Two essays on transparency in corporate finance Mergers and repurchases

Two essays on transparency in corporate finance Mergers and repurchases

THE FLORIDA STATE UNIVERSITY

COLLEGE OF BUSINESS

TWO ESSAYS ON TRANSPARENCY IN CORPORATE FINANCE:

MERGERS AND REPURCHASES

By Yung Ling Lo

A Dissertation submitted to the Department of Finance

in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded:

Summer Semester, 2007

Copyright © 2007 Name All Rights Reserved

UMI Number: 3282641

3282641 2007

Copyright 2007 by

Lo, Yung Ling

UMI Microform Copyright

All rights reserved This microform edition is protected against unauthorized copying under Title 17, United States Code.

ProQuest Information and Learning Company

300 North Zeeb Road P.O Box 1346 Ann Arbor, MI 48106-1346 All rights reserved.

by ProQuest Information and Learning Company

The members of the Committee approve the dissertation of Yung Ling Lo defended on May16th,

2007

James Ang Professor Directing Dissertation

Bruce Billings Outside Committee Member

William Christiansen Committee Member

Ying-Mei Cheng Committee Member

Approved:

William Christiansen, Chair, Finance Department

Caryn L Beck-Dudley, Dean, College of Business

The Office of Graduate Studies has verified and approved the above named committee members

ACK NOWLEDGEMENTS

I appreciate all the suggestions and comments from my dissertation committee, which consists of James Ang, William Christiansen, Bruce Billings, and Yingmei Cheng; from James Doran; and from finance faculty members at Florida State University, University of New Hampshire,

Western Kentucky University, Central Michigan University, and Bowling Green University I also thank the Dissertation Research Grant Committee for the Dissertation Grant in 2006-2007

TABLE OF CONTENTS

LIST OF TABLES……….v

LIST OG FIGURES……….………ix

ABSTRACT………x

ESSAY 1: ESSAY ON TRANSPARENCY IN THE MERGERS MARKET 1

INTRODUCTION 1

LITERATURE REVIEW 3

HYPOTHESES 5

METHODOLOGY 12

RESULTS 28

CONCLUSIONS 37

ESSAY TWO: ESSAY ON MARKET-TIMING ABILITY OF LOW TRANSPARENCY FIRMS THROUGH STOCK REPURCHASE 38

INTRODUCTION 38

LITERNATURE REVIEW 39

HYPOTHESES 42

METHODOLOGY 55

RESULTS 71

CONCLUSIONS 82

APPENDIX A: PROBABILITY OF ANNOUNCING STOCK REPURCHASE 167

APPENDIX B: SIZE OF ANNOUNCED AND ACTUAL REPURCHASE 168

APPENDIX C: THE PROBABILITY OF CARRYING OUT REPURCHASE AND TENDER OFFER 169

APPENDIX D: MARKET REACTION OF LT FIRMS VERSUS HT FIRMS 170

REFERENCES 171

BIOGRAPHICAL SKETCH 179

LIST OF TABLES

TABLE 1: CHARACTERISTICS OF THE TARGETS BASED ON TOTAL

TRANSPARENCY 84

TABLE 2: PROBIT ANALYSIS OF BIDDING 86

TABLE 3: UNIVARIATE ANALYSIS OF ACQUISITION PREMIUM 88

TABLE 4: MULTIVARIATE ANALYSIS ACQUISITION PREMIUM 89

TABLE 5: PROBABILITY OF STOCK PAYMENT 92

TABLE 6: UNIVARIATE ANALYSIS ON ACQUIRER’S MARKET REACTION 94

TABLE 7: OLS ANALYSIS ON ACQUIRER’S MARKET REACTION 95

TABLE 8: LONG-TERM BHARS OF THE MERGED FIRM 96

TABLE 9: LONG-TERM CARS OF THE MERGED FIRM 97

TABLE 10: CALENDAR-TIME APPROACH FACTOR ANALYSIS 98

TABLE 11: REGRESSION ANALYSIS OF POST-ACQUISITION LONG-TERM BUY-AND-HOLD ABNORMAL RETURNS 99

TABLE 12: REGRESSION ANALYSIS OF POST-ACQUISITION LONG-TERM CUMULATIVE ABNORMAL RETURNS 101

TABLE 13 MARKET REACTION OF ACQUIRERS AND LONG-TERM PERFORMANCE OF MERGED FIRMS BASED ON THE NEGOTIATING POWER OF THE TARGETS 104

TABLE 14: CALENDAR-TIME APPROACH FACTOR ANALYSIS 105

TABLE 15: MARKET REACTION OF ACQUIRERS AFTER CONTROLLING FOR CHARACTERISTICS OF THE TARGETS 107

TABLE 16: LONG-TERM PERFORMANCE AFTER CONTROLLING FOR

NEGOTIATING POWER, BOOK-TO-MARKET, AND SIZE OF THE TARGETS 108

TABLE 17: CHARACTERISTICS OF THE OPEN MARKET REPURCHASE FIRMS BASED ON TOTAL TRANSPARENCY 109

TABLE 18: PROBABILITY OF OPEN MARKET REPURCHASE ANNOUNCEMENT 110

TABLE 19: MARKET REACTION OF OPEN MARKET REPURCHASE FIRMS 112

TABLE 20: SIZE OF THE REPURCHASE ANNOUNCEMENT 114

TABLE 21: PROBABILITY OF ACTUAL REPURCHASE RELATIVE TO

ANNOUNCED REPURCHASE 116

TABLE 22: SIZE OF THE ACTUAL REPURCHASE RELATIVE TO MARKET VALUE

OF EQUITY PRIOR TO ANNOUNCEMENT 119

TABLE 23: PROBABILITY OF ANNOUNCING TENDER OFFER VERSUS OPEN MARKET REPURCHASE 122

TABLE 24: LONG-TERM CARS OF THE OPEN MARKET REPURCHASE FIRM 126

TABLE 25: LONG-TERM BHARS OF THE OPEN MARKET REPURCHASE FIRM 128

TABLE 26: CALENDAR-TIME APPROACH FACTOR ANALYSIS OF COMPLETED OPEN MARKET REPURCHASE FIRMS 130

TABLE 27: CALENDAR-TIME APPROACH FACTOR ANALYSIS OF CANCELLED

OPEN MARKET REPURCHASE FIRMS 132

TABLE 28: LONG-TERM PERFORMANCE OF OPEN MARKET REPURCHASE FIRMS 133

TABLE 29: CHARACTERISTICS OF THE DUTCH AUCTION FIRMS BASED ON TOTAL TRANSPARENCY 137

TABLE 30: PROBABILITY OF DUTCH AUCTION ANNOUNCEMENT 138

TABLE 31: MARKET REACTION OF DUTCH AUCTION FIRMS 139

TABLE 32: LONG-TERM CARS OF THE DUTCH AUCTION FIRM 140

TABLE 33: LONG-TERM BHARS OF THE DUTCH AUCTION FIRM 141

TABLE 34: CALENDAR-TIME APPROACH FACTOR ANALYSIS OF DUTCH AUCTION FIRMS 142

TABLE 35: LONG-TERM PERFORMANCE OF DUTCH AUCTION FIRMS 143

BASED ON THREE-FACTOR MODEL RATS PROCEDURE 143

TABLE 36: CHARACTERISTICS OF THE FIXED-PRICE TENDER OFFER FIRMS BASED ON TOTAL TRANSPARENCY 145

TABLE 37: PROBABILITY OF FIXED-PRICE TENDER OFFER ANNOUNCEMENT 146

TABLE 38: MARKET REACTION OF FIXED-PRICE TENDER OFFER FIRMS 148

TABLE 39: LONG-TERM CARS OF THE FIXED-PRICE TENDER OFFER FIRMS 150 TABLE 40: LONG-TERM BHARS OF THE FIXED-PRICE TENDER OFFER FIRMS 152

