Stock returns in mergers and acquisitions

This study analyzes the empirical literature concerning the influence of various factors on shareholder wealth creation in mergers and acquisitions using a multivariate framework.. In t

Strategic Management Journal, Vol 13, 67-84 (1992)

FACTORS INFLUENCING WEALTH CREATION FROM

DEEPAK K DATTA and GEORGE E PINCHES

School of Business, University of Kansas, Lawrence, Kansas, U.S.A

V K NARAYANAN

Graduate School of Management, Rutgers-The State University of New Jersey, Newark, New Jersey, U.S.A

This study analyzes the empirical literature concerning the influence of various factors on

shareholder wealth creation in mergers and acquisitions using a multivariate framework Overall, results indicate that while the target firm’s shareholders gain significantly from mergers and acquisitions, those of the bidding firm do not Findings also indicate that the use of stock financing has a signijicant impact on the wealth of both the target and bidding firms’ shareholders The presence of multiple bidders and the type of acquisition inpuence the bidders’ return, while regulatory changes and tender offers influence the targets’ returns The paper also provides a comparison of our findings with that of previous narrative reviews and discusses their implications from the viewpoint of managers and researchers

The performance implications of mergers and

acquisitions have been of considerable interest

to researchers over the last couple of decades

Recent studies have focused on the market

performance of individual mergers and acqui-

sitions both in the area of strategic management

(e.g Barney, 1988; Chatterjee, 1986; Lubatkin,

1983, 1987; Shelton, 1988; Singh and Montgom-

ery, 1987) and in financial economics (Jarrell,

Brickley and Netter, 1988 and Jensen and

Ruback, 1983 provide narrative reviews of this

literature) In this paper we provide a meta-

analytic synthesis of the findings of studies on

wealth creation or the market peformance of

mergers and acquisitions where ‘market perform-

ance’ or shareholder wealth gains refers to

the stock market appraisal of specific merger

transactions Unlike traditional literature reviews

we employ a multivariate framework and

Key words: wealth creation, mergers and acquisitions,

meta-analysis

regression analysis using observations from 41 studies The results indicate that a select set of factors explain a substantial proportion of the wealth gains to both bidding and target firm shareholders By ‘bidding’ firm we mean the firm initiating the acquisition transaction; the object

of their interest is the ‘target’ firm

The paper is structured as follows First, we provide an overview of the literature on wealth creation in mergers and acquisitions, discussing five key factors which have been hypothesized

to influence shareholder wealth Next we describe the method employed in this study, the studies used in our analysis and the way the variables are coded from the studies Third, we present the results of the regression analysis assessing the influence of various factors on shareholder gains Finally, in the concluding section, we discuss the findings and compare them with those identified in two prior narrative reviews of the literature We also raise important theoretical questions and issues for future research, and discuss their implications for both practice and research in strategic management

014~2095/9uo10067-18$09.00

0 19!Z by John Wiley & Sons, Ltd

Received 27 July 1990

Revised 15 July 1991

68 D K Datta, V K Narayanan and G E Pinches

THEORETICAL OVERVIEW

Measuring wealth creation

Research on wealth creation from mergers and

acquisitions, both in strategic management and

financial economics has been primarily under-

taken using the event study methodology (see

Brown and Warner, 1980; 1985 for a detailed

description of the market and other models used

in such methodology) This approach is based on

the proposition that in an efficient market the

immediate wealth effect reflects the capital

market’s overall unbiased assessment of the

present value of the future benefits of the merger

or acquisition This focus on present value

incorporates both:

1

2

The actual flow of cash or securities incurred

by the bidder during the transaction; and

The expected incremental cash inflow gener-

ated during the integration phase by combining

the operations of the target and bidder (the

expectation being formed by an unbiased

assessment of all available information)

The merger transaction itself may span several

weeks The negotiations may be started before

any public announcement, and the acquisition is

consummated some time after the initial public

announcement (if it is consummated at all)

Although negotiations may be conducted in strict

secrecy, often there is speculation, news begins

to leak, and/or the market begins to anticipate

the acquisition However, the market’s expec-

tations are almost fully formed by the announce-

ment date of the merger or acquisition with

wealth effects being insignificant around the

consummation date (Asquith, 1983; Dodd, 1980)

