Does Corporate International Diversification Destroy Value?

Evidence from Cross-Border Mergers and Acquisitions

Marcelo B. Dos Santos, Vihang Errunza and Darius Miller*

This Draft: July 30, 2003

ABSTRACT

This paper investigates the valuation effects of corporate international

diversification by examining cross-border mergers and acquisitions of U.S.
acquirers over the period 1990-1999. We find that, on average, acquisitions
of “fairly valued” foreign business units do not lead to value discounts.
Consistent with industrial diversification discount literature, unrelated cross-
border acquisitions result in a significant diversification discount of about 24
percent after accounting for valuation of foreign targets. Furthermore,
significant wealth gains accrue to foreign target shareholders regardless of the
type of acquisition. Overall, our results suggest that international
diversification does not destroy value while industrial diversification leads to
discounts even after controlling for the pre-acquisition value of the target.

Keywords: Corporate international diversification; Mergers and acquisitions;

Diversification discount.
JEL Classification: F30, G30 and G34

*
Dos Santos and Errunza are from McGill University and Miller is from Indiana University. Contact
author, Dos Santos, Faculty of Management, McGill University, 1001 Sherbrooke Street West, Montréal,
QC H3A 1G5, Canada. Phone: (514) 398-4000 ext. 00820, Fax: (514) 398-3876, or E-mail:
marcelo.dos_santos@mail.mcgill.ca. We thank Kris Jacobs and Sergei Sarkissian for many helpful
comments. Errunza acknowledges financial support from the Bank of Montreal Chair at McGill University
and SSHRC. Miller acknowledges financial support from a DailmerChrysler Faculty Fellowship.
1. INTRODUCTION

Over the past two decades, the integration of global financial and product markets has
been accompanied by increases in the number and fraction of firms that operate in
international markets, and to a large extent cross-border mergers and acquisitions
(M&As) have been an important driving force to propel such a globalization1. Despite its
prevalence, research on the valuation effects of corporate international diversification has
been relatively sparse and has yielded mixed results. Although earlier studies report
evidence of a significantly positive relation between internationalization and firm value
[see, for example, Errunza and Senbet (1981, 1984), Kim and Lyn (1986), and Morck and
Yeung (1991)], recent empirical evidence is mixed. For example, Bodnar, Tang and
Weintrop (1999) corroborate the prior evidence on the positive effects of corporate
international diversification on firm value, whereas Christophe (1997) and Denis, Denis
and Yost (2002) find evidence that international operations are associated with value
destruction.
To shed further light on this important issue, we examine a sample of U.S. firms
that expand globally via cross-border M&As. There are at least four reasons for focusing
on cross-border M&As. First, many divisions of multinational corporations (MNCs) arise
via cross-border takeovers. Second, it will allow us to observe the firm’s diversification
status resulting directly from the act of adding new business units rather than subjective
segment reporting by the firm’s managers2. Third, we will be able to measure the market
value of the foreign target firms as well as compare their underlying characteristics with
those of U.S. benchmarks3. Since a number of recent studies have argued that the
diversification discount can be a biased result if there are significant differences between
divisions of conglomerates and the stand-alone firms to which they are benchmarked, it is

1
According to the Mergers and Acquisitions Annual Almanac (1992-2000), the number of cross-border
deals in the world increased from 7,096 in 1991 to 12,899 transactions in 1999. In terms of U.S. dollar
value, the volume of international deals was equivalent to about $304 billion in 1991, while it totaled
$1.376 trillion in 1999. In particular, the number of U.S. acquisitions of non-U.S. companies during the
1990s increased sharply as well, from 532 in 1991 to 1,034 transactions in 1999. In terms of dollar value,
the volume of such acquisitions was about $57 billion in 1991, while it totaled $247 billion in 1999.
2
For example, Hayes and Lundholm (1996) argue that segment reporting, as an accounting event, is often
subject to strategic managerial motives.
3
Differences in firm characteristics are likely to be more pronounced since target operations depend on
different trade policies, products and factor markets, corporate governance systems, and imperfections and
informational asymmetries of foreign capital markets.

1
important to gauge the extent to which foreign targets might differ from the domestic
benchmarks4. Fourth, our approach will build on the work of Graham, Lemmon and Wolf
(2002) to determine how much of the post-merger change in the excess value of the
multinational acquirer can be traced directly to the valuation status of the target firm prior
to the takeover event. This is particularly important since recent work of Moeller and
Schlingemann (2002) suggests that from the perspective of U.S. acquirers, cross-border
acquisitions differ from domestic transactions.
We study 136 cross-border M&As involving U.S. acquirers of foreign target firms
over the period 1990-1999. We find no evidence of a significant decrease in excess
values of the U.S. acquiring firms in the two-year period surrounding the acquisition5.
The valuation analysis also indicates that, on average, the foreign targets are not
significantly valued at either a discount or a premium in their last year of operation as
stand-alone firms. Taken together, our results suggest that the act of cross-border
acquisition does not lead to any value destruction.
Furthermore, we find that merging firms in related cross-border M&As does not
destroy value, while U.S. acquirers involved in unrelated cross-border M&As experience
a significant post-merger change in their excess values of approximately – 24 percent.
This reduction in value occurs even after controlling for the relative value of the target
prior to the acquisition. This result suggests that unrelated cross-border acquisitions lead
to value destruction and is, therefore, consistent with the large evidence on “industrial
diversification discount”6. Finally, we also document significantly positive changes in

4
For example, Chevalier (2000) investigates how the hypothesis of inefficient cross-subsidization among
divisions of conglomerates could be related to the selection bias and finds that some of the investment
patterns commonly attributed to cross-subsidization within conglomerates may be explained by systematic
differences between the investment opportunities of conglomerates and those of stand-alone firms. Whited
(2001) addresses the measurement error in Tobin’s Q related to the hypothesis of inefficient cross-
subsidization among divisions of conglomerates. Graham, Lemmon and Wolf (2002) show that in U.S
M&As, acquisitions of already-discounted targets significantly reduce the average excess value of the
acquiring firms, even if the act of combining firms does not necessarily lead to any value destruction.
Villalonga (2002) also reports that conglomerates and stand-alone firms differ in multiple characteristics.
She uses propensity score matching to reduce selection bias and finds that the diversification discount
vanishes, or even turns into a significant premium.
5
We use the same valuation methodology as in Bodnar, Tang and Weintrop (1999) and Denis, Denis and
Yost (2002), which represents a variation of the industry-matched multiplier approach originally developed
by Berger and Ofek (1995).
6
For an extensive literature on industrial diversification discount, see Wernerfelt and Montgomery (1988),
Lang and Stulz (1994), Berger and Ofek (1995), Servaes (1996), Bodnar, Tang and Weintrop (1999),
Denis, Denis and Yost (2002), and Lins and Servaes (1999, 2002).

2
foreign target shareholder wealth, regardless of the type of acquisition under
consideration.
We also conduct robustness tests that control for endogeneity as in Campa and
Kedia (2002) by pooling our globally diversified firms with their stand-alone U.S.
benchmarks in a two-way fixed-effects framework, which includes both firm-specific and
year fixed effects7. The results of this analysis suggest that the introduction of firm fixed
effects does not support the hypothesis that valuation effects of corporate international
diversification can be explained by unobserved firm-specific characteristics.
In summary, we find that international diversification does not destroy value in
that acquisitions of “fairly valued” foreign firms, on average, do not lead to an
international diversification discount. In addition, our valuation results are consistent with
past evidence that MNCs tend to trade at a premium relative to industry-matched
uninational firms. However, we find that cross-border M&As that occur in unrelated
industries do indeed lead to value destruction, which is consistent with the existence of an
industrial diversification discount. Overall, our results suggest that the underlying
characteristics of the foreign acquired firms are important for valuing MNCs.
The remainder of this paper is organized as follows. Section 2 briefly surveys the
literature regarding the valuation effects of corporate international diversification. Section
3 describes the sample selection procedures, data sources and the valuation methodology.
Section 4 analyzes the valuation effects of cross-border M&As and then investigates the
issues of selection bias and endogeneity. Section 5 concludes the paper.

2. THEORY AND EVIDENCE ON INTERNATIONAL DIVERSIFICATION

2.1. Corporate International Diversification Motives

The internationalization theory has its roots in Caves (1971), who advocates that
substantial foreign direct investment (FDI) occurs mainly in industries characterized by

7
Campa and Kedia (2002) show that firm value and diversification discount are endogenously related.
Once the endogeneity is accounted for, the observed diversification discount is significantly reduced and in
some cases, it turns out to be a premium. Hyland and Diltz (2002) find that diversified firms were already
valued at a discount prior to diversifying. On the other hand, Lamont and Polk (2002) show that the
observed discount is not a consequence of endogeneity problems. They find that exogenous changes in

3
certain market structures such as product differentiation and oligopoly. The firm can
increase its market value by internalizing the market imperfections for valuable firm-
specific assets.
Errunza and Senbet (1981, 1984) advance theoretical explanations for the
valuation effects of corporate international diversification based on the idea of capital
market imperfections. Due to the presence of investment barriers that are costly and
whose costs are not uniform across firms and individual investors, the shares of MNCs
should trade, in equilibrium, at a premium because these firms provide investors with
indirect portfolio diversification services through their operations abroad. In this case,
MNCs act as a costly financial intermediary and investors recognize this by bidding up
their stock price vis-à-vis those of purely domestic firms. According to Senbet (1979) and
Errunza and Senbet (1981), imperfections characterized by differences in taxation across
countries may also explain value creation. Kogut and Kulatilaka (1994) explore the idea
that MNCs have the opportunity to exploit a variety of market conditions, especially due
to the increasing uncertainty in international markets, based on the functionality of their
production networks and the flexibility of their cost structure. MNCs possessing such
specific characteristics (i.e., flexibility options) should be more valuable than comparable
domestic firms. Thus, multinational corporate expansion should give rise to valuation
premium if the benefits outweigh the agency costs of international diversification8.

2.2. Empirical Evidence on the Valuation Effects of International Diversification

The early studies on the value of international diversification were based on either a risk-
adjusted performance analysis [Hughes, Logue and Sweeney (1975), and Brewer (1981)]
or international analogues of return-generating processes [Agmon and Lessard (1977),
and Jacquillat and Solnik (1978)]. Table 1 presents a comprehensive summary of the
empirical findings and methodological features of the main studies on the valuation
effects of corporate international diversification.

investment diversity are negatively related to changes in excess value, giving support to the causal effect of
industrial diversification.
8
Viewing MNCs as complex structures, the costs to corporate international diversification can also arise
from agency problems when managers have incentives to cause the firm to grow beyond its optimal size
[Jensen and Meckling (1976), Jensen (1986)].

4
 The empirical results of Errunza and Senbet (1981) show that the degree of
international involvement is significantly positively related to excess market value, and
such a relationship is even stronger during periods characterized by restrictions on capital
flows. Other studies also corroborate the evidence of positive effects of corporate
international diversification on firm value. For example, Errunza and Senbet (1984)
expand their previous valuation analysis by controlling for firm size and introducing
different measures of the degree of corporate international involvement, and find
evidence of a positive relation between firm value and corporate international
diversification. Kim and Lyn (1986), and Morck and Yeung (1991) find support for the
positive valuation effects of international diversification under the hypothesis that MNCs
can internalize the market imperfections for their intangible assets abroad. On the other
hand, Christophe (1997) shows that MNCs have significantly lower Q ratios than
domestic firms do, and the largest discounts occur during the first half of the 1980s when
the U.S. dollar is strongest. However, he does not control the MNC subsample for
industrial diversification, nor does he compare the Q ratios of MNCs with those of
(specialized) domestic firms matched in the same industry.
Two recent studies examine the independent valuation effects of industrial and
international diversification in a framework that simultaneously controls for both forms
of corporate diversification in order to avoid the correlated omitted variable problem. The
omitted variable problem is a natural issue of concern due to the fact that many
conglomerates are also “globally” diversified firms. By segmenting the data into four
subsamples of firms (single-segment domestic firms, multi-segment domestic firms,
single-segment MNCs, and multi-segment MNCs), in which the single-segment domestic
firms represent the benchmark group used to calibrate the valuation effects of the three
other subsamples, Bodnar, Tang and Weintrop (1999) find that the average
diversification premium for MNCs is statistically significant and equivalent to 2.7
percent9. Denis, Denis and Yost (2002) develop an empirical valuation analysis similar to

9
Bodnar, Tang and Weintrop (1999) do not provide separate estimates of the valuation effects for the two
categories of MNCs, that is, single-segment MNCs and multi-segment MNCs. Their multivariate valuation
model incorporates only an indicator variable for firms that are internationally diversified, regardless of
whether the firms are specialized or diversified. Thus, the estimated diversification premium of 2.7 percent
represents an average value for the two subsamples of MNCs.

