USF Libraries
USF Digital Collections

Two essays on the conflict of interests within the financial services industry-financial industry consolidation

MISSING IMAGE

Material Information

Title:
Two essays on the conflict of interests within the financial services industry-financial industry consolidation the motivations and consequences of the Financial Services Modernization Act (FSMA) and "down but not out" mutual fund manager turnover within fund families
Physical Description:
Book
Language:
English
Creator:
Bryant, Lonnie Lashawn
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Agency issues
Banking
Management
Acquisitions
Replacement
Dissertations, Academic -- Finance -- Doctoral -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: The objective of this paper is to examine the impact the Financial Services Modernization Act (FSMA) of 1999 has on the consolidation of the banking industry. The FSMA allows banks to simultaneously offer commercial banking, investment, and insurance services. I find a strong positive market response to the announcement of bank acquisition of brokerage firms (10.2 percent) and insurance companies (9.3 percent), but no significant response to bank acquisitions. I also find support for two complimentary hypotheses that explain the long-run returns to the acquiring banks. The "product-market spillover hypothesis" states that the post-consolidation returns of the acquirer are directly related to the banks' ability to cross market their products and services to a more diverse client base, while the efficiency hypothesis states that banks acquire financial services companies to realize efficiency gains resulting from exploiting economies of scale.Finally, I show that the premiums paid in the post-FSMA acquisitions increases with the diversity of the transaction. In addition, this study is the first to link managerial turnover to mutual fund managerial structure in a manner that indicates the strong presence of a conflict of interests between investors and fund sponsors in an area of fund governance where we have been led to believe there are strong and well-functioning mechanisms to guard against the exploitation of investors. I utilize the unique characteristics of mutual funds where managers sometimes manage multiple "firms" simultaneously, something not generally observed in industrial firms. I test the governance mechanisms using the mutual fund complexes management structure; unitary and multiple fund management (UFM and MFM). This study shows that UFMs tend to have higher asset growth rates and higher fees than MFMs, suggesting that sponsors can benefit more from keeping them intact.I find that changing managers under the UFM is more costly to sponsors making them more reluctant to fire poor performers. I document that underperforming UFM are -2.77 percent less likely to be replaced than their underperforming MFM counterparts. In addition, the conflict of interests affect the replacement decision, as high expense ratio fund managers have a lower probability of replacement for a given level of underperformance.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2008.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Lonnie Lashawn Bryant.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 123 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002001266
oclc - 319691731
usfldc doi - E14-SFE0002624
usfldc handle - e14.2624
System ID:
SFS0026941:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 002001266
003 fts
005 20090430105750.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 090430s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002624
035
(OCoLC)319691731
040
FHM
c FHM
049
FHMM
090
HG173 (Online)
1 100
Bryant, Lonnie Lashawn.
0 245
Two essays on the conflict of interests within the financial services industry-financial industry consolidation :
b the motivations and consequences of the Financial Services Modernization Act (FSMA) and "down but not out" mutual fund manager turnover within fund families
h [electronic resource] /
by Lonnie Lashawn Bryant.
260
[Tampa, Fla] :
University of South Florida,
2008.
500
Title from PDF of title page.
Document formatted into pages; contains 123 pages.
Includes vita.
502
Dissertation (Ph.D.)--University of South Florida, 2008.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
520
ABSTRACT: The objective of this paper is to examine the impact the Financial Services Modernization Act (FSMA) of 1999 has on the consolidation of the banking industry. The FSMA allows banks to simultaneously offer commercial banking, investment, and insurance services. I find a strong positive market response to the announcement of bank acquisition of brokerage firms (10.2 percent) and insurance companies (9.3 percent), but no significant response to bank acquisitions. I also find support for two complimentary hypotheses that explain the long-run returns to the acquiring banks. The "product-market spillover hypothesis" states that the post-consolidation returns of the acquirer are directly related to the banks' ability to cross market their products and services to a more diverse client base, while the efficiency hypothesis states that banks acquire financial services companies to realize efficiency gains resulting from exploiting economies of scale.Finally, I show that the premiums paid in the post-FSMA acquisitions increases with the diversity of the transaction. In addition, this study is the first to link managerial turnover to mutual fund managerial structure in a manner that indicates the strong presence of a conflict of interests between investors and fund sponsors in an area of fund governance where we have been led to believe there are strong and well-functioning mechanisms to guard against the exploitation of investors. I utilize the unique characteristics of mutual funds where managers sometimes manage multiple "firms" simultaneously, something not generally observed in industrial firms. I test the governance mechanisms using the mutual fund complexes management structure; unitary and multiple fund management (UFM and MFM). This study shows that UFMs tend to have higher asset growth rates and higher fees than MFMs, suggesting that sponsors can benefit more from keeping them intact.I find that changing managers under the UFM is more costly to sponsors making them more reluctant to fire poor performers. I document that underperforming UFM are -2.77 percent less likely to be replaced than their underperforming MFM counterparts. In addition, the conflict of interests affect the replacement decision, as high expense ratio fund managers have a lower probability of replacement for a given level of underperformance.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Co-advisor: Scott Besley, DBA
Co-advisor: Christos Pantzalis, Ph.D.
653
Agency issues
Banking
Management
Acquisitions
Replacement
690
Dissertations, Academic
z USF
x Finance
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2624



PAGE 1

Two Essays on the Conflict of Interests within the Financial Services Industry-Financial Industry Consolidation: The Motivations and Consequences of the Financial Services Modernization Act (FSMA) and Down but Not Out Mutual Fund Manager Turnover within Fund Families by Lonnie Lashawn Bryant A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Finance College of Business Administration University of South Florida Co-Major Professor: Scott Besley, D.B.A. Co-Major Professor: Christos Pantzalis, Ph.D. Delroy Hunter, Ph.D. Michael Loewy, Ph.D. Date of Approval: June 16, 2008 Key Words: Agency Issues, Banking, Management, Acquisitions, Replacement Copyright 2008, Lonnie L. Bryant