As a consequence, studies examining the wealth

creation in mergers and acquisitions have typically

employed relatively short event periods surround-

ing the announcement date This is consistent

with both the efficient market hypothesis as well

as available empirical evidence

Factors influencing wealth creation

Evidence from previous narrative reviews indi-

cates that the wealth created in mergers and

acquisitions accrues almost exclusively to the

target firms’ shareholders (Jarrell, Brickley and

Netter, 1988; Jensen and Ruback, 1983) Analysis

of the empirical estimates examined in this study indicate that while the average shareholder gains for target firms in the month of the merger announcement is about 22 percent, gains to bidding firms in the same period are less than one-half of one percent To provide an indication

of the magnitude of wealth creation and variability therein we have plotted on a month-to-month basis the means and variation (mean & one standard deviation) of the prediction errors (the indicator of weath gains) for both bidders and targets during and around the month of the announcement of the merger or acquisition (Figure 1).2 Note that only in month zero (the month of the announcement) is there any indication of substantial wealth increases, and then only for targets Figure 1 also highlights that

Month

Man

-

f one standud Dcviltirn

Figure 1 Mean and standard deviation of prediction errors for bidders and targets around the announcement

month (t = 0)

‘cumulative abnormal returns’ (CARS) has been prevalent in

the literature to connote the wealth effects of an event However, more recently, this has been redefined as ‘prediction

errors’ (PEs) or ‘cumulative prediction errors’ (CPEs) to be

more congruent with what is actually measured

The data for Figure 1 came from previous studies using monthly data which reported wealth effects in and around the announcement month

Wealth Creation from Mergers and Acquisitions 69

there is considerable cross-sectional variability in

the wealth effects for both bidders and targets

This cross-sectional variability has been obscured

in previous narrative reviews

What contributes to this variation in wealth

creation as measured in the observations across

various studies? While the question has been

addressed in both the strategic management and

the financial economics literature, the two have

adopted somewhat different frames of reference

in identifying the sources of shareholder wealth

in mergers and acquisitions The ‘market for

corporate control’ perspective, where mergers

and acquisitions have been viewed as contests

between competing management teams for the

control of corporate entities, is common among

financial economists The proponents of this

perspective argue that the total economic value

created in a merger or acquisition, and its

partitioning between the bidding and target

firms’ shareholders is determined by market

characteristics, including its competitiveness (e.g

number of bidders and regulatory changes affect-

ing the market) On the other hand, strategic

management researchers, while sometimes adopt-

ing the financial economists’ methodological

tools, have typically emphasized factors which

are management controlled For example, based

on the existing literature on diversification

strategies, they have argued that the type of

merger (i.e whether it involves the acquisition

of related or unrelated business units) is an

important determinant of performance Other

factors, such as the bidders’ approach (i.e merger

or tender offer) and the mode of financing, while

also being management controlled, have been

examined primarily in the finance literature from

a market for corporate control perspective

A review of the theoretical and empirical

literature on shareholder wealth creation in

mergers and acquisitions identifies a number of

factors which may explain differences in wealth

creation Primary among these are regulatory

changes, the number of bidders, the bidder’s

approach (i.e merger vs tender offer), the mode

of financing (i.e cash vs stock), and the type

of merger or acquisition (i.e conglomerate vs

non-conglomerate) This study is limited to these

five factors-factors that have been examined in

a sufficient number of previous studies so their

impact (or, lack thereof) can be systematically

assessed It should be emphasized that additional

factors may also impact wealth creation, however, they have not been examined with sufficient frequency in previous research to be considered

in this empirical analysis and ~ynthesis.~

Regulatory changes

Changes in the environment over time, including regulatory changes, play a key role in influencing the selection of a firm’s strategy and also determining the consequences of various strategic decisions In the case of mergers and acquisitions Jarrell and Bradley (1980) and Schipper and Thompson (1983), among others, point to two important regulatory changes, namely, the 1968 Williams Amendment and the 1969 Tax Reform Act Two specific mechanisms have been postu- lated in relation to their impact on wealth creation First, the 1968 Williams Amendment required bidding firms to provide targets adequate time (10 trading days) to evaluate tender offers, thereby allowing additional bidders to enter the process and increasing the competitiveness of the market for corporate control This, in turn, can

be expected to increase the bid price, benefiting the target firms’ shareholders at the expense of the bidding firms’ shareholders Second, the 1969 tax reform disallowed the interest deduction on convertible bonds issued to finance a merger and also taxed negotiable bonds given to the seller as installment payments Since both these regulatory changes imposed additional costs on the bidder, the returns to the bidding firms’ shareholders should be lower in the post-1969 period; the opposite being true for targets

