Abnormal returns associated with selected financial profiles of leveraged buyouts, an empirical analysis

... ABNORMAL RETURNS ASSOCIATED WITH SELECTED FINANCIAL PROFILES OF LEVERAGED BUYOUTS: AN EMPIRICAL ANALYSIS A dissertation submitted in partial fulfillment of the requirements for the degree of. .. date? 2) the financial profiles of these LBOs provide ex ante identification of this event? and 3) is it possible to invest in companies with similar financial profiles and also earn abnormal re­... rod uction prohibited w ith o u t perm ission ABNORMAL RETURNS ASSOCIATED WITH SELECTED FINANCIAL PROFILES OF LEVERAGED BUYOUTS: AN EMPIRICAL ANALYSIS R eproduced w ith perm ission o f the cop

INFORMATION TO USERS The most advanced technology has been used to photograph and reproduce this manuscript from the microfilm master. U M I films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send U M I a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact U M I directly to order. University M icrofilm s International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 313761-4700 800 521-0600 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r reprod uction prohibited w ith o u t perm ission. R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. O rder N um ber 0111260 A bnorm al re tu rn s associated w ith selected financial profiles of leveraged buyouts: A n em pirical analysis Richardson, Linda Lush, Ph.D. University of Arkansas, 1989 UMI 300 N. Zecb Rd, Ann Arbor, MI 48106 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. NOTE TO USERS T H E O R IG IN A L DOCUM ENT RECEIVED BY U .M .I. C O N TAINED PAGES W ITH SLANTED AND POOR PRINT. PAGES WERE FILM ED AS RECEIVED. THIS REPRODUCTION IS THE BEST AVAILABLE COPY. R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. ABNORMAL RETURNS ASSOCIATED WITH SELECTED FINANCIAL PROFILES OF LEVERAGED BUYOUTS: AN EMPIRICAL ANALYSIS R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. ABNORMAL RETURNS ASSOCIATED WITH SELECTED FINANCIAL PROFILES OF LEVERAGED BUYOUTS: AN EMPIRICAL ANALYSIS A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy By LINDA L. RICHARDSON, B.A., M.S. Connecticut College, 1970 Montana State University, 1974 December 1989 University of Arkansas R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. This dissertation is approved for recommendation to the Graduate Council Dissertation Adviser: /William F. Hardin Dissertation Committ Joe Jones R e pro duce d w ith perm ission of th e cop yrig ht ow ner. F urthe r rep rod uction pro hibited w ith o u t perm ission. ACKNOWLEDGMENTS I have many people to thank. My first debt of gratitude is to W. Clint Johnson, who encouraged me to en­ ter the field of finance after many years in economics. Next, to Jonathan Willard, my highest regards, respect, and many thanks for making this possible. Without technical support from Wayne Persons, Wayne Barber, Tom MacDonald, Bruce Andrews and Carl Helms, life would have been even more difficult. My official unofficial editors were Marie Dickson, Alice Persons, and Gary Lombardo, who spent many sleep-filled nights reading my copy. Thanks also to Lynne Cote and Anne LaLime for their typing assistance and to Jim Brady and Casandra Fitzherbert for library assistance. And literally in my hour of need, Tom Sanders graciously filled in for me. Thank-you Dr. Hardin for persevering and giving me the hope and encouragement needed to complete the task. I dedicate this endeavor to Bob and Jeremiah. They both make everything and anything possible in my life. iii R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. TABLE OF CONTENTS ACKNOWLEDGMENTS .................................... iii LIST OF TABLES ..................................... vii Chapter I Introduction ................................ Definition of a Leveraged Buyout II III ........ 1 6 Abnormal Returns ......................... 8 Potential Candidates ..................... 10 History .................................. 12 A Hybrid in Theory ....................... 15 Summary .................................. 17 Literature Review ............................ 19 Studies Testing for Abnormal Returns ...... 19 Motivations for Going Private ............. 21 Potential Gains for Shareholders .......... 23 Gains are Transitory ..................... 27 Gains Measurements May BeInappropriate ..... 31 Impetus For This Study ................... 32 Possible Profiles for Managementbuyouts .... 34 Summary .................................. 35 Methodology ................................. 36 Definition of the Population and Sample .... 36 Financial Ratio Analysis .................. 41 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Possible Predictors ........................ 42 Data Collection ........................... 47 Testing for Any Industry Effect ............ 48 Factor Analysis to Eliminate Multicollinearity ............. 49 Classification and Multiple Discriminant Analysis .......... 50 Intertemporal Effects .............. IV 55 Probit and Logit .......................... 56 Selection of Likely Candidates ............. 59 Efficient Markets and Event Studies ......... 60 Models Used for Residual Analysis .......... 63 The Test .................................. 64 Summary of Methodology ..................... 66 Empirical Analysis and Findings ............... Initial Empirical Question ................. 68 68 Multivariate Analysis of Variance ......... 68 Reasons for Factor Analysis ............... 70 Results of Factor Analysis ............ 76 Stepwise Procedures ...................... 94 Results of Logistic Regression (LOGIST) .... 101 Abnormal Returns From the LBO Sample ....... 104 Abnormal Returns for Predicted LBOs ........ 110 Summary of Results ........................ 112 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. VSummary and Conclusions ........................... 113 Research Methodology....................... 113 Empirical Findings......................... 116 Conclusions and Implications............... 120 Limitations of This Study.................. 121 Contributions of This Study................ 122 ENDNOTES ........................................... 124 WORKS CITED ............................ 128 R epro duce d w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. LIST OF TABLES Table Page 1. 25 LEVERAGED BUYOUTS OF ALL TIME BY DOLLAR VALUE .......................... 2 2. TOP 10 DIVESTITURE LBOs OF ALL TIME BY DOLLAR VALUE .......................... 3 3. THE LEVERAGED BUYOUT MARKET 1981-1988 .......... 4 4. FORMS OF RESTRUCTURING BUSINESS FIRMS .......... 7 5. TEN-YEAR MERGER COMPLETION RECORD 1978-1988 .... 30 6. SAMPLE OF LEVERAGED BUYOUTS BY ANNOUNCEMENT DATE ........................ 38 7. INDUSTRY REPRESENTATION OF LEVERAGED BUYOUT ACTIVITY .................... 40 8. DESCRIPTIVE STATISTICS for 1983 ............... 71 9. DESCRIPTIVE STATISTICS for 1984 ............... 72 10. DESCRIPTIVE STATISTICS for 1985 ............... 73 11. DESCRIPTIVE STATISTICS for 1986 ............... 74 12. DESCRIPTIVE STATISTICS FOR ALL LBOs USING FISCAL YEAR PRIOR TO ANNOUNCEMENT DATE ..... 75 13. FACTOR PATTERN for the INITIAL UNROTATED SOLUTION for 1983 ................ 80 14. FACTOR PATTERN for the INITIAL UNROTATED SOLUTION for 1984 ................ 81 15. FACTOR PATTERN for the INITIAL UNROTATED SOLUTION for 1985 ................ 82 16. FACTOR PATTERN for the INITIAL UNROTATED SOLUTION for 1986 ................ 83 vii R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. 17. VARIANCE EXPLAINED BY EACH FACTOR............. 84 18. FACTOR PATTERN FOR VARIMAX ROTATED SOLUTION FOR 1983 .................. 85 19. FACTOR PATTERN FOR VARIMAX ROTATED SOLUTION FOR 1984 .................. 86 20. FACTOR PATTERN FOR VARIMAX ROTATED SOLUTION FOR 1985 .................. 87 21. FACTOR PATTERN FOR VARIMAX ROTATED SOLUTION FOR 1986 .................. 88 22. INTERCORRELATIONS FOR FINANCIAL RATIOS FOR 1983 . 89 23. INTERCORRELATIONS FOR FINANCIAL RATIOS FOR 1984 . 90 24. INTERCORRELATIONS FOR FINANCIAL RATIOS FOR 1985 . 91 25. INTERCORRELATIONS FOR FINANCIAL RATIOS FOR 1986 . 92 26. VARIABLES SUBMITTED TO STEPWISE DISCRIMINANT ANALYSIS BASED UPON FACTOR ANALYSIS RESULTS .................... 93 27. SUMMARY OF PREDICTOR VARIABLES ................ 96 28. 1983 MDARESULTS ............................. 97 29. 1984 MDARESULTS ............................. 98 30. 1985 MDARESULTS ............................. 99 31. 1986 MDARESULTS ............................. 100 32. AVERAGE RESIDUALS FOR LBO SAMPLES ............. 107 33. CUMULATIVE RESIDUALS FOR LBO SAMPLES ........... 108 34. PREDICTED LBO SAMPLE .......................... 109 viii R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Chapter I Introduction After its completion, the Kohlberg Kravis Roberts & Co. acquisition of RJR Nabisco Inc. will represent a record book entry as the largest leveraged buyout (LBO) in history at $24.7 billion.1 Leveraged buyouts were among the most lucrative financial deals of the past decade. Previously topping the list of companies taken private through a le­ veraged transaction was Beatrice Co. Inc., taken private by Kohlberg Kravis Roberts & Co. for an estimated $6.25 bil­ lion in 1986 (see Table 1). Allied-Signal Inc. headed the list of divestiture type leveraged buyouts in the sale of 50% of Union Texas Petroleum Holdings Inc. for $1.7 billion in 1985 (see Table 2).a Motivated by high yields, favorable tax treatment of debt, and a period of unprecedented economic growth, the number and dollar value of deals increased for the period beginning in 1984 (see Table 3). Falling interest rates provided an environment well suited to deals made possible by significant amounts of debt. R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. TABLE 1 25 LEVERAGED BUYOUTS OF ALL TIME BY DOLLAR VALUE Acquiring Company Kohlberg Kravis Roberts & Co. Kohlberg Kravis Roberts & Co. AV Holdings Corporation Kohlberg Kravis Roberts & Co. Macy Acquiring Corporation Kiewit-Murdock Investment Corp. SCI Holdings Incorporated Inc. Private investor group led by Kohlberg Kravis & Roberts L Acquisition Corp. Aancor Holdings Inc. Eckerd Holdings Inc. HHF Corp. Anac Holding Corp. Griffith Acquisition Corp. Reliance Acquisition Corp. PACE Group Holdings Inc. Playtex Holdings Inc. JSC/MS Holdings Inc. LMC Holdings Inc. Farley Industries Inc. J.W.K. Acquisition Corp. Private investor(s ) Private investor(s) ARA Holding Co. Occidental Petroleum and Drexel Burnham Lambert Inc. Sources Company Taken Private Beatrice Cos. Inc. Safeway Stores Inc. Borg-Warner Corp. (90%) Owens-Illinois Inc. R.H. Macy & Co., Inc. Continental Group Inc. Storer Communications Union Texas Petroleum Holdings (50%) Lear Siegler Inc. National Gypsum Jack Eckerd Corp. Levi Strauss & Co. Revco D .S . Inc. National Car Rental