PAGE 2

Table of Contents List of Tables iii Abstract v Essay 1Financial Industry Consolida tion: The motivations and consequences of the Financial Servic es Modernization Act Introduction 1 Financial Services Industr y Legislation Background 5 Banking Industry Literature Review 8 Literature on the Depository Instituti ons Deregulation and 11 Monetary Control Act of 1980 Literature on Interstate Banking and Br anching Efficiency 12 Act of 1994 Literature on the Financial Services Modernization Act of 1999 14 Literature Review on Non-Financial Co rporate Acquisitions 15 and Conglomerates Hypotheses Development: Efficiency Hypothe sis versus Spillover Hypothesis 19 Efficiency Hypothesis 20 Product Market Spillover Hypothesis 22 Data Description and Empirical Methodology 25 Data Description 25 Methodology 27 Empirical Analysis 29 Descriptive Statistics for financial institutions 29 Financial Services Industry M&A Transaction 32 Value and Premium Financial Services Industry M&A Transaction 37 Premium Regression Financial Services I ndustry Announcement-Day Returns 40 Financial Services Acquirers Efficiency Measurement and 41 Correlations Univariate Efficiency Regression Results 43 Financial Services Industry Pr oduct Market Spillover 45 Measurement Financial Services Industr y Product Market Spillover 46 Multivariate Regression Results Conclusion 51 Essay 2- Down but Not Out Mutu al Fund Manager Turnover within Fund Families Introduction 54 i

PAGE 3

Related Literature/ Hypotheses Development 58 Agency Issues 58 Past Performance 60 Management Structure 61 Expense Ratio and Management Fee 62 Fund Size 63 Manager Tenure and Reputation 64 Style Drift/ Tracking Error 65 Fund Styles/ Competency 66 Data and Descriptive Statistics 66 Sample Selection Procedure 66 Description of full sample 69 Methodology 80 Empirical Results 82 Performance-Replacement Relationship 82 Univariate Logistic Analysis 88 Logistic Regressions 92 Managerial Turnover from De motion 103 Conclusion 107 Reference 109 About the Author End Page ii

PAGE 4

List of Tables Table 1 Description of Financial Industry Transactions 31 Table 2 Financial services Industry M&A Transaction Value and 33 Premium Table 3 Correlation analysis of Financia l Service Industry Acquisition 36 from 1999-2002 Table 4 Financial Service Industry Premiu m regression Analysis 39 Table 5 Financial Services Industr y M&A Announcement Day Returns 41 Table 6 Correlation of Efficiency Measures 42 Table 7 Univariate Analysis of Financ ial Service Industry M&As with 44 Efficiency Measures Table 8 Financial Service Industry Post-C onsolidation Changes 46 Table 9 Multivariate Analysis of Fina ncial Service Industry M&As with 48 the Profit Efficiency Measure Table 10 Multivariate Analysis of Fina ncial Service Industry M&As with 50 the Operation Efficiency Measure Table 11 Managerial Replacement Distribution 70 Table 12 Descriptive Statistics 72 Table 13 Multiple Fund and Multiple Objective Preliminary Statistics 77 Table 14 Managerial Replacement Samp le Distribution 79 Table 15 Performance in the years Prea nd PostManager Turnover: 83 Full Sample Table 16 Performance in the years Prea nd PostManager Turnover: 85 Multiple Fund Management (MFM) Sample Table 17 Performance in the years Prea nd PostManager Turnover: 86 Unitary Fund Management (UFM) Sample Table 18 Mutual Fund Manager Replacement Univariate Regressions 89 iii

PAGE 5

iv Table 19 Multivariate Regression results for all Mutual Fund Manager 93 Replacements: Unitary Fund Management Specific Table 20 Multivariate Regression results for all Mutual Fund Manager 97 Replacements: High Management Tenure Specific Table 21 Multivariate Regression results for all Mutual Fund Manager 101 Replacements: Total Expenses Specific Table 22 Multivariate Regression results fo r Poor Performing Funds 105

PAGE 6

Two Essays on the Conflict of Interests within the Financial Services Industry-Financial Industry Consolidation: The Motivations and Consequences of the Financial Services Modernization Act (F SMA) and Down but Not Out Mutual Fund Manager Turnover within Fund Families Lonnie L. Bryant ABSTRACT The objective of this paper is to examine the impact the Financial Services Modernization Act (FSMA) of 1999 has on th e consolidation of the banking industry. The FSMA allows banks to simultaneously o ffer commercial bank ing, investment, and insurance services. I find a strong positive market response to the announcement of bank acquisition of brokerage firms (10.2%) and insurance companies (9.3%), but no significant response to bank acquisitions. I also find support for two complimentary hypotheses that explain the long-run returns to the acquiring banks. The product-market spillover hypothesis states that the post-consol idation returns of the acquirer are directly related to the banks ability to cross market their products and services to a more diverse client base, while the efficiency hypothesis states that banks acquire financial services companies to realize efficiency gains resulting from exploiting economies of scale. Finally, I show that the premiums paid in the post-FSMA acquisitions increases with the diversity of the transaction. In addition, this study is the first to link managerial turnover to mutual fund managerial structure in a manner that indi cates the strong pres ence of a conflict of interests between investors and fund sponsor s in an area of fund governance where we have been led to believe there are strong and well-functioning mechanisms to guard against the exploitation of investors. I utilize the unique characteristics of mutual funds where managers sometimes manage multiple firms simultaneou sly, something not v

PAGE 7

vi generally observed in industrial firms. I test the governance mechanisms using the mutual fund complexes management structure; unita ry and multiple fund management (UFM and MFM). This study shows that UFMs tend to have higher asset growth rates and higher fees than MFMs, suggesting that sponsors can benefit more from keeping them intact. I find that changing managers under the UFM is more costly to sponsors making them more reluctant to fire poor perf ormers. I document that underperforming UFM are -2.77% less likely to be replaced than th eir underperforming MFM counterparts. In addition, the conflict of interest s affect the replacement decision as high expense ratio fund managers have a lower probability of replacement for a given level of underperformance.

PAGE 8

Essay 1 Financial Industry Consolidation: Th e motivations and consequences of the Financial Services Modernization Act Introduction On November 12, 1999, the United States Congress passed the GrammLeach-Bliley Financial Services Moderniz ation Act (FSMA) allowing competition between commercial banks, brokerage firm s and insurance companies. The FSMA repealed the Glass-Steagall Act of 1933 wh ich prohibited banks from simultaneously offering commercial banking, investment, and insurance services. The FSMA allowed commercial and investment banks to consolid ate; cross-selling ba nking services with insurance services, brokerage services, a nd other financial services. The combined industries are now known as the financial services industry. The implementation of the new legislation and hence, the incepti on of the financial services industry has resulted in unprecedented merger and acquisition activities (M&A) because financial companies are electing to purchase existing expertise in diverse financial services versus growing theses services organically. The ratification of the FSMA has led to intense competition between financial services companies to manage a diverse por tfolio of financial service products. These regulatory changes have mixed implications regarding whether diversified financial service firms will provide better products and services than specialized financial services firms and, therefore, whether the FSMA adds to shareholder value. On one hand, diversification reduces costs due to economies of scale (Kwan and Laderman (1999)). However specialization produces a greater quality and/or variety of financial services potentially increas ing sales (Berger, Demsetz and Strahan (2000)). A second implication is that the FSMA a llows financial services companies to participate in commercial banking, investment brokering, and insurance activities, providing customers with the convenience of having all their financial service needs 1