Number of bidders

The extent to which the market for corporate control is competitive is ultimately an empirical question However, in a situation where there are a number of bidders, there is an increase in the level of competitiveness resulting in a negative impact on the stockholder gains of bidding firms

On the other hand, target firms are likely to benefit when there are multiple bids At the extreme, the rivalry among bidding firms might even result in the price of the target being bid

Only factors where we had a minimum of five observations

each of the factors are available from the authors

70 D K Datta, V K Narayanan and G E Pinches

up until the takeover is a zero net present value

(or at times, even a negative) investment for

bidders This argument is the essence of Barney’s

(1988) theoretical analysis and Roll’s (1986)

‘hubris’ hypothesis Ruback’s (1982) case study

of the DuPont-Conoco merger also illustrates

that restricting the number of bidders (in this

case the court ordered withdrawal of Mobil from

the bidding process) can have an impact on

stockholder gains

BidderS approach

The approach used by the bidding firm (i.e

merger or tender offer) has figured fairly promi-

nently in the empirical literature on mergers and

acquisitions (e.g Bradley, 1980; Huang and

Walking, 1987; Kummer and Hoffmeister, 1978;

Kusewitt, 1985) Mergers, by definition, are

negotiated directly with the target firm’s manage-

ment andor the board of directors and approved

by them before going to a shareholder vote On

the other hand, in tender offers, which may be

friendly or unfriendly, the offer is made directly

to the target firm shareholders It is then up to

the shareholders (and not the management) to

decide whether or not to tender their shares to

the bidding firm

There are at least two reasons why the target

firms’ shareholders are likely to benefit more in

tender offers First, the announcement of a

tender offer can alert other firms to the intent

of the bidding firm, thereby attracting other

bidders and initiating a competitive (auction-

type) process for the target firm This increased

competition should result in higher bid premiums

being paid by the bidding firm, benefiting the

target firm at the expense of the bidder Second,

as Bradley, Desai and Kim (1988) suggest,

mergers permit payment of a ‘control premium’

diectly to target firm’s management in the form

of favorable post-acquisition contracts In tender

offers such premiums go directly to the target

firms’ shareholders, who thereby stand to benefit

more from the transaction In either case, tender

offers allow targets to benefit at the expense of

bidders

Mode of payment

In making an acquisition, the bidding or acquiring

firm can choose either cash and stock financing

(or some combination of the two) While the choice depends on how the transaction is nego- tiated, the mode of payment can influence shareholder wealth creation in a variety of ways First, the mode of payment affects the speed, and, with it, the cost of the transaction In stock offers, the bidding firm must obtain the approval

of the Securities and Exchange Commission (SEC) before target shareholders can exchange their shares This process tends to increase the competitiveness of the acquisition market and benefits targets at the expense of the bidders Second, financial theory (e.g Myers and Majluf, 1984) sugg6sts that the issuance of stock is viewed negatively by the capital markets DeAngelo, DeAngelo and Rice (1984) provide empirical evidence with respect to stock issues that is in line with the theoretical position of Myers and Majluf Alternatively, a common stock offer may also lead to wealth transfer from stockholders to bondholders, implying a fall in stock prices (e.g Eger, 1983; Travlos, 1987) Third, unlike stock offers, cash transactions impose an immediate tax liability on the shareholders of the target firm, who, obviously, seek compensation in the form of higher premiums (e.g Franks, Harris and Mayer, 1988; Hayn, 1989) The above arguments suggest that both bidders and targets are likely to be better off in cash-financed rather than in stock-financed transactions

Type of acquisition

The type of acquisition as a factor in wealth creation has been of concern to researchers in strategic management given their interest in linking wealth effects of mergers and acquisitions

to the diversification literature Thus, it has been argued that synergistic benefits through a transfer

of core skills between the bidding and target firms in related acquisitions should result in greater wealth creation than in unrelated or conglomerate transactions (Salter and Weinhold, 1979) In extending Salter and Weinhold’s work, Lubatkin (1983,1987) and Singh and Montgomery (1987) hypothesized three additional mechanisms: merger-related economies of scale, economies of scope, and market power economies, all of which help enhance the total value of related acquisitions However, others have hypothesized there are factors which favor conglomerate acquisitions: cheaper access to capital (Steiner,