System Inc. Reliance Electric Co. Three units of City Investing Co. International Playtex Inc. Container Corp. of America Fruehauf Corp. Northwest Industries Metromedia Inc. MGIC Investment Corp. Wometco Enterprises ARA Services Inc. Diamond Shamrock Chemicals Co. Mergers and Acquisitions. (November/December 1987), 48. R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. TABLE 2 TOP 10 DIVESTITURE LBOs OF ALL TIME BY DOLLAR VALUE SELLER BUYER Allied-Signal Inc. Private investor group led by Kohlberg Kravis Roberts & Co. Household International Inc. Griffith Acquisition Corp. Exxon Corp. Reliance Acquisition Corp. City Investing Co. PACE Group Holdings Inc. BCI Holdings Corp. Playtex Holdings Inc. Mobil Corp. JSC/MS Holdings Inc. Baldwin-United Corp. Management group led by Northwestern Mutual Ins. Diamond Shamrock Corp. Occidental Petroleum and Drexel Burnham Lambert Union Carbide Corp. First Brands Corp. City Investing Co. Motel Holding Co. Inc. Source: Mergers and Acquisitions. (November/December 1987), 52. R e pro duce d w ith perm ission of the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. TABLE 3 THE LEVERAGED BUYOUT MARKET 1981-1988 Year 1981 1982 1983 1984 1985 1986 1987 1988 Source: No. of Deals 99 164 230 253 254 335 270 318 Value $ mil 3,093.1 3,451.8 4,519.0 18,697.3 19,633.8 45,159.9 36,069.2 42,914.0 Mergers and Acquisitions. (May/June 1989), 64. R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. These lucrative deals have provided returns to in­ vestors, as well as facilitators, which have been unobtainable elsewhere. The classic benchmark of promised riches is William Simon's 20,000% profit on his $330,000 personal investment in the leveraged buyout of Gibson Greeting Cards in 1983.3 This study answers three empirical questions: 1) do abnormal returns accrue to actual LBOs for a trading period surrounding their announcement date? 2) do the financial profiles of these LBOs provide ex ante identification of this event? and 3) is it possible to invest in companies with similar financial profiles and also earn abnormal re­ turns ? R e pro duce d w ith pe rm issio n o f th e cop yrig ht ow ner. F urthe r re p rod uction prohibited w ith o u t perm ission. Definition of a Leveraged Buyout A leveraged buyout involves the purchase of the as­ sets or stock of either a privately or publicly owned com­ pany, subsidiary or division. Its distinguishing charac­ teristic is the amount of debt used in the acquisition. Up to ninety-five percent of the purchase price is financed by layers of loans secured by the target company's assets and is to be repaid with the target company's operating cash flow. These transactions represent restructuring which could be categorized as either a sell-off or change in own­ ership activity (see Table 4). Both types of activity, when facilitated through the use of debt financing in ex­ cess of ninety percent, have been referred to as a lever­ aged buyout. With divestitures, the purchase of a portion of a firm by an outside third party does not create a new legal entity. However, in a going-private transaction, a pub­ licly owned company is taken private by a group of inves­ tors which often includes existing management. This is re­ ferred to as a management buyout, or pure going-private transaction. The focus of this study is leveraged management buyouts, going-private transactions financed by significant amounts of debt. R e pro duce d w ith pe rm issio n o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. TABLE 4 FORMS OF RESTRUCTURING BUSINESS FIRMS I. EXPANSION A. B. C. II. SELL-OFFS A. B. C. III. Spin Offs Divestitures Equity Carveouts CORPORATE CONTROL A. B. C. D. IV. Mergers and Acquisitions Tender Offers Joint Ventures Premium Buybacks Standstill Agreements Antitakeover Amendments Proxy Contents CHANGES IN OWNERSHIP STRUCTURE A. B. C. D. Source: Exchange Offers Share Repurchases Going-Private Transactions/Mgmt. Buyouts Leveraged Buyouts/LeveragedMgmt. Buyouts J. Fred Weston and Thomas E. Copeland, Managerial Finance. 8th ed. (Chicago: Dryden Press 1986), 901. R e pro duce d w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Abnormal Returns The technique used to assess the impact of an­ nouncement type news is the event study methodology. The normal relationship between a particular stock's return and the "market" is established for a period unaffected by this news. With these predicted returns estimated, actual re­ turns surrounding the announcement date are compared, and the difference is a measure of abnormal or excess returns. Halpern reviewed event studies applied to corporate acquisitions in his 1983 paper.4 He concluded that studies conducted to identify acquisition specific influences in merger activities using accounting data and stock price data were flawed because: 1) accounting data do not provide information on the expected long-run impacts; 2) serious problems exist in defining comparable control groups; and 3) the comparison group does not provide information on how specific firms would have performed without an acquisition. Of particular interest in this study is the ex­ amination of stock price movements for the targeted LBO prior to the announcement. Capital markets react to rumors of possible takeovers by driving up the stock price of tar­ get companies in anticipation of premiums that must be paid to existing shareholders in order to buy out their equity position. It has been estimated that in a management buyout, this premium may average approximately 56% relative to the open market share price for up to 40 trading days prior to the proposal.3 8 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. After an announcement is made, the ability to earn this premium dissipates. In order to earn these returns in full, an investor would have to possess information about these events in advance of their announcement or be able to predict their occurrence. If an investor could predict these events based upon the existence of a unique financial profile, then direct knowledge of the event would not be necessary. Merger and acquisition related research took this hypothesis one more step by testing for abnormal returns for firms with similar financial profiles. Would firms with similar financial profiles, which never were acquired, also produce abnormal returns? Are these firms subject to a "spillover effect" because investors consider them to be potential candidates? Studies conducted by Stevens,6 Simkowitz and Monroe,7 and Wansley, Roenfeldt and Cooley,0 concluded that this was possible. Specifically, Wansley, Roenfeldt and Cooley found that abnormal returns of up to 29.1 percent accrued in the seven months prior to the actual takeover. This represents a return which is 29.1 percent above its risk-adjusted expected rate of return. However, the port­ folios of the potential candidates also earned abnormal re­ turns of 17.1 percent for a 21 month test period. They concluded that abnormal returns could be earned by inves­ tors under both scenarios. 9 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. These studies testing for abnormal returns for eco­ nomic events such as mergers and acquisitions, as well as pure going-private transactions, have limited applications for leveraged buyout transactions. The motivation for leveraged management buyouts is similar to a pure going-private transaction. But debt has significant tax benefits, and therefore has greater poten­ tial gains for equity participants. In a recent working paper on management buyouts, Kaplan noted that the largest positive returns accrued to those buyouts financed with at least 70% debt. He estimated that the tax benefits gener­ ated by debt financing were worth between 30% and 130% of the premium paid to the pre-buyout shareholders.9 Potential Candidates The establishment of a statistically significant LBO profile could serve as a basis for predicting potential candidates. Practitioners have proposed different characteristics which would produce an ideal LBO candidate. Some suggested criteria include (1) a steady, reliable cash flow with a solid base for projecting its continuance; (2) a proven, reliable product; (3) a low to medium technology field; (4) key managers with proven track records; (5) key managers who will have their own money at stake after the buyout; (6) an up to date physical plant; (7) an ability to avoid higher than normal capital expenditures during the payback period; (8) a low level of pre-existing debt;10 (9) 10 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. substantial dividends; (10) operating in a noncyclical in­ dustry; and (11) general financial strength.11 Those companies which held the most allure prior to the Tax Reform Act of 1986 were those operating in basic industries which had lackluster appeal in the stock market or had fallen into disfavor with investors.12 These estab­ lished industrial firms were often found to be undervalued. A great deal of emphasis has been placed upon screens designed to find undervalued assets. A 1984 ar­ ticle in Forbes suggested choosing companies with (1) sales of 20 million to 5 billion; (2) a long-term debt to equity ratio of 35% or less; (3) a cash flow to debt ratio of .30 or greater; (4) a 12 month return on equity of 7% or more; (5) a positive 4 year average annual growth in earnings; (6) a current ratio of 1.43 or more; and (7) a price-earnings ratio of less than ten.13 Although LBOs have managed to survive and flourish in various economic climates, the new tax code has changed the complexion of the potential candidate. Since the abil­ ity to write off those assets which were sold in order to service the debt was eliminated, bust-up LBOs have become less popular. The asset rich firm, once an attractive can­ didate, has been replaced by a profile whose emphasis is on internally generated cash flows.1*1 11 R e pro duce d w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. History An understanding of the significance of leveraged buyout activity during the last decade can be enhanced by an historical perspective of merger activity in the United States over the last century. The first wave of merger activity occurred at the turn of the century from 1895 to 1904. Horizontal merger activity, or the combination of a number of companies in the same industry to form a large, single enterprise, pro­ duced such giants as U.S. Steel, E.I. DuPont and American Tobacco. From 1925 to 1931 the direction of merger activ­ ity changed from horizontal to vertical with the combina­ tion of manufacturers, suppliers and distributors into single economic entities, an example being General Foods. The third wave, beginning after World War II and ending in the late 1960s, was characterized by the combination of un­ related businesses, the era of conglomeration which pro­ duced Litton Industries and Textron.13 Leveraged buyout activity emerged in the early 1960s as a form of interim financing, and was frequently used by the owners of small, privately held businesses to attract competent owner-managers. Financing was confined to a select group of investment bankers and venture capital firms. During the era of deconglomeration, which began in the early 1970s, divestiture of marginally acceptable busi­ nesses acquired in the merger and acquisition era was made viable using significant amounts of debt financing. 