TABLE 41: CALENDAR-TIME APPROACH FACTOR ANALYSIS OF COMPLETED FIXED-PRICE TENDER OFFER FIRMS 154

TABLE 42: CALENDAR-TIME APPROACH FACTOR ANALYSIS OF CANCELLED FIXED-PRICE TENDER OFFER FIRMS 156

TABLE 43: LONG-TERM PERFORMANCE OF FIXED-PRICE TENDER FIRMS

BASED ON THREE-FACTOR MODEL RATS PROCEDURE 157

LIST OF FIGURES

FIGURE 1: THE PROBABILITY OF RECEIVING OFFER 161

FIGURE 2: SUMMARY OF LOW TRANSPARENCY FIRMS 162

FIGURE 3: SUMMARY OF HIGH TRANSPARENCY FIRMS 163

FIGURE 4: PROBABILITY OF ANNOUNCING OPEN MARKET REPURCHASE 164

FIGURE 5: PROBABILITY OF ANNOUNCING DUTCH AUCTION 165

FIGURE 6: PROBABILITY OF ANNOUNCING FIXED-PRICE TENDER OFFER 166

ABSTRACT

In this dissertation, I examine whether low transparency (LT) firms have more timing opportunities and are able to earn higher market-timing profits than high transparency (HT) firms in two separate essays: 1) Essay on Transparency in Mergers Market and 2) Essay on Market-Timing Ability of Low Transparency Firms through Stock Repurchase Activities

market-LT firms have more information asymmetry problems (Diamond and Verrechia (1991)), while information asymmetry can make stock valuation of LT firms more difficult1 As some investors underestimate the value of the LT firms, others may overestimate the value of the LT firms Therefore, LT firms can have higher price dispersion and larger magnitude of mispricing

In addition, the risk-averse investors will discount the stock of LT firms when the perceived information risk increases Therefore, the magnitudes of price dispersion and price discount are both positively correlated with the level of information asymmetry problems In essay one, I examine whether the combination of price dispersion and price discount caused by higher

information asymmetry will allow LT firms to have more market-timing opportunities and to demand higher acquisition premiums in the acquisition market In essay two, I examine whether the price discount and larger magnitude of mispricing provide LT firms with more market-timing opportunities and higher profits through stock repurchase activities

Based on my analysis in essay one, I find that LT firms in fact are more likely to receive takeover offers during cold acquisition season However, LT targets without enough negotiating power are often acquired at discount, while LT targets with enough negotiating power can

demand higher acquisition premiums Most importantly, the long-term performance indicates that buying (selling) high negotiating power, high book-to-market, and large HT (LT) targets can earn significant long-term profits Therefore, while majority of existing studies have found HT

to be more advantageous, I find poor performers, proxied by book-to-market ratio, with enough negotiating power can sell the firms at higher value than their true intrinsic value by remaining less transparent

Based on the analysis in essay two, LT firms are more likely to announce open market repurchases and fixed-price tender offers than HT firms In addition, LT firms are more likely to

1

Mensah et al (2004) find high corporate disclosure can reduce analyst forecast dispersion

announce larger repurchase programs and to carry out the announced repurchase programs This indicates less immediate and less complete price adjustment at announcement for LT firms and therefore more market-timing opportunities and profits for the LT firms The long-term

performance analysis confirms the hypothesis that LT firms are more successful at timing the repurchase of undervalued stock Results based the open market repurchase sample are the strongest of all

ESSAY 1: ESSAY ON TRANSPARENCY IN THE MERGERS MARKET

transparency can be more beneficial to the shareholders in certain circumstances The rationale behind the basic idea is that low transparency can generate greater uncertainty and thus, higher volatility, while higher volatility can create a potentially valuable timing option

In this paper, I examine the value of the timing option of LT targets in the acquisition market More specifically, I investigate whether LT firms have more market-timing

opportunities and profits in the acquisition market Low transparency can be advantageous if the

LT targets are able to use the information asymmetry problem to attract acquirers’ interests, to receive a higher number of bids, and more importantly, to receive higher acquisition premiums

In my study, I assume that the investors are risk averse and are aware of the information

asymmetry problem of the LT firms As a result, investors will discount the price of LT targets

as a form of information compensation discount Acquirers that mistake the price discount as a bargain will be more likely to bid on LT targets On the other hand, since HT firms have little or

no information asymmetry problem, their prices are more likely to be traded at or close to their fair market value Therefore, when the stock price of HT firms deviates from its intrinsic value, the magnitude of mispricing will be smaller than that of LT firms Consequently, such small magnitude of mispricing predicts fewer market-timing opportunities and smaller market-timing profits for HT firms as potential targets

A common question may be raised here Why would acquirers overpay during

acquisition? Based on the existing literature, acquirer managers may overpay during acquisition when they are affected by hubris (Roll (1986)), empire-building mentality (Jensen (1986)), or free cash flow problems during acquisition If LT targets are able to use the information

asymmetry problem and larger scaled mispricing to take advantage of such acquirer managers,

by accepting offers that are higher than the true value and rejecting offers below the true value as

in optimal option exercise, they should be able to receive higher final acquisition premiums than

HT targets do In addition, such findings will indicate that the information asymmetry problem can facilitate wealth transfer from the acquirer shareholders to the target shareholders More importantly, the result will also explain why some managers choose to remain low transparency when majority studies in the transparency literature have found high transparency to be more advantageous

This study provides several contributions First, and most importantly, I examine if low transparency does in fact provide managers with more market-timing opportunities and allow LT targets to negotiate higher acquisition premiums Second, I examine if LT firms are more likely

to receive stock as the primary method of payment Based on optimal contract theory (Hansen (1987)), when the target has information asymmetry problems, it is difficult for the acquirer to accurately estimate the fair market value of the LT firm This information asymmetry problem predicts that the acquirers are more likely to experience unanticipated negative shocks post acquisition Therefore, it will be crucial for the acquirer to use stock payment when acquiring

LT targets, while the stock payment in this case serves as a risk-sharing tactic in the acquisition market It is important to note that the market-timing hypotheses in this study do not require or assume acquirers to be irrational However, when there is information asymmetry problems, the less informed acquirers may make sub-optimal decisions by overpaying Therefore, rational acquirers that are aware of the information asymmetry problem will prefer stock payment when acquiring targets at the estimated fair market value Third, I examine if the market reacts

differently to the same payment method when the targets have different transparency levels Fourth, I examine if the merged firms consisting of LT targets do underperform in the long run because LT targets are able to demand higher acquisition premiums

In addition, while many may draw conclusions based on the public status of the targets, since private targets in general have higher information asymmetry problems than HT targets do, such assumption that private targets and LT targets always have the same firm characteristics may be unrealistic Therefore, to measure corporate transparency, I use more direct transparency measures provided by the Association for Investment Management and Research (AIMR) report

in this study To provide more robust results, I use the analyst forecast dispersion obtained from

the IBES2 Summary Tapes, scaled by the price and absolute value of earnings, as additional proxies of corporate transparency measures By doing so, the dispersion measures allow me to include mergers and acquisitions from January 1, 1980 to December 31, 2005 to ensure that the results in this study are consistent through time Furthermore, instead of only including

successful mergers as in many mergers and acquisitions studies, I include unsuccessful mergers

in the hypotheses as well when applicable

LITERATURE REVIEW

In the corporate transparency literature, several researchers have identified the

advantages of high transparency while very few have acknowledged advantages of low

transparency Even though existing studies in transparency have found advantages of higher transparency to far outweigh the advantages of lower transparency, managerial choice of low transparency in the corporate world indicates further examination of potential benefits of low transparency is justified Lang and Lundholm (1993) find that the level of transparency is

positively related to firm size and firm performance Corporate transparency is also found to have a positive relationship with profitability in the US market (Singhvi and Desai (1971)); with responsiveness to earning (Price (1998)); and with stock performance, institutional ownership, analyst following, and stock liquidity (Healy, Hutton, and Palepu (1999)) In addition, Welker (1995) and Leuz and Verrecchia (2000) also find the increased disclosure can reduce bid-ask spreads and increase share turnover Amihud and Mendelsohn (1980), Callahan, Lee, and Yohn (1997) suggest that increased liquidity can lower a firm’s cost of capital Heflin, Shaw, and Wild (2005) find that higher quality accounting disclosures, measured by total disclosure rating, can increase market liquidity Sengupta (1998) finds that a higher quality of disclosure is related to lower effective interest costs in debt issuances Botosan (1997) finds a negative (no) relation between the cost of equity and disclosure when the firm has low (high) analyst following