Wealth Creation from Mergers and Acquisitions 71

1975); improved income stability, lower bank-

ruptcy probabilities and increased market value

of debt of the combined firm (Higgins and

Schall, 1975; Leontiades, 1986; Lewellen, 1971)

Recently, the relatedness argument has also been

extended to targets: in related acquisitions, the

targets should gain more since such acquisitions

result in greater overall value and most of the

gains in these transactions accrue to the target

firms’ shareholders Singh and Montgomery

(1987) provide some empirical support for this

hypothesis On the whole, this literature views

related acquisitions more favorably than conglom-

erate acquisitions in terms of shareholder wealth

creation for both bidders and targets

In summary, a review of the theoretical and

empirical research suggests a number of factors

which influence shareholder wealth creation in

mergers and acquisitions Table 1 summarizes the

nature of the influence and also the hypothesized

direction for these five factors It is important to

note that previous studies have investigated the

influence of these factors on shareholder wealth

creation within a univariate framework, i.e one

factor at a time The central thrust of this study

is to assess the significance of various factors

simultaneously in a multivariate framework, one

that allows the impact of individual factors to be

examined after partialling out the effects of

others

METHOD

In this study, we utilize a meta-analytic procedure

to estimate the significance of the hypothesized

independent variables on merger returns for both

bidders and targets Meta-analytic procedures

have been extensively used in areas such as

organization theory (e.g Gooding and Wagner,

1985), marketing (e.g Assmus, Farley and

Lehman, 1984), and more recently in strategic

management (Datta and Narayanan, 1989) The

procedure is useful for synthesizing an existing

body of evidence and is particularly appropriate

when there is a substantial body of empirical

evidence already available

In a meta-analysis, empirical studies of a

phenomenon are viewed as a natural unplanned

experiment Since it is often difficult to account

for all relevant factors within a single study, a

number of studies are necessary to shed light on

various aspects of a phenomenon ‘me studies naturally differ along a number of dimensions; hence, the evidence from them is not directly comparable without controlling for interstudy differences Study observations are viewed as estimates in a multi-factor analysis of variance, with the experimental factors corresponding to the interstudy and/or intercase differences This allows us to exploit the empirical evidence available from previous merger and acquisition studies to draw statistically meaningful con- clusions In contrast, narrative reviews adopt ad hoc procedures to cumulate their findings

In this study we adopt a form of meta-analytic procedure called replication analysis (Farley, Lehman and Ryan, 1981) We array the estimates

of merger gains (i.e prediction errors) from previous empirical studies as observations in

a multi-factor natural experiment, with the experimental factors corresponding to the factors hypothesized to influence wealth creation Thus the factor levels for each estimate are identified, and the estimate of wealth created (the prediction error) is the dependent variable Using multiple regression analysis the impact of each factor on the dependent variable is assessed

As in any meta-analytic review, we are forced

to identify the factor levels based on the data available in the existing studies For some factors the data precludes finer refinements of factor levels For example, although the ‘mode of payment’ may vary from ‘all cash’ to ‘all stock’ along a continuum, we could identify only two levels for analysis: cash or stock The factor level

‘mixed mode of payment’ was not generally represented in the studies This is common in meta-analysis, since individual studies are driven less by a planned program and more by the specific researchers’ concerns; consequently, the studies form an unplanned experiment

Key methodological issues to be resolved in a meta-analysis are: sample selection, oper- ationalization of dependent and independent variables, and analytic procedures for drawing conclusions

Sample

The sample consisted of studies identified through

a comprehensive search of articles that have examined the issue of shareholder wealth creation

in mergers and acquisitions using the event study

Table

Regulatory changes

transaction (tender

acquisition (conglomerate

transactions Stimulates

n 3 a

Wealth Creation from Mergers and Acquisitions 73

methodology Using the market model, or other

similar models, researchers have tried to assess

whether mergers and acquisitions result in

increases in shareholder wealth The search was

initiated by a systematic examination of articles

appearing in major journals over the past 15

years.4 Next, additional studies published in

books or other journals were identified from the

references provided in these articles For the

purpose of analysis we used four screens in

selecting studies First, the study had to be based

on either daily or monthly returns Second, we

confined our analysis to studies on acquisitions

in the United States so as not to confound the

value gains due to country specific factors and

cross-national differences Third, we omitted a

small number of studies which focused on unique

or very specific industries (e.g REITs and

banking) Finally, only those studies which

reported wealth effects based on the announce-

ment (and not the outcome or consummation)