12 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. In 1976, LBO activity was institutionalized with the establishment of the first commingled fund solely dedicated to financing leveraged buyouts. The fund, estab­ lished by Kohlberg Kravis Roberts & Co., was used in 1978 to finance the going-private transaction of Houdaille In­ dustries. With a newly established trend in place, commer­ cial banks began in 1982 to look to available and projected cash flows for securing credit. This accelerated LBO ac­ tivity with banks adding credibility and credit.16 LBOs have, in general, been financed by layers of debt and equity. Senior debt held by commercial banks has represented approximately 40%-65% of total financing needs. Approximately 15%-30% has been subordinated debt held by insurance companies and 5%-15% held as equity by various participants. The list of participants has included Wall Street firms, pension funds, wealthy individuals, venture capital firms and incumbent management. Venture capital firms in­ creased their commitment to LBOs beginning in 1984 by put­ ting up $481.46 million, an increase over the previous year of 71.9%.17 Wall Street firms such as Drexel Burnham & Lambert Inc., First Boston Corp., Shearson Lehman Bros., Merrill Lynch, and Bear Stearns and Co., have returned to merchant banking functions by putting their own money into these deals. Interest in these deals has spread to thrifts, as well as regional banks.10 13 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Senior debt held by banks earned, on the average, one percent over the prime rate without any additional risk beyond that of commercial loans• Returns on subordinated debt averaged above 20% and equity returns ranged from 35%-70%. Investment banking firms have charged a commis­ sion of approximately 1 percent on setting up these deals, with Kohlberg Kravis Roberts & Co. earning $45 million as advisory fees for the Beatrice buyout. Attractive second­ ary markets make debt liquid and encourage investors to participate, with an average holding period of less than two years.19 For those remaining equity owners, these high risk ventures have been compensated by returns on equity in ex­ cess of 35%, often in excess of 70%. This was accomplished when these firms returned to the public domain three to five years after having gone private. The restructuring, accomplished by management in a private setting, made this new entity more competitive and cost effective. The com­ pany emerged as one more highly valued by the marketplace. 14 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. A Hybrid in Theory The theory which describes the combining and recom­ bining of real assets differs from the theory which de­ scribes rearranging ownership and maintaining corporate control. Leveraged buyouts represent a hybrid in theory in the sense that the activity results in both restructuring and rearranging. Restructuring may be defined as any activity that is motivated by the desire to expand, as in the case of mergers, acquisitions, and joint ventures, or by the desire to sell off a subsidiary or division, as in the case of di­ vestitures . The theory of mergers, acquisitions, tender offers, and joint ventures places an emphasis on the benefits to shareholders that result from restructuring activity. These activities may be driven by the increased efficiency resulting from synergy, the benefits from the information or signalling of the event itself, a minimization of agency costs, increased market power, or possible tax benefits. A divestiture represents the transfer of assets to a higher valued use or to a more efficient user, a type of reverse merger, with synergistic benefits resulting from streamlin­ ing efforts.20 15 w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. The theory of maintaining corporate control seeks to explain different methods used to maintain the status quo ownership by preventing possible takeovers and dilu­ tion. Repurchase agreements, antitakeover amendments, and ways to deal with "greenmail" are included in corporate control theory or behavior.ax Changes in ownership affect the capitalization of the firm. In this context, theory must deal with changes in financial risk and expected changes in shareholder wealth. The purchase of a previously publicly held company through the use of debt, and little or no equity, consti­ tutes a change in ownership structure. Pure going-private transactions restructure corporate ownership by replacing the entire public stock interest with full equity ownership by an incumbent management group. With LBO type management buyouts, management proposes to share equity ownership in the subsequent private firm with third-party investors. Leveraged buyouts can be divestiture type transac­ tions or going-private transactions. With divestitures, the sell-off is accomplished by debt financing; the motiva­ tion is to eliminate inefficiencies. Going-private trans­ actions represent a different motivation. Presumably, man­ agement is motivated to arrange these deals because the potential exists for substantial gains to accrue to a pri­ vate company. These gains may be realized as management eventually takes the company public again. 16 R epro duce d w ith pe rm issio n of th e co p yrig h t ow ner. F urthe r re p rod uction prohibited w ith o u t perm ission. Therefore, LBO activity can be motivated by either synergy, as in the case of divestitures, or by potential wealth gains resulting from going-private transactions. It is important to distinguish between the two types of LBO activity since the term has taken on a generic context as applied to two entirely different economic events. Summary Leveraged buyouts represent a dominant financial trend over the last decade. They have been a significant force in financial markets. Past empirical studies have been confined to management buyouts without examining le­ verage driven deals. The available literature on mergers and acquisitions does not address the issues of go­ ing-private and impacts of debt on agency costs or changes in shareholder wealth. LBOs remain, in theory, a hybrid lacking empirical investigation. Practitioners, always in­ terested in creating more efficient markets, have suggested characteristics and conditions necessary for LBOs to flour­ ish. However, these remain speculative due to the lack of scholarly activity in the area. This study investigates those characteristics con­ sidered to be significant in differentiating firms which have changed to private ownership through a leveraged man­ agement buyout from those which remain publicly owned. A multivariate framework, using both multiple discriminant analysis and logistic regression, was developed to deter17 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. mine which attributes best distinguished these leveraged management buyouts from similar firms not going private. A test for abnormal returns for the actual LBO sample indicated that abnormal returns of up to 21.48 per­ cent accrued in the eighty day period surrounding the an­ nouncement date. For those firms predicted to be LBOs be­ cause of their statistically similar profile, abnormal returns of 32.68 percent were reported using 1986 as the test period. These findings imply that financial characteristics can provide a means by which firms going private through a leveraged buyout can be separated from others. In addi­ tion, they argue that an investor would have been able to earn abnormal profits by investing in both the actual LBOs, as well as those predicted to be LBOs. 18 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Chapter IX Literature Review A review of the literature revealed an absence of any empirical studies dealing specifically with leveraged buyout events. Most journal citations dealt with mergers, acquisitions, tender offers and pure going-private transac­ tions. LBO specific literature was confined to business periodicals whose readers are practitioners rather than re­ search oriented academicians. However, they have been in­ cluded in the bibliography and will be summarized in this chapter in an attempt to serve a broader based need. Studies Testing for Abnormal Returns The fundamental assumption of the existing lit­ erature is that capital markets are efficient with respect to publicly available information such as a merger, acqui­ sition, or going-private announcement. By efficient, it is meant that stock prices will instantaneously and fully re­ flect the information or signal that an announcement of this type will initiate. The difficulty is in separating the effect of these signals on security prices from the un­ derlying motivation of the event itself. Merger and acquisition studies related to pre-buyout shareholder gains for the owners of target firms 19 R epro duce d w ith perm issio n o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. indicate that it is possible to earn abnormal returns for the period preceding the announcement. Mandelker reported abnormal returns of 14% for a period of 7 months before the merger.22 Elgers and Clark found cumulative average re­ siduals of 43% for 2 years prior to the acquisition.23 Torabzadeh and Bertin, in a 1987 study of 48 lever­ aged buyouts occurring between 1983 and 1985, found excess returns of 23.26% accruing from 12 months prior to 4 months after the announcement date.2,4 With mergers and acquisitions, the motivation for increased value may be attributed to such synergistic ben­ efits as increased market power, economies of scale in pro­ duction, financial economies of scale, replacement of inef­ ficient management, and a change in systematic or unsystematic risk. These studies suggest that pre-buyout returns to the owners of the target firm represent a shar­ ing of these potential post-buyout gains by the acquiring firm. 20 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r reprod uction prohibited w ith o u t perm ission. Motivations for Going Private These synergistic benefits may not be present with a leveraged buyout. In a leveraged buyout, a public firm is reorganized into a private concern. This going-private transaction represents a change in both the ownership and capitalization of the firm. These