thus, assigning a country dummy variable and an
industry dummy variable to each industry portfolio return, generates an identification
problem due to perfect collinearity. The isolation of industry and country effects can only
be accomplished in terms of measuring cross-sectional differences between industries and
cross-sectional differences between countries. Thus, industry and country effects will be
measured relative to the average M&A-induced valuation effect on rivals as the
benchmark.
Setting up a regression model that incorporates a dummy for each class of
independent variables creates computational difficulties.7 As no unique set of coefficients
minimizes the sum of the squared disturbances 2, any constant K can be subtracted from
the value of each of the coefficients and added to the intercept without altering any
statistical properties of the model. As equation (12) incorporates two sets of dummy
variables (representing industries and countries), any constant K1 can be subtracted from
each of the coefficients of the set of industry dummies while any constant K2 can be
subtracted from the coefficients of the set of country dummies. The sum of the two
constants (K1+K2) can then be added to the equation’s intercept.
Implementing the constraints that the average European-wide industry effect is zero
and that the average European-wide country effect is zero, formally:
n
(13)

j 1
j
16
(14)

k 1
k
 0 , and
 0,
a set of unique values of the coefficients can be obtained.8 Equation (12) can be
interpreted as a means to measure the degree economic integration and market
harmonization among European countries and across industries.
The least squares estimate of the intercept term  is equal to the average M&Ainduced valuation effect on European rivals, measured as the CAR at time t on European
rivals in industry j, since the average residual is zero in every country and in every
7
As each class of qualitative predictor variable is represented by an indicator variable, the columns of the
X’X matrix become linearly dependent. Thus, the X’X matrix has no inverse and no unique estimators of
the regression coefficients can be found.
8
See also Suits (1984) and Kennedy (1986) for a detailed discussion.
industry.9 Adding these disturbances over all countries and industries, it can be assumed
that the estimated disturbance for the aggregate CAR is also zero. With the European
market return represented in form of the OLS regression’s intercept , both the country
effect and the industry effect are estimated net of the average M&A-induced valuation
effect.
B. Characteristics That May Affect Inter-Country Effects
The six factors specific to the merger that are hypothesized to influence the
magnitude of intra-industry valuation effects of M&A transactions may also affect the
magnitude of M&A-induced inter-country valuation effects. The performance
characteristic is measured with a different proxy as explained next. One additional
characteristic (Euro Participant) is also included when assessing the variation in intercountry valuation effects.
Relative European industry performance The performance characteristic that was used
for intra-industry effects is also applied to the assessment of inter-country effects, but the
proxy is modified. Just as intra-industry effects could be conditioned on national industry
performance (relative to that country’s market), inter-country effect could be conditioned
on
European
industry performance
(relative
to
the
European
market).
An
underperforming industry could be associated with a high potential for such efficiency
gains (and therefore acquisitions). However, a counter argument is that mergers will
proliferate in those industries that have experienced the most growth in Europe, where
many of the firms have experienced high performance.
The European industry performance variable (Perf) is measured as the average
daily return difference between the target’s European industry index
j and the
corresponding European market index over the one-year period prior to the M&A
announcement.
That is, the performance proxy in the intra-industry cross-sectional
model is revised for the inter-country model by measuring the performance of the
corresponding European industry (rather than a corresponding national industry) relative
9
By construction, in OLS estimation of a regression the estimated disturbances are orthogonal to all
industries and countries.
to the overall European market (rather than a single European country market).
Euro Participant
Just as closer ties between countries are expected to yield more
pronounced inter-country announcement effects of corporate control transactions,
participation of both the target’s and the acquirer’s country of origin in the first round of
the European Monetary Union, introducing the single currency on January 1, 1999, is
likely to generate more pronounced rival effects in the most recent years. Acquisitions of
targets based in participant member countries may emit stronger signals about the
prospects of mergers within the same industry throughout European countries. The fixed
exchange rates enforced by the euro allowed for more transparent valuations of European
firms in those countries. Thus, investors may have more confidence in mergers involving
these firms than mergers involving firms whose valuations are less transparent. This
translates into the expectations of more favorable prospects within the corresponding
industry throughout Europe.
The single currency dummy variable will indicate whether the countries in which
a target and an acquirer is located are participants in the euro. The dummy variable
(Euro) is set equal to one if both countries are introducing the Euro on January 1, 1999
and zero otherwise.
C. Results of Cross-Sectional Analysis of Inter-Country Effects
According to the inter-country model, the average abnormal returns on the event
day t0, were the highest for targets located in Norway, France and Switzerland and lowest
for targets located in Ireland, Belgium, and Sweden. European rivals experienced the
highest positive AAR if the target belonged to the oil, diversified financial, beverage,
telephone utilities, or software industries, and lowest if the target was a member of the
aerospace, steel, transportation equipment, textiles, and entertainment industry.
M&A-induced valuation effects on European rivals is regressed on a set of country
and industry dummy variables (eq.12) to isolate country and industry components. The
isolation of industry and country effects was accomplished by measuring cross-sectional
differences between industries and cross-sectional differences between countries. Thus,
industry and country effects were measured relative to the average M&A-induced
valuation effect on European rivals as the benchmark.