5
that in Bodnar, Tang and Weintrop (1999) and find evidence of significant diversification
discounts for both specialized MNCs (18 percent) and diversified MNCs (32 percent).
In summary, the evidence suggests a significant positive relationship between
multinationality and firm value, that is, MNCs trade at an international diversification
premium. Although more recent results based on industry-matched benchmarks are
inconclusive with respect to change in premia from international diversification, these
studies [e.g., Bodnar, Tang and Weintrop (1999) and Denis, Denis and Yost (2002)] have
not explicitly considered, for example, how selection bias could affect their valuation
results. Therefore, in this paper, we reexamine the value of international diversification
using M&A activity, a setting where Graham, Lemmon and Wolf (2002) argue is natural
to study these issues. Furthermore, we apply the methodology of Campa and Kedia
(2002) to address any potential endogeneity in our findings. Overall, our analysis takes
into account selection bias, the underlying characteristics of the foreign acquired firms in
valuing MNCs, and the role of endogeneity in international diversification decision.

3. SAMPLE SELECTION AND VALUATION METHODOLOGY

3.1. Sample Selection and Data Sources

The sample selection procedure begins by identifying an initial sample of 11,392 cross-
border transactions involving U.S. acquirers of foreign target firms registered on the
Securities Data Corporation (SDC) Mergers and Acquisitions database over the period
1990-1999. The data sources and the series of data filters used to refine the sample are
described more fully in Table 2. We eliminate uncompleted transactions and completed
transactions for which the U.S. bidder did not acquire a majority ownership stake in the
foreign target. As in Graham, Lemmon and Wolf (2002), we eliminate both U.S.
acquirers and foreign target firms reporting main operations in the financial services
industry (SIC codes between 6000 and 6999). We also disregard completed transactions
representing divestitures of private business units or certain assets, concessions, joint
ventures and management buyouts, mainly because such transactions may not
characterize a merging event, or they commonly involve privately held firms/business

6
units for which no market data is available. It follows that the final SDC sample of
merging firms consists of 1,363 cross-border M&A transactions.
We collect both market and accounting data for the pairs of merging firms10. It is
also important to determine the diversification status of the merging firms in the year
prior to the acquisition as well as the diversification status of the U.S. acquirers in the
year following the acquisition. We search for data on share prices for the foreign acquired
firms from the Datastream database. A total of 222 foreign acquired targets have
historical stock prices available in the Datastream database. We then collect pre-merger
accounting data to calculate the valuation measures for these 222 publicly traded foreign
target firms, which reduces the sample to 150 foreign targets that have data available in
the Worldscope database. We identify a foreign target’s industrial diversification status
from the segmental data provided by the Worldscope database. Most of the foreign target
firms in our sample happen to be single-segment firms. A total of only four foreign target
firms report business operations in different two-digit SIC codes.
We collect the data for the 150 U.S. acquiring counterparts from the Compustat
database. We begin by identifying the U.S. acquirers’ industrial diversification status
from the Compustat Industrial Segment (CIS) database. Under the disclosure
requirements imposed by the Statement of Financial Accounting Standards Board
(FASB) No. 14 and the Security Exchange Commission (SEC) Regulation S-K, firms
must report segment information for fiscal years ending after December 15, 1977 when
sales, identifiable assets, or operating income relative to their business segments exceed
10 percent of consolidated totals. In this case, firms are required to provide segment
information on five accounting elements such as net sales, identifiable assets, operating
income, capital expenditures, and depreciation. We follow previous studies on the
valuation effects of corporate diversification and disregard firms whose total sales are not
completely allocated among their reported business segments (i.e., the sum of the
segment sales deviates from firm’s total sales by more than one percent). Unlike sales,
assets are not usually completely allocated among the firm’s business segments. In this
case, we prorate the unallocated assets among the reported segments based on the relative

10
We require data on sales, total assets, share price, number of shares outstanding at the end of the fiscal
year, book value of total debt, and liquidating value of preferred stocks available from Compustat (for U.S.
acquirers) and Worldscope and Datastream (for foreign targets) databases.

7
asset size of the segments. It follows that two U.S. firms are disregarded because they
report business segments with SIC codes within the 6000-6999 range. In addition, twelve
U.S. acquiring firms are eliminated because they do not have either share or accounting
data available in Compustat. Therefore, our final sample of paired merging firms consists
of 136 cross-border M&As.
Table 3 summarizes the sample of 136 cross-border M&As according to the
origin of the foreign targets and the main characteristics of the acquisition deals. Panel A
presents the sample distribution across the several countries of the foreign acquired firms
and, not surprisingly, it turns out that the sample is largely dominated by target firms
acquired in the United Kingdom (about 34 percent) and Canada (30 percent). We find
that, consistent with the overall trend of corporate internationalization, the number of
firms in our sample increases thru time. Cross-border acquisitions are fairly large deals,
with a mean (median) deal value of $741 ($213) million, and U.S. firms acquired a
substantial majority of ownership stake in the foreign targets, with a mean (median)
acquisition of 94 percent (100 percent) of the shares of the foreign target (Panel B). Panel
C presents the sample distribution of the deal characteristics in terms of the type of
acquisition, medium of payment, accounting method, and bidder attitude. As expected, a
substantial proportion of the acquisition deals is characterized by tender offers (74
percent) and the use of cash (62 percent), given the fact that several U.S. firms primarily
use cash bids in cross-border takeovers because they are not cross-listed in foreign
markets. On the other hand, the use of stock swaps is largely associated with cross-border
mergers. Though not reported in the table, 33 out of a total of 35 cross-border mergers
report stock swaps as the main method of payment used. Finally, the vast majority of the
cross-border deals treat the acquired target as a purchase (about 87 percent), rather than a
pooling of interests (13 percent), while approximately 90 percent of the transactions
represent friendly takeovers.

3.2. Sample Segmentation and Diversification Status

The overall sample of 136 cross-border M&As is a heterogeneous group with respect to a
few features such as the degree of business relatedness between the U.S. acquiring firms
and the foreign targets, the degree of international involvement of the U.S. acquiring

8
firms, and target firms’ corporate governance systems. This allows us to segment the
sample firms according to two different criteria that may help contrast the subsample
conclusions with those drawn from the overall (heterogeneous) sample.
First, we consider the business relatedness between U.S. acquiring and foreign
target firms at the time the takeover takes place. If the U.S. acquirer and the foreign target
firm report operations with the same two-digit SIC code, we classify the acquisition as
related, and unrelated otherwise. This type of relatedness classification is also used by
Berger and Ofek (1995), Servaes (1996), Chevalier (2000) and Lamont and Polk (2002).
Based on this classification scheme, we can assess how these two different forms of
corporate international diversification affect the value of the acquiring firms as well as
how the selection bias is specifically related to the valuation of specialized and
diversified MNCs. In our sample, 88 out of a total of 136 cross-border M&As are
classified as related, whereas 48 are classified as unrelated cross-border M&As.
Second, we group the sample firms according to whether the cross-border
acquisition represents the “premiere” way through which the U.S. acquiring firm
establishes operations abroad. Going abroad for the first time, in terms of establishing
production operations in a foreign location for the first time, might induce significant
changes in the firm’s operations status quo and, accordingly, the value of the acquiring
firm might sharply change as well. For the U.S. acquiring firms with established
operations abroad, we expect a less drastic change in firm value as compared to the
situation in which a firm goes abroad for the first time. Hence, if the U.S. acquiring firm
has no operations abroad at the time the takeover occurs, we then classify the acquisition
as premiere cross-border operation, and non-premiere otherwise. We identify the
international involvement status of the U.S. acquiring firms from the Compustat
Geographic Segment (CGS) database, by checking the CGS code for the firms two years
backwards until the effective year of the acquisition. We find that 31 out of a total of 136
cross-border M&As are classified as premiere cross-border operations, whereas 105 are
classified as non-premiere operations.

9
3.3. Excess Value Measures

We use the excess value measure as defined in Bodnar, Tang and Weintrop (1999) and
Denis, Denis and Yost (2002), which represents a variation of the industry-matched
multiplier approach originally developed by Berger and Ofek (1995). The excess value
(EV) compares a firm’s market value (the market value of common equity plus the book
value of total debt plus the liquidating value of preferred stock) to its imputed value (IV).
For each business segment of a diversified firm, we search for a group of matching
single-segment domestic firms with the same two-digit SIC code, and then calculate the
median ratio of the market value to sales (assets) for this matching group of specialized
domestic firms, that is, the industry-matched sales (asset) multiplier. The imputed value
for each business segment of a diversified firm then equals the business segment’s
reported sales (assets) multiplied by the industry-matched sales (asset) multiplier. The
excess value measure for a given firm i at time t can be represented as follows,

æ MVi, t ö
EVi, t = Ln ç ÷, [1.1]
ç IVi , t ÷
è ø
where
n

j =1
n
( { })
IVi , t = å wi , j , tθ j , t = å wi , j , t Median θ1 ,θ 2 ,...,θ K j , t ,
j =1
[1.2]

where MVi, t stands for the market value of firm i measured at time t, IVi, t represents the

imputed value of firm i at time t, wi, j, t stands for the reported amount of sales (assets) of

the j-th business segment of firm i at time t, and θ j, t indicates the industry-matched sales

(asset) multiplier calculated from K j , t stand-alone domestic firms operating in the business

segment j at time t such that K j , t ≥ 5 . We use the industry definition based on the two-

digit SIC code grouping that includes at least five stand-alone firms so as to ensure that
the industry-matched multipliers are representative. Moreover, in calculating the
industry-matched multipliers, we disregard firm-years with total sales less than $20
million and whose Compustat incorporation code is greater than zero (FINC > 0), that is,
non-U.S. incorporated firms.

10
 As shown in Graham, Lemmon and Wolf (2002), the choice of the accounting
method in M&As (e.g., the choice of purchase versus pooling accounting) can lead to
potential valuation problems when one uses industry-matched asset multipliers. More
specifically, since all identifiable assets acquired and liabilities assumed in a business
combination using the purchase accounting should be assigned a portion of the cost of the
acquired target, normally equal to their face values at the date of acquisition, the book
value of total assets of the newly combined firm will reflect the purchase price of the
target. In other words, the book value of total assets is “marked-to-market”, which
induces a negative bias into Equation [1.1] (i.e., larger asset-based imputed values) when
U.S. acquirers are valued in the year following the acquisition. This accounting effect is
particularly important in this paper because, as reported in Table 3, approximately 87
percent of the cross-border M&As in our sample use purchase accounting. Therefore, we
will focus on sales-based figures when analyzing our results, even though we report the
results for the asset-based calculations as well.