Furthermore, Verrecchia (2001) suggests that increasing disclosure can reduce investment

inefficiency by reducing information asymmetry and agency problems Beatty and Ritter (1986)) believe that disclosure can reduce the ex-ante uncertainty and ex-ante underpricing Durnev and Kim (2002) find that firms with more profitable investment opportunities (proxied by higher Q),

2

Institutional Brokers Estimate System

more concentrated ownership, and rely more on external financing tend to disclose more and have better governance Last but not least, Mensah et al (2004) find that increase in disclosure can reduce analyst forecast dispersion

However, high transparency is not always more advantageous than low transparency Botosan and Plumlee (2002) find that while cost of equity is negatively related to annual

transparency rating, it is also positively related to timeliness of disclosure, such as quarterly report disclosure rating Therefore, timely disclosure can increase the cost of equity capital Bushee and Noe (1999) argue that increasing total disclosure can potentially reduce future stock volatility when attracting long-term investors However, when transient institutions trade on short-term earnings news, the increase in volatility induced by transient investors will completely offset the reduction in volatility brought by long-term investors Therefore, timely disclosure attracts transient investors and increases stock volatility In addition, Verrecchia (1983), Healy, Hutton, and Palepu (1999), Darrough and Stoughton (1990), and Wagenhofer (1990) argue that disclosure can also be costly to a firm when it reveals information to competitors, which may cause the firm to lose its competitive advantage or bargaining power Therefore, based on the existing literature in corporate transparency, the verdict on the pros and cons of greater corporate transparency is not yet settled or definitive This allows some firms to choose high transparency, while other firms to choose low transparency for different reasons and advantages

Mergers and acquisitions is a promising area in which to conduct empirical tests of

whether there is a rational basis for the choice of low transparency in certain situations

Acquirers are known to misprice targets during acquisitions and make value-reducing decisions Roll (1986) argues that managers are often overconfident and over-optimistic Acquirers tend to overestimate the potential synergistic gains and are likely to overpay while making acquisitions Numerous studies have found consistent evidence indicating that acquirers often overpay during acquisitions Asquith, Bruner, and Mullins (1983) find that bidders on average lose Varaiya (1985) finds that when there are rival bidders, the successful bidders’ loss is significantly greater

In addition, Loughran and Vijh (1997) find that acquirers in stock mergers underperform in the long run Consequently, if LT targets are able to utilize the advantage of the price discount and price dispersion caused by information asymmetry, LT targets may be able to time the market more successfully and to receive higher acquisition premiums, provided with enough negotiating power to reject offers below their reservation value

To summarize, the purpose of this study is to provide a more extensive and more direct examination of LT targets to determine (1) if LT targets are more likely to receive bids or attract acquirers’ interest, (2) if LT targets are more likely to receive higher acquisition premiums than

HT targets, (3) if acquirers are more likely to use stock as the primary payment method when acquiring LT targets than when acquiring HT targets, (4) if the market reacts more negatively to the acquirer’s merger announcement when the target has low transparency, (5) and most

importantly, if LT targets are in general more successful in market timing based on the long-term performance of the merged firms

LT firms On the other hand, since HT firms have little or no information asymmetry problem, the stock of HT firms is more likely to be traded at or close to the equilibrium The equilibrium pricing of HT targets predicts fewer ex ante profitable opportunities for acquirers As a result, acquirers that mistake price discount of LT targets as a bargain will be more likely to bid on LT firms than on HT firms

Acquirer managers can often make value-reducing decisions when they are affected by hubris (Roll (1986)), empire-building mentality (Jensen (1986)), or free cash flow problems Acquirer managers who are affected by hubris (Roll (1986)) tend to over-estimate their ability,

3

While some may argue that it does not seem rational for target managers to choose low transparency and accept the price discount caused by the information asymmetry problem, managers may still choose low transparency if the estimated benefit of doing so is greater than the cost Furthermore, the undervaluation may not always be crucial to the managers except when the firm needs to raise external capital, when the insiders need to sell their share holdings,

or when the manager needs to exercise his or her stock options Moreover, the undervaluation of the low

transparency firms may be acceptable when releasing proprietary information to the outside investor can reduce the company’s competitive advantage

overestimate the potential synergy in the merger, and overestimate the fair market value of the firm; therefore, they are more likely to ignore the information risk and to overpay during

acquisition On the other hand, acquirer managers who are affected by empire-building

mentality are more likely to make bad acquisition decisions not because they over-estimate their ability, but because they tend to knowingly invest in negative NPV projects solely to expand the size of the firm for their own fame and image, or simply because their compensation is tied to the size of the firm In addition, managers may also make bad decisions when the firms have excess free cash flows Excess cash flows can increase agency problems, provide managers with more opportunities to waste the money, and cause managers to invest in negative NPV projects Therefore, acquirer managers who are affected by hubris, empire-building mentality, or free cash flow problems are more likely to make bad decisions and to overpay during acquisitions More importantly, acquirers that mistake the information discount as a bargain will be more likely to bid on LT targets and be willing to pay higher prices

In my next hypothesis, I examine if LT targets are able to receive higher acquisition premiums than HT targets under certain circumstances Since LT targets have higher price dispersion, the price dispersion predicts that LT targets should receive some offers at the lower end of the price range as well as some at the higher end of the price range Therefore, when the

LT target has higher price dispersion, the negotiating power of the target will play a key role on the acquisition premium that the target will receive The higher price dispersion of the LT

targets predicts that those targets with enough negotiating power to reject offers below their reservation price should receive higher acquisition premium, while those without enough

negotiating power may be acquired at discount when there is no competition in the acquisition market

valuation more difficult to outside investors, the LT target should have higher price dispersion than the HT target In other words, some investors may overestimate the value of the LT target while others may underestimate the value of the LT target as a result of information asymmetry problems On the other hand, the HT firm has less or no information asymmetry problem Therefore, the HT target will have lower price dispersion As a result, the price range of the two targets is as the following:

In the analysis, since higher competition in the acquisition market will cause acquirer managers to overpay, I also make sure to control for competition in the acquisition market when examining the acquisition premium in this study In addition, related mergers can occur when the acquirer and the target are in the same industry Since related mergers can create more synergy and therefore higher acquisition premiums, an industry dummy is created in the analysis

to control for such potential synergistic gains Furthermore, when the acquirer and the target are

in the same industry, whether the target has HT or LT, information asymmetry becomes less important from the acquirer’s perspective If reduced information asymmetry in related mergers can reduce the market-timing ability of the LT target, acquirers should be less likely to overpay

in related mergers than in unrelated mergers when acquiring LT targets To determine if

acquirers in related mergers are able to reduce the information asymmetry problem of LT targets and therefore are less likely to overpay, I create an interaction term between the industry dummy and the LT dummy in the analysis as well