transactions reduce the often conflicting interests of management and owners, as ownership and control emerge as the right and responsibility of management, the new majority interest. This stronger link between managerial performance and reward results in greater managerial incentives which are now internalized. In addition, the need to monitor management for owner related interests is all but eliminated. In a leveraged buyout, management increases its residual claim and establishes a stronger monitoring link with third-party investors. These third-party inves­ tors, who specialize in arranging these debt driven deals, are better at monitoring managerial decisions than public shareholders. The reduction of these costs, referred to as agency costs, provides a dimension to these activities that merger related theory does not possess. DeAngelo, DeAngelo, and Rice considered the capital structure effect of increased leverage to be a complemen­ tary monitoring effect. Additional leverage reduces the firm's cost of capital because of the tax advantage of debt. An optimal capital structure would substitute debt for equity financing in an attempt to minimize the cost of 21 R e pro duce d w ith perm ission o f th e co p yrig h t ow ner. F urthe r rep rod uction pro hibited w ith o u t perm ission. capital. Because of existing long-term institutidhal ties with lenders, these sawy third party equity investors are encouraged to borrow directly from these sources. Optimal borrowing strategies result in the complementary effect of minimizing the cost of capital.25 The resurgence of management stock, with the in­ creased level of LBO activity in the mid 1980s, provided evidence that agency costs can be reduced when management perceives their own best interest to be more closely aligned with those of existing shareholders. This class of stock is similar to an option, as management buys the right to trade this class of stock for voting privileged common stock. This option may be exercised provided certain goals are achieved, and management remains with the company. It is often referred to as a "golden handcuff" because it ne­ cessitates management's continued participation. The value of this class of stock is a direct func­ tion of management's ability to increase sales, cut costs, acquire assets, build up market share, increase profits, and most importantly, pay off the newly acquired debt. Lenders are receptive to this idea since rewards are distributed only after the debt is repaid. The only drawback is the potential dilution to existing shareholders upon issuance of the new common stock. 22 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Potential Gains to Shareholders Jensen hypothesized that the benefits that accrue with a leveraged takeover will exceed those accomplished by an exchange of stock.as The assumption is that debt is an effective substi­ tute for dividends. Dividends are paid out when the inter­ nal rate of return on existing projects is less th=n o>: equal to the cost of capital. Shareholder wealth would be reduced if earnings were reinvested at a rate less than the cost of capital, and would remain unchanged if they were equal. If all projects that had a positive net present value could be funded, and excess cash was available, Jensen argued that this "free cash flow" could be used to service debt instead of paying dividends. This would re­ duce the amount of discretionary income available to manag­ ers . If agency costs measure the costs of conflict be­ tween managers and owners, these costs could be reduced when managers were forced to disburse the cash rather than investing it at a rate below the cost of capital.27 This control function of debt is most important in organizations that have low growth prospects but generate large cash flows. Therefore, desirable LBO candidates will be those that exhibit these characteristics. The benefits that accrue to the remaining shareholders after a leveraged takeover are a direct result of the role of debt in the re23 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. duction of agency costB. The necessity to use future cash flows to pay down the existing debt eliminates the need to monitor these funds and their use. DeAngelo, DeAngelo, and Rice investigated the moti­ vation and source of gains in going-private transactions.28 They hypothesized that "pure ownership structure changes can create significant productive gains."29 They gave rec­ ognition to a fundamental premise of this study — corpo­ rate combinations differ fundamentally from going-private transactions which simply restructure ownership of a single existing entity. DeAngelo et al. argued that the potential gains from going private can be passed on or shared with current owners. in The purpose of their study was to test for changes shareholder wealth at the time of the initial proposal, and to test for a net wealth effect in the event that a proposal was withdrawn. In estimating the wealth effects of going-private transactions, DeAngelo et al. approached the empirical question as an event study. They measured the announcement effect over a two day trading period, since it was not clear whether the announcement occurred during the day's trade or after the close of trade. Using the Wall Street Journal to be the primary source of announcement dates for these events between 197 3 and 1980, the full sample in­ cluded 72 firms, twenty-eight of which had undergone a le­ veraged buyout. 24 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. The empirical results showed an average change in stockholder wealth for going-private proposals at the an­ nouncement date to be 22.27%, and when including the pre-proposal period, the net wealth effect increased to 33.40%. The evidence seems to be consistent with their gains-sharing hypothesis.30 In a recent working paper, Kaplan cited four sources responsible for premiums paid to the existing shareholders in a management buyout: 1) a reduction in agency costs and improved incentives; 2) tax benefits; 3) wealth transfers from bondholders or employees to the in­ vestor group; and 4) privileged information held by manag­ ers concerning the value of the company.31 Tax benefits cited by Kaplan included increased in­ terest deductions generated by debt, higher depreciation expense due to the election of the stepped-up basis for ex­ isting assets, and the use of Employee Stock Ownership Plans (ESOPs) to finance a takeover. Both principal and interest are tax deductible when an ESOP loan is used to facilitate the buyout. The potential disadvantage of using an ESOP loan is the dilution of ownership when employees retain a large equity stake. Wealth transfers from other contracting parties could be expensive contracts broken with employees or the increased risk of pre-buyout debt. 25 R e produced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Finally, undisclosed information possessed by man­ agement concerning the proper valuation of the firm repre­ sents a potential source of value which could be trans­ ferred to the new equity owners. In a sample of 76 management buyouts, Kaplan found that pre-buyout shareholders earned a median premium of 42.3% over the market price two months prior to the an­ nouncement.32 Kaplan found efficiency gains and tax benefits to be the most important sources of value. He attributed less importance to wealth transfers and asymmetric information possessed by management. Will the existing shareholders participate in the potential gains that may result from a buyout transaction? If management's control of the transaction was absent and perfect competition in bidding existed, stockholders would stand to share these gains. With a leveraged buyout, the incumbent management is assumed to possess significant con­ trol over the event. Management may not have every incen­ tive to exercise their fiduciary responsibility to the ex­ isting shareholders. However, certain stockholder rights such as a possible minority veto, the right to refuse to sell an existing share, and the right to challenge the transaction with private litigation all serve to assure ex­ isting equity holders a share in potential gains.33 26 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Gains Axe Transitory Lowenstein, in his study of 28 management buyout proposals, viewed incentives and shareholder involvement from a slightly different perspective.34 Included in the sample were twenty-eight buyouts with large interest deduc­ tions, and an aggressive writeup of existing assets coupled with the accelerated depreciation of these assets. The cash flow made possible by the existing tax laws fa­ cilitated management purchase of these companies. The promise of future gains to management equity provided the incentive, and the tax law actually encouraged financing with ever increasing amounts of debt. Inherent in this arrangement was increased business risk, and a future dedicated to the repayment of this debt burden. Instead of providing flexibility, and room for growth and increased productivity, management found itself restricted by dedication to debt-equity ratios. The tax laws, as seen by Lowenstein, encouraged the government to become a 46% business partner, and gave management a short-term objective and time frame within which to oper­ ate. This "marriage of convenience" would be dissolved within three to five years, when tax advantages were ex­ hausted. Pressure to once again go public terminated the relationship without some companies having established or achieved any long-term goals or commitments. 27 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. It is within this context that Lowenstein viewed management buyouts (MBOs) as just financial transactions with little social value. In fact, he argued that they can be added to the list of hostile takeovers in which the com­ petitive bidding process may benefit shareholders, but does not contribute to any tangible long-term gain. The larger the dollar value of the transaction, the more likely it be­ came that a third party entered into the competitive bid­ ding process. stock The median premium over pre-announcement prices was 76% when the number of competitive bids was three or more, with a premium of 48% when fewer than three bids occurred.35 Billion dollar deals in the mid 1980s have replaced the million dollar deals of the mid 1970s. It is highly improbable that a billion dollar company goes private to eliminate the costs associated with the public disclosure, as closing costs on these deals would pay for a "generation of compliance costs."36 These companies, therefore, are not in the same class as those companies with depressed or insider ma­ nipulated earnings reflected in undervalued stock prices. In all 28 deals, all participants won with a median gain of 58% over pre-announcement market prices for existing share­ holders of the target company, cash and an increased equity stake for managers, average annual returns over 50% for in­ vestors , and above average market returns for commercial bankers. 