date of the merger or acquisition were considered

In our analysis, we examined a total of 41

studies with a total of 409 usable observations

(i.e estimates of wealth creation) for both bidders

and targets, each over four event period^.^ The

Appendix provides a list of the studies used along

with the number of usable observations contained

in each study.) The 41 studies have drawn their

samples from 16 different sources, and have used

different time periods (dating as far back as 1948)

and widely different screens or sampling criteria.6

Although the population of acquisitions from

which the studies have drawn their ‘samples is

finite, the sampling criteria are so different that

observations are treated as being independent of

one another Similarly, while multiple obser-

vations from the same study have been used,

they represent wealth estimates from samples

unrelated to each other.’ However, as with any

authors

We have 75 observations for bidders and 79 for targets for

* Less than 25 percent (10 out of 41) of the studies used

from the authors

For example, Asquith (1983) provides prediction errors

both ‘completed’ and ‘not completed‘ mergers Therefore,

meta-analytic study, the data used suffers from

a ‘publication bias,’ due to the use of studies that have passed a reviewing process

Dependent variable: wealth effects

Wealth creation in mergers and acquisitions has typically been estimated using the technique commonly referred to as an ‘event study.’ The procedure entails analyzing common stock returns

of a firm that is hypothesized to have been affected by a particular ‘information event’ (in this case the announcement of the merger or acquisition).* The return for any period is defined

as the change in the value of the stock plus any dividends (if appropriate) as a percentage of the stock price at the beginning of the period In the absence of new information which impacts the price (and returns) for the stock, the difference between the actual and expected (i.e ‘normal’) returns will, over a large number of firms, be approximately zero However, when unusual new information is present, the actual returns might exceed or be less than expected giving rise to positive or negative prediction errors, respec-

t i ~ e l y ~ In an efficient capital market these prediction errors represent the investment com- munities’ unbiased expectation of the present value of the long-run benefits of the event Thus, they capture both immediate and anticipated longer term impacts of the merger Of course,

as additional information about the merger and its success or failure becomes known, it is assimilated by the market and the value of the firm may be further affected

In this study we operationalize shareholder

wealth creation as the prediction errors (PEs) at

and two for targets) are available from this study

Virtually all studies use the date on which the announcement

the expected return Under the market model the expected return is estimated via

over some relevant time period (often 100 to 200 trading

estimated the returns

74 D K Datta, V K Narayanan and G E Pinches

and around the transaction announcement date

Our analysis focused on wealth creation separ-

ately for both bidders (or acquiring firms) and

targets For studies using daily return data this

involved examining the gains in the (- 10,lO)

period In other words, the total prediction error

over the 21-day period beginning 10 days before

the announcement date, and continuing for

10 days after the announcement date was

examined lo For studies which utilized monthly

data, we used the prediction error during month

zero (i.e the month of the announcement)

Secondary analysis involved employing only daily

data and using the prediction errors in the (-10,

-2), (-1,O) and (1,6) time periods representing

pre-announcement, announcement and post-

announcement periods relative to the specific

date the merger or acquisition was announced

Variables: factors influencing wealth creation

From the available studies, we coded the factor

levels for the five independent variables identified

in the literature review Four were identified

directly: number of bids (single vs multiple),

bidder’s approach (merger vs tender offer), type

of financing (cash vs stock), and type of

acquisition (conglomerate vs non-conglomerate)

The number of studies and/or cases available did

not permit us to code multiple bidder cases in

finer detail nor to consider mixed modes of

financing (i.e cash and stock) Similarly, studies

did not permit us to code the type of acquisition

using more complex classifications such as the

FTC scheme, or Salter and Weinhold’s (1979)

For the fifth variable, namely, regulatory

change, direct identification of factor levels was

not possible However, the empirical work of

Asquith, Bruner and Mullins (1983), Jarrell and

Bradley (1980), and Schipper and Thompson

(1983) suggest the regulatory changes in 1968

and 1969 (Williams Amendment and the Tax

Reform Act) had a significant impact on wealth

typology

’OTwenty-one trading dates is approximately equal to 1

month The window was allowed to vary for the (-10,lO)

window, depending on the data available in a study The

ending day between 5 and 20, providing there are a minimum

from the authors) indicate this variation does not effect the

results

effects in mergers and acquisitions The year of the transaction was used as a proxy for regulatory change and studies were classified based on whether the mergers and acquisitions in the study pertained to the 1969 and after period, or not Control variables: methodological artifacts Since studies differed along methodological dimensions, we also controlled for four methodo- logical artifacts: the source of the merger sample, type of sample employed, the kind of data employed, and the outcome of the proposed merger or acquisition