H3: Low transparency targets are more likely to receive stock payment during acquisition The importance of stock payment increases as the relative size between the target and the acquirer increases

When the target is too small to have a significant impact on the merged firm’s merger operation, the acquirer may have no specific preference over the different payment methods because the stake for the acquirer is small Since the information asymmetry problem

post-of LT targets predicts a higher probability that acquirers may not discover all crucial information during negotiation, acquirers that have to acquire LT targets at the full estimated fair market value should be more likely to use stock payment as a risk-sharing tactic between the acquirer shareholders and target shareholders In this case, the stock payment is used to protect the

acquirer shareholders from the potential problems that may occur after the acquisition4 In addition, as the relative size between the target and the acquirer increases, the stakes involved in acquisition increase as well, from the acquirer’s perspective Therefore, the risk-sharing function

of the stock payment becomes more important when the target has LT and when the target is large enough to provide a significant impact on the merged firm In addition, while the

acquirer’s undervalued stock may discourage the acquirer managers from using stock as the primary method of payment, stock payment may still be preferred by the acquirers for as long as the target value plus synergy is greater than the acquirer’s stock price per share times the number

of shares paid in acquisition, whether the stock is overvalued or undervalued Lastly, the

importance of the risk-sharing function diminishes when the LT target is acquired at discount Therefore, the probability of stock payment should also decrease when acquirers are able to acquire LT targets at discount, while everything else being equal

On the other hand, since HT targets are more likely to be traded at or close to their fair market value, acquirers are less likely to acquire HT targets at discount In addition, since HT targets have less information asymmetry problems, the risk-sharing function of stock payment becomes less important when the target has high transparency In this case, the method of

payment may simply be a function of the stock valuation of the acquirer If the acquirer believes that its stock is overvalued, the acquirer will choose stock as the primary method of payment

Otherwise, the acquirer may prefer to use cash payment Therefore, when the transparency level

of the target firm is not an issue, the acquirer should be more willing to use cash as the primary method of payment when acquiring the target at the fair market value

H4a: If low transparency targets can efficiently time the market, acquirers of low

transparency targets should receive worse market reaction for overpaying during

acquisition

A target firm may be able to time the acquisition market when the target (a) knows its fair market value and (b) has the ability to reject offers that are below the target’s reservation price and accept those that are at or above the target’s reservation price Theoretically, when the target

is acquired at discount or fair market value, the market is more likely to react positively to the acquirer’s announcement5 On the other hand, if LT does allow the manager of LT firms to demand higher acquisition premiums, the acquirers of LT target should receive worse market reaction than the acquirers of HT targets Since the price dispersion of LT targets predicts that

LT targets without enough negotiating power to reject offers below their reservation price are likely to be acquired at discount, negotiating power of the target, more importantly of the LT target, is used in the multivariate analysis6 In addition, since it may not be clear if the market will react only to the raw unadjusted acquisition premium relative to the price prior to the

negotiation or the industry-adjusted premium, both unadjusted and adjusted premiums are

examined in the analysis Furthermore, since acquirers in related mergers can mitigate the information asymmetry problems of the LT targets, acquirers in related mergers less likely to overpay As a result, acquirers that acquire LT targets in related mergers should receive more positive market reaction than acquirers that acquire LT targets in unrelated mergers

H4b: When acquiring low transparency targets, acquirers that choose stock as the primary method of payment will receive more positive market reaction at announcement than acquirers that use cash as the primary method of payment

The choice of payment should have different impacts on the acquirer when the targets have different transparency levels When the target firm has more information asymmetry problems, valuation of the target can be more difficult As a result, optimal contract theory predicts that using stock as the method of payment can alleviate the information disadvantage that the acquirer faces when the target firm has more information asymmetry problems (Hansen (1987)) Note that the importance of risk-sharing function also increases with the relative size of the target Since LT targets are more likely to have information asymmetry problems, which indicates that it is more likely for the acquirers to face negative shocks after the negotiation or after the merger is completed, acquirers will be better off to use stock as the primary method of payment to provide the risk-sharing protection for the acquirer shareholders when acquiring LT targets Therefore, when acquiring LT targets, acquirers that choose stock as the primary

payment method will receive a more positive market reaction than those that choose cash as the primary payment method More importantly, the higher the proportion of stock payment used when acquiring LT targets and the larger the relative size of the target, the more positive the market reaction that the acquirer should experience, while everything else being equal It is important to note that while the market-timing hypotheses predict acquirers may make

suboptimal decisions by overpaying during acquisitions, the hypotheses do not assume or argue acquirers to be irrational Therefore, when acquirers are aware of the potential information asymmetry problem, they should prefer stock payment over cash payment when acquiring targets

at the estimated fair market value

On the other hand, this risk-sharing function becomes trivial when the acquirer is

acquiring a HT target In addition, the risk-sharing function may also become trivial when the acquirer and the target are in the same industry, since related mergers may reduce the

information asymmetry problem between the LT target and the acquirer Therefore, in this case, overvaluation signal7 plays the dominant role when the role of risk-sharing function of stock payment becomes insignificant As a result, acquirers that choose stock as the primary payment method when there are fewer information asymmetry problems will receive more negative market reactions than those that choose cash payment, since stock payment in this case signals that the stock of the acquirer is overvalued Furthermore, the higher proportion of stock payment

7

Myers and Majluf (1984) believe that acquirers tend to issue stock when they believe that their stocks are

overvalued; therefore, the market tends to react to the overvaluation signal negatively

used in acquisition and larger relative size of the target will trigger more negative market

reaction to the acquirer’s stock price

Several studies have provided consistent support for this hypothesis Chang (1998) finds that acquirers that use stock to acquire private targets tend to experience positive market reaction

on average Officer, Poulsen, and Stegemoller (2006) find that when the private target

experiences more information asymmetry problems, proxied by high intangible assets, the

acquirers receive a more positive market reaction when the primary method of payment is stock rather than cash In addition, Travlos (1987), and Houston and Rayngaert (1997) find that

acquirers that use stock to acquire public targets experience negative market reaction on average Therefore, if public firms in general have higher transparency than private firms, then the above findings are consistent with this market reaction hypothesis

H5: If low transparency targets with strong negotiating power are more successful at

market timing in the acquisition market, merged firms consist of low transparency targets should underperform in the long run

To determine if LT targets are more successful in market timing, I examine the long-term performance of successful mergers, in addition to the acquisition premiums received by the LT targets and the market reaction of the acquirers If LT provides target firms with more market-timing opportunities and allows target managers to demand higher acquisition premiums, the acquirer of LT targets should experience worse market reaction If the market reaction is

incomplete, merged firms consisting of LT targets should underperform the non-merged firms and merged firms consisting of HT targets in the long run This result will indicate wealth

transfer from the acquirer shareholders to the target shareholders of LT firms However, if the merged firms of LT targets outperform their counterparts in the long run, the result will indicate that LT targets have weaker negotiating power than the acquirers in general; therefore, the

acquirers are able to take advantage of the price discount of LT targets while keeping the

synergistic gains to the acquirer shareholders On the other hand, if the merged firms of LT targets have zero abnormal performance in the long run, the result will indicate that LT still allows the target managers to demand a fair acquisition premium, while the target shareholders are able to capture the complete synergistic gains Note that since overvalued acquirers are

likely to experience negative long-term performance because of price correction, book-to-market

of the acquirer is used as one of the control variables in the analysis

In summary, if LT combined with strong negotiating power provide target managers more market-timing opportunities and market-timing profits, those LT targets should be able to not only receive more offers but also to negotiate higher acquisition premiums As a result, acquirers that overpay when acquiring those LT targets should receive worse market reaction than acquirers that do not overpay If the market reaction is incomplete at announcement, the merged firm consisting of LT targets with strong negotiating power should underperform merged firms consisting of HT targets or of LT targets with weak negotiating power in the long run