4. VALUATION ANALYSIS

4.1. Comparing Firm-Specific Characteristics

We first consider the differences in underlying characteristics among U.S. acquirers,

foreign targets and U.S. benchmarks11. Much of the literature argues that some firm-
specific characteristics might affect firm value. Caves (1971) develops a theoretical
argument emphasizing that corporate value can increase when firms are able to
internalize market imperfections for their intangible assets abroad. Kim and Lyn (1986),
Morck and Yeung (1991), Bodnar, Tang and Weintrop (1999) and Denis, Denis and Yost
(2002) provide empirical support for research and development (R&D) and advertising
expenditures as proxies for firm-specific (intangible) assets. Capital structure (debt) is
also used to control for the valuation effects that may result from financial leverage.
From the industrial diversification literature, Lang and Stulz (1994), Berger and Ofek
(1995), and Servaes (1996) show the importance of controlling for firm size. Other

11
Throughout this paper, we assess the statistical significance of the difference in mean values based on the
parametric t-statistics, while the significance of the difference in median values is assessed using the
nonparametric Wilcoxon rank sum test statistics.

11
factors such as growth opportunities (investment) and profitability are also included as
additional corporate control variables.
Table 4 reports the descriptive statistics of the underlying characteristics for our
three sets of firms (U.S. acquirers, foreign targets and U.S. benchmarks)12. Panel A
shows that in the first year following the acquisition event (t = +1), U.S. acquirers of
foreign target firms experience an overall, significant increase in their mean and median
sizes as proxied by total capital, sales, or assets. In general, the q-ratio proxies for a
firm’s incentive to invest in new assets, and it will supposedly do so as long as q > 0 (i.e.,
as long as the firm is still overvalued). We can observe that, even though the U.S.
acquirers’ q-ratios are all positive in the year prior to the acquisition (t = −1), the mean
(median) q-ratio based on sales displays an insignificant decline of 7.04 percent (9.92
percent) following the acquisition. On the other hand, the mean (median) q-ratio based on
assets significantly decreases by 15.64 percent (16.64 percent). However, as previously
reported in Table 3, since the vast majority of U.S. acquirers use purchase as the
accounting method, in which the book value of total assets is usually increased to reflect
the purchase price of the target, a negative bias is likely to affect the asset-based q-ratio at
t = +1. It is important to note that these q-ratios are not adjusted for industry effects since
the market value of the firm is not being compared to its imputed value regarding each of
its business segments. As for the remaining underlying characteristics of U.S. acquirers,
we can also observe a significant mean (median) rise in financial leverage of about 5
percent (5 percent), while profitability, firm-specific assets and investment levels display
no significant change over the two-year period surrounding the acquisition.
In Panel B, we compare U.S. firms to their foreign acquired counterparts at t = −1.
Overall, we find strong evidence that U.S. acquiring firms are significantly bigger than
the foreign targets across all proxies used and at any conventional levels of statistical

12
Total capital is the sum of market value of common equity, book value of total debt, and the liquidating
value of preferred stock. The q-ratio_Sales (q-ratio_Asset) is computed as the natural logarithm of the ratio
of total capital to sales (total assets). Leverage is calculated as the ratio of book value of total debt to total
assets. EBIT/Sales stands for earnings before interest and tax expenses normalized by sales. R&D/Sales
represents research and development expenditures divided by sales, and CAPEX/Sales is the ratio of capital
expenditures to sales. The q-ratios for the foreign target firms are calculated using the market value of
common equity based on the last stock price available prior to delisting. In addition, the market and
accounting data for the foreign targets have been converted to U.S. dollar values using the corresponding
National Exchange Rate Quotes provided by Datastream.

12
significance. For instance, the mean (median) sales for foreign targets correspond to
approximately 10 percent (12 percent) of the U.S. acquirers’ mean (median) sales before
they combine their business operations, and such results are indicative of the relatively
small size of the acquired foreign targets in our sample. U.S. acquirers are significantly
overvalued relative to the foreign target firms; their mean (median) sales-based q-ratio is
about 30 percent (35 percent) larger than that of the foreign targets at the one percent
level. In addition, U.S. acquiring firms are more financially levered as well as more
profitable than the foreign targets, but there is no significant evidence that the merging
firms differ in terms of firm-specific advantages or growth opportunities.
Panel C reports contemporaneous differences in mean and median values between
136 foreign target firm-years and 5,635 U.S. stand-alone firm-years specifically used to
value the foreign target firms, without matching these two groups of firms to their
particular industries. Due to skewness in the distributions, we emphasize medians rather
than means. Overall, foreign target firms are significantly bigger than U.S. stand-alone
firms at the one percent level. In addition, there is weak evidence that foreign targets may
have less firm-specific advantages and higher growth opportunities than U.S. stand-alone
firms. In Panel D, we calculate the contemporaneous industry-adjusted differences in
firm-specific characteristics between foreign targets and U.S. specialized firms by
subtracting the median value for a group of U.S. specialized firms from a foreign target’s
actual value when they share the same two-digit SIC code. After adjusting for industry-
median effects, we still find evidence that foreign targets are significantly bigger than
U.S. stand-alone firms. In addition, besides having lower median R&D expenses, foreign
target firms do not seem to differ from U.S. benchmarks in their other underlying
characteristics. Particularly interesting is the evidence that there is no significant median
difference in the q-ratios between foreign target and U.S. benchmark firms. Overall, the
simple univariate analysis suggests that foreign target firms are relatively small in size,
but otherwise have similar financial characteristics as their U.S. benchmark firms. We
next examine the valuation effects of cross-border M&As over the two-year period
surrounding the acquisition.

13
4.2. Examining the Valuation Effects of Cross-Border M&As

We calculate excess values of U.S. acquirers in the year prior to the acquisition (t = −1)
as well as excess values of the combined firms (U.S acquirer + foreign target) in the year
following the acquisition (t = +1), using industry-matched stand-alone U.S. firms as
benchmarks. We disregard the effective year of the acquisition (t = 0) because excess
value measures based on accounting data as of t = 0 may not properly represent year-long
value performance for many of our sample firms after they combine their operations.
Table 5 reports excess values for U.S. acquirers in the year prior to the acquisition
(EV-1), in the first year following the acquisition (EV+1), as well as the actual change in
excess values from t = −1 to t = +1, which is defined as follows,

∆EV+1 = EV+1 − EV−1 [2]

In Panel A, we present the valuation measures for the full sample of 136 U.S.
acquiring firms. In the year prior to the acquisition, U.S. acquiring firms are valued at a
mean (median) premium of 31.86 percent (33.02 percent) relative to the stand-alone U.S.
firms matched in the same two-digit SIC code industries, and these excess values are
significantly different from zero at the one percent level. At t = +1, the U.S. acquirers are
still significantly overvalued relative to the industry-matched benchmarks, trading at a
mean (median) premium of 25.26 percent (24.97 percent). Thus, consistent with past
results, MNCs trade at a premium relative to uninational firms. Further, the cross-border
acquisitions are associated with an insignificant mean (median) actual change in excess
values of −6.59 percent (−7.99 percent). Hence, cross-border acquisitions do not lead to
value destruction.
In Panels B through E, we calculate the actual change in excess values by
segmenting the sample according to the business relatedness between the merging firms
(Panels B and C) as well as the international involvement status of the U.S. acquiring
firms (Panels D and E). Panel B tells a largely similar story as the evidence for the full
sample reported in Panel A. U.S. acquirers involved in related cross-border acquisitions
are significantly overvalued relative to the industry-matched benchmarks in both years
surrounding the acquisition, but the mean (median) actual change in excess values of 2.80

14
percent (−0.72 percent) is not statistically different from zero. On the other hand,
unrelated cross-border M&As (Panel C) are associated with a large mean (median)
decline in excess values of 23.82 percent (21.53 percent), and such a mean (median)
discount is statistically significant at the one (ten) percent level. Taken together, these
results are consistent with the evidence on “industrial diversification discount”. However,
these results contrast with those reported by Graham, Lemmon and Wolf (2002) for
domestic M&As. Although they also find evidence that U.S. acquiring firms always trade
at a significant premium in both years surrounding the effective year of the acquisition,
they document a significant decline in excess values across the full sample of domestic
acquisitions as well as the related and unrelated acquisition subsamples. On the other
hand, Moeller and Schlingemann (2002) compare announcement wealth effects for U.S.
bidders across domestic and cross-border M&As and report that unrelated cross-border
acquisitions have lower announcement returns and operating performance than any other
form of acquisition. For cross-border acquisitions in which 105 U.S. acquiring firms have
already established operations abroad (Panel D), the mean (median) actual change in
excess values is −3.83 percent (−5.48 percent), which is not statistically different from
zero. However, for a much smaller subsample of 31 U.S. acquiring firms that are
establishing business operations abroad for the first time through cross-border M&As,
there is weak evidence of economic loss from t = −1 to t = +1, even though the mean
(median) actual change in excess values of −16 percent (−20.93 percent) is not
statistically significant.
In summary, there is strong evidence that MNCs tend to trade at a significant
premium relative to industry-matched uninational firms in the two-year period
surrounding the acquisition. Further, we find no evidence of a significant decline in
excess values in our full sample of cross-border M&As. Related cross-border acquisitions
are not associated with a reduction in firm value, however, unrelated acquisitions lead to
significant discounts consistent with the effect of industrial diversification. Finally, going
abroad for the first time through the acquisition of foreign firms seems to result in a
greater economic loss than that for globally established acquirers, but the small size of
our subsample does not allow a meaningful statistical test. We next examine the valuation
status of the foreign target firms in their last year of operations as stand-alone firms.

15
4.3. Examining the Valuation Status of the Foreign Target Firms

We follow Graham, Lemmon and Wolf (2002) and calculate two measures of excess
value for the foreign target firms in their last year of operations as stand-alone firms. The
first measure stands for the pre-announcement excess value ( EV−A1 ) and is calculated using
the market value of common equity observed one month before the announcement of the
acquisition. The second measure represents the pre-effective excess value ( EV−E1 ) and is
computed using the market value of common equity based on the last stock price
available prior to the date on which the target firm is delisted. The main difference
between these two measures is that the former does not take into account the valuation
effects due to the acquisition announcement. Thus, the difference between these two
measures of excess value can reflect the change in foreign target shareholders’ wealth
( ∆EV−1 ) associated with the cross-border acquisition such that,

∆EV−1 = EV−E1 − EV−A1 [3]

In Table 6, we report the two excess value measures as defined above and the
change in target shareholders’ wealth as given by Equation [3]. In addition to looking at
the valuation status of the foreign targets prior to the acquisition, it is also useful to have
an idea about the target size relative to the overall combination of the merging firms, and
then relate both information to the actual change in the excess values of U.S. acquiring
firms. We define the relative size of the acquisition as the ratio of the foreign target’s
sales (assets) to the combined sales (assets) of both foreign target and U.S. acquirer in the
year prior to the acquisition.
For the full sample (Panel A), foreign target firms are invariably valued at a
significant average discount relative to the U.S. stand-alone firms matched in the same
two-digit SIC code industries 30 days prior to the announcement of the acquisition. The
mean (median) pre-announcement excess values based on both sales and asset multipliers
are −19.77 percent (−19.21 percent) and −13.31 percent (−12.38 percent), respectively,
and all of these value discounts are significantly different from zero at the one percent
level. However, once the excess value measures are adjusted for the announcement of the