PAGE 9

met at a single location. However, the Glas s-Steagall Act was initially passed due to improper banking activity. The Act prohibited banks from participating in diverse security activities to protect depositors from the additional risk associated with security transactions. It was initially thought that banks that offer investment banking services and mutual funds were subject to conflicts of in terest and other abuses. For instance, a commercial banks financial in terest in the ownership, price, or distribution of securities w ould lead to increased pressure on banking customers to invest in securities that the bank sells. The repeal of the Glass-Stegall Act with the Financial Services Modernization Act s uggests that these ba nking conflicts of interests are now of minimal concern1 and the consolidation of the financial industry will enhance social welfare. Thus, it is an empirical question whether or not the regulatory changes and the resulting changes in the structur e of the financial industry add value to shareholders. The purpose of th is paper, therefore, is to investigate whether and how financial services industry acquisitions in the post-FSMA period affect shareholder value. I examine two hypothesesthe efficiency hypothesis and the product market spillover hypothesisto explain the source of any increase in shareholder value resulting from financial services industr y acquisitions in th e post-FSMA period. Theoretical arguments predict that merger s and acquisitions are motivated by target efficiency improvements that can be achieved by the acquirer (Calomiris and Karceski (1998) and Rhoades (1998)). Be rger and Humphrey (1992) find that acquiring banks are more cost efficient than target banks. Efficiency may also be improved by M&A if increased diversifica tion improves the risk -return relationship. Thus, the efficiency hypothesis states that acquiring banks increase shareholder value by purchasing other financial service indus try companies and effectively managing the combination of the two firms more e fficiently than the target management. The corporate M&A literature suggests that diversifying acquisitions are generally value-reducing, and that increases in corporat e focus are value-enhancing 1 This might be because there are mechan isms in place to minimize the conflicts. 2

PAGE 10

(Lang and Stulz (1994), Jensen and Ruback (1984), Berger and Of ek (1995) and John and Ofek (1995)). Due to the uniqueness of financial services industry assets, examining the financial service industry co nsolidation will provide a new perspective into the diversification benefits, or lack thereof, of mergers and acquisitions. Since the primary asset acquired is information in the form of a client list, the acquirers market returns will be directly related to the acquirers ability to cross-market its products and services to a mo re diverse customer base. Un like the general result for industrial firms, we could observe positive wealth effects for acquirers resulting from the purchase of diverse financial products and services.2 In addition, the acquirers ability to tailor a variety of new products and services to existing customers should lead to increased sales and shareholder value.3 Diamond (1984) states that banks use private information about clients to make a profit. The FSMA allows for increased access to clients and therefore a greater ch ance to cross-market. Thus, the product market spillover hypothesis stat es that the acquirers retu rns primarily depend on the banks diversification of cust omers, products and services. While there has been considerable res earch on mergers and acquisitions, this is the first study to examine bank mergers and acquisitions after the creation of the financial services industry by the FSM A, when banks can merge with non-bank financial firms to obtain othe r than geographic diversifi cation. Using the bank as the base organization, the FSMA allows us to examine three consolidation options bank mergers with 1) banks, 2) investment fi rms or 3) insurance companies. It is reasonable to believe that there is a monotone increase in the organizational differences between banks as an acquirer a nd these three firm-types as targets. This then allows for an assessment of whether and how the degree of diversifi cation in the financial services industr y matters to investors. 2 Laeven and Levine (2006) find that banks that di versify trade at a value discount utilizing measures of Tobins Q. However, their study is based on bank to bank M&As. I consider the financial services industry acquisitions which includes bank to bank, bank to investment firm and bank to insurance company M&As. I employ univariate and multivariate analysis to examine fina ncial service industry M&As. 3 The acquirer can sell services to a larger client base including those of the target and it can sell the services of the target to its established customers. 3

PAGE 11

The specific issues I address are as follows: Are there differences in the market reactions to (announ cement effects of) the differe nt consolidation options exercised by acquiring banks? Do the effici ency and/or spillo ver hypotheses explain the post-acquisition (long-term ) returns of acquiring banks? In considering these two issues, I gauge investors short-term and long-run view s on the benefits, or lack thereof, that the FSMA provides. Finall y, I address a related issue: Does the acquisition premium reflect the degree of dive rsification in financ ial services industry acquisitions? Benston et al. (1995) find that U.S. acquiring banks bid more for targets when the resulting combinati on leads to significant geographical diversification gains. The FSMA provides a unique opportunity to examine bank managers perception of the benefits of merging with the various firm-types, as reflected in their willingness to pa y a higher premium for diversity. Utilizing all U.S. financial servi ces industry affiliated mergers and acquisitions from 1999 to 2002, I find support for two complimentary hypotheses that explain the long-run returns to the acquiring banks. Using tw o measures of efficiency and various spillover variables to test the hypotheses, I find support for both hypotheses. I find that both profit efficien cy and operation efficiency determines future returns. In addition, there is evid ence that when banks e ngage in diversifying mergers with brokerage firms the effects of profit efficiency on post-consolidation returns is larger. After evaluating various spillover characteristics, the results indicate that acquiring financial firms take advant age of the product and service diversity. I document that on a year-by-year basi s the acquisition premium ranges from 17% to 23%. It appears that bank managers do not hold financial services diversification in high rega rds as witnessed by the decreasing premium for bank diversification mergers and acquisitions (ba nk to investment firm/ insurance company acquisitions). I find a larger premium for bank to bank acquisitions (18.7%) than bank to investment firm acquisitions (16.9%) and bank to insurance company acquisitions (8.5%). 4