The first control variable relates to whether or not the study sample was drawn from the FTC

data base (the FTC merger series was last published in 1981, covering mergers through 1979) Since the FTC only listed mergers in the manufacturing sector that are valued in excess

of $10 million, samples selected from the FTC data base are biased in favor of larger mergers Second, the criteria used in selecting the sample has also differed across studies Some studies have used only ‘clean’ samples (i.e excluding mergers where the announcement occurred along with other significant firm-specific announce- ments) On the other hand, in studies which did not systematically consider the possibility of other announcements (e.g earnings releases, dividend announcements, capital expenditures, etc.), the samples are contaminated We distinguish between ‘clean’ and ‘contaminated’ samples Third, as previously mentioned, studies have differed on the kind of data used in assessing wealth effects Some studies, especially the earlier ones, used monthly return data while the more recent studies have predominantly used daily return data We distinguish between whether returns data is daily or monthly Finally, we control for ‘outcome’, i.e whether the trans- actions were completed or not completed ‘Com- pleted’ acquisitions are those which were eventu- ally consummated while ‘not completed’ relates

to those which were not Theoretically, there is

no reason to suspect that the market can anticipate the outcome at the time of the announcement; however, we included this factor

to enable us to meaningfully compare our conclusions with previous narrative reviews The above variables, namely, the source of merger sample (FTC or other), type of sample

Wealth Creation from Mergers and Acquisitions 75

(contaminated or clean), kind of data (daily

or monthly) and outcome (completed or not

completed), were included to control for potential

variability across studies in estimates of wealth

gains arising from methodological artifacts rather

than substantive factors

Analysis

To assess the impact of the five independent

factors on the wealth effects in mergers and

acquisitions, we employed a multiple regression

approach using dummy variables The following

regression model was estimated:

PEW = f(TIME, MBID, TEND, MERG, CASH,

STOCK, CONG, NCONG, FT'C, CONTAM,

DAILY, OUTCOME)

where:

PEW

TIME

MBID

TEND

MERG

CASH

STOCK

CONG

NCONG

FTC

CONTAM

DAILY

= Prediction error for event period

= Time period of merger (1 = 1969

and after, 0 = Pre-1969)

= Number of bidders (1 = Multiple

bidders, 0 = Otherwise)

= Tender offer (1 = Tender offer,

0 = Otherwise)

= Merger (1 = Merger, 0 =

Otherwise)

= Cash-financed transaction (1 =

Cash, 0 = Otherwise)

= Stock-financed transaction (1 =

Stock, 0 = Otherwise)

= Unrelated or conglomerate merg-

ers (1 = Conglomerate, 0 =

Otherwise)

= Related or non-conglomerate

mergers (1 = Non-conglomerate, 0

= Otherwise)

= Sample drawn from FTC data

base (1 = FTC, 0 = Other sources)

= Type of sample (1 = Contami-

nated, 0 = Non-contaminated)

= Kind of data (1 = Daily, 0 =

Monthly)

W

OUTCOME = Outcome of merger (1 = Not

completed, 0 = Completed)

the effects of the other independent and control variables This is a form of multi-factor analysis

of variance with the interaction terms suppressed; the regression intecept is the main effect and each zerolone variable adds to or subtracts from the main effect

RESULTS

Our primary focus was on shareholder wealth creation for bidders and targets as assessed by examining the PEs in the combined (-10,lO) time period (for daily data) and month zero (for monthly data) Based on 75 observations for bidders and 79 for targets, the mean PEs were 0.388 percent for bidders and 21.814 percent for targets with standard deviations of 2.105 and 7.256, respectively (see Table 2).Thus, bidders had, on average, a gain of less than half of one percent when the merger was announced, while target firms' shareholders experienced over a 20 percent increase in value The overall results for targets are consistent with those arrived at by

Table 2 Regression analysis of factors influencing

wealth creation as measured by prediction errors

Bidders Targets

Mean prediction errors 0.388 21.814"'