28 R e pro duce d w ith pe rm issio n o f th e co p yrig h t ow ner. F urthe r re p rod uction prohibited w ith o u t perm ission. Lowenstein attributed the existence and success of these leveraged MBOs to the tax laws then in effect and predicted that change would occur with new legislation.37 The Tax Reform Act of 1986 has not produced the ex­ pected impact suggested by many. Statistics to date show a decline in both merger and LBO activity for 1987, followed by a resurgence in 1988 (see Tables 3 and 5). This is in spite of laws which now tax the capital gains of sellers at a maximum ordinary rate of 28%, as opposed to the previous maximum rate of 20%. Buyers have also lost use of the General Utilities Doctrine, which allowed generous depre­ ciation write-offs on the purchase price of assets. In ad­ dition, the ability of acquired companies to carry forward tax losses has been strictly limited. 29 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. TABLE 5 TEN-YEAR MERGER COMPLETION RECORD 1978-1988 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 Source: No. of Transactions % Change 1,530 1,560 2,329 2,298 2,391 3,164 3,437 4,381 3,920 3,487 +2.0 +49.3 -1.3 +4.1 +32.3 +8.6 +27.5 -11.8 -11.1 Value $ mil 34,197.2 32,882.8 70,064.4 60,697.8 52,691.4 126,073.7 145,464.3 204,894.6 177,203.3 226,642.6 % Change -3.8 +113.1 -13.4 -13.1 +139.3 +15.4 +40.9 -13.5 +27.9 Mergers and Acquisitions. (May/June 1989), 53. R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. GainB Measurements May Be Inappropriate The 1985 Annual Report to the Council of Economic Advisers concluded that buyouts are beneficial because of the increased shareholder wealth of the target company without a "commensurate decrease" in the value of the ac­ quiring company's shares.38 If stock prices are an "accu­ rate and unbiased estimate" of the true intrinsic value of a firm, then this belief may be true.39 Warren Law, in his 1985 article "Management versus the Wild Bunch," argued that management, operating in an environment which places such emphasis on instant success and gratification, would sacrifice worthwhile goals such as long-term growth and stability. As evidence of the "hidden costs" of takeovers, Law cites the migration of many of the brightest graduates of business and law schools to higher paying service jobs such as investment banking. It is his belief that this represents a great long-term cost to the production sector, which faces a crisis in attracting top management. These financial activities which revolve around business restructuring have contributed to stock price fluctuations which cannot be justified by a change in "ra­ tionally formed expectations."*° Therefore, the notion that takeover activities have contributed to social produc­ tivity as reflected in higher stock prices can only be believed if one accepts the proposition that prices do re­ flect a true intrinsic value, and therefore, these ac31 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. tivities have contributed to the fundamental value of a company based on earnings projections. Impetus For This Study The interest in this particular topic was generated by current events and by a 1983 published article by Wansley, Roenfeldt, and Cooley.41 The study involved test­ ing for abnormal returns for acquired companies, as well as those which exhibited a high degree of resemblance to these companies based on a statistically significant financial profile. Linear discriminant analysis was used to develop a classification function that provided a total classifica­ tion accuracy of 75.0 percent. The sample used in the pro­ file derivation included 44 firms which merged during 1975-1976, and 44 randomly selected non-merged firms whose fiscal years matched those of the acquired firms. Sig­ nificant predictor variables included the price-earnings ratio, long-term debt to total assets, natural log of net sales, compound growth on net sales, and market value of equity to total assets. The classification function was then applied to a sample of 754 firms, using financial data from fiscal years prior to 1977. Posterior probabilities were calculated, and the top 50 most likely candidates based on these prob­ abilities were placed in portfolios of 25 stocks each. The first portfolio represented firms with a posterior 32 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. probability of .92 or above. With the inclusion of the second portfolio, the probability dropped to .8b or better. Although 18 merged firms were included in the sample of 754 firms, none of the eighteen was included in these two top portfolios based on posterior probabilities. For these derived portfolios, abnormal returns for the top 50 likely candidates were estimated using a 21 month test period. The cumulative average residuals (CARs) of the top 25 candidates ranged from a high of 28.0 percent to a low of 17.1 percent at the close of the 21 month test period. The range for the top 50 showed a high of 22.3 percent to a low of 14.1 percent at closing. Maximum returns available from purchasing prior to an acquisition were then estimated using a sample of 101 firms which had merged during 1973-1977. Predicted returns were estimated using the two-factor capital asset pricing model (CAPM) developed by Black. Actual returns were pro­ vided by the Center for Research in Security Prices. The results showed that had investors been able to successfully identify the 101 companies used in the sample, and had the investor purchased the stock of these acquired companies from 30 months prior to the merger and up to one month immediately preceding the event, abnormal returns of up to 38.0 percent could have been captured. 33 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Possible Profiles for Management Buyouts Does the financial profile for a firm taken private and ex-quoted differ statistically from the company that continues to be listed as a public firm? Will the method or means by which the company is taken private provide unique characteristics which identify it as a leveraged transaction? The motivation for going private, as well as the method, has been a topic of past empirical studies. The buyout of publicly quoted companies by their management is seen as an important event representing a "remarriage of ownership and control."42 In a survey con­ ducted in 1984, financial officers provided clues to sig­ nificant predictors of management buyouts.43 With a pairing based on industry and asset size, 63 ex-quoted and still public firms were surveyed to identify predictors relevant to the buyout decision. The research showed that in the two years prior to the buyout, the most significant variables included the following: 1. Concentration of ownership - On the average, 56% of outstanding shares were in the hands of management and the Board of Directors for ex-quoted firms versus 38% for those remaining public. The significance is the ease of the task being accomplished with less resistance and lower premiums required. 2. Cash flow to net worth and cash flow to total assets where cash flow is defined as net income plus depreciation, depletion and amortization Here the test is for the capacity to finance the transaction. Both ratios were found to be statistically significantly higher for ex-quoted firms. 34 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r reprod uction prohibited w ith o u t perm ission. 3. Price/book value - The market price/book value per share was much lower for ex-quoted firms with the perception of the stock being highly undervalued. This would provide management with the greatest potential for gains measured by a return on their equity. 4. Dividend yield - The cash dividend as a per­ centage of stock price being much higher for ex-quoted firms which are viewed, a. priori. as mature, low-growth companies with high, stable cash flows. Here the conflict between owners and management becomes apparent and resolved with a change to private ownership status. Shareholders of the public company valued the stock for ordinary dividend income where management looked to long-term capital gains. In an attempt to maximize long-term capital gains, maximum future cash flows become all-important. Upon going-private, cash dividends for a majority of these ex-quoted firms dropped to zero. Summary Empirical studies related to mergers, acquisitions, and going-private activity have confirmed the existence of abnormal returns for pre-buyout shareholders. These stud­ ies have also identified significant predictors for each of these events. Using these financial predictors, it has been shown that abnormal returns also accrue to firms with similar financial profiles. However, there has been no comprehensive treatment of leveraged buyouts, a hybrid transaction. This study ad­ dresses itself to the proper placement of this business re­ structuring in theory, measures potential gains to pre-announcement shareholders, and identifies potential candidates. In conclusion, the existence of abnormal returns for these potential candidates is confirmed. 35 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Chapter III Methodology The purpose of this chapter is to identify the population and period of time studied, and to describe the sampling and data collection procedures, as well as the em­ pirical models and hypotheses to be tested. the A review of literature relevant to each topic is also included. Definition of the Population and Sample The population was defined as all companies that announced their intent to go private through a leveraged transaction during the period from January 1, 1984 to De­ cember 31, 1986. A list compiled by IDD Information Services, Inc.4'® provided 470 leveraged buyouts listed chronologically by announcement date. IDD used the first public announcement in The Wall Street Journal as the official announcement of intent. The list included both divestiture and going-private type leveraged buyouts. Since the focus of interest was going-private type leveraged transactions, the next step was to identify only those companies taken pri­ vate. This information was found using the COMPUSTAT Re­ search Status Report.45 36 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. When a company is deleted from the COMPUSTAT Indus­ trial Tape,46 its historical data is transferred to the COMPUSTAT Industrial Research Tape4"7 and is subsequently referenced in the Research Status Report. Codes for dele­ tion include a merger/acquisition, bankruptcy, liquidation, reverse acquisition, change of format, and now a private company. If the IDD supplied company was located on the Status Report and its deletion code indicated that it had gone private, it was retained as part of the sample. process yielded fifty-nine companies (see Table 6). This Pre­ sumably, the majority of the remaining 470 IDD identified companies were divestiture type LBOs. By four digit industry classification (DNUM) codes, this sample of 59 companies represented 41 industries. The largest representation by any single industry had a fre­ quency of seven (see Table 7). In order to select a sample to represent the non leveraged buyout population, all companies with matching four digit DNUMs were chosen from the COMPUSTAT Primary, Supplementary, and Tertiary Files. No match could be found for six leveraged buyouts and these LBOs were dropped from the sample. 