16
acquisition, all of these discounts tend to vanish. In fact, there is weak evidence of an
average economic increase in excess values; the mean (median) sales-based pre-effective
excess value is 4.15 percent (4.28 percent), even though it is not significantly different
from zero. The corresponding mean (median) asset-based gain in value is 10.61 percent
(11.58 percent), which is statistically significant at the ten (five) percent level. This
sizeable turnaround in the excess values actually reflects the wealth gains accruing to the
foreign target shareholders, and such wealth gains are significantly different from zero
(always above 20 percent) at the one percent level. There is also evidence supporting the
relatively small size of the cross-border acquisitions. The mean (median) ratio of the
target’s sales to the merging firms’ combined sales is 19.10 percent (14.70 percent), and
even lower proportions are documented for the asset-based figures. Overall, the results in
Panel A suggest that target firms trade at a discount prior to the acquisition
announcement, but when the announcement effect is incorporated into the stock price, are
“fairly valued”.
The results for the subsample of related cross-border acquisitions (Panel B) are
similar to those reported in Panel A. For both sales and asset multipliers, foreign targets
in related M&As trade at significant mean (median) discounts of 23.23 percent (23.70
percent) and 13.38 percent (13.61 percent), respectively, prior to the acquisition
announcement. The corresponding mean (median) sales-based pre-effective excess value
is also negative, −1.07 percent (−0.51 percent), but is not significantly different from
zero. The asset-based measures are all positive, 8.78 percent (8.38 percent), but such
value premia are not significant at any conventional level. Moreover, there is no evidence
of significant differences in terms of target shareholder wealth gains and relative size of
acquisitions as compared to the full sample. For the subsample of unrelated cross-border
acquisitions (Panel C), although all measures of pre-announcement excess value are
negative, they are not significantly different from zero. The mean (median) sales-based
pre-effective excess values are rather positive, 13.74 percent (11.75 percent), but also
statistically insignificant, while the corresponding asset-based figures are 13.98 percent
(17.62 percent), and they are significantly greater than zero at the ten (five) percent level.
Shareholder wealth gains are larger in unrelated acquisitions, given the mean (median)
gain of 27.17 percent (24.10 percent). The mean (median) sales-based relative size of

17
14.86 percent (8.75 percent) indicates that foreign targets in unrelated acquisitions are
smaller than foreign targets involved in related acquisitions.
Finally, Panels D and E present the results for the subsamples based on the
international involvement status of U.S. acquiring firms. Overall, there is no qualitative
change in the results for either subsample as compared mainly to those reported in Panels
A and B. In particular, given the mean (median) sales-based relative size of acquisitions,
15.51 percent (11.50 percent), U.S. firms with operations already established overseas
tend to acquire smaller target firms relative to U.S. acquiring firms establishing business
operations abroad for the first time. It is interesting to note that the later group of U.S.
firms acquires not only much larger foreign targets but also more deeply discounted
foreign targets, even though such discounts are statistically insignificant. Overall, the
results from Table 6 indicate that targets tend to trade at discounts prior to the M&A
announcement, but are “fairly valued” by the time the acquisition is consummated.

4.4. Examining the Relationship Between Excess Values of the Merging Firms

We next examine whether the post-merger change in excess values of U.S. acquirers is
related to the pre-merger valuation status of the newly acquired foreign target firms. We
first need to compute the projected excess value, as originally defined in Graham,
Lemmon and Wolf (2002), which represents the excess value measure the merging firms
would have if they combined their operations instantaneously in the year prior to the
acquisition event. The projected excess value ( EV+P1 ) can be calculated in the following
way,

1 + MV −1
æ MV−US FT
ö
EV+P1 = Ln çç ÷÷ [4]
è IV−1 + IV−1
US FT
ø

where MV−US1 and MV−FT

1 stand for the ex ante market values of the U.S. acquiring and the

foreign target firms at t = −1, respectively, while IV−US1 and IV−FT

1 are their corresponding

imputed values. MV−FT

1 is based on the pre-effective market value of common equity

because this quantity incorporates the valuation effects due to the announcement of the

18
acquisition. In other words, it takes into account the investors’ assessment of the value of
the target as one of the parts in the newly combined firm. Once we compute EV+P1 , we
then compare it to the excess value of the U.S. acquiring firm in the year prior to the
acquisition (EV-1) in order to define the projected change in excess value as follows,

∆EV+P1 = EV+P1 − EV−1 [5]

Based on the actual change in excess values ( ∆EV+1 ), as defined in equation [2],

and the projected change in excess values ( ∆EV+P1 ), as defined in Equation [5], we can
therefore measure the unexplained change in excess values of the U.S. acquiring firms by
the following,

∆EV+U1 = ∆EV+1 − ∆EV+P1 = EV+1 − EV+P1 [6]

The unexplained change in excess value ( ∆EV+U1 ) measures the additional value
gain or loss (or nothing) that occurs beyond the effect of adding overvalued or
undervalued (or “fairly valued”) target firms to the acquiring firms. In other words, it
indicates whether there are additional valuation effects related to the acquisition event
after accounting for the underlying characteristics of the acquired firms.
Table 7 shows the univariate results for Equation [6] as well as provides a
breakdown of its main components. For the full sample (Panel A), the mean (median)
actual change in excess values based on sales multipliers is −6.59 percent (−7.99
percent), and the corresponding projected change in excess value is −4.76 percent (−5.34
percent), but only the later mean change is significantly less than zero. It follows that the
mean (median) sales-based unexplained change in excess values is −1.84 percent (−2.48
percent), but neither difference is significantly different from zero. These findings
suggest that the act of international diversification does not destroy value, even after
accounting for the value of the target prior to the acquisition. That is, adding “fairly
valued” targets does not cause any impact on the excess values of U.S acquirers. Indeed,

19
adding “fairly valued” foreign targets explains a substantial fraction, about 72.08 percent
[1− (−0.0184/−0.0659)], of the mean actual change in excess values of the U.S. acquirers.
As for the subsample of related cross-border acquisitions (Panel B), the mean
(median) difference between the actual and projected changes in excess values based on
sales multipliers is 7.47 percent (6.14 percent), but neither difference is significantly
different from zero. As for the subsample of unrelated cross-border M&As (Panel C), the
mean (median) sales-based unexplained change in excess values is −18.89 percent
(−16.97 percent), and only the mean change is significantly less than zero, at the five
percent level. These value differences suggest that unrelated cross-border acquisitions
result in value loss beyond that which can be explained by the pre-merger valuation status
of the foreign target firms. For example, 79.30 percent (−0.1889/−0.2382) of the actual
change in excess values based on sales multipliers cannot be explained by simply
combining the value of the foreign target with that of the acquiring firm at t = −1. For the
subsample of U.S. acquiring firms with established operations abroad (Panel D), the
mean (median) unexplained change in excess values based on sales multipliers is
statistically insignificant −2.53 percent (−3.57 percent). As for the subsample of U.S.
acquiring firms establishing business operations abroad for the first time (Panel E),
acquisitions of discounted targets lead to a very small and insignificant mean (median)
unexplained change in excess values of 0.5 percent (−0.87 percent). In summary, the
mechanical valuation effect of combining the firms explains all merging cases except the
subsample of unrelated cross-border M&As. Therefore, our results suggest that global
M&As that result in industrial diversification do indeed destroy value, even after taking
into account the pre-acquisition value of the target.
Based on Equation [6], we extend our analysis on the relationship between the
excess values of the merging firms to a regression framework so as to determine how
much of the cross-sectional variation in the actual change ( ∆EV+1 ) can be explained by
the projected change in excess values ( ∆EV+P1 ) such that,

∆EV+1 = α + β ∆EV+P1 + γ I (⋅) + ε +1 [7]

20
 We first estimate a linear regression model for the full sample of merging firms in
which the projected change in excess values is the only predictor variable. We then
expand this model by alternately incorporating an indicator variable I (⋅) in order to infer
on the characteristics of the subsamples. An indicator variable for business relatedness of
the acquisition, I (Relatedness), equals one if the acquisition is classified as related, and
zero otherwise. In addition, an indicator variable for the international involvement status
of U.S. acquiring firms, I (Premiere), is set to equal one if the U.S. firm already has
business operations overseas prior to the acquisition, and zero otherwise. Under the null
hypothesis, we test whether α = 0, β = 1, and γ = 0. The intercept α captures the
unexplained change in excess values.
The results for the regression model above are presented in Table 8. We
emphasize mostly the sales-based results (Panel A) and report asset-based results for
completeness. Regression 1 shows that the estimated intercept of the regression
(−0.0292) is not significantly different from zero (P-value = 0.467). This result is
consistent with that reported in the first panel of Table 7, since it indicates that no
additional, significant value loss occurs after accounting for the valuation status of the
foreign targets. The estimated regression coefficient on the projected change in excess
values is 0.5071, which is reliably different from one (P-value = 0.009). In this case, the
projected change in excess values explains only 4.30 percent of the cross-sectional
variation in the actual change in excess values.
With respect to the differential valuation effects according to the business
relatedness of cross-border acquisitions (Regression 2), the intercept term now captures
the additional valuation effects of unrelated cross-border acquisitions that cannot be
accounted for by the underlying characteristics of the foreign targets. It turns out that the
estimated intercept (−0.1769) is significantly less than zero (P-value = 0.007), indicating
that unrelated cross-border acquisitions result in a significant decline in excess values
beyond the valuation effects from simply combining foreign targets with U.S. acquiring
firms. This result confirms those reported in the third panel of Table 7. Moreover, related
cross-border acquisitions exhibit an incremental positive impact on the U.S. acquirers’
excess values according to the estimated coefficient on the related-acquisition indicator
variable (0.2282), which is significantly greater than zero (P-value = 0.005). Not

21
surprisingly, the projected change in excess values is still a poor predictor of the actual
changes in excess values. The estimated beta coefficient (0.5039) is significantly
different from one (P-value = 0.008), and the model can now explain about 9.30 percent
of the cross-sectional variation in the actual change in excess values.
Finally, the results from Regression 3 also corroborate those reported in Table 7.
The estimated intercept (−0.0170) is not significantly different from zero (P-value =
0.426), indicating that no incremental, significant decline in excess values remains after
accounting for the underlying characteristics of the foreign targets, when U.S. firms
establish operations abroad for the first time by means of cross-border acquisitions. The
negative coefficient on the indicator variable (−0.0659), which is statistically
insignificant (P-value = 0.512), suggests that cross-border acquisitions involving U.S.
acquiring firms already operating overseas does not result in additional value loss beyond
the valuation effects accruing from simply adding foreign targets. The estimated beta
coefficient of the projected change in excess values (0.4627) is significantly different
from one at the five percent level (P-value = 0.025).
Summarizing our valuation results, we first find that acquisitions of “fairly
valued” foreign firms, on average, do not lead to diversification discount. This result is
consistent with the combined evidence that foreign acquired targets do not significantly
differ from the corresponding industry-matched U.S. stand-alone firms, and that they are
significantly smaller than their U.S. acquiring counterparts. Thus, accounting for the
underlying characteristics of the target firms has important implications on the actual
change in excess values of U.S. acquiring firms. Therefore, the results support the view
that selection bias hypothesis may play a role in valuing divisions of MNCs based on
industry-multiplier approaches. Second, we provide new evidence that unrelated cross-
border M&As result in significant value loss after accounting for the valuation of the
foreign targets, thus corroborating the industrial diversification discount. On the other
hand, related cross-border M&As do not destroy value. Finally, we document
significantly large wealth gains for the foreign target shareholders in cross-border
acquisitions, regardless of the type of acquisition considered. In the next section, we
explore the issue of whether excess firm value and the international diversification
decision are endogenously related.