METHODOLOGY

Mergers and acquisition data are obtained from Securities Data Company, SDC, while financial data are obtained from Compustat and returns data are obtained from CRSP Only domestic mergers and acquisitions announcement dated between January 1, 1980 and December

31, 2005 are included in the sample In this study, I include both successful and unsuccessful merger and acquisition offers in order to provide a more complete study

A takeover is defined as successful when it results in a completed transaction Since AIMR report and IBES only cover public firms, only public targets are included in the study All observations with deal value of less than $1 million8 or deal value that is less than 1% of the market value of the acquirer are eliminated from the sample (Moeller, Schlingemann, and Stulz (2002)) Deal value is defined by SDC as the total value of consideration paid by the acquirer, excluding fees and expenses yet including amount paid for all common stock, common stock equivalents, preferred stock, debt, options, assets, warrants, and stake purchases made within six months of the announcement date of the transaction Financial and utility firms are excluded from the study Targets with stock price of less than $2 per share are eliminated from the study since Ball, Kothari, and Shanken (1995) find that lower-priced stocks are more likely to be affected by market microstructure effects, such as large proportional bid-ask spreads

While some may conclude the potential effects of the information asymmetry problem on mergers and acquisitions by comparing the empirical results of public targets with those of

8

Other studies, like Schwert (1996), choose $10 million as the cut off when they examine public targets

private targets9, assuming that public targets are always more transparent and therefore have less information asymmetry than private targets are, such assumption may not be valid10 Therefore, studies using direct transparency measures may provide more reliable conclusions than studies using the target’s public status to proxy for the information asymmetry problem In this study, I use the Annual Reviews of Corporate Reporting Practices provided by the Association for

Investment Management and Research (AIMR) to directly measure corporate transparency However, since the AIMR Report has been discontinued since 1996, I also use the analyst

forecast dispersion from IBES Summary Tapes for robustness check

While some researchers use other proxies to measure corporate transparency or

information asymmetry problem, AIMR scores have been widely accepted as appropriate

transparency measures by Lang and Lundholm (1993, 1996), Welker (1995), Bamber and Cheon (1998), Healy, Hutton, and Palepu (1999), Bushee and Noe (2000), Botosan and Plumlee (2002), and Yu (2003) The AIMR scores are calculated based on a set of wide-ranged criteria rather than only financial information, managerial forecasts, or analysts’ evaluation Each industry is examined by a different subcommittee, while the transparency score is calculated based on annual reports, 10-Ks, quarterly reports, proxy statements, and press releases and communication with the analysts In most cases, the subcommittee will classify the transparency measures of a firm into annual transparency, quarterly and other published information transparency, and investor relation transparency Note that while the Corporate Information Committee, CIC,

9

Officer (2006) and Chang (1998) use private target samples to proxy for targets with higher information

asymmetry problem in their study

10

Private (public) targets and low (high) transparency targets may have very different firm characteristics For example, ownership of private firms cannot be bought or sold as easily as that of public low transparency firms Public firms have publicly-traded stocks, which allow investors to transfer their ownership easily, while private firms do not have publicly-traded stocks to provide investors with similar benefits As a result, while the lack of liquidity in private firms may allow the acquirers to purchase private targets at discounts, such liquidity discounts may not be equally applicable to publicly-traded low transparency targets In addition, private firms in general have more concentrated ownership when the stock of the firm is owned by a smaller number of shareholders Such concentrated ownership in private firms often brings about better monitoring and fewer agency problems On the other hand, concentrated ownership and fewer agency problems are not the typical characteristics of low

transparency firms Furthermore, regulations often affect public targets and private targets differently For example, the William Act of 1968 only makes tender offers more costly and more time-consuming for the acquirer of public targets, while private targets and subsidiaries are not covered or protected by this regulation Moreover, private firms are likely to have more limited sources of funds than low transparency firms, since funds of private firms are limited to the personal wealth of the small number of shareholders, in addition to debt financing and IPOs Besides, assuming that all public firms are transparent is also unrealistic since managers often can choose the level, type, and timing of information disclosure to the outside investors, for as long as they still meet the basic disclosure

requirements

provides a set of standard criteria for the disclosure evaluation for all firms, the subcommittees in different industries are allowed to modify the criteria to fit the characteristics of the industry As

a result, the raw scores provided by AIMR Report are not directly comparable across different industries Therefore, I follow the methodology used in Lang and Lundholm (1993, 1996) and Healy et al (1999) to convert the raw scores provided by AIMR Report to the industry/year ranking in order to make cross-industry comparison

I will compute relative industry ranking (RIRit) for each firm i and each year t This RIR

For robustness checks, I also use the analyst forecast dispersion obtained from the IBES Summary Tapes The analyst forecast dispersion has been commonly used to measure

transparency or information asymmetry11 To measure analyst forecast dispersion, the standard deviation of forecast is scaled by the stock price to facilitate comparisons across firms12

Industry median will be subtracted from the scaled dispersion measure to adjust for the industry variation in scores due to differences in subcommittee composition and in industry characteristics Since I am using the forecast dispersion to measure transparency rather than announcement effect, I follow Lang and Lundholm (1996) by averaging the dispersions across the twelve monthly reporting periods on the IBES tape during the company’s fiscal year This

average analyst forecast dispersion will be used to proxy for corporate transparency of the firm13 The price-scaled (earnings-scaled) dispersion provides 6,209 (6,275) target events14

I first examine the firm characteristics of LT and HT targets I report the mean and

median of the size, book-to-market ratio, Q, cash flow, quick assets, and leverage of the targets

to determine if LT targets have different firm characteristics from HT targets Target size is measured by the market value of common stock at the end of the fiscal year before the first bid Book-to-market, is calculated as book value of equity divided by market value of equity in year t-1 Book value of equity is calculated as book value of common stock equity plus deferred taxes and investment tax credit Q is calculated as market value of assets divided by book value of assets Cash Flow is calculated as operating income before depreciation - total tax income + change in deferred taxes from the previous year to the current year -gross interest expense on short- and long-term debt, total preferred dividend requirement on cumulative preferred stock and dividend paid on noncumulative preferred stock, and total dollar dividends declared on common stock before scaled by total assets (Lehn and Poulsen (1989))15 Quick assets are the sum of cash, receivables, marketable securities divided by the market value of common stock Leverage is calculated as long-term debt divided by the market value of common stock T-test and Wilcoxon-Mann-Whitney test are used to test the hypothesis

A Number of Bids

To determine if LT targets are more likely to receive takeover offers from acquirers, I perform both univariate and multivariate analyses To examine the number of offers of a target, I use only the first announced offer from the same acquirer for each acquisition event Therefore, when competing bids from a different acquirer occur, I include the first offer of the competing acquirer in the sample as well

CF = INC - TAX – INTEXP - PRDDIC – COMDIC, where INC is operating income before depreciation

(Compustat item #13), TAX is total tax income minus change in deferred taxes from the previous year to the current year (Compustat item #16 - item 35), INTEXP is gross interest expense on short- and long-term debt (Compustat item #15), PFDDIV is the total preferred dividend requirement on cumulative preferred stock and dividend paid on noncumulative preferred stock (Compustat item #19), and COMDIV is total dollar dividends declared on common stock (Compustat item #21)

In addition, I identify industry dummy based on all SIC codes of the acquirer and the

target If any of the acquirer’s SIC codes matches with any of the target’s SIC codes, the

industry dummy is equal to one; otherwise, it is equal to zero The complete AIMR and IBES

firm year samples are used separately to determine if LT targets are more likely to receive

acquisition offers T-test and Wilcoxon-Mann-Whitney test are used to test the hypothesis I

also examine each of the event years separately to see if the result is consistent through out the

sample period Total transparency and annual transparency at 40% and 60% cutoffs are

examined separately

After the above univariate analysis, I also perform multivariate analysis to determine if