37 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. TABLE 6 SAMPLE OF LEVERAGED BUYOUTS BY ANNOUNCEMENT DATE NAME ANNOUNCEMENT DATE Cone Mills Corporation Ti-Caro Incorporated Kane-Miller Corporation Axia Incorporated Cellu-Craft Incorporated Harte-Hanks Communications Company Blue Belle Incorporated After Six Incorporated Denny's Incorporated Malone and Hyde Incorporated Amerace Corporation Guardian Industries Corporation ARA Services Incorporated Palm Beach Incorporated CGA Computer Incorporated Brooks Fashion Stores Incorporated Miller Bros. Industries Incorporated SFN Cos. Incorporated American Sterilizer Company Parsons Corporation Cook International Incorporated Conair Corporation Caressa Group Incorporated Glosser Brothers Incorporated Jupiter Industries Incorporated Papercraft Corporation Calton Incorporated Storer Communications Perfect Fit Incorporated Mary Kay Cosmetics Incorporated Standun Incorporated Levi Strauss & Company First National Supermarkets Rival Manufacturing Beatrice Companies Swift Independent Corporation RH Macy Warnaco Incorporated Amsted Industries Incorporated 01--01-■84 01--25-■84 02--29-■84 03--06-■84 03--23-■84 03--28--84 05--03-•84 05--14-■84 05--30-•84 06--08--84 06--19-•84 07--79--84 07--12-•84 07--24-•84 07--25--84 08--06--84 08--16--84 08--23--84 09--19--84 09--21--84 11--09--84 12--12--84 12--31--84 01--28--85 03--06--85 04--09--85 04--11--85 04--22--85 04--13--85 05--30--85 05--30--85 07--11--85 09--12--85 09--24--85 10--16--85 10--17--85 10--21--85 lO--31--85 ll--04--85 38 R e pro duce d w ith perm issio n o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. TABLE 6 (cont.) NAME Republic Health Sybron Corporation Avondale Mills Sheller-Globe Corporation Revco DS Incorporated Revere Copper & Brass Ceco Industries Incorporated Riblet Products Corporation Kroy Incorporated Safeway Stores Incorporated Ealing Corporation Star-Glo Industries Incorporated Carrols Corporation Mayflower Group Incorporated Pawnee Industires Incorporated Lehigh Press Incorporated American Bakeries Aloha Incorporated Owens-Illinois Pandick Incorporated Source: ANNOUNCEMENT DATE 01-03-86 01-13-86 01-15-86 02-14-86 03-10-86 04-11-86 04-16-86 05-08-86 07-14-86 07-27-86 07-28-86 07-28-86 09-25-86 09-26-86 09-30-86 10-26-86 10 -22-86 11-17-86 12 -11-86 12-17-86 IDD Information Services Inc. R e pro duce d w ith pe rm issio n o f th e co p yrig h t ow ner. F urthe r re p rod uction prohibited w ith o u t perm ission. TABLE 7 INDUSTRY REPRESENTATION OF LEVERAGED BUYOUT ACTIVITY COMPUSTAT DNUM 1700 2000 2010 2050 2200 2300 2640 2711 2731 2750 2844 3041 3069 3079 3140 3211 3221 3320 3350 3420 3550 3630 3714 3790 3832 3842 3843 4210 4511 4891 5140 5311 5411 5621 5812 5912 6552 7379 7399 8060 8911 NAME Construction-Special Trade Food and Kindred Products Meat Products Bakery Products Textile Mill Products Apparel and Other Finished Pds. Convrt, Paprbrd Pd, Ex Contain Newspaper: Pubg, Pubg & Print Books: Pubg, Pubg & Printing Commercial Printing Perfume, Cosmetics, Toilet Prep Rubber and Plastic Hose, Belting Fabricated Rubber Products, NEC Misc Plastics Products Footwear, Except Rubber Flat Glass Glass Containers Iron and Steel Foundries Rolling and Draw NonFer Metal Cutlery, Hand Tools, Gen Hrdwr Special Industry Machinery Household Appliances Motor Vehicle Part, Accessory Misc Transportation Equip Optical Instruments and Lease Ortho, Prosth, Surg Appl, Supply Dental Equipment & Supplies Trucking, Local, Long Distance Air Transportation, Certified Cable Television Operators Groceries and Related Pds - Whsl Department Stores Grocery Stores Women's Ready-To-Wear Stores Eating Places Drug and Proprietary Stores Subdivid, Develop, Ex Cemetery Computer Related Svcs, NEC Business Services, NEC Hospitals Engr, Architect, Survey Svcs FREQUENCY 1 1 1 3 7 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Financial Ratio Analysis Financial ratio analysis and the use of financial ratios in a predictive capacity can be traced back to the early 1930s. In Horrigan's40 comprehensive review of the history of financial ratio analysis, he identifies Winakor and Smith as pioneers in the use of ratios for predicting financial behavior. In their early study of firms that ex­ perienced financial difficulties during the period 1923-1931, they studied twenty-one ratios over a period of ten years, concluding that net working capital to total as­ sets was the best predictor over time of business failure. The major limitation of their study was the lack of a con­ trol group. Horrigan cites their study and subsequent studies by Fitzpatrick, Ramser and Foster as significant contributions because of their scientific approach to the subject.49 One of the most comprehensive studies done com­ paring industry averages for "discontinued" firms against industry norms was published by Merwin in the 1940s and is still considered to be valid. He concluded that three ra­ tios were sensitive predictors to discontinuance: net working capital to total assets, net worth to debt and the current ratio. 41 R e produced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Possible Predictors Assuming that a leveraged management buyout trans­ action has characteristics which it shares with management buyouts or pure going-private transactions, measures sug­ gested as a result of a study done by Maupin, Bidwell and Ortegren were considered.30 Their study tested whether publicly quoted firms which change to closely-held ownership through management buyouts (ex-quoted firms) possess characteristics prior to the change which differentiate them from firms that remain publicly quoted. Using a paired sample of 63 ex-quoted firms, they surveyed 43 financial officers for a list of the most important factors in the management buyout deci­ sion. Pairing was based on similar industries and asset size. With the identification of the last year in which the ex-quoted firm was publicly quoted, all firms which re­ mained public and possessed the same SIC code were then as­ sessed for similar asset size. The quoted firm closest in total asset size was then paired with its ex-quoted coun­ terpart. Using discriminant analysis and data for the ten years prior to the buyout, the five most significant vari­ ables for the two years immediately prior to the buyouts were found to be concentration of ownership, cash flow to net worth, cash flow to total assets, price to book value and the dividend yield. A discussion of the significance of these will follow. 42 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. The limitations of the study, as cited by the au­ thors, included (1 ) the inability to generalize to a population of publicly-quoted firms because of the non-random paired sampling procedure; (2 ) testing of impor­ tant predictors was limited to the list submitted by the financial officers; (3) no consideration was given to the intertemporal effect over ten years; and (4) the actual date of the buyout, not the announcement date, was used as the initial point of reference. Some of the following measures were selected from the Maupin, Bidwell and Ortegren study. In addition to these, measures which may be specific to a leveraged event were included in an attempt to determine those predictors that would identify a management buyout consummated with debt. Also included were representative predictors of merger and acquisition studies. A. Measures of Cash Flow (where cash flow is defined as net income plus depreciation, deple­ tion and amortization) 1. Cash flow to sales, 2. Cash flow to total assets, 3. Cash flow to net worth, and 4. Cash flow to total debt. 43 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r reprod uction prohibited w ith o u t perm ission. These measures suggest that a more likely candidate for a buyout would have high, stable cash flows— those character­ istics possessed by mature, slow growth firms. B. Measures of Excess Cash (where cash is defined as cash plus marketable securities or short-term investments) 1. Cash to net worth, 2. Cash to sales, and 3. Cash to total assets. These measures suggest that a firm about to purchase all outstanding equity in order to take the company private would have a significant amount of cash on hand. C. Measures of Liquidity 1. Net working capital to sales, 2. Net working capital to total assets, and 3. Current assets to current liabilities. Presumably a firm about to take on significant amounts of debt would display a higher degree of liquidity. 44 R e pro duce d w ith perm ission o f th e co p yrig h t ow ner. F urthe r rep rod uction pro hibited w ith o u t perm ission. D. Measures of Leverage 1. Long-term debt to total stockholders' equity, 2. Total stockholders' equity to total debt, and 3. Earnings before interest and taxes to interest. The assumption would be that a company which is to become a leveraged buyout would ex-ante display a greater debt ca­ pacity. E . Measures of Profitability 1. Net income to net worth, 2. Net income to total assets, and 3. Price earnings ratio. The assumption is that a firm would more likely be a candi­ date in an undervalued situation. ratiowould A higher priceearnings make a buyout more difficult to achieve by re­ quiring larger amounts of capital. 45 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. F. Miscellaneous 1. Concentration of ownership - Institutional holdings to total shares outstanding - A buyout would be more easily accomplished with a greater concentration of ownership. 2. Price of common to book value per share Using high, low, and closing prices to common equity per share - Again, this is a measure of a possible undervalued situa­ tion . 3. Market share-Sales to total industry sales. 4. Dividend policy - High dividend yield rate and low dividend yield rate using dividends per share to high and low priceA company taken private through a leveraged transaction should have a higher dividend yield indicating the cash availability and low retention rate. Companies then taken private will divert this cash flow to debt servicing, as management focuses on debt pay down and capital gains rather than dividend income. 