22
4.5. Investigating the Endogeneity of the International Diversification Decision

We next reexamine our findings in a setting that is mindful of the potential endogeneity
between firm attributes and value. To the extent that firm-specific characteristics are
likely to be correlated with firm value, we model U.S. acquirers’ excess value as a
function of both the observable and unobservable firm-specific characteristics over the
two-year period surrounding acquisition. We first estimate ordinary least squares (OLS)
regressions of excess value on an indicator variable denoting the international
diversification status of the U.S. acquiring firms (INTL_DIV) as well as on a set of
observable control variables including (i) firm size, proxied by the natural logarithm of
total assets (LTA), (ii) leverage, measured as the ratio of book value of total debt to total
assets, (iii) profitability, defined as earnings before interest and tax expenses normalized
by sales (EBIT/Sales), (iv) intangible assets, measured as the ratio of research and
development expenditures to sales (R&D/Sales), and (v) investment, calculated as capital
expenditures divided by sales (CAPEX/Sales). We also control for the possibility of a
nonlinear effect of firm size on excess firm value by including the square of the natural
logarithm of total assets (SQ_LTA)13. The regression coefficient on the indicator variable
INTL_DIV captures the average difference in excess values between internationally
diversified acquiring firms and domestic stand-alone firms used as benchmarks.
The results for the OLS regressions are based on both sales and asset multipliers,
and are presented in columns 1 and 3 of Table 9, respectively. Panel A reports the
coefficient estimates for the year prior to the acquisition event, while Panel B presents the
corresponding results for the year following the acquisition. In both years surrounding
acquisition, the sales-based OLS regressions indicate that U.S. acquiring firms that are
internationally diversified are not significantly undervalued or overvalued relative to the
industry-matched stand-alone firms used as benchmarks; in year t = −1 (Panel A), there is
weak evidence of a diversification premium (7.82 percent), while at t = +1 (Panel B),
there is weak evidence of a diversification discount (−8.89 percent), even though both
point estimates are not statistically significant. However, the post-merger change in the

13
Because excess values are measures of firm value relative to domestic stand-alone firms operating in the
same industry, all the control variables in our valuation analysis are also relative measures, calculated as
the difference between the actual value for the firm and the median value for the domestic stand-alone
firms matched with the same two-digit SIC code.

23
sign of the point estimates of the valuation effects from positive to negative somewhat
reconciles with the univariate results previously reported in the first panel of Table 5
regarding the insignificant average decline in excess values for U.S. acquirers. The
results for the asset-based OLS regressions virtually show the same evidence, except that,
in the year before the acquisition, U.S. MNCs are valued at a premium (13.30 percent)
that is significant at the one percent level. Overall, both the statistical significance and the
signs of the coefficient estimates for the other control variables conform with those
reported by Bodnar, Tang and Weintrop (1999), Denis, Denis and Yost (2002), and
Campa and Kedia (2002). For instance, we find evidence that excess values are positively
related to firm size, capital expenditures, and R&D expenditures, and negatively related
to leverage. We also find evidence that the coefficient estimate of the square of firm size
is significantly negative (except for the asset-based result at t = +1), indicating that the
positive effect of firm size on excess value decreases as firm size increases.
We next control the valuation analysis for unobservable firm-specific
characteristics that might affect excess firm value by estimating a fixed-effects model in a
panel data framework as defined in Hausman and Taylor (1981). At the same time, we
also introduce year dummies to control for time effects that might affect the international
diversification decision. The basic idea is to assign a unique intercept for each firm in our
sample in order to capture an average latent firm-specific effect on excess value that
might potentially be correlated with the other control variables included in our regression
framework. Hence, in the presence of such correlations, the coefficient estimates from the
OLS regressions may be biased and inconsistent. In introducing firm-specific fixed
effects, we order our sample firm-year observations by firm and then by time to define
unbalanced panel data for both the year prior to and the year following the acquisition14.
We report the estimates of our multivariate valuation analysis with firm-specific
and year fixed effects (i.e., a two-way fixed effects) for both sales and asset multipliers in
columns 2 and 4 of Table 9, respectively. The introduction of two-way fixed effects at t =
−1 (Panel A) does not alter the statistical insignificance of the sales-based diversification

14
At t = −1, the sample consists of 6,649 firm-year observations for 2,747 domestic single-segment firms
and 136 firm-year observations for 122 U.S. acquiring firms. At t = +1, there are 5,857 firm-year
observations for 2,700 domestic single-segment firms and 136 firm-year observations for 122 U.S.
acquiring firms.

24
premium for internationally diversified U.S. acquiring firms, even though the economic
magnitude of such a premium now declines from 7.82 percent (OLS estimate) to 4.24
percent. On the other hand, the point estimate on INTL_DIV based on asset multipliers
substantially changes from a significant premium of 13.30 percent (OLS estimate) to a
statistically insignificant discount of −7.84 percent. As for the other control variables, the
signs of the sales-based coefficient estimates on these variables are invariably robust to
the inclusion of fixed effects, but the statistical significance in some cases is sensitive to
the fixed-effects specification. For example, the positive effect of EBIT/Sales on excess
values now becomes statistically significant, whereas the negative effect of both
Leverage and SQ_LTA turns out to be insignificant. The signs of the asset-based point
estimates on LTA, Leverage and SQ_LTA significantly change in the presence of fixed
effects, while the positive impact of CAPEX/Sales on excess value becomes statistically
insignificant. At t = +1 (Panel B), after the completion of the acquisition, the coefficient
estimates on INTL_DIV based on both types of industry multipliers continue to be
insignificantly negatively related to excess values as compared to the OLS estimates.
Furthermore, the signs and statistical significance of the sales-based coefficient estimates
on the other control variables are now more robust to the inclusion of fixed effects. Thus,
the post-merger results from pooling the globally diversified U.S. firms with their stand-
alone U.S. benchmarks in a fixed-effects framework support the evidence that
international diversification does not appear to be correlated with unobserved firm-
specific characteristics.

5. CONCLUSIONS

In this paper, we examine the valuation effects of corporate international diversification

within the context of cross-border mergers and acquisitions. Consistent with theoretical
expectations, MNCs are valued at a significant premium relative to industry-matched
U.S. benchmarks. We find no evidence of a significant decline in excess values of U.S.
acquirers, given that the foreign targets are “fairly valued” relative to the industry-
matched benchmarks. Indeed, the foreign target firms do not differ from the U.S.
benchmarks in most of their underlying characteristics, and accounting for such

25
characteristics is important for ex post valuation status of U.S. acquiring firms. Overall,
our results suggest that the act of cross-border acquisition does not lead to value
destruction.
Consistent with industrial discount literature, our results suggest that unrelated
cross-border acquisitions lead to additional value loss after accounting for the underlying
characteristics of the foreign targets. On the other hand, we find weak evidence that
related cross-border acquisitions may be value-enhancing beyond simply adding “fairly
valued” foreign firms to U.S. acquirers. We also find weak evidence of an incremental
decline in excess values when U.S. firms establish operations abroad for the first time.
However, this value reduction occurs because such acquisitions involve foreign targets
that are valued, on average, at an economically sizeable discount, and not because
premiere cross-border acquisitions destroy value. Thus, the selection bias seems to affect
the post-merger valuation of U.S. “first-timers”. U.S. firms with operations already
established abroad do not experience any significant change in excess value after
accounting for the characteristics of the target firms. Our results do not suggest that
excess firm value and the international diversification decision are endogenously related.
Finally, we document significantly positive changes in foreign target shareholder wealth
regardless of the type of acquisition.

REFERENCES

Agmon, Tamir and Donald Lessard, 1977, “Investor Recognition of Corporate

International Diversification”, Journal of Finance, 32, 1049-1055.

Berger, Philip and Eli Ofek, 1995, “Diversification’s Effect on Firm Value”, Journal of
Financial Economics, 37, 39-65.

Bodnar, Gordon, Charles Tang, and Joseph Weintrop, 1999, “Both Sides of Corporate
Diversification: The Value Impacts of Geographic and Industrial Diversification”,
Working Paper, Paul H. Nitze School of Advanced International Studies, Johns
Hopkins University, 35p.

26
Brewer, H.L., 1981, “Investor Benefits from Corporate International Diversification”,
Journal of Financial and Quantitative Analysis, 16, 113-126.

Cakici, Nusret, Chris Hessel, and Kishore Tandon, 1996, “Foreign Acquisitions in the
United States: Effect on Shareholder Wealth of Foreign Acquiring Firms”, Journal of
Banking and Finance, 20, 307-329.

Campa, Jose Manuel and Simi Kedia, 2002, “Explaining the Diversification Discount”,
Journal of Finance, 57, 1731-1762.

Caves, Richard, 1971, “International Corporations: The Industrial Economics of Foreign

Investments”, Econometrica, 38, 1-27.

Cebenoyan, A. Sinan, George Papaioannou, and Nickolaos Travlos, 1992, “Foreign

Takeover Activity in the U.S. and Wealth Effects for Target Firm Shareholders”,
Financial Management, 21, 58-68.

Chevalier, Judith, 2000, “What Do We Know about Cross-Subsidization? Evidence from

the Investment Policies of Merging Firms”, Working Paper, University of Chicago,
36p.

Christophe, Stephen, 1997, “Hysteresis and the Value of the U.S. Multinational
Corporation”, Journal of Business, 70, 435-462.

Denis, David, Diane Denis, and Keven Yost, 2002, “Global Diversification, Industrial
Diversification, and Firm Value”, Journal of Finance, 57, 1951-1979.

Dewenter, Kathryn, 1995, “Does the Market React Differently to Domestic and Foreign
Takeover Announcements? Evidence from the U.S. Chemical and Retail Industries”,
Journal of Financial Economics, 37, 421-441.

Doukas, John and Nickolaos Travlos, 1988, “The Effect of Corporate Multinationalism
on Shareholders’ Wealth: Evidence from International Acquisitions”, Journal of
Finance, 43, 1161-1175.

27
Errunza, Vihang and Lemma Senbet, 1981, “The Effects of International Operations on
the Market Value of the Firm: Theory and Evidence”, Journal of Finance, 36, 401-417.

Errunza, Vihang and Lemma Senbet, 1984, “International Corporate Diversification,

Market Valuation, and Size-Adjusted Evidence”, Journal of Finance, 39, 727-743.

Eun, Cheol, Richard Kolodny, and Carl Scheraga, 1996, “Cross-Border Acquisitions and
Shareholder Wealth: Test of the Synergy and Internalization Hypotheses”, Journal of
Banking and Finance, 20, 1559-1582.

Fatemi, Ali, 1984, “Shareholder Benefits from Corporate International Diversification”,

Journal of Finance, 39, 1325-1344.

Graham, John, Michael Lemmon, and Jack Wolf, 2002, “Does Corporate Diversification
Destroy Value?”, Journal of Finance, 57, 695-720.

Harris, Robert and David Ravenscraft, 1991, “The Role of Acquisitions in Foreign Direct
Investment: Evidence from the U.S. Stock Market”, Journal of Finance, 46, 825-844.

Hausman, Jerry and William Taylor, 1981, “Panel Data and Unobservable Individual
Effects”, Econometrica, 49, 1377-1398.

Hayes, Rachel and Russell Lundholm, 1996, “Segment Reporting to the Capital Market
in the Presence of a Competitor”, Journal of Accounting Research, 34, 261-279.

Hughes, John, Dennis Logue, and Richard Sweeney, 1975, “Corporate International
Diversification and Market Assigned Measures of Risk and Diversification”, Journal
of Financial and Quantitative Analysis, 10, 627-637.

Hyland, David C. and J. David Diltz, 2002, “Why Firms Diversify: An Empirical
Examination”, Financial Management, 31, 51-81.

Jacquillat, Bertrand and Bruno Solnik, 1978, “Multinationals are Poor Tools for
Diversification”, Journal of Portfolio Management, 4, 8-12.

28
Jensen, Michael and William Meckling, 1976, “Theory of the Firm: Managerial
Behavior, Agency Costs and Ownership Structure”, Journal of Financial Economics,
3, 305-360.

Jensen, Michael, 1986, “Agency Costs of Free Cash Flow, Corporate Finance, and
Takeovers”, American Economic Review, 76, 323-329.

Kim, Wi Saeng and Esmeralda Lyn, 1986, “Excess Market Value, the Multinational
Corporation, and Tobin’s Q-Ratio”, Journal of International Business Studies, 17, 119-
125.