LT targets are more likely to receive offers I use Probit model for the multivariate analysis In the multivariate analysis, all firm year observations are included in the sample The dependent

variable is equal to one if the firm receives an offer and zero otherwise

OFFER = f (TRANSPARENCY, SIZE, BTM, Q, QUICK ASSETS, (2)

CASH FLOW, LEVERAGE, COMPETITION, TIME DUMMIES, Q*QUICK ASSETS, Q*CASH FLOW, Q*LEVERAGE)

where TRANSPARENCY is the transparency of the target firms, measured by

transparency score or dispersion, rather than by the transparency dummy SIZE is the size of the transparency firm, which is the natural log of the market value of common stock, measured at the end of the fiscal year before the first bid It is included in the analysis since smaller firms are

less likely to be covered by financial analysts and have more information asymmetry problems BTM is the book-to-market ratio of the transparency firm Book value of equity is calculated as the book value of common equity plus deferred taxes and investment tax credits in fiscal year t-1 Book-to-market is then calculated as book value of equity divided by the market value of equity QUICK ASSETS of the firm are calculated as the target’s quick assets16 divided by the market

value of the common stock17 LEVERAGE of the firm is the debt-to equity of the target, and it

is calculated as book value of long-term debt divided by the market value of the target’s common stock outstanding as of the last balance sheet date before the acquisition Quick assets, cash flow, and leverage of the targets are included in the model because Jensen’s (1986) free cash flow theory suggests that targets with excess free cash flow and lower leverage are more likely to have agency problems As a result, acquirers may prefer to acquire such targets in anticipation of more gains Book-to-market, quick assets, cash flow, and leverage will all be adjusted by the industry median COMPETITION is used to measure the competition in the acquisition market

To measure competition, I follow Schlingemann, Stulz, and Walkling (2002) to measure

competition as the liquidity index18 in specific industry as the value of all corporate control

transactions of $1 million or more reported by SDC for each year and two-digit SIC code divided

by the total book value of assets of all Compustat firms in the same two-digit SIC code and year 1980s and 1990s DUMMIES are added to the regression to allow change in time series data19

B Acquisition Premium

I use several different measures to estimate acquisition premium in this analysis The first premium measure is the percentage premium provided by SDC, which is calculated as the premium offered with respect to the target’s trading price four week prior to the first

announcement date However, since Officer (2006) finds that the premium information provided

by SDC is only available for about half of the observations, I follow Moeller et al (2002) to use the value of cash, stock and other securities of the offer as the premium measure, since this

variable provides the highest number of observations and is often available at announcement The premium is then scaled by the market value of equity of the target 50 days prior to the

announcement day (Moeller et al (2002) and Officer (2003)) In addition, I follow Schwert (2000) to calculate the premium measure as the sum of price runup prior to the first

announcement and the price markup from the announcement More specifically, the abnormal

performance is estimated by the market model residual for the target firm cumulated over the period (-63, 126) from the first announcement day20

mt i i

where Rit is the return to target firm i on trading t, Rmt is the return to the CRSP

NYSE/AMEZ/Nasdaq value-weighted portfolio return on day t, while the α and β are estimated

by using market model and return data during the (-316 and -64) event window from the first announcement date

Rit = α i + βi Rmt + εit where t = -316 to -64 (4)

I first use univariate analysis to examine if LT targets are more likely to receive higher acquisition premiums based on the above premium measures, both industry-adjusted and unadjusted measures After the univariate analysis, I also apply multivariate analysis to control for other variables I perform a regression for each of the premium measures mentioned above, both adjusted and unadjusted

Premium T = α+ β1TRANSPARENCYT + β2 SIZET + β3SIZEA + β4 BTM T (5)

+ β5BTM A + β6 COMPETITIONT + β7 INDUSTRY DUMMY + β8 % of STOCK + β9 POISON PILL + β10 QUICK ASSETST + β11 LEVERAGET + β12 POOLING INTERESTS

+ β13 COMPETITION T * LT DUMMY T

+ β14 ADVISING FEES PAID T * LT DUMMY T

+ β15 INDUSTRY DUMMY * LT DUMMY T + β16 1980s DUMMY + β17 1990s DUMMY

20

Price runup is included as a part of premium calculation because of potential insider trading activities prior to announcement (Meulbroek (1992)) and information leakage prior to the first announcement (Jarrell and Poulsen (1989)) Announcement of 13D filings with the Securities and Exchange Commission (SEC) when an investor acquires more than 5% of the target’s stock can often provide a clue to the market of potential takeover offer In addition, they find that price runrup prior to the first announcement does not substitute for post announcement markup in their sample Therefore, price runup should be considered as a part of acquisition premium

where TRANSPARENCYT is the transparency measure of the target, SIZET and SIZEA are the size of the target and acquirer respectively, while RELATIVE SIZE TA, the relative size between target and acquirer, is some times used as the substitute BTMT and BTMA are the book-to-market ratio of the target and acquirer firms respectively The book-to-market ratios are included here to proxy for misevaluation COMPETITIONT is measured by the liquidity index

of the target firm Competition is included here to proxy for the growth and competition in the acquisition market, since competition is likely to be stronger in the hot market INDUSTRY DUMMY is equal to one if any of the target’s SIC codes matches with any of the acquirer’s SIC codes, while the % of STOCK is the percentage of acquisition payment made in stock Based on the definition of SDC, POISON PILL 21 is equal to one if the target invokes the poison pill or if the existence of the poison pill discourages the potential acquirer, and zero otherwise22 QUICK ASSETST are the quick assets of the target, POOLING INTERESTS is equal to one if the

accounting method of the corporate combination is pooling of interests method, or zero

otherwise23 Two interaction terms, COMPETITIONT * LT Dummy and ADVISING FEES PAIDT * LT Dummy, are used to examine if competition in the target firm’s industry and

advising fees paid by the LT target will increase the negotiating power of the LT firm The interaction term between poison pill and LT dummy is not included because it is not a full rank

ADVISING FEES PAIDT is the investment banking fees or advisor fees24 that the target pays upon completion of the merger Hi-Tech target dummy is included because Kohers and Kohers (2000) find that Hi-tech targets are more likely to receive higher acquisition premium

INDUSTRY DUMMY * LT DUMMY is included to examine if related merger can reduce the information asymmetry problem and market-timing premium of LT targets Time dummies are used to allow variables to change in time

C Method of Payment

The hypothesis predicts that when the target is acquired at discount, the acquirer may not have preference in specific method of payment On the other hand, when the LT target is not acquired at discount, acquirers will prefer to use stock as the primary method of payment,

especially when the target and acquirer are not in the same industry

Since small targets are less likely to provide significant impact on the merged firms, I separate the successful merger sample into small and non-small target sub-samples Following Schlingemann and Stulz’s paper (2002), I define small firms as firms in the lowest NYSE size quartile by using market value of equity25 I then examine the payment method on the whole sample and the non-small target sample to determine if LT targets are more likely to receive stock payment than HT targets are T-test and Wilcoxon test are used to test the difference

Next I use Probit model to examine if LT targets are more likely to receive stock payment than HT targets are Therefore, the dependent variable will be equal to one if the payment

method is stock and zero otherwise STOCK payment will be defined in two ways as the

following First, I follow Fuller, Netter, and Stegemoller (2002) and define the method of

payment as stock when at least 50% of the consideration is paid in stock and the method of payment as cash when at least 50% of the consideration is paid in cash; otherwise, the method of payment are defined as others In addition, I also follow Martin (1996) to classify the method of payment into three categories The first category is cash financing which includes combinations

of cash, debt, and liabilities The second category is stock financing, which includes payments with common stock or a combination of common stock and options or warrants The third category is the combination (others) financing which comprises combinations of common stock, cash, debt, preferred stock, convertible securities, and method classified as “others” by SDC