PAGE 12

The remainder of the paper is organi zed as follows. Section II provides a background of the relevant legi slation that has affected th e financial services industry in the United States. Section III provides a review of the literature on both M&A in the banking industry and the impact the FSM A has on banks. Section IV discusses the non-financial industry mergers and acquisi tions literature and develops the hypotheses. Specifically, Section IV presen ts the motives for combining banking services with other financial services. S ection V describes the data and methodology and provides a sample description and prelim inary statistics of the financial services industry over the 1999 to 2002 period. Sect ion VI presents the announcement day abnormal return results of this study. The determinants of post-acquisition returns are presented in Section VI. The paper conclu des with a summary of the findings in Section VII. Financial Services Industry Legislation Background4 The United States government often plays an important role in constraining or encouraging financial industry consolidation activity by directly approving or disapproving individual merger s and acquisitions or by cha nging explicit or implicit regulatory restrictions on consolidation. In 1933, duri ng the aftermath of the 1929 stock market crash and the Great Depre ssion, the Glass-Steag all Act (GSA) was enacted. As a collective reaction to the worst financial crisis at the time, the GSA set up regulatory walls between commercial ba nk and investment bank activities. The 1933 Glass-Steagall Act had two basic objectives to: 1) require that investors receive significant or material information concerni ng securities being offered for public sale and 2) prohibit deceit, misrepresentations, and other fraud in the sale of securities. At this time, "improper banking activity" by overzealous commercial banks involved in stock market investment, was deemed the ma in culprit of the financial disaster. 4 Information pertaining to banking legislation was provided by the Federal Deposit Insurance Corporation (FDIC) at www.fdic.org 5

PAGE 13

The provisions of the GlassSteagall Act were directed at specific abuses by financial service companies. First, Congre ss was concerned with banks investing their own assets in securities with consequent risk to commercial and savings deposits. Second, there was an issue with banks providing unsound loans to companies in which the bank had invested its own assets Third, there was a concern that bank officials may be tempted to press their banki ng customers into investing in securities which the bank itself was under pressure to se ll because of its own pecuniary stake in the transaction. The GSA was put in place to protect against commercial banks with financial interest in the ownership, price, or distribution of securities. Financial service companies were accused of being t oo speculative in the pre-Depression Era, not only because they were investing their as sets in risky equities but also because they were buying new issues for resale to the public. As a result of their speculative disposition, the financial services industry objectives became blurred. The Glass-Steagall Act was the first major federal legislation to regulate the offer and sale of securities. Prior to the enactment of the GSA, the Blue Sky laws governed by the state regulated the securities market. The GSA of 1933 left in place the patchwork of existing state securities laws to supplem ent the federal law. Under the 1933 GSA, the company offering securitie s is required to di sclose significant information about themselves and the terms of the securities to potential investors to assist in making an informed investment d ecision. In addition to the disclosure rule, the GSA required that banks separate comm ercial banking from investment banking services. Banks were given a year to deci de on whether they would specialize in commercial or in investment banking. By cr eating this separation, the GSA attempted to prevent the banks' use of deposits to o ffset the losses of a failed underwriting division. Thus, the purpose of the Glass-Steag all Act of 1933 was to place restrictions against the improper banking activity that resu lted from the conflict of interests of offering both commercial and investment banking service. In addition to the GSA, Congress passed the Bank Holding Company Act (BHCA) on May 9, 1956 to regulate the ba nking sector. This Act required Federal 6

PAGE 14

Reserve Boards approval for the esta blishment of a bank holding company and prohibited bank holding companies headquarter ed in one state from acquiring a bank in another state. The BHCA also prohib ited a bank holding company from engaging in most non-banking activities or acquiring voting securities of certain co mpanies that are not banks. Legislators were concerne d that huge banking conglomerates would monopolize the banking industr y. In addition, many non-bank businesses feared that firms affiliated with banks would gain a competitive advantage over unaffiliated competitors in the same industry. These non-banking businesses were concerned that firms affiliated with banks would receive preferential credit treatment from the banks and would have access to low-cost funds provided by them from non-interest-bearing deposits. There was also a concern that a bank would combine the access to credit with the purchase of services provided by its non-bank affiliates. For example, if all of the bank's commercial borrowers were required, as a condition of obtaining credit, to buy their business travel services from the bank holding company's travel agency, the independent travel agencies would be unable to compete. Given these concerns, Congress restricted non-bank activities and on ly permitted those activities incidental to banking or performing services for banks. Approved activities by the BHCA of 1956 include ownership of the bank's prem ises, auditing and appraisal, and safe deposit services. The law required that nonconforming non-bank businesses be divested. However, the interstate acquisi tion restrictions of the BHCA were eradicated by the Riegle-Neal Interstate Banking a nd Branching Efficiency Act of 1994 (IBBEA). IBBEA allowed interstate mergers between banks, subject to concentration limits, state laws and Community Reinvestment Act (CRA)5 evaluations. 5 The Community Reinvestment Act of 1977, revised in 1995, encourages depository institutions to help meet the credit needs of communities in which they operate, including lowand moderate-income neighborhoods. The CRA requires federal agencies re sponsible for supervising such institutions to evaluate their compliance periodically and to take their records into account in considering applications for deposit facilities. 7

PAGE 15

The Gramm-Leach-Bliley Financial Services Moderni zation Act (FSMA) of 1999 had the most influential impact on the financial services industry6. The FSMA repealed the Glass-Steagal Act, removing many of the remaining restrictions on combining commercial banking, secur ities underwriting, and insurance in consolidated organizations. This act opened competition between banks, securities companies and insurance companies to own and operate comparable financial services. Under the FSMA, individuals w ho would put money in investments when the economy is good can now put money in to a saving account with the same company when the economy is bad. With the ratification of the FSMA, Congress was concerned with the sharing of customers private information between divisions of the financial service firm. Thus, included in the FSMA, the Financial Privacy Rule requires financial institutions to provide th eir clients a privacy notice that explains what information the company gathers about the client, where th is information is shared, and how the company safeguards th at information. The privacy notice must also explain the opt-out policy; whic h allows customers to not permit their information to be shared with affiliated parties. By eliminating the restrictions on the separation of commercial ba nking from securities and insurance activities, the Financial Services Modernization Act allo ws for the consolidation of financial services and the formation of financial conglomerates. Banking Industry Literature Review In the seminal literature of the banking industry, Diamond (1984) analyzed the monitoring and information ga thering role of banks. Di amond (1984) theorized that under asymmetric information banks are ab le the extract positive profits from the private information about borrowers. Von Th adden (1998) shows that this theory is 6 The FSMA amends, among other laws: the Banking Act of 1933 (Glass-Steagall), the Bank Holding Company Act of 1956, the Interstate Banking and Branching Efficiency Act of 1994, the Investment Company Act of 1940, the Securities Exchange Act of 1934,the International Banking Act of 1978, the Depository Institutions Deregulation and Monetary Co ntrol Act of 1980, the Federal Reserve Act, the Federal Deposit Insurance Act, the National Bank Consolidation and Merger Act, and the Home Owners Loan Act. 8