Kogut, B. and N. Kulatilaka, 1994, “Operating Flexibility, Global Manufacturing, and the
Option Value of a Multinational Network”, Management Science, 40,123-139.

Kuipers, David, Darius Miller, and Ajay Patel, 2003, “The Legal Environment and
Corporate Valuation: Evidence from Cross-Border Takeovers”, Working Paper, Rawls
College of Business Administration, Texas Tech University, 44p.

Lamont, Owen and Christopher Polk, 2002, “Does Diversification Destroy Value?
Evidence from Industry Shocks”, Journal of Financial Economics, 63, 51-77.

Lang, Larry and Rene Stulz, 1994, “Tobin’s q, Corporate Diversification, and Firm
Performance”, Journal of Political Economy, 102, 1248-1280.

Lins, Karl and Henri Servaes, 1999, “International Evidence on the Value of Corporate
Diversification”, Journal of Finance, 54, 2215-2239.

Lins, Karl and Henri Servaes, 2002, “Is Corporate Diversification Beneficial in Emerging
Markets?”, Financial Management, 31, 5-31.

Marr, M. Wayne, Sanjeev Mohta, and Michael Spivey, 1993, “An Analysis of Foreign
Takeovers in the United States”, Managerial and Decision Economics, 14, 285-294.

Mergers and Acquisitions: Almanac and Index, Washington, D.C., Issues 1992-2000.

29
Moeller, Sara, and Frederik Schlingemann, 2002, “Are Cross-Border Acquisitions
Different from Domestic Acquisitions? Evidence on Stock and Operating Performance
for U.S. Acquirers”, Working Paper, Cox School of Business, Southern Methodist
University, 44p.

Morck, Randall and Bernard Yeung, 1991, “Why Investors Value Multinationality”,
Journal of Business, 64, 165-187.

Senbet, Lemma, 1979, “International Capital Market Equilibrium and the Multinational
Firm Financing and Investment Policies”, Journal of Financial and Quantitative
Analysis, 14, 455-480.

Servaes, Henri, 1996, “The Value of Diversification During the Conglomerate Merger
Wave”, Journal of Finance, 51, 1201-1225.

Villalonga, Belén, 2002, “Does Diversification Cause the “Diversification Discount”?”,

Working Paper, Harvard Business School, 46p.

Wernerfelt, Birger and Cynthia Montgomery, 1988, “Tobin’s Q and the Importance of
Focus in Firm Performance”, American Economic Review, 78, 246-250.

Whited, Toni, 2001, “Is It Inefficient Investment that Causes the Diversification
Discount?”, Journal of Finance, 56, 1667-1691.

30
 Table 1
Summary of the Empirical Studies on the Valuation Effects of Corporate International Diversification
Study Period Valuation Measure Diversification Measure Control Variables Average Effect

Errunza and 1968-1973 Ratio of excess market value (i) Fraction of sales generated abroad; (a) Total variability of stock returns; (i) Positive;
Senbet (1981) 1974-1977 of equity to sales. (ii) Fraction of net earnings from abroad; (b) Industrial concentration. (ii) None;
(iii) Fraction of net assets from abroad. (iii) None.

Errunza and 1970-1978 Ratio of excess market value (i) Fraction of sales generated abroad; (a) Systematic risk; (i) Positive;
Senbet (1984) of equity to sales. (ii) Absolute value of sales generated abroad; (b) Price-earnings ratio. (ii) Positive;
(iii) Entropy measure; (iii) Positive;
(iv) Number of foreign subsidiaries. (iv) Negative.

Kim and Lyn 1974-1978 Tobin’s Q. (i) Fraction of sales generated abroad; (a) R&D/sales; (i) Positive
(1986) (ii) Number of foreign subsidiaries; (b) Advertising expenses/sales; (ii) None;
(iii) Interaction (i) versus (ii). (c) Past sales growth rate; (iii) None.
(d) Monopoly power.

Morck and 1978 Industry-adjusted (i) Number of foreign subsidiaries; (a) R&D/replacement cost of tangibles; (i) Positive/Negative(2);
Yeung (1991) Tobin’s Q(1). (ii) Number of hosting countries. (b) Advertising exp./replacement cost; (ii) Positive.
(c) Leverage.

Christophe (1997) 1978-1986 Tobin’s Q. Fraction of sales generated abroad. (a) R&D/assets; Negative.
(b) Advertising expenses/assets;
(c) Natural log of total assets;
(d) Leverage;
(e) Diversification measure.

Bodnar, Tang and 1984-1997 Industry-adjusted Indicator variable for firms reporting (a) R&D/sales; 2.7%
Weintrop (1999) excess value(3). more than 10 percent of their total sales (b) Advertising expenses/sales;
from operations abroad. (c) Leverage;
(d) Diversification dummy;
(e) Time dummies.

Denis, Denis and 1984-1997 Industry-adjusted Fraction of sales generated abroad. (a) R&D/Sales; - 18%(4)
Yost (2002) excess value(3). (b) Advertising expenses/sales; - 32%(5)
(c) Leverage;
(d) Diversification dummies.

(1) Adjustment includes three-digit SIC code indicator variables in the multivariate regression models; (2) Evidence of negative valuation effect when firms have more than twenty subsidiaries abroad;
(3) Industry-median multipliers are based on single-segment domestic firms; (4) Diversification discount for single-segment MNCs; (5) Diversification discount for multi-segment MNCs.

31
 Table 2
Sample Selection Procedure
Panel A: Selection of Cross-Border M&A Transactions from Securities Data Corporation (SDC) Database

Initial SDC sample of transactions involving cross-border M&As (U.S. bidder/foreign target) between 1990 and 1999 11,392

Subtract:
(a) Intended, pending, rumored, or withdrawn transactions, and completed transactions whose U.S. bidder acquired less than 51% of the foreign target shares (3,596)

Intermediate Sample:
(b) Number of completed transactions representing cross-border acquisitions of a majority ownership stake in a foreign target firm 7,796

Subtract:
(c) Transactions involving either U.S. bidder or foreign target with operations in the financial services industry (SIC code between 6000 and 6999) (1,148)
(d) Transactions involving either U.S. bidder or foreign target with undisclosed identity ( 39)
(e) Transactions representing divestitures of private business units, concessions, joint ventures, or acquisitions by an investor group (2,483)
(f) Transactions representing bankruptcy acquisitions associated with divestitures of certain assets or private business units (2,763)

Final SDC sample of cross-border M&As 1,363

Panel B: Collection of Market and Accounting Data for Foreign Targets from Worldscope and Datastream Databases

Initial sample of foreign targets to U.S. acquirers (from the final SDC sample) 1,363

Subtract:
(a) Foreign targets for which no data on market price/shares outstanding is available from Datastream (i.e., privately held firms) (1,141)
(b) Publicly traded foreign targets for which no pre-merger accounting data is available from Worldscope ( 72)

Final Worldscope/Datastream sample of publicly traded foreign target firms to U.S. acquirers 150

Panel C: Collection of Market and Accounting Data for U.S. Bidders from Compustat Database

Initial sample of U.S. acquirers of foreign target firms (from the final SDC sample) 1,363

Subtract:
(a) U.S. acquirers for which foreign targets are either privately held firms or public companies without pre-merger accounting data available (1,213)
(b) U.S. acquirers for which neither data on stock price/shares outstanding nor accounting data is available from Compustat ( 12)
(c) U.S. acquirers reporting business divisions in the financial services industry according to the Compustat Industry Segment Files ( 2)

Final Compustat-Worldscope-Datastream sample of U.S. acquirers and foreign target firms 136

32
 Table 3
Sample Distribution and Descriptive Statistics for Completed Cross-Border M&As
Panel A: Sample Distribution According to the Origin of the Target Firm and the Calendar Year for the 136
Cross-Border M&As Consummated Over the Period 1990-1999

Target Country Total Number % of Total Period Total Number % of total

Argentina 1 0.73 1990 1 0.73

Australia 7 5.15 1991 1 0.73
Belgium 1 0.73 1992 3 2.21
Bermudas 2 1.47 1993 5 3.68
Brazil 1 0.73 1994 5 3.68
Canada 41 30.15 1995 13 9.56
Denmark 3 2.21 1996 10 7.35
France 9 6.62 1997 18 13.24
Germany 5 3.68 1998 42 30.88
Hong Kong 1 0.73 1999 38 27.94
Ireland 1 0.73
Israel 6 4.41
New Zealand 1 0.73
Norway 2 1.47
South Africa 2 1.47
South Korea 1 0.73
Spain 1 0.73
Sweden 5 3.68
United Kingdom 46 33.82

Panel B: Descriptive Statistics of the Deal Characteristics for the 136 Cross-Border M&As in Terms of
Deal Value and Ownership Stake Acquired in the Foreign Target by the U.S. Bidder

Deal Characteristics Mean Median Minimum Maximum

Deal Value (US$ Million) 741.17 213.00 5.40 11,070.00

Target Stake Acquired (%) 94.06 100.00 51.00 100.00

Panel C: Sample Distribution of the of the Deal Characteristics for the 136 Cross-Border M&As in Terms
of Type of Acquisition, Medium of Payment, Accounting Method, and Bidder Attitude

Deal Characteristics Total Number % of Total

Type of Acquisition:
Merger 35 25.74
Tender Offer 101 74.26
Medium of Payment:
Cash 84 61.76
Stock 40 29.41
Mix Cash/Stock 6 4.41
Mix Cash/Debt 6 4.41
Accounting Method:
Pooling of Interest 18 13.24
Purchase 118 86.76
Bidder Attitude:
Friendly 121 88.97
Hostile 15 11.03

33
 Table 4
Descriptive Statistics for U.S. Acquirers, Foreign Targets and U.S. Valuation Benchmarks
Descriptive statistics for a sample of 136 U.S. acquiring firm-years (Panel A), 136 foreign target firm-years (Panels B, C, and D), and 5,635 U.S. single-segment firm-years
specifically used as benchmarks to value the foreign targets (Panels C and D) over the period 1990-1999. Total capital is the sum of market value of common equity, book value of
total debt, and the liquidating value of preferred stock. The q-ratio_Sales (q-ratio_Asset) is the natural logarithm of the ratio of total capital to sales (total assets). Leverage is the
ratio of book value of total debt to total assets. EBIT/Sales represents earnings before interest and tax expenses divided by sales. R&D/Sales stands for research and development
expenditures normalized by sales. CAPEX/Sales represents capital expenditures divided by sales. STD refers to the standard deviation and N stands for the number of yearly
observations available to calculate each of the firm-specific measures described above. The q-ratios for the foreign target firms are calculated using the market value of common
equity based on the last stock price available prior to delisting. In Panel D, the industry-adjusted difference in firm-specific characteristics between foreign target and U.S.
benchmark firms is calculated as the contemporaneous difference between a foreign target’s actual value and the median value for a group of matching stand-alone domestic firms
with the same two-digit SIC code. The significance of the difference in means is based on the parametric t-statistics, while the significance of the difference in medians is assessed
using the nonparametric Wilcoxon rank sum test statistics. The asterisks *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
Panel A: U.S. Acquirer Sample Characteristics Over the Two-Year Period Surrounding the Acquisition