Stock T = f (TRANSPARENCY T, SIZET, SIZEA, RELATIVE SIZETA, BTM T, BTM A, (6)

PREMIUMT, INDUSTRY DUMMY, LEVERAGE A, POOLING INTERESTS, INDUSTRY DUMMY * LT DUMMY T,

LT DUMMY T * HT DUMMY A, Hi-TECH T, TIME DUMMIES)

25

Book value of the assets is also used in Schlingemann and Stulz (2002) for robustness checks They find similar results with both size measures

Note that most of the independent variables are defined the same way as in the previous equations New variables include the following: LEVERAGEA, the acquirer’s leverage ratio, is included in the model because acquirers with higher leverage ratio than the optimal capital structure can use stock payment to reduce the leverage ratio and to bring it closer to its optimal capital structure Optimal capital structure is proxied by the industry median of the year

POOLING INTERESTS is equal to one if the accounting method used is pooling of interests method26 INDUSTRY DUMMY * LT DUMMY T is the interaction term between the industry dummy and the LT target dummy, since when the target and acquirer are in the same industry, information asymmetry problem becomes less important Therefore, they are less likely to choose stock as the primary method of payment as a risk-sharing tactic LT DUMMY T * HT DUMMYA is included in the model because Finnerty and Yan (2006) find that the acquirer is more likely to offer stock when the target, but not the acquirer, has information asymmetry problem In addition, they find the acquirers are more likely to offer cash payment when the acquirer-side information asymmetry problem dominates the information asymmetry problem of the target Furthermore, when both parties have information asymmetry problems, the acquirer

is more likely to use convertible security as the payment27 Hi-TECH T dummy is included because Kohers and Kohers (2000) argue that high-tech targets have higher growth, risk, and uncertainty Therefore, if high-tech targets have more uncertainty and risk, they are more likely

to receive stock payment during acquisition

D Acquirer’s Market Reaction: Acquisition Premium and Payment Method

I follow Brown and Warner (1985) standard event study methodology to calculate CARs for the three-day period (-1, 1) around each announcement date provided by SDC I also

measure the CARs from the date before the first acquirer announcement to the date of merger completion28 or withdrawal and also the CARs from the date before the first acquirer

announcement to the end of the month of merger completion or withdrawal I examine the

26

Acquirers that prefer pooling of interests method will have to use its own voting common stock in exchange for at least 90 percent of the voting common stock of the target based on accumulated postretirement benefit obligation (APBO) No.16 (Robinson and Shane (1990))

successful and unsuccessful mergers separately Note that the end-of-month period is included

in the short-term study because the long-term study relies on monthly returns data which start in the month following the relative merger completion date Therefore, examine short-term returns

up to the merger completion date or withdraw date leaves a gap between the short-term and term event windows not examined in this study The parameters of the market model are

long-estimated over the (-205, -6) event window, while the p-value is calculated using the time-series and cross-sectional variation of abnormal returns The acquirer’s dollar gain is calculated as CAR times the acquirer’s market value of equity five days before the announcement (Officer, Poulsen, and Stegemoller (2006)) The abnormal dollar return divided by the total transaction value reported by SDC represents the dollar gain per dollar spent on acquisitions T-test and Wilcoxon-test are used to test the difference between the abnormal performance between

acquirers of LT targets and acquirers of HT targets Mean and median abnormal performances are both reported

Abnormal return is calculated as the following:

Next, I also examine if the market reacts differently to the acquirer when the same

method of payment is used for targets with different transparency levels Therefore, I first

classify the targets into LT and HT target portfolios before further dissecting each of the

portfolios into stock and cash payment portfolios

However, to provide a more thorough analysis, I use the following multivariate analysis29:

29

Fuller et al (2002) use stock dummy, and combination payment dummy which include some cash and some stock payments, and interaction term between the stock dummy and relative size and between the combination dummy and relative size of the target, log of target size, log of acquirer size, dummy if the target and acquirer are in the same industry, dummy for first acquisition, and dummy for repeated acquirer in their multivariate analysis Moeller,

CAR i = α+ β1 TRANSPARENCYT + β2 SIZET + β3SIZEA + β4 BTM T (8)

+ β5 BTM A + β6 INDUSTRY DUMMY + β7 % of STOCK + β8 PREMIUMT

+ β9 POISON PILL+ β10 LEVERAGEA + β11 QUICK ASSETST

+ β12 POOLING INTERESTS + β13 Hi-TECH T

+ β14 COMPETITION + β15 COMPETITION*LT DUMMY

+ β14 STOCK * SIZET30 + β15 INDUSTRY DUMMY * STOCK

+ β16 STOCK * LT TARGET + β17 PREMIUM * LT DUMMY

+ β18 1980s DUMMY + β19 1990s DUMMY

Transparency of the target is included for it is the focus of this study, and if LT targets are able to receive higher premiums from the acquirers, the acquirers should experience worse market reactions However, since information asymmetry problems of LT target may also

prevent the market from estimating the value of LT target accurately, the market reaction may not accurately reflect the overpayment in the short-term study Such complete price reaction will occur in the long term INDUSTRY DUMMY is equal to one when the target and acquirer are

in the same industry It is included because acquirers should receive more positive reaction when the merger creates higher synergy more likely to create higher synergy, acquirer should receive more positive reaction % of STOCK is included since higher percentage of stock

payment is more likely to cause acquirers to receive worse reaction Premium paid is included because higher the premium paid indicates more wealth transfer from the acquirer to the target; therefore, higher premiums will trigger worse market reactions for the acquirer, while everything else being equal POISON PILL dummy is included because Schwert (2000) finds that acquirers

of hostile takeovers are more likely to experience worse market reactions Maloney, McCormick, and Mitchell (1993) find acquirers with higher leverage have better market reactions Jensen (1986) suggests that acquiring targets with more free cash flow problem or agency problem may

Schlingemann, and Stulz (2002) find smaller acquirers tend to earn positive returns while large acquirers tend to lose

In addition, larger acquirers in general pay higher premium and that they are more likely to compete with other bidders Myers and Majluf (1984) believe that acquirers tend to issue stock when they believe that their stocks are overvalued; therefore, such news announcement signals stock overvaluation and causes market to react to the news negatively

30

STOCK * RELATIVE SIZE T may be used

allow acquirers to receive more gains by reducing the agency problem Therefore, QUICK

ASSETST andLEVERAGET are also included to proxy for target’s agency problem31

POOLING INTERESTS is included in the model because the choice of reporting method may affect the market reaction when it can impact the reported income of the merged firm

tremendously in large merger deals Hi-TECH Dummy is used because Kohers and Kohers (2000) argue that high-tech targets have higher growth, risk, and uncertainty Various interaction terms are used Note that it may be possible that the market only reacts to the raw, unadjusted premium rather than the industry-adjusted premium, so I use each of the premium measures in different regressions

E Long-term performance

Majority mergers and acquisition studies find negative abnormal performance following mergers (Loderer and Martin (1992), Anderson and Mandelker (1993), Loughran and Vijh

(1997), and Gregory (1997)) However, no study has specifically examined if the overall

negative post-merger performance is caused by LT targets’ success in market timing Therefore,

I first separate the mergers and acquisitions sample into two portfolios based on the transparency level of the targets If LT provides targets with more market-timing opportunities, the LT

portfolio should underperform the HT portfolio and the non-merged firm sample in the long run However, since not all LT targets should have strong negotiating power, I will control for the firm characteristics