PAGE 16

robust even when markets are characterized by pure price competition. In a more recent study, Martinez (2002) develops a framework that measures the information asymmetry across banks. Martinez deve loped a borrower turnover measure that focuses on how changes in customer informa tion affect the banks ability to compete. This framework differentiates between e ffects on bank profits stemming from the banks relative size versus those profits from superior information. On the other hand, information asymmetry between investors and bank managers may limit a banks ability to raise funds (Ste in (1998)) or banks may have limited ability to process information or monitor loans (Gale ( 1993) and Almazan (1996)).These studies suggest that banks utilize private informati on to earn a profit and that any change in regulation or industry struct ure that affects the use or generation of private information will impact the banks ability to compete. Other researchers have evaluated the im pacts of legislation changes on the financial services industry. Kashyap and Stein (1995), for instance, find that the lending behavior of banks seems to be quite sensitive to exogenous changes in monetary policy. Thakor (1996) finds that regulations that increase capital requirements for banks decrease the aggreg ate lending affecting a banks ability to earn profits. However, Stiroh and Strahan (2003) find the link between a banks relative performance and its subsequent mark et share growth strengthens significantly after deregulation as competitive reallocation effects transfer assets to better performers. They conclude th at earlier regulation of U.S. banks blunted this market mechanism and seriously hindered the competitive process. The changes in government regulations have led to dynamic changes in the financial industry structure. Berger and Humphrey (1992 ) and Rhoades (1993) both document significant structur al and organizational changes in the banking industry following regulation amendments. However, th ese industry changes have resulted in mixed results. DeYoung (1997) shows si gnificant losses for acquiring U.S. commercial banks from 1984 through 1994. Peri stiani (1997) finds substantial gains for acquiring banks following the consolid ation and reconstructing during the 1980s. 9

PAGE 17

Akhavein, Berger and Humphrey (1997) s uggests that the improvements in profit efficiency can be linked to improved dive rsification of risks. Berger and Mester (2003) concludes that industry diversification allows instit utions to make additional high-risk, high-expected retu rn investments without addi tional productivity declines during the 1990s. The bank holding company conglomerates ar e able to navigate the restrictions and geographical limitations of bank legisl ation. Bank holding company activity in states with limitations on intrastate branching allowed these companies to diversify into non-banking industries. Bank Holding Company (BHC) diversification decreases the companys firm-specific risk while unaffecting its systematic risk.7 However, the diversification is not the only factor affecting a Bank Holding Companys firmspecific risk. The uncertainty of the i ndividual components of a BHCs assets, leverage, and liabilities influence it stock returns. Demsetz and Strahan (1997) finds that larger BHCs are able to operate wi th higher leverage and engage in riskier lending practices without increasing overall risk because of their diversification advantage. Similarly, Akhavein, Berger a nd Humphrey (1997) suggest that better diversification allows the me rged banks to hold riskier a nd more profitable portfolios. These findings suggest that diversificatio n may be an important motivation for bank consolidation. Measuring th e performance of BHCs fo llowing geographic banking deregulation, Liang and Rhoades (1991) show s that larger BHCs can take advantage of wide branch networks to more e ffectively diversify loan portfolios. The Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994 (IBBEA) eradicated the Bank Holding Co mpany Act restrictions and permitted bilateral agreements among states made expansion across states by bank holding companies possible.8 Examining these two banking deregulation legislations, Clarke (2004) finds a significant linkage does not exist between banking markets as defined 7 Under the United States law, a bank holding compan y is any entity that directly or indirectly owns, controls or has the power to vote 25% or mo re of a class of securities of a U.S. bank. 8 Initially, most states' bilateral agreements permitted entry only from states in a surrounding region, with the boundaries of the region determined by the passing of state law. These regions expanded to include more states. Some states even passed state legislation allowing national interstate banking. 10

PAGE 18

by deregulation and economic growth in th e state. Thus, diversification due to deregulation appears to have no influenc e the economy. The Financial Services Modernization Act of 1999 allows banks the opportunity to expand into non-banking financial activities increasing financia l service industry c onsolidation. This consolidation will enhance diversificati on, however, the resulting change in shareholder value will depend on the extent to which consolidation is accompanied by changes in banks activities. Literature on the Depository Institutions Deregulation and Monetary Control Act of 1980 In March of 1980, the Depository Inst itutions Deregulati on and Monetary Control Act (DIDMCA) was enacted to eliminate the distinctions among different types of depository institutions9 and remove interest rate limitations on deposit accounts. Essentially, this deregulation of the Glass Steagall Act allowed credit unions and savings and loan firms to offer checkable deposits, thus competing with commercial and savings banks for custom er deposits. The DIDMCA has two main sections, Title 1 and Title 2. Title 1, the Monetary Control Act, extends the monetary reserve requirements to all U.S. banking institutions.10 The Depository Institutions Deregulation Act of 1980, Title 2, eradicated th e Federal Reserve depos it interest rate ceilings. The Depository Institutions Deregula tion and Monetary Control Act was passed to deal with the pr oblems facing depository institutions. Due to the high interest rate level experienced in the United States, the depository institutions were forced to pay higher rates to attract consum er funds than the rates they were earning on their portfolios of assets. The DIDMCA allowed depository institutions to relax 9 The Depository Institutions Deregulation and Monetary Control Act defines depository institutions as banks, savings banks, Savings and Loans firms and credit unions. 10 The mandatory reserve requirements that banks keep in non-interest earning accounts at Federal Reserve Banks were lowered. State chartered bank s that are not members of the Federal Reserve System and thrift institutions we re required to maintain reserve account balances. And the mandatory reserves requirements for all depository institutions were phased in over an eight-year period ending in 1988. 11