Table 6 presents the country effect estimates net of the average M&A-induced
valuation effect. The average abnormal return (AAR) on the event day t0, over the entire
four-year period is the highest for targets located in Norway. Targets in France and
Switzerland experience the second and third highest valuation impulse to European
rivals. Norway and France sustained this valuation premium over all four event windows.
On the other hand, target companies located in Ireland, Belgium, or Sweden experience
relatively low abnormal returns.
Table 7 shows the AARs and ACARs net of the mean M&A-induced valuation effect
on European rivals. An announced merger or takeover of an oil company triggers the
most pronounced ripple effects through a European industry (with an AAR on the event
day of 2.66% above the mean). Industry effects net of the mean AAR are also positive if
targets belong to the diversified financial, beverage, telephone utilities, or software
industries. ACARs in excess of the mean ACARs are sustained over the four event
windows by all but the beverage industry. The least pronounced inter-country effects are
in the aerospace industry, followed by the steel, transportation equipment, textiles, and
entertainment industry.
Results of the cross-sectional analysis of inter-country effects for the four-year
period are reported in Table 8. To isolate the M&A-induced valuation effects on
European rivals after the introduction of the euro, the results of the cross-sectional
analysis of the year 1999 are reported separately in Table 9. The mean abnormal return
was measured on the day of the M&A announcement, t0. The cumulative abnormal
returns were measured over a five-day, ten-day, twenty-day, and thirty-day event window
( t0 – t4, t0 – t9, t0 – t19, and t0 – t29, respectively).
Horizontal M&As
The Horizont coefficients for the event day t0 and the four event
windows in the period of 1996-1999 were positive. Statistical significance could only be
found for the 5-day event window (at the 10%-level). For the year 1999, all five Horizont
coefficients are positive but not significant. Thus, there is no support for the hypothesis
that the announcement of horizontal M&As induces a more pronounced spillover effect
then the announcement of vertical M&As.
Cross-border transaction
For the period of 1996 – 1999, all five 2 Cross coefficients
were positive. Statistical significance only exists for the 30-day event window. The more
impressive results, however, were found in 1999: The coefficient estimates for the event
day t0, the 5-day, the 10-day, and the 20-day event window are all significantly positive
(at the 5%, 1%, 5%, and 5%-level, respectively). Thus, inter-country effects are stronger
for industries in which there are cross-border mergers.
Non-European acquirer
For the four-year period and the year 1999, all Non-EU
coefficients are negative, as hypothesized, with the exception of the 20-day event window
in the period of 1996 – 1999. None the coefficients display statistical significance.
Transaction value
The variable Vol displayed significant positive correlations for the
event day t0 (at the 1%-level) and the 5-day as well as the 10-day event window (at the
5%-level) for the four-year observation period. In 1999, the 4 coefficients were positive
but not significant. Signaling is therefore be more intense in case of high value mergers
and acquisitions. However, markets seem to have become accustomed to vastly
increasing M&A transaction values.
Tax treatment differential
In 1999, the tax variable coefficient is significantly
negative for t0 (at the 5%-level) and the 10-day and 20-day event window (at the 10%level). Although the sign of all 5 coefficients is negative, as hypothesized, it wass not
significanct over the four-year period. Thus, the tax environment of the acquirer seems to
have little impact on the magnitude of the M&A-induced inter-country effect.
Relative European industry performance
For the period of 1996 – 1999, the
coefficient associated with the variable Perf was positive for the event day and positive
and significant at the 1%-level for all four subsequent event windows. In 1999, all four
event windows displayed significant positive coefficents (at the 10%, 10%, 5%, and 5%level, respectively). The M&A announcement-induced valuation effects on European
rivals are more pronounced if the target company belong to an industry segment that had
already been outperforming the market.
EMU membership
The single currency dummy variable Euro, indicates whether the
countries in which a target and the acquirer are located are participating in the first round
of the EMU. In the first two years of the observation period, it was not clear which
countries would join the Euro, under what conditions, and when they would do so. Even
in 1998, financial assets were still priced in local currencies. With the exception of the
20-day event window, the Euro variable is not significant for the period of 1996-1999.
With the introduction of the single currency on January 1, 1999, national stock
exchanges of EMU-participating countries converted all stock prices into Euro, along
with the corresponding indices, and prices of financial derivatives. This had a clear effect
on the inter-country valuation of M&A announcements in 1999. The coefficient of the
Euro variable is significantly positive (at the 5%-level) for the event day t0, the 5-day and
10-day event window, and at the 10%-level for the 20-day and the 30-day event window.
The hypothesis that the introduction of the Euro facilitates the transmission of private
information across geographic markets boundaries, is clearly supported by these findings.
6. SUMMARY
Our objective is to measure intra-industry and inter-country effects of merger
announcements in Europe, and to explain the variation in these effects among mergers.
We find favorable intra-industry effects in response to European mergers. We find
evidence that the magnitude of the intra-industry valuation effect is more favorable when
the merger involves large targets.