Given that long-term abnormal return estimation can be very sensitive to the model

specifications because small errors in the short-horizon studies can be compounded in the long term and cause mis-specified results, I use several different measures such as cumulative

abnormal returns (CARs) estimated based on the Brown and Warner approach (1980), hold abnormal returns (BHARs) based on size and book-to-market matching bootstrapping

buy-and-methodology (Brock, Lakonishok, and LeBaron (1992)), Calendar-time approach three-factor model (Fama and French (1993)), and Calendar-time approach four-factor model (Carhart (1997))

to examine the long-term performance I use one-year, three-year, and five-year event windows

31

Hartford (1999) finds acquirers tend to receive lower market reactions when the acquirer has large cash reserve (Compustat item #1 / Sales) However, because of data restriction, this variable is not used

for my long-term study since many long-term post merger studies examine return beyond the three-year period32

Several researchers advocate CARs over BHARs in long-term study Lyon et al (1999) argue that CARs are less skewed than buy-and-hold abnormal returns, and that conventional t-statistics for CARs are well specified In addition, Fama (1998) also argues that CARs are a lot more reliable long-term performance measure than BHARs because the latter is not normally distributed The lack of normality can weaken the power of the test, while the long-term

abnormal performance is exaggerated

Therefore, for the monthly CARs in the long-term study, I use methodology described by Brown and Warner (1980) Standard errors are calculated using month -36 to +48 from the announcement Note that the Brown and Warner method is only used for the CAR calculations, but not for BHAR calculations Therefore, 12-month, 36-month, and 60-month CARs from the month after the completion month are calculated for the high and LT samples, based on both CRSP equally-weighted and value-weighted index returns I use the t-test to determine if HT firms and LT firms have statistically different long-term performance based on the CARs In addition to the t test, the sign test and Wilcoxon ranked-sum test are used to determine if LT firms under-perform HT firms in the long run

On the other hand, many others prefer BHAR because it measures the true investor experience, even though common procedure can provide biased buy-and-hold abnormal return caused by the new listing bias Lyon, Barber, and Tsai (1999) find that bootstrapped skewness-adjusted t-statistics can eliminate skewness bias, while Ikenberry et al (1995) suggest that average BHAR generated by bootstrapping procedure or large sample size can provide reliable estimates and normal distribution Therefore, for the buy-and-hold abnormal returns, I use the bootstrapping method described by Brock, Lakonishok, and LeBaron (1992) rather than the traditional BHAR calculation To find the matched firms, I use size and book-to-market two-way matching I calculate size and book-to-market equity for each event firm and each purged sample firm The purged sample will eliminate any firm that has either become an acquirer or a

32

Mandelker (1971) examine CAARs over the 40-months period following merger completion Dodd and Ruback (1977) use CAAR in two event windows, (1, 12) and (13, 60) relative to the first public announcement of the bid Langetieg (1978) examines up to 70 months after merger Malatesta (1983) examines abnormal returns during (1, 12) months after the merger announcement Dodd and Quek (1985) examine CAAR over the 60-months period following the month of merger announcement

takeover target in the past five years Size is the market value of firm equity as of June 30 Book value of equity is calculated as the book value of common equity plus deferred taxes and investment tax credits for fiscal year t-1 Book-to-market equity is this book value divided by the June 30 market value of equity I then use NYSE stocks to calculate the size breakpoint as the 50th percentile of the NYSE sample in order to segment small and large firms NYSE stocks are used to calculate the book-to-market breakpoints as the 30th and 70th percentiles of the NYSE sample in order to segment firms into book-to-market terciles All event firms and purged sample firms listed on NYSE, AMEX, or NASDAQ are assigned to one of the six portfolios based on the NYSE breakpoints To perform the bootstrapping procedure, I randomly match each of the event firm with a firm from the same size and book-to-market portfolio until each of the targets and acquirers has a randomly matched firm This creates a pseudo sample Next, I use the sum of the returns of the randomly-matched acquirer and target as the benchmark return The abnormal performance are calculated as the actual merged firm return minus the sum of the returns of the matched firms

I then compute the mean BHAR for the pseudo-sample

BHARit = ∏

=

+ t

t

it R 1

] 1 [ - ∏

R

1

,]1

where BHAR it represents buy-and-hold abnormal return for firm i at time t, while Rit

represents return on the sample firm i at time t and Rbenchmark represents benchmark return at time

t The returns for each of the 6 portfolios are calculated from the month after the acquisition announcement (t = 1) for one year and three years When a matched firm is delisted within the event window, the return of another randomly chosen matched firm are spliced in to substitute for the missing firm returns (Ikenberry et al (1995))

The mean buy-and-hold return,BHAR is the weighted average of the individual BHAR ,

where wi represents the weight for stock i, while N represents the number of firms in the portfolio Both equal-weighted BHARs and value-weighted BHARs based on the market

capitalizations are calculated The same procedures are repeated 1,000 times to generate a large sample distribution Then a p-value are computed for the low and HT portfolios separately

However, Mitchell and Stafford (2000) argue that the bootstrapping procedure assumes event-firm abnormal returns are independent, while corporate announcements and actions are not always random events Clusters in industry or the economy, such as merger waves, can occur

In addition, Fama (1998) argues that BHAR suffers the bad model problem by compounding the errors through time Furthermore, BHAR also ignores the cross-sectional dependence of event-time abnormal returns overlapping in calendar time can inflate test statistics Therefore, Fama (1998) recommends calendar-time portfolio approach to examine the long-term performance because a) monthly returns suffer less bad model problem than daily returns do, b) monthly calendar-time portfolios can account for all cross-correlations of event-firm abnormal returns in the portfolio variance, and c) calendar-time approach can better approximate the normal

distribution and provide more reliable statistical inference

With the calendar-time approach Fama and French (1993) three-factor model and Carhart (1997) four-factor model, I first separate successful mergers into those consist of HT targets and those consist of LT targets Since Ang and Zhang (2004) find the WLS provides more reliable results than the OLS does, I use WLS calendar-time factor models33 I buy stocks in the HT target portfolio and stocks in LT targets portfolio separately at the end of the merger completion month then hold the companies for one year, three years, and five years, etc The intercept from the regression represents the abnormal performance that cannot be predicted by the factor model

Fama and French (1993) three-factor model can be stated as the following:

r it = αiT + b i MKT t + s i SMB t + h i HML t + e it (11)

where i represents the LT or HT portfolio, while rit represents the monthly return on the

LT and HT portfolios, respectively, in excess of T-bill rate at month t, starting at t = 1, the month following the merger completion date MKT represents the excess monthly return on the value-

33

Since Ang and Zhang (2004) also find the calendar-time approach four-factor model has type I error (rejects the null when it is true), I focus more on the three-factor model in this study

weighted market proxy at time t SMB and HML represent monthly returns on value-weighted zero-investment portfolios, which are calculated as the small portfolio return minus the large portfolio return and the high book-to-market return minus low book-to-market return,

In addition, a zero-investment portfolio are formed to determine if a long position in HT target portfolio and a short position in LT target portfolio will provide positive long-term

abnormal returns Again, the intercept represents the monthly abnormal return obtained from the zero-investment portfolio

RESULTS

In Table 1, I examine the characteristics of the targets based on the total transparency measure of AIMR LT and HT targets are determined based on the 40% and 60% cutoffs Firms with 60% or higher industry-adjusted total transparency scores are classified as HT targets, while targets with 40% or lower industry-adjusted total transparency scores are classifies as LT

targets35 With 5846 AIMR firm/year observations from 1980 to 1995, the 40% and 60% cutoffs generate 2300 LT and 2343 HT AIMR firm year observations As additional transparency measures, I use the standard deviation of one-year monthly analyst earnings forecast scaled by price and earnings before adjusting by industry median Results are shown in different panels of