PAGE 19

deposit rate limitation, enabling these institu tions to earn higher returns due to the reduced applicability of the state usur y laws. Depository institutions were now permitted to lend money and charge borrowers exorbitant interest rates. The new lending powers were, however, extende d to only individu als and nonprofit organizations and not to businesses. Alle n and Wilhelm (1988) states that bankers recognize the desirability of regulatory si mplification among de pository institutions, such as homogeneous reserve requirements, and the abolition of interest rate ceilings on deposits. However, James (1983) finds that bank deregulation of deposit rate ceilings resulted in gains for credit unions and savings and loans but losses for commercial banks. Similarly, Allen and Wilhelm (1988) find evidence that the Depository Institutions Deregulation and Monetary Control Act provided a wealth transfer from non-federal Reserve System member banks and savings and loans to Federal Reserve member banks. Cornett a nd Tehranian (1989) find that the DIDMCA banking deregulation benefited stockholders of large banks and savings and loans but produced negative abnormal stock return s for small banks. Timberlake (1985) suggests that the DIDMCA increased the powers of the Federa l Reserve System, benefiting large established banking inst itutions. The implementation of the DIDMCA made it possible for depository in stitutions to compete for funds regardless of the level of interest rates in the ec onomy. However, Cornett and Tehranian (1990) states that the authority for Federal Reserv e member institutions to make risky loans was expanded, which ended up with the savings and loan crisis in 1985. Literature on Interstate Banking and Branch ing Efficiency Act of 1994 Prior to 1994, banks and bank holding companies were prohibited from acquiring banks across state lines. The Inte rstate Banking and Branching Efficiency Act of 1994 extends interstate banking in two formal steps. First, as of September 1995, bank holding companies were allowed to acquire banks in any state. Secondly, in June 1997, holding companies were able to convert out-of-state bank affiliates to branches of the lead bank. However, bank acquisitions and convers ions are subject to the approval of the Community Reinvest ment Act (CRA) of 1977, concentration 12

PAGE 20

limits and state laws. Previous research indi cates that financial reforms, such as the Interstate Banking and Branching Efficiency Act of 1994, have had important effects on the structure of banking markets. Br ook, Hendershott and Lee (1998) document a value gain of $85 billion for the financial industry post ratific ation on the IBBEA. However, Carow and Kane (2002) state that the passing of the IBBEA resulted in the redistribution of wealth rather than the creation of value. Carow and Kane (2002) find that returns are positive for some fina ncial sectors and negative for others. Although Amel and Liang (1992) and Ca lem (1994) both find that banking market structure changed little after the ratification of the Interstate Banking and Branching Efficiency Act Amel and Liang (1992) find significant entry into local markets after intra-state branching restric tions of the BHCA we re repealed. Calem (1994) show that many small banks are acqui red and incorporated as branches into large bank holding companies after branching reform. McLaughlin (1994) finds that multi-bank holding companies convert existing and acquired bank subsidiaries into branches following the IBBEA. Furthermore, Savage (1993) finds that over the 19801993 period the market share of large bank s grew, while concen tration at both the state and national level rose. Jayatne and Strahan (1996) also show that banking deregulation increased state-level growth a nd Jayatne and Strahan (1998) show that deregulation increased efficiency resulting from costs and prices of banking services. Overall, the evidence suggests that larger mo re efficient banks emerge post-inter-state deregulation. In addition, increases in size are associated with better inter-state diversification (Demsetz and Strahan (1995 )). Stiroh and Strahan (2003) find these beneficial results are the result of competitive dynamics. Nippani and Green (2002) show that bank performance improved in the post-IBBEA period, but when they controlled for general economic conditions and interest rate movements, the impact of IBBEA on bank performance appears to be insignificant. Howeve r, based on several studies (Demsetz and Strahan (1995), Jayatne and Strahan (1996, 1998), Nippani and Green (2002), Stiroh and Stra han (2003)), deregulation of the financial services industry appears to be beneficial to only some s ections of the industry. 13

PAGE 21

Literature on the Financial Serv ices Modernization Act of 1999 Prior to 1999, consolidation occurred primarily through intra-industry mergers and acquisitions. Under the FSMA, while the intra-industry M&A activity will continue, much of the consolidation will be in the form of cross-industry mergers involving banks, brokerage firms, and insura nce companies. There is an extensive research literature on the motives for and consequences of consolidation resulting from the ratification of the Financial Services Modernization Act. According to Covington and Burling (1999) and Macey (2000) the FSMA provides commercial banks with strong incentiv es to expand into investment banking and insurance services, investment banks into commerc ial banks and insurance companies, and insurance companies into investment banking. Barth, Brumbaugh and Wilcox (2000) investigate the major provisions of the Act. They argue that the Act favors big banks. Brewer et al. (1988), Wall et al. (1993), a nd Boyd et al. (1993) hypothesize that banks will receive greater benefits from involve ment in insurance activities than from participation in other nonbanking activities, without increasing their organizational risk significantly. Kwan and Laderman (1999) also find similar results after surveying the literature on the effect s of combining banking and nonbank financial activities on banking organizations risk and return. They conclude that expanding banking services to include insuran ce and investment activities can provide diversification benefits to banking organizations. Whalen (2000) shows that banking firms are likely to improve, or at least not unfavorably alter their risk/return opportunities by engaging in both banking and insurance act ivities-particularly life insurance underwriting activities. Alte rnatively, Kwast (1989) and Apilado et al. (1993) find that adding underwriting services will lead to an increase in individual bank risks and little reduction in total risk. Several studies examine the effect of the passage of the FSMA on the return of financial services firms. Akhigbe and Whyte (2001) examine the legislative events leading up to and the passage of the FSMA on the stock returns of banks, brokerage 14

PAGE 22

firms, and insurance companies. They find that the impact is positive for all institutions. Bank gains are positively rela ted to size (bank assets) and capitalization level (customer deposits). Brokerage firms and insurance companies gain regardless of their size. Insurance firms gain regardle ss of their capital position, but brokerage firms gains are inversely re lated to their capita l position. Alternatively, Carow and Heron (2002), find insignificant returns for banks, negative returns for foreign banks, thrifts and finance companies and positive returns for investment banks and insurance companies. Similarly, Hendershott, Lee and Tompkins (2002) document strong positive response among investment banks and insurance companies and insignificant response among commercial banks. Neale and Peterson (2003) show that the market reactions to key events related to passi ng of the FSMA for insurance companies are positive. Akhigbe and Whyte (2004) report that the systematic risk of all types of financial institutions decreases after the FSMA, while the total and unsystematic risks increase for banks and insurance companies and decreases for investment firms. More recently, Yildirim, Kwag and Collins (2006) fi nd that investment banks and insurance firms are better positioned to exploit the benefits of product-line diversification opportunities allowed by the legislation compared to commercial banks that experience no significant market reactions to the key legislative events leading to the passage of the FSMA. Literature Review on Non-Financial Co rporate Acquisitions and Conglomerates The financial literature a dvances two main strands of research related to the motivations for mergers and acquisitions. Th e first commonly found is the synergistic effects from the combination of targets a nd acquirers. The synergistic hypothesis can take form in three distinct ways. One, the target shareholders as well as the bidder shareholders would benefit from the me rger if the bidders management can effectively manage the combination of the tw o firms more efficiently than the target management can manage the target firm. Tw o, a target firm can benefit from being absorbed by an acquirer, gaining access to outside funds at the lowest attainable rates. 15