5. Natural logarithm of net sales - This is a proxy for size. 46 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Based on practitioner speculation, a composite profile would be a mature, slow growth company with high and stable cash flows, liquid and undervalued assets and little or no existing debt. Included in this scenario would be an experienced management team with a significant personal equity commitment to the future. Data Collection Thirty-five data items were taken from the Primary, Supplementary, and Tertiary Industrial Tapes for the years 1983, 1984, 1985, and 1986 for the non LBOs. Since the going-private event occurred throughout the period of 1983-1986, it was impossible to obtain all four years of public data for the LBO sample. Therefore, only the last year of public data was chosen for each LBO. In most cases, this was the year prior to the announcement date. The data for the LBO sample was taken from the COMPUSTAT Industrial Research Tapes. Some measure of concentration of ownership was also desired. As a proxy, institutional holdings as a percent­ age of total shares outstanding was used. Institutional holdings data was collected using the Standard and Poor's Stock Guide end of the year data published in the January edition. 47 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Testing for Any Industry Effect The first hypothesis that was tested assessed the likelihood of any industry effect present in the data. Thirty-six industries represented by the four digit COMPUSTAT industry classification scheme were used. The em­ pirical question is whether or not a particular industry is more likely to produce leveraged buyout candidates. If so, then the financial profile of that particular industry may differ significantly from the remainder, and some control or adjustment must be made to factor out the industry in­ fluence on its member firms. Multivariate analysis of variance (MANOVA) is a statistical technique that is concerned with the differ­ ences between groups. It tests whether or not samples come from populations with equal means. With MANOVA, the depen­ dent variable is a vector of mean responses. Each treat­ ment group is observed on two or more dependent variables simultaneously, and the hypothesis tested is the equality of dependent mean vectors.51 To test for this industry effect using the MANOVA procedure, industry means were calculated for each finan­ cial ratio for each industry. These industry means became the dependent variables with each industry representing the treatment effect. At this stage both the LBOs and non LBOs were grouped by industry. The industry effect was found to be statistically significant for each of the four years of this study. The 48 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. thirty-five financial ratios for each company were then ad­ justed to eliminate any industry specific influence on the data. Factor Analysis to Eliminate Multicollinearity Financial ratios are highly susceptible to multicollinearity since some linear combination of the nu­ merator and denominator can be expressed for sets of ra­ tios. Factor analysis is a statistical technique whose ob­ jective is to represent a set of variables in terms of a smaller number of hypothesized variables. Donald Stevens, in order to examine the financial characteristics of merged firms, used multiple discriminant analysis (MDA) to develop a linear model that discriminated the acquired from non-acquired firm.3* Because the se­ lected predictive ratios were expected to exhibit a high degree of multicollinearity, he found it necessary to re­ duce the data in such a way that the independent variables were mutually uncorrelated. The data was first subjected to factor analysis to simplify or discover group patterns or factors representa­ tive of sets of ratios. The original ratios were reduced to a smaller number of factors which were independent or orthogonal, with each factor becoming a linear combination of the original ratios. Using the factor loadings as a ba­ sis for his selection of a ratio to represent these fac­ tors, Stevens chose earnings before interest and taxes to 49 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. sales as a measure of profitability, net working capital to total assets as a measure of liquidity, sales to total as­ sets as a measure of efficiency, and long-term liabilities to total assets as a measure of leverage. Eventually, MDA found these four factors to be statistically significant in the classification of firms into two distinct groups. Because the financial ratios chosen for this study were not mutually uncorrelated, factor analysis was em­ ployed. The original data set of thirty-five ratios was reduced to nine ratios. STEPWISE multiple discriminant analysis was then used to identify those characteristics that differentiate the 53 LBOb from the sample of non LBOs. Classification and Multiple Discriminant Analysis Multiple discriminant analysis is a classification technique that utilizes the a, priori identification of group memberships to classify entities into mutually exclu­ sive groups by the statistical decision rule that maximizes the ratio of among-group variance to within-group variance using the set of independent variables. This multivariate classification technique also provides esti­ mates of the contribution of each independent variable to group discrimination. The purpose of multiple discriminant analysis (MDA) is both to identify the variables that are important in separating group centroids and to place original or unclassified observations into one of these groups. 50 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. In a comprehensive review of classification results and multiple discriminant analysis, Pinches reviewed the methodological and statistical problems encountered in business studies and offered some solutions.33 The problems encountered by business researchers include a lack of knowledge of the statistical properties of the methodology, problems associated with business data, and a lack of solutions provided by statisticians. Pinches stressed the importance of recognizing these limitations. Factors that may affect classification results by affecting the group centroids and/or dispersion matrices include multivariate non-normality, the number and indepen­ dence of variables, sample size, missing values and initial misclassification. The presence of multivariate non-normality inter­ acts with other factors to influence classification re­ sults . Tests for the equality of group centroids and/or dispersion matrices, error rates for both the linear and quadratic discriminant function, and the number and inde­ pendence of predictor variables may all be influenced by the presence of non-normality. The number of predictor variables and sample size also affect classification results. Pinches concluded that seeking a reduction in the number of predictor variables makes good sense statistically in order to avoid the prob­ lem of multicollinearity so often present in business data. The larger the sample size, the more likely it is that the 51 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. researcher will reject the null hypothesis of equality of group means, and therefore accept classification as an ap­ propriate methodology. However, the larger the sample size, the higher the probability the researcher will also conclude that the dispersion matrices are unequal, indicat­ ing quadratic classification techniques may be preferred to linear discriminant analysis.34 At the time of this Btudy, no significant research had been conducted on the effect of missing values since most statistical packages or studies had screened out ob­ servations with missing variables. With respect to sample sizes and a priori probabilities, the study concluded that a priori probabilities assigned directly affect classifica­ tion results. Therefore, if unequal sample sizes dictate a priori probabilities, the classification results will be influenced.33 The last factor, initial misclassification, pre­ dicts that results will be biased if initial misclassification is non-random because its classification is based on its closeness to the established group norms.36 The effects of these factors cannot be influenced by the researcher once the sample and variables are se­ lected. However, Pinches placed error rates, a priori probabilities, and equal/unequal dispersion matrices into a category that he considered to be under the control of the researcher and which could be accounted for before classi­ fication was undertaken. A priori probabilities can be ad52 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r reprod uction prohibited w ith o u t perm ission. justed to reflect the population priors, error rates can be estimated and accounted for, and the equality of dispersion matrices can be estimated and quadratic classification used if necessary. Overall, these factors, as Pinches concludes, "are easy to control for and conceptually easier to deal with than those factors that affect the group centroids or dis­ persion matrices."37 Several financial studies involving mergers and ac­ quisitions have been conducted using MDA. tions have been mentioned in these studies. Certain limita­ However, no conclusive evidence has been submitted which renders this method obsolete. Press and Wilson continued the discussion of non-multivariate normality and unequal variance-covariance matrices, concluding that MDA should not be used under these conditions.se According to Eisenbeis, violations of these assump­ tions bias the tests of significance for each independent variable and estimation of the error rates. In addition, with MDA, only the contribution of the i^1* independent variable to the Mahalanobis distance between the group means is considered. Therefore, as compared to regression techniques, there is no meaning to the sign or absolute value of these variables. With regression techniques, in­ terpretation of the coefficients, as well as their sig­ nificance, can be tested.39 53 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Methods proposed for testing the importance of the variables using discriminant analysis include: (1) sig­ nificance of their univariate F-statistics; (2) stepwise forward methods based upon their contribution of the multivariate F-statistic used in testing for the difference in group means; (3) backward stepwise methods; and (4) con­ ditional deletion, which measures the additional contribu­ tion of the variable to the multivariate F-statistic.60 Wansley, Roenfeldt and Cooley used linear discriminant analysis in their study measuring abnormal re­ turns from merger profiles. In using linear discriminant analysis rather than quadratic when the test for equal variance-covariance matrices was rejected at the .01 level, the authors justified the method used based on a higher correct classification rate with smaller samples. They be­ gan with twenty variables representing ten dimensions of a firm's financial profile and made no test or adjustment for multicollinearity.61 If robustness of classification is the only crite­ rion, then the methodology is most appropriate. However, using the results beyond the classification stage may give misleading conclusions regarding predictor variables and their interdependent relationship. 