We also find strong evidence of inter-country effects. This is a major finding, as it
suggests that signals about rivals can transcend national borders. M&A-induced intercountry effects are stronger if target and acquirer are not located in the same country.
Companies that share the target’s operational and financial attributes face an increasing
likelihood to be taken over, as the number of potential domestic acquirers is raised by a
much larger number of potential non-domestic acquirers. There is no evidence that
markets distinguish between EU and non-EU acquirer. Over the four-year period, intercountry effects are larger for larger mergers.
Since the introduction of the single currency on January 1, 1999, inter-country
effects have become more pronounced. The Euro coefficient was significantly positive
for the event day t0 and all four event windows assessed in response to M&As during
1999. This offers strong evidence that the more transparent markets caused by the Euro
facilitate the transmission of private information across country boundaries in Europe.
References
Akhigbe, A. and Jeff Madura. 1999a. Intra-Industry Signals Embedded in Bank
Acquisition Announcements, Journal of Banking and Finance 23: 1637 - 1655.
Akhigbe, A. and Jeff Madura. 1999b. The Industry Effects Regarding the Probability of
Takeovers, The Financial Review 34: 1 - 18.
Berger, P. and E. Ofek. 1995. Diversification’s Effect on Firm Value, Journal of
Financial Economics 37: 39-65.
Berkovitch, E. and N. Khanna. 1990. How Target Shareholders Benefit From ValueReducing Defense Strategies in Takeovers, Journal of Finance 45: 137 - 156.
Cheng, L. and J. McDonald. 1996. Industry Structure and Ripple Effects of Bankruptcy
Announcements, Financial Review 31: 783 - 807.
Dennis, D. and J. McConnell. 1986. Corporate Mergers and Security Returns, Journal
of Financial Economics 16: 143 - 187.
Dodd, P. 1980. Merger Proposals, Management Discretion, and Stockholder Wealth,
Journal of Financial Economics 8: 105 - 137.
Eckbo, B. 1983. Horizontal Mergers, Collusion and Stockholder Wealth, Journal of
Financial Economics 11: 241 - 273.
Grullon, G., R. Michaely, and I. Swary. 1997. Capital Adequacy, Bank Mergers, and the
Medium of Payment, Journal of Business Finance and Accounting 24: 97 - 124.
Houston, J. and M. Ryngaert. 1997. Equity Issuance and Adverse Selection: A Direct
Test Using Conditional Stock Offers, Journal of Finance 52: 197 - 219.
Kaplan, S. and M. Weisbach. 1992. The Success of Acquisitions: Evidence from
Divestitures, Journal of Finance 47: 107-38.
Kennedy, P. 1986. Interpreting dummy variables, Review of Economics and Statistics
68:
174 - 175.
Lang, L. and R. Stulz. 1992. Contagion and Competitive Intra-Industry Effects of
Bankruptcy Announcements: An Empirical Analysis, Journal of Financial
Economics 32: 45 - 60.
Lee, C. and F. Choi. 1992. Effects of Alternative Goodwill Treatments on Merger
Premia: Further Empirical Evidence. Journal of International Financial
Management and Accounting, (Autumn): 220 - 236.
Lewellen, W. 1971. A Pure Financial Rationale for the Conglomerate Merger, Journal
of
Finance: 521 - 535.
Mead, W.J. 1969. Instantaneous Merger Profit as a Conglomerate Merger Motive.
Western Economic Journal (December): 295 - 306.
Mikkelson, W. and M. Partch. 1988. Withdrawn Securities Offerings, Journal of
Financial and Quantitative Analysis: 119 - 133.
Mitchell, M. and J. Mulherin. 1996. The Impact of Industry Shocks on Takeover and
Restructuring Activity, Journal of Financial Economics 41: 193 - 229.
Pyo, Y. and S. Lustgarten. 1990. Differential Intra-Industry Information Transfer
Associated with Management Earnings Forecasts, Journal of Accounting and
Economics 13: 365 - 379.
Scanlon, K., J. Trifts, and R. Pettway. 1989. Impact of Relative Size and Industrial
Relatedness on Returns to Shareholders of Acquiring Firms, Journal of Financial
Research 12: 103 - 112.
Schwert, W. 1996. Markup Pricing in Mergers and Acquisitions, Journal of Financial
Economics 41: 153 - 192.
Slovin, M., M. Sushka, and Y. Bendeck. 1991. The Intra-Industry Effects of GoingPrivate Transactions, Journal of Finance 46: 1537 - 1550.
Slovin, M., M. Sushka, and J. Polonchek. 1992. Informational Externalities of Seasoned
Equity Issues: Differences Between Banks and Industrial Firms, Journal of
Financial Economics 32: 87 - 101.
Suits, D. 1984. Dummy variables: Mechanics vs interpretation, Review of Economics
and
Statistics 66: 177 - 180.
Sun, H. and A. Yang. 1998. The Intra-Industry Information Effect of Downsizing
Announcements, American Business Review 16: 68 - 76.