Before Acquisition (t = −1) After Acquisition (t = +1) Post-Merger Difference

Firm-Specific Characteristics
Mean Median STD N Mean Median STD N Mean Median STD N
*** **
Total Capital (US$ Million) 9753.0 2769.0 23781.0 136 13840.0 4609.0 30288.0 136 4088.0 976.6 14831.0 136
Sales (US$ Million) 5160.0 1378.0 13905.0 136 6954.0 2087.0 18472.0 136 1794.0*** 541.1** 5125.0 136
Total Assets (US$ Million) 5721.0 1742.0 11457.0 136 8276.0 2960.0 14093.0 136 2555.0*** 829.8*** 5777.0 136
q-Ratio_Sales 0.7367 0.6959 0.7660 136 0.6664 0.5506 0.8516 136 -0.0704 -0.0992 0.5215 136
q-Ratio_Asset 0.5169 0.4286 0.5365 136 0.3605 0.2390 0.5710 136 -0.1564*** -0.1664** 0.4931 136
Leverage 0.2456 0.2312 0.1891 136 0.2945 0.2864 0.2057 136 0.0490*** 0.0494** 0.1445 136
EBIT/Sales 0.0936 0.1190 0.1492 135 0.0752 0.1083 0.1687 135 -0.0184 -0.0110 0.1522 135
R&D/Sales 0.0706 0.0279 0.1120 73 0.0632 0.0312 0.0752 73 -0.0074 0.0011 0.0501 73
CAPEX/Sales 0.1365 0.0559 0.2318 136 0.1221 0.0583 0.1906 136 -0.0144 -0.0014 0.1645 136

Panel B: Paired-Sample Characteristics of U.S. Acquiring and Foreign Target Firms in the Year Prior to the Acquisition

U.S. Acquirer (t = −1) Foreign Target (t = −1) Pre-Merger Difference (Acquirer − Target)
Firm-Specific Characteristics
Mean Median STD N Mean Median STD N Mean Median STD N

Total Capital (US$ Million) 9753.0 2769.0 23781.0 136 829.0 227.0 1857.0 136 8924.0*** 2192.3*** 23207.0 136
Sales (US$ Million) 5160.0 1378.0 13905.0 136 536.0 167.0 1223.0 136 4625.0*** 1058.8*** 13221.0 136
Total Assets (US$ Million) 5721.0 1742.0 11457.0 136 541.0 158.0 1273.0 136 5180.0*** 1368.8*** 11199.0 136
q-Ratio_Sales 0.7367 0.6959 0.7660 136 0.4329 0.3395 0.9437 136 0.3039*** 0.3503*** 0.8404 136
q-Ratio_Asset 0.5169 0.4286 0.5365 136 0.4170 0.3620 0.6570 136 0.1000* 0.0880 0.6419 136
Leverage 0.2456 0.2312 0.1891 136 0.2076 0.1823 0.1720 136 0.0380** 0.0281 0.2108 136
EBIT/Sales 0.0936 0.1190 0.1492 135 0.0574 0.0834 0.3331 136 0.0362** 0.0286*** 0.2131 135
R&D/Sales 0.0706 0.0279 0.1120 73 0.1960 0.0290 0.8980 54 -0.1220 0.0000 0.9410 47
CAPEX/Sales 0.1365 0.0559 0.2318 136 0.1368 0.0560 0.2468 130 0.0030 0.0044 0.2392 130

34
 Table 4 (Continued)

Panel C: Overall Sample Characteristics of the Foreign Target and U.S. Stand-Alone Firms

Foreign Target Firm (t = −1) U.S. Stand-Alone Firm (t = −1) Unadjusted Difference
(Target − Stand Alone)
Firm-Specific Characteristics
Mean Median STD N Mean Median STD N Mean Median STD N

Total Capital (US$ Million) 829.0 227.0 1857.0 136 871.0 143.2 4094.0 5635 -42.0 67.7*** − 5771
Sales (US$ Million) 536.0 167.0 1223.0 136 387.0 84.8 1511.0 5635 149.0 58.4*** − 5771
Total Assets (US$ Million) 541.0 158.0 1273.0 136 591.0 89.6 2515.0 5635 -50.0 46.6*** − 5771
q-Ratio_Sales 0.4329 0.3395 0.9437 136 0.4140 0.3420 1.1300 5635 0.0189 0.0263 − 5771
q-Ratio_Asset 0.4170 0.3620 0.6570 136 0.3970 0.2813 0.7390 5635 0.0202 0.0711 − 5771
Leverage 0.2076 0.1823 0.1720 136 0.2410 0.1888 0.2740 5635 -0.0335** -0.0006 − 5771
EBIT/Sales 0.0574 0.0834 0.3331 136 0.0290 0.0683 0.3040 5635 0.0084 0.0122 − 5771
R&D/Sales 0.1960 0.0290 0.8980 54 0.1500 0.0840 0.2508 2326 0.0470 -0.0348*** − 2380
CAPEX/Sales 0.1368 0.0560 0.2468 130 0.1430 0.0462 0.3770 5583 -0.0062 0.0067* − 5713

Panel D: Paired-Sample Characteristics of Foreign Target and Industry-Matched Stand-Alone U.S. Firms

Foreign Target Firm (t = −1) U.S. Benchmark Firm (t = −1) Industry-Adjusted Difference
(Target − Benchmark)
Firm-Specific Characteristics
Mean Median STD N Mean Median STD N Mean Median STD N

Total Capital (US$ Million) 829.0 227.0 1857.0 136 223.0 131.9 270.0 136 606.0*** 78.1*** 1815.0 136
Sales (US$ Million) 536.0 167.0 1223.0 136 126.0 73.4 150.0 136 409.8*** 79.0*** 1151.4 136
Total Assets (US$ Million) 541.0 158.0 1273.0 136 151.0 75.0 172.0 136 390.0*** 51.4*** 1220.0 136
q-Ratio_Sales 0.4329 0.3395 0.9437 136 0.3910 0.3040 0.6285 136 0.0419 0.0087 0.8235 136
q-Ratio_Asset 0.4170 0.3620 0.6570 136 0.3099 0.3202 0.2700 136 0.1070* 0.0711 0.6325 136
Leverage 0.2076 0.1823 0.1720 136 0.1922 0.1501 0.1443 136 0.0154 0.0023 0.2009 136
EBIT/Sales 0.0574 0.0834 0.3331 136 0.0723 0.0714 0.0680 136 -0.0353 0.0082 0.3387 136
R&D/Sales 0.1960 0.0290 0.8980 54 0.0679 0.0758 0.0488 114 0.1281 -0.0132** 0.9290 49
CAPEX/Sales 0.1368 0.0560 0.2468 130 0.1253 0.0395 0.2067 136 0.0115 0.0044 0.2219 130

35
 Table 5
Excess Value Measures for U.S. Acquirers Based on Sales and Asset Multiples, and U.S. Shareholder
Wealth Changes in Cross-Border Mergers and Acquisitions
This table displays excess value (EV) measures for U.S. acquirers in the year prior to (EV-1) and the year following
(EV+1) the acquisition, as well as the actual change in EV from year t = − 1 to year t = + 1 (∆EV+1). Excess value,
calculated as in Bodnar, Tang and Weintrop (1999), and Denis, Denis and Yost (2002), which represents a variation of
the industry-matched multiplier approach originally developed by Berger and Ofek (1995), is defined as the natural
logarithm of the ratio of a firm’s actual market value to its imputed value. A firm’s imputed value is the sum of the
imputed values of its business units, with each business unit’s imputed value equal to the business unit’s sales (assets)
multiplied by the median ratio of market value to sales (assets) for all single-segment domestic firms in the same
industry. The full sample (Panel A) consists of 136 non-financial, publicly traded U.S. firms that completed an
acquisition of a majority ownership stake in a non-financial, publicly traded foreign target firm over the period 1990-
1999. Acquisitions are classified as related cross-border M&As (Panel B) when the U.S. acquirer and the foreign target
report the same two-digit SIC codes in the year prior to the acquisition, and as unrelated cross-border M&As (Panel C)
otherwise. Acquisitions are classified as non-premiere cross-border operations (Panel D) when the U.S. acquirer
already has operations overseas in the year prior to the acquisition, and as premiere cross-border operations (Panel E)
otherwise. N refers to the number of yearly observations in each subsample. The significance of the mean values and of
the difference in means is based on the parametric t-statistics. The significance of the median values is based on the
nonparametric Wilcoxon signed-rank test statistics, while the significance of the difference in medians is assessed using
the nonparametric Wilcoxon rank sum test statistics. The asterisks *, **, and *** indicate statistical significance at the
10%, 5%, and 1% levels, respectively.
Sales Multiples Asset Multiples

Mean Median Mean Median

Panel A: Full Sample (N=136)

EV-1 0.3186*** 0.3302*** 0.1962*** 0.1791***
EV+1 0.2526*** 0.2497*** 0.1068** 0.0908*

∆EV+1 -0.0659 -0.0799 -0.0893** -0.0850

Panel B: Related Cross-Border M&As (N=88)
EV-1 0.2828*** 0.2914*** 0.1999*** 0.1632***
EV+1 0.3108*** 0.2842 ***
0.1484** 0.1290**

∆EV+1 0.0280 -0.0072 -0.0515 -0.0346

Panel C: Unrelated Cross-Border M&As (N=48)
EV-1 0.3841*** 0.4085*** 0.1894*** 0.1982***
**
EV+1 0.1460 0.1886 0.0307 0.0283

∆EV+1 -0.2382*** -0.2153* -0.1587** -0.1725*

Panel D: Non-Premiere Cross-Border Operations (N=105)
EV-1 0.3197*** 0.3243*** 0.1957*** 0.1821***
*** ***
EV+1 0.2814 0.2762 0.1548*** 0.1354***

∆EV+1 -0.0383 -0.0548 -0.0408 -0.0434

Panel E: Premiere Cross-Border Operations (N=31)
EV-1 0.3150** 0.3479** 0.1978* 0.1703*
EV+1 0.1550 0.1685 -0.0560 -0.0873

∆EV+1 -0.1600 -0.2093 -0.2537** -0.2320*

36
 Table 6
Excess Value Measures for Foreign Targets Based on Sales and Asset Multiples, Target Shareholder
Wealth Changes, and Relative Sizes of Cross-Border Merger and Acquisitions
This table reports two measures of excess value (EV) for foreign target firms in their last year of operations as stand-
A
alone firms prior to the acquisition. EV -1 denotes the pre-announcement EV and is calculated using the market value
E
of common equity one month prior to the announcement of the acquisition. EV -1 stands for the pre-effective EV and
is computed using the market value of common equity based on the last stock price available prior to delisting. ∆EV-1
E A
stands for the difference between EV -1 and EV -1. Excess value is calculated as in Berger and Ofek (1995). The
relative size of the acquisition (Rel_Size-1) is measured as the ratio of the target’s sales (asset) to the combined sales
(assets) of both the target and the U.S. acquirer in the year prior to the acquisition. The full sample (Panel A) consists
of 136 non-financial, publicly traded foreign targets acquired by non-financial, publicly traded U.S. firms over the
period 1990-1999. Acquisitions are classified as related cross-border M&As (Panel B) when the U.S. acquirer and the
foreign target report the same two-digit SIC codes in the year t = − 1, and as unrelated (Panel C) otherwise.
Acquisitions are classified as non-premiere cross-border operations (Panel D) when the U.S. acquirer already has
operations overseas in the year prior to the acquisition, and as premiere cross-border operations (Panel E) otherwise. N
refers to the number of yearly observations in each subsample. The significance of the mean values and of the
difference in means is based on the parametric t-statistics. The significance of the median values is based on the
nonparametric Wilcoxon signed-rank test statistics, while the significance of the difference in medians is assessed using
the nonparametric Wilcoxon rank sum test statistics. The asterisks *, **, and *** indicate statistical significance at the
10%, 5%, and 1% levels, respectively.
Sales Multiples Asset Multiples

Mean Median Mean Median

Panel A: Full Sample (N=136)

EVA-1 -0.1977*** -0.1921*** -0.1331*** -0.1238***
EVE-1 0.0415 0.0428 0.1061* 0.1158**
∆EV-1 0.2393*** 0.2329*** 0.2393*** 0.2401***