Standard deviation 2.105 7.256

Regression analysis"

Factor MBID (1 = multiple bidders) -1.133+ 3.176 TIME (1 = 1969 and after) -0.547 4.913' TEND (1 = tender offer) -0.310 8.060"'

CASH (1 = cash-financed) 0.909 3.159 STOCK (1 = stock-financed) -2.738"L5.606' NCONG (1 = non-conglomerate) 1.789' 2.488 FTC (1 = sample drawn from FIT) 0.024 -0.443 CONTAM (1 = contaminated) 1.121t 1.539 OUTCOME (1 = not completed) -0.980 0.285

Intercept 0.565 14.743"

CONG (1 = conglomerate) -0.950 -1.866

DAILY (1 = daily data) -0.786 2.546

This procedure allows us to test the impact of

each independent variable after controlling for

76 D K Datta, V K Narayanan and G E Pinches

Jarrell et af (1988) and Jensen and Ruback

(1983); however, our results for bidders are, on

net, more pessimistic than theirs In addition, an

examination of the standard deviations indicates

there is considerable variability in the prediction

errors across studies, a fact that is obscured in

the previous reviews Multiple regression analysis

was employed to assess the extent to which the

five independent and four control factors explain

the variability in the bidder and target gains

Multiple regression analyses

Table 2 also presents the results of the multiple

regression analyses As shown, the set of indepen-

dent variables explained a significant amount of

variation in the PEs for both bidders and

targets (with adjusted R2 being 0.414 and 0.373,

respectively") One factor, namely, the mode of

payment (STOCK) was significantly related to

changes in shareholder wealth for both bidders

and targets Its direction of influence is consistent

with the theoretical predictions identified in

Table 1 This finding substantiates that both

bidders and targets are worse off in stock

transactions

In addition, the number of bids (MBID)

and the type of acquisition (NCONG) were

significantly related to changes in shareholder

wealth for bidders Multiple bids had a negative

effect on the wealth of bidding firms, while non-

conglomerate mergers provided higher returns

for bidders All these results were consistent with

the theoretical arguments presented earlier Only

one control variable was significant in the case

of bidders: the use of contaminated data provided

a higher estimate of returns than where no other

firm-specific announcements occurred around the

announcement of the merger or acquisition

As hypothesized, the target firm's shareholders

registered higher gains in transactions that took

place during or after 1969 (as indicated by

TIME) Also, as hypothesized, target firm

shareholders experienced substantial wealth gains

when tender offers were employed The other key

independent variable, namely, type of acquisition

The possible existence of multicollinearity was assessed by

which is explained by the other independent variables Results

available from the authors indicate multicollinearity is not a

problem

(i.e conglomerate vs non-conglomerate), was not significantly related to wealth creation for targets None of the control variables, i.e the source of the sample (FTC), type of sample (CONTAM), kind of data (DAILY) or outcome

of the transaction (OUTCOME) contributed significantly to the variation in estimates of shareholder wealth effects for targets

Secondary analysis Secondary analysis was undertaken to obtain a better understanding of the influence of the hypothesized factors on wealth creation in the pre- announcement (- 10, -2), announcement (- 1,O) and post-announcement (1,6) periods.'2 Table 3 provides the means and standard deviations of the prediction errors for each of these periods for both bidding and target firms For bidders the mean PE for the 2-day announcement period was 0.167 percent compared to a mean of PE of 0.388 percent (from Table 2) for the combined sample using both monthly and daily data Similarly, for targets, the mean PE for the 2-day announcement period was 14.596 percent vs 21.814 percent for the combined daily and monthly sample This is consistent with Jarrell et

al's (1988) results which indicate that as the time period for estimating returns is extended over

more than the 2-day period (-l,O), the estimates

are higher

We also estimated separate regression estimates for the 2-day announcement period (-l,O), as well as pre-announcement (- 10,-2), and post- announcement (1,6) periods In the regression model the type of data (DAILY) control variable was not used since only daily data were relevant Also, because of insufficient data, we could not estimate the effects of MBID in these regressions The results of these secondary analyses are also presented in Table 3

A detailed examination of Table 3 points to four important findings First, for both bidders and targets, stock transactions depressed the gains significantly around the announcement date, a result that is consistent with that of the

period could vary between -18 and -5 and the ending day

in the (1,6) period between 5 and 18 Statistical tests available from the authors indicate this variation does not effect the results