54 R eproduced w ith perm ission o f the co p yrig h t ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. Intertemporal Effects An article written by Joy and Tollefson,62 and later critiqued by Altman and Eisenbeis,63 discussed the difference between the stability of the discriminant model and its predictive power. The original article by Joy and Tollefson was in reference to Altman's application of discriminant analysis to bankruptcy data.64 It is Joy and Tollefson's belief that in studying events that occur over time and which use time series data as explanatory or predictive variables, the stability of the relationships between and among these variables may change. Therefore, discriminant analysis using time series data is only valid for predictive purposes when these rela­ tionships and parameters are stable over time. Even the use of a holdout sample will not resolve this problem, un­ less the same conditions exist for the period of the hold­ out sample as exist for the period which was used to de­ velop the discriminant function. Inferences about the role of variables would only be valid for the period during which the sample has been taken. Therefore, MDA is incor­ rectly used as a predictive tool unless stationarity is as­ sumed.63 Because this study is concerned with the ability to correctly predict LBO activity, logistic regression (logit) was used as a methodology of choice in addition to multiple discriminant analysis. In order to avoid the problems of stability of the discriminant model over time, each of the 55 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. four years was treated separately. A discriminant function was developed for each year and then applied to a holdout sample for the same year. Probit and Logit Probit and logit are both statistical methodologies which deal with predicting the level of a dependent vari­ able that is non-metric or possesses an upper and lower limit. Both are regression techniques that have relaxed assumptions concerning the error term. Their use does not require the assumption that the expected value of the error term is zero and is normally distributed with constant variance (homoschedasticity). Because these assumptions are not necessary, probit and logit can be used under con­ ditions of non-normality; the resulting regression coeffi­ cients will have meaning and can be used to make probabil­ ity statements about given responses. The difference between probit and logit lies in the assumption made about the frequency distribution which relates the qualitative variables to the independent variables. Logit or logistic regression assumes these are related through a logistic cu­ mulative density function; probit assumes a normal cumula­ tive density function. According to Doyle, the choice is a matter of opinion and makes no difference in terms of re­ sults .66 56 R eproduced w ith perm ission o f the cop yrig ht ow ner. F urthe r rep rod uction prohibited w ith o u t perm ission. In statistical models whose dependent variable can take on limited values such as "0 " or "1 " to indicate yes or no, success or failure, multiple discriminant analysis has traditionally been used to classify or discriminate ob­ servations into a priori determined groups. Probit or logit has been used as a regression technique which is in­ terested in establishing a statistical relationship between the qualitative or dependent variable and one or more inde­ pendent variables. Multiple discriminant analysis is theoretically correct only when grouping populations which have normal density distributions and identical covariance matrices. It is apparent that a dependent variable whose only values are "0 " or ”1 " will therefore have error terms that cannot be assumed to exhibit the normality assumption. However, the classification results do not appear to be sensitive to non-normality. Ingram and Frazier, in a 1982 study, assessed al­ ternative multivariate tests in limited dependent variable models by comparing the results of a study in mortgage lending discrimination using both logit and MDA .67 Diver­ gent statistical results would involve serious legal impli­ cations in mortgage lending. The results showed that there was a small, insignificant difference in the classification accuracy using each methodology. However, there was a sub­ stantial difference in the conciusions reached as to the significance of the predictor variables. Not only were 57 R e produced w ith perm ission o f the cop yrig ht ow ner. F u rth e r rep rod uction prohibited w ith o u t perm ission. different variables included depending upon which methodol­ ogy was used, but the direction of correlation was found to be reversed in some instances.60 Anderson provides additional support for this choice by pointing out that the rationale for a logistic formulation results from a variety of underlying assump­ tions.69 The assumptions concern the distribution of the explanatory variables. Logistic regression can be used when these variables are assumed to be multivariate nor­ mally distributed with equal covariance matrices, or when they are independent and dichotomous, or a mix. the Therefore, use ofcategorical or dummy variables isbetter by this served technique, as the model is relatively robust. In this study, the dependent variable took on a value of "1 " as a member of the leveraged buyout sample and a value of "0" as a member of the non-buyout sample. Lo­ gistic regression produced a probability statement which is a linear function of the predictor variables and the re­ gression coefficients. Consider a sequence of n indepen­ dent observations n , r 2 .... in where ri= { "1 " if a leveraged buyout rj.= { "0 " if not With each ri is associated a vector of P predictor variables (l,xn,xi2 ,...xiP ) = Xj.'. P[ri=l] = ______ 1______ = We may postulate that where 1' = (0o,Pi,.. .0*,) 1 + e “ *■* o ia f 5 *? 3n 8 82 sr for 1983 FACTOR PATTERN for the INITIAL UffiOTATED SOLUTION 8I 5OO k H n h o S5~ S8^ d o ill s o o s s i o oo ' » o o o S 5! d «j» « odd * 5 ft «?[...]... the leveraged buyout of Gibson Greeting Cards in 1983.3 This study answers three empirical questions: 1) do abnormal returns accrue to actual LBOs for a trading period surrounding their announcement date? 2) do the financial profiles of these LBOs provide ex ante identification of this event? and 3) is it possible to invest in companies with similar financial profiles and also earn abnormal re­ turns... Testing for Any Industry Effect 48 Factor Analysis to Eliminate Multicollinearity 49 Classification and Multiple Discriminant Analysis 50 Intertemporal Effects IV 55 Probit and Logit 56 Selection of Likely Candidates 59 Efficient Markets and Event Studies 60 Models Used for Residual Analysis 63 The Test 64 Summary of Methodology 66 Empirical Analysis and Findings Initial Empirical. .. official unofficial editors were Marie Dickson, Alice Persons, and Gary Lombardo, who spent many sleep-filled nights reading my copy Thanks also to Lynne Cote and Anne LaLime for their typing assistance and to Jim Brady and Casandra Fitzherbert for library assistance And literally in my hour of need, Tom Sanders graciously filled in for me Thank-you Dr Hardin for persevering and giving me the hope and encouragement... Merger and acquisition related research took this hypothesis one more step by testing for abnormal returns for firms with similar financial profiles Would firms with similar financial profiles, which never were acquired, also produce abnormal returns? Are these firms subject to a "spillover effect" because investors consider them to be potential candidates? Studies conducted by Stevens,6 Simkowitz and... Analysis of Variance 68 Reasons for Factor Analysis 70 Results of Factor Analysis 76 Stepwise Procedures 94 Results of Logistic Regression (LOGIST) 101 Abnormal Returns From the LBO Sample 104 Abnormal Returns for Predicted LBOs 110 Summary of Results 112 R eproduced w ith perm ission o f the co p yrig h t ow ner F urthe r rep rod uction prohibited w ith o u t perm ission VSummary and... financial risk and expected changes in shareholder wealth The purchase of a previously publicly held company through the use of debt, and little or no equity, consti­ tutes a change in ownership structure Pure going-private transactions restructure corporate ownership by replacing the entire public stock interest with full equity ownership by an incumbent management group With LBO type management buyouts,. .. u t perm ission History An understanding of the significance of leveraged buyout activity during the last decade can be enhanced by an historical perspective of merger activity in the United States over the last century The first wave of merger activity occurred at the turn of the century from 1895 to 1904 Horizontal merger activity, or the combination of a number of companies in the same industry... legal entity However, in a going-private transaction, a pub­ licly owned company is taken private by a group of inves­ tors which often includes existing management This is re­ ferred to as a management buyout, or pure going-private transaction The focus of this study is leveraged management buyouts, going-private transactions financed by significant amounts of debt R e pro duce d w ith pe rm issio n... possible to earn abnormal returns for the period preceding the announcement Mandelker reported abnormal returns of 14% for a period of 7 months before the merger.22 Elgers and Clark found cumulative average re­ siduals of 43% for 2 years prior to the acquisition.23 Torabzadeh and Bertin, in a 1987 study of 48 lever­ aged buyouts occurring between 1983 and 1985, found excess returns of 23.26% accruing... era of conglomeration which pro­ duced Litton Industries and Textron.13 Leveraged buyout activity emerged in the early 1960s as a form of interim financing, and was frequently used by the owners of small, privately held businesses to attract competent owner-managers Financing was confined to a select group of investment bankers and venture capital firms During the era of deconglomeration, which began