Rel_Size-1 0.1910 0.1470 0.1701 0.1044

Panel B: Related Cross-Border M&As (N=88)
EVA-1 -0.2323** -0.2370** -0.1338* -0.1361**
EVE-1 -0.0107 -0.0051 0.0878 0.0838
∆EV-1 0.2216*** 0.2243* 0.2216*** 0.2169**

Rel_Size-1 0.2142 0.1616 0.1941 0.1211

Panel C: Unrelated Cross-Border M&As (N=48)
EVA-1 -0.1343 -0.1321 -0.1319 -0.1014
EVE-1 0.1374 0.1175 0.1398* 0.1762**
∆EV-1 0.2717*** 0.2410 *
0.2717*** 0.2701***

Rel_Size-1 0.1486 0.0875 0.1261 0.0538

Panel D: Non-Premiere Cross-Border Operations (N=105)
EVA-1 -0.1638** -0.1471* -0.0800 -0.0720
EVE-1 0.0782 0.0909 0.1620** 0.1740***
∆EV-1 0.2419 ***
0.2339 **
0.2419*** 0.2454***

Rel_Size-1 0.1551 0.1150 0.1347 0.0833

Panel E: Premiere Cross-Border Operations (N=31)
EVA-1 -0.3130** -0.3475** -0.3130*** -0.3300***
EVE-1 -0.0820 -0.1214 -0.0831 -0.0810
∆EV-1 0.2302*** 0.2258 0.2302*** 0.2428

Rel_Size-1 0.3126 0.2548 0.2901 0.2126

37
 Table 7
Actual, Projected and Unexplained Changes in Excess Values for U.S. Acquirers of Foreign Targets
This table presents actual and projected changes in excess values for U.S. acquirers of foreign targets from the year
prior to the acquisition to the year following the acquisition. The actual change in excess value (∆EV+1) is measured as
P
the difference between EV+1 and EV-1, and the projected change in excess value (∆EV +1) is calculated as the
P
difference between the projected excess value (EV +1) and EV-1. The projected excess value is based on the pre-
effective market value of the target firm, and represents the excess value the merging firms would have if they were
combined instantaneously in the year prior to the actual acquisition. ∆EV +1 represents the unexplained change in
U

excess value, and is calculated as the difference between ∆EV+1 and EV +1. Excess value is calculated as in Bodnar,
P

Tang and Weintrop (1999), and Denis, Denis and Yost (2002), which represents a variation of the industry multiplier
approach originally developed by Berger and Ofek (1995). The full sample (Panel A) consists of 136 non-financial,
publicly traded U.S. firms that completed an acquisition of a majority ownership stake in a non-financial, publicly
traded foreign target firm over the period 1990-1999. Acquisitions are classified as related cross-border M&As (Panel
B) when the U.S. acquirer and the foreign target report the same two-digit SIC codes in the year t = − 1, and as
unrelated cross-border M&As (Panel C) otherwise. Acquisitions are classified as non-premiere cross-border operations
(Panel D) when the U.S. acquirer already has operations overseas in the year prior to the acquisition, and as premiere
cross-border operations (Panel E) otherwise. N refers to the number of yearly observations in each subsample. The
significance of the mean values and of the difference in means is based on the parametric t-statistics. The significance
of the median values is based on the nonparametric Wilcoxon signed-rank test statistics, while the significance of the
difference in medians is assessed using the nonparametric Wilcoxon rank sum test statistics. The asterisks *, **, and
*** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
Sales Multiples Asset Multiples

Mean Median Mean Median

Panel A: Full Sample (N=136)

∆EV+1 -0.0659 -0.0799 -0.0893** -0.0850
∆EVP+1 -0.0476*** -0.0534 -0.0155 -0.0126

∆EVU+1 -0.0184 -0.0248 -0.0739* -0.0708

Panel B: Related Cross-Border M&As (N=88)
∆EV+1 0.0280 0.0000 -0.0515 -0.0346
∆EVP+1 -0.0466** -0.0589 -0.0144 -0.0112

∆EVU+1 0.0747 0.0614 -0.0371 -0.0239

Panel C: Unrelated Cross-Border M&As (N=48)
∆EV+1 -0.2382*** -0.2153* -0.1587** -0.1725*
∆EVP+1 -0.0492** -0.0431 -0.0173 -0.0158

∆EVU+1 -0.1889** -0.1697 -0.1414** -0.1514

Panel D: Non-Premiere Cross-Border Operations (N=105)
∆EV+1 -0.0383 -0.0548 -0.0408 -0.0434
∆EVP+1 -0.0130 -0.0167 0.0111 0.0062

∆EVU+1 -0.0253 -0.0357 -0.0520 -0.0494

Panel E: Premiere Cross-Border Operations (N=31)
∆EV+1 -0.1600 -0.2093 -0.2537** -0.2320*
∆EVP+1 -0.1647 ***
-0.1823 -0.1055** -0.0817

∆EVU+1 0.0050 -0.0087 -0.1480 -0.1297

38
 Table 8
Relation Between Actual and Projected Changes in Excess Value for U.S. Acquirers of Foreign Targets
This table shows the results for the regressions of the actual change in excess values (∆EV+1) on the projected change
in excess values (∆EV +1) for 136 U.S. acquirers of foreign targets. ∆EV+1 is measured as the difference between
P

EV+1 and EV-1, and ∆EVP+1 is calculated as the difference between the projected excess value (EVP+1) and EV-1.
Excess value is calculated as in Bodnar, Tang and Weintrop (1999), and Denis, Denis and Yost (2002), which
represents a variation of the industry multiplier approach originally developed by Berger and Ofek (1995). Panel A
reports the sales-based results, while Panel B presents the regression results based on asset multiples. In Regression 1,
∆EV +1 is the only predictor variable in the model. An indicator variable for business relatedness of the acquisition, I
P

(Relatedness), equals one when the acquisition is classified as related, and zero otherwise (Regression 2). An indicator
variable for the international involvement status of U.S. acquiring firms, I (Premiere), equals one if the U.S. acquirer
already has operations overseas prior to the acquisition, and zero otherwise (Regression 3). Standard errors are
presented in parentheses, and the two-sided P-values for the null hypothesis that α = 0, β = 1, γ = 0 are reported in
brackets.

Panel A: Sales Multiples

Regression 1 Regression 2 Regression 3

Intercept -0.0292 -0.1769 -0.0170

(0.0400) (0.0645) (0.0442)
[0.4670] [0.0070] [0.7020]

∆EVP+1 0.5071 0.5039 0.4627

(0.1923) (0.1872) (0.2042)
[0.0090] [0.0080] [0.0250]

I (Relatedness) 0.2282
(0.0793)
[0.0050]

I (Premiere) -0.0659
(0.1002)
[0.5120]

Adjusted R2 0.0430 0.0930 0.0390

Panel B: Asset Multiples

Regression 1 Regression 2 Regression 3

Intercept -0.0874 -0.1566 -0.2587

(0.0433) (0.0727) (0.0940)
[0.0460] [0.0330] [0.0070]

∆EVP+1 0.1263 0.1235 -0.0477

(0.2687) (0.2683) (0.2788)
[0.0010] [0.0010] [0.0000]

I (Relatedness) 0.1068
(0.0901)
[0.2380]

I (Premiere) 0.2184
(0.1067)
[0.0430]

Adjusted R2 0.0001 0.0001 0.0180

39
 Table 9
Multivariate Valuation Analysis of Cross-Border Merger and Acquisitions
This table presents the results for the regressions of excess values on an indicator variable denoting international
diversification (INTL_DIV) and a set of control variables, based on ordinary least squares (OLS) estimation and fixed-
effects estimation. Under the fixed-effects specification, we introduce both firm-specific and year-specific fixed effects
(i.e., a two-way fixed effects model). The valuation analysis of the 136 cross-border M&As is carried out in two stages:
in the year prior to the acquisition (Panel A), the sample consists of 6,649 domestic single-segment firm-year
observations and 136 U.S. acquiring firm-year observations. In the year following the acquisition (Panel B), the sample
consists of 5,857 domestic single-segment firm-year observations and 136 U.S. acquiring firm-year observations.
Excess value, calculated as in Bodnar, Tang and Weintrop (1999), and Denis, Denis and Yost (2002), is defined as the
natural logarithm of the ratio of a firm’s market value to its imputed value. The variable INTL_DIV takes the value one
when the Compustat Geographic Segment (CGS) database reports foreign sales for the firm and zero otherwise. The
remaining control variables are relative measures calculated as the difference between the actual value for the firm and
the median value for the domestic single-segment firms matched with the same two-digit SIC code. LTA stands for the
relative natural logarithm of total assets. Leverage is calculated as the relative ratio of book value of total debt to total
assets. EBIT/Sales represents the relative ratio of earnings before interest and tax expenses to sales. R&D/Sales stands
for the relative ratio of research and development expenditures to sales. CAPEX/Sales is the relative ratio of capital
expenditures to sales. SQ_LTA represents the relative square of the natural logarithm of total assets. The regression
coefficients are estimated based on robust standard errors, and the corresponding t-statistics are reported in parentheses
below.
Panel A: Valuation Analysis in the Year Prior to the Acquisition

Sales Multiples Asset Multiples

OLS Fixed Effects OLS Fixed Effects

Intercept 0.0797 -0.4468 0.0875 -0.2159

(6.487) (-12.290) (8.704) (-6.556)

INTL_DIV 0.0782 0.0424 0.1330 -0.0784

(1.178) (0.070) (2.609) (-0.143)

LTA 0.2188 0.2571 0.0212 -0.0617

(23.335) (14.475) (2.969) (-3.836)

Leverage -0.2085 -0.0414 -0.2250 0.0620

(-4.436) (-0.933) (-2.916) (1.540)

EBIT/Sales 0.1100 0.3051 0.3152 0.5374

(1.389) (6.239) (6.156) (12.130)

R&D/Sales 1.4615 0.4698 0.7174 0.2606

(10.491) (6.458) (9.217) (3.954)

CAPEX/Sales 0.6257 0.2221 0.1327 0.0384

(9.582) (7.130) (5.251) (1.362)

SQ_LTA -0.0417 -0.0108 -0.0065 0.0049

(-10.869) (-1.556) (-2.321) (0.772)

Year Controls Included Included

F-statistic 148.22 41.30 22.75 22.56

Adjusted R2 0.2261 0.1526 0.0549 0.0896
No. of Observations 6,785 6,785 6,785 6,785

40
 Table 9 (Continued)

Panel B: Valuation Analysis in the Year Following the Acquisition

Sales Multiples Asset Multiples

OLS Fixed Effects OLS Fixed Effects

Intercept 0.0753 -0.3615 0.0736 -0.0698

(5.431) (-7.980) (6.802) (-1.739)

INTL_DIV -0.0889 -0.0512 -0.0261 -0.0351

(-1.136) (-0.072) (-0.448) (-0.555)

LTA 0.2406 0.2996 0.0349 -0.0248

(25.247) (14.080) (4.749) (-1.315)

Leverage -0.1925 -0.0065 -0.2626 0.0740

(-3.877) (-0.096) (-5.967) (1.239)

EBIT/Sales 0.0021 0.0264 0.1751 0.1911

(0.046) (0.729) (5.740) (5.960)

R&D/Sales 1.5647 0.3538 0.7220 0.0566

(9.478) (3.496) (7.970) (0.631)

CAPEX/Sales 0.5067 0.2421 0.1237 0.0424

(9.855) (6.272) (4.778) (1.238)

SQ_LTA -0.0348 -0.0201 -0.0012 -0.0063

(-8.310) (-2.659) (-0.368) (-0.942)

Year Controls Included Included

F-statistic 153.40 27.03 24.13 8.64

Adjusted R2 0.2470 0.1272 0.0548 0.0445
No. of Observations 5,993 5,993 5,993 5,993

41