PAGE 23

Finally, the reduction of risk through the me rger of target and acquirer is the third justification of an acquisi tion. Even if the target a nd acquirer are in unrelated industries, the risk surrounding their earnings streams will be reduced when these earnings streams are pooled. The second hypothesis advanced related to the motivations for mergers and acquisitions is the agency cost hypothesis. The agency cost hypothesis asserts that managers pursue pecuniary and non-pecuniary rewards that are closely related to the growth rate of their firm. Je nsen (1986) and Stulz (1990) both assert that the power and prestige associated with managing a larger firm may lead to firm diversification. Jensen and Murphy (1990) maintain that ma nagerial compensation is related to firm size. Amihud and Lev (1981) note that dive rsification also reduces the risk of managers undiversified personal portfolios. Finally, Shleifer a nd Vishny (1989) state that diversification helps make the manager indispensable to the firm. Several empirical studies have investig ated the validity of both the synergy hypothesis and the agency co st hypothesis. Initially, Mandelker (1974) finds a positive 0.6% abnormal return of merged firms from a month after through 12 months after the effective merger date. Empirical results by Bradley, De sai and Kim (1983) show positive but statistically insignifi cant total dollar gains of $17.2 million to acquirers and targets in 162 tender offers. They also found that the average percentage change in total value of the combined target and acquirer firms is a significant 10.5%. Asquith (1983) and Eckbo (1983) report slightly positive, but statistically insignificant, abnormal return s. Servaes (1991) supports the notion that abnormal returns are higher when well managed, high q firms take over poorly managed low q firms. Finally, Chevalier (1999 ) finds that the mark et reacts positively to the announcement of diversification merg ers. This evidence in dicates that changes in corporate control increase the combined market value of assets of the acquiring and target firms. In stark contrast, there is a body of rese arch that states th at acquisitions are negative present value investments. Dodd ( 1980) finds a significant abnormal return 16

PAGE 24

of -1.09% for 60 acquirers on the day before and the day of the first public announcement of the merger, indicating that me rger bids are, on av erage, negative net present value investments for acquirers. Similarly, Malatesta (1983) reports an average loss of about $28 million in the peri od four months before through the month of announcement of the merger outcome. He finds significant negative abnormal post-outcome returns of -13.7% for merg ers. Recently, Graham, Lemmon and Wolf (2002) found that acquiring firms experien ce a reduction in exce ss value following the acquisition. They state that the addition of a poor perfor ming target firm explains most of the decline in value. Several papers discuss the effects of diversification on firm value. Stein (1997) suggests that conglomerates transfer capital from divisions with lower growth opportunities to those with hi gher growth opportunities but facing capital restrains. Thus, the Stein hypothesis suggests that internal capital mark et benefits the conglomerate. Alternatively, Sc harfstein and Stein (2000) su ggests that in a two-tier agency model, headquarters executives w ill over-allocate capital to rent-seeking divisional managers, theref ore destroying firm value. There is a general consensus in the em pirical literature validating the losses resulting from diversification. Morck, Sh leifer, and Vishny ( 1990), Lang and Stulz (1994), Berger and Ofek (1995), Servaes (1996), and Lamont and Polk (2001) document significant value losses associated with diversification related acquisitions. Similarly, Loughran and Vijh (1997) finds that acquirer stock returns, on average, are smaller than matching stock returns in cases where a merger is made and stock is used for payment. Empirical results by Rau and Vermaelen (1998) also show that acquirers in mergers underperform in th e three years following the acquisition. The theory that investors overreact to anticipated and unanticipated information, resulting in exaggerated moveme nts in stock prices have been used to explain the post-acquisition performance of acquirers. DeBondt and Thaler (1985) establishes the stock market overreaction hypoth esis that asserts that stock prices take temporary fluctuations away from their f undamental values due to investor optimism 17

PAGE 25

and pessimism. Numerous studies find evidence for stock price overreaction. DeBondt and Thaler (1987) document that over and underperforming firms exhibit extreme price performance over long-term period s. Fama (1991) state that part of the response of prices to information announcements occur slowly over time. Studies analyze the differential risk iness of over and underperforming firms. Chan (1988) and Ball and Kothari (1989) argue that differe nces in risk can explain the abnormal performance of these firms. These studi es are evidence of overreaction to the accumulation of information. Chopra, Lakoni shok and Ritter (1992) provide evidence which suggests that differential risk cannot explain the asymmetr ic overreaction of under and over performing firms. However, Conrad and Kaul (1993) report that most of the long-term overreaction reported in DeBondt and Thaler (1985) can be attributed to a combination of the bid-ask spread effect and the use of price rather than returns to calculate cumulative abnormal returns. The overreaction excess returns following acquisitions might be due to the change in the type of stockholders owning the company (Black and Scholes (1974) and Shefrin and Statman (1984). Stiglitz ( 1989) hypothesize that the overreaction is caused by speculative trading. One might e xpect similar overreaction for acquiring financial service industry firms following th e ratification of the Financial Services Modernization Act of 1999, especially si nce these merged firms will form new financial conglomerates and the novelty and uncertainty of such a firm can lead to investor speculation. The consolidation efforts of commercial banks can provide insights into the true value of an acquisition and the bene fits conglomerates receive from diverse divisions. Financial services industry acqui sitions provide an empirical laboratory with relatively little asymme tric information or moral h azard compared to industrial firms in which the aforementioned hypothese s can be tested. Therefore, banking industry acquisitions provide an opportunity to test an aspect of the synergy 18

PAGE 26

hypothesis11 (Mandelker (1974)), and agency cost hypotheses (Jensen (1986)) and provide new insights about the effects of diversifying acquisitions (Morck, Shleifer, and Vishny (1990)). I use the reorganization of banks to comprehensively examine the value of an acquisition. Specifically, I quantify and then examine the determinants of the returns of the various ba nking reformation combinations. This study contributes to the de bate on bank returns post-FSMA, by examining the consolidation options of banks following th e ratification of the FSMA. Specifically, I compare the market reactio ns of the consolid ation options; bank mergers with 1) banks, 2) investment firms or 3) insurance comp anies. Additionally, I analyze the returns post-FSMA using competing hypotheses; efficiency hypothesis and spillover hypothesis. Finally, I examine the acquisition premium to see if bank managers pay premium for financial service industry diversification. Hypotheses Development: Efficiency Hypothesis versus Spillover Hypothesis In this section, I present a set of sub-hypotheses under two broad hypotheses, efficiency and product-market spillover hypot heses, to explain the cross-sectional variation in stock market retu rns of banks that acquire fi nancial services companies. The efficiency hypothesis states that acqui ring banks purchase other financial service industry companies because the acquiring ba nk can manage the combination of the two firms more efficiently than the target management can manage the target firm. Therefore, the returns to the acquiring banks are partly driven by the acquirers ability to utilize the targets existing assets in a more efficient and effective manner. To test the efficiency hypothesis, I implement both of the banking profit efficiency measure utilized by Berger (1993) and Berger, De Young, Genay and Udell (2000) as well as the operating efficiency measure deve loped by Harris and Robinson (2002). The product market spillover hypothesis states that the acquirers postacquisition stock market returns are related to the acquirers ability to cross market its products and services to a more diverse client base, wh ich now includes the clients of 11 In relation to the banking industry, the synergy hypothesis is defined as the synergistic use of information on clients of both the target and the acquirer. 19

PAGE 27

the newly acquired target firm. Similarly, th e targets products and services can be marketed to the acquirers clients. Thus according to the product market spillover hypothesis, the acquirers returns primarily depend on the banks diversification of customers, products and services. The efficiency hypothesis and spillover hypot hesis are distinctly different. The influence behind the efficiency hypothesis is the gains from exploiting economies of scale. The efficiency hypothe sis centers around the cost func tion of a firm in that it primarily captures the reduction of cost re sulting from increase in efficiency. These operational efficiencies could result from a variety of syne rgies including a reduction in production or redistribution cost vertical integrat ion and adoption of more efficient production or organizat ional technology. The spillover hypothesis captures the gains from improvements in fina ncial standing that result from increased profits. The changes in the financial struct ure and standing of the acquirer or the combined firm resulting from excess sales, cash/cash flow and financial leverage will arise because of spillover from one division into other divisions.12 Specifically, the spillover hypothesis states that the acquiring financial institution purchases a target financial company to cross market its products and services to a more diverse client base. The implications of each hypothesi s are very different The efficiency hypothesis is different in that it implies that the new consol idated financial entity will be able capture sustainable efficiencies resulting from economies of scale, whereas the spillover hypothesis implies that spi llover benefits will disappear once the acquirer has completed the cross-marketing of its products and services to the expanded client base. Efficiency Hypothesis Efficiency gains from exploiting econom ies of scale are often cited as a motivation for financial services industry consolidation (Mester (1993), Berger and 12 For instance, the relatively low correlation between the cash flows of a regular bank and those of a target insurance firm will reduce the probability of default and, therefore, lead to increased debt capacity. This is the coinsurance argument of Lewellen (1971)). The latter may then be exploited to increase the combined firms revenues. 20

PAGE 28

Hannan (1998), Berger, Demsetz and Straha n (1999)). Using data from the 1990s, Berger and Mester (1997) s uggest that there may be substantial scale economies even for the largest banks. They suggest that thes e results are due in part to technological advancement. Alternatively, several studi es find little evidence of substantial economies or diseconomies of scale within banking (Ke llner and Mathewson (1983) and Mester (1987)), securiti es (Berger, Hanweck and Hu mphrey (1987)) or insurance industries (Mester (1993)). Th ese findings suggest that losses, or at least no gains in efficiency, will result from banking industr y consolidation. Consistent with these findings, Altunbas, Molyneux, and Thornton (1997) find that simulated pro forma mergers between banks in the European Union are more likely to increase costs rather than decrease them. However, Berger, Hancock and Humphrey (1993) and Berger, Cummins, Weiss, and Zi (1999) find that joint production within both banking and insurance companies is more efficient for some firms and specialization is more efficient for others, depending on the size of the firm. In the United States, domestic banks are on average slightly less cost efficient than foreign banks by 2.8% of costs (Berger, Demsetz and Strahan (1999)). This finding suggests that diversification reduces costs whereas specialization increases cost. Berger et al. (1999) conclude that th e higher expenses are more likely incurred to produce a greater quality and/or variet y of financial services that generate substantially greater revenues. Berger, DeYoung, Genay and Udell (2000) find that foreign banks are less efficient on average than domestic banks, suggesting that efficiency considerations may limit the global consolidation of the financial services industry and leave substantial market share for domestic institutions. As it relates to industry consolidation, Berger and Hu mphrey (1992) and Pilloff and Santomero (1998) find that acquiring banks appear to be more cost efficient, on average, than their banking industry peers. These cost e fficiencies could result from a variety of synergies including a reduction in production or redistribution cost, vertical integration and adoption of more efficient production or organizational technology (Jensen and Ruback (1984)). In light of the above discussion, I hypothesize that the 21

PAGE 29

post-consolidation returns of the consolidated entity are positively related to the efficiency of the acquiring bank. Furt hermore, I hypothesi ze that the postconsolidation returns are increasing in th e difference between the efficiency of the acquirer and the target. Due to the difficulty of realizing efficiency gains across diverse industry divisions, I also hypothesize that the postconsolidation returns of the new financial services entity will be larg er when the bank acquires other bank(s) relative to when it acquires an inve stment bank or insurance company. Product Market Spillover Hypothesis The basis of the product market sp illover hypothesis is that banks pay substantial premiums to acquire financial co rporations so that the acquiring financial company can cross market its products and services to a more diverse client base. This hypothesis posits that the acquirer would like to obtain other benefits from the acquisition other than just the income aris ing from improving the efficiency of the targets operations (Efficiency Hypothesis). This may result from a spillover effect to the acquirers other financial products and services due to the financial services industry acquisition. Whereas efficiency gain s arise from the improved use of existing target assets in association with the ac quirers assets, the product market spillover hypothesis holds that post-consolidation returns are expected to increase as a result of increased use of the existing client base of both the target and acquirer firms to market the consolidated entitys services. For instance, after a bank acquires an insurance company the bank will cross-ma rket its banking products to the new insurance clients while selling its insu rance products to the banks clients. Most tests of the Efficiency Hypothesi s use data on financial institutions from the 1990s, and it is possible that r ecent technological progress might have increased product spillover in financial se rvices and thus created opportunities to improve returns through consolidation. A numb er of studies have examined firms that provide multiple products within the financia l services industry with mixed results. Some research finds that consolidated ba nks and consolidated insurance companies 22

PAGE 30

may lower costs by using one consolidated cu stomer database and cross-selling their products. Greenbaum, Kanatas, and Venezi a (1989) and Rajan (1996) state that information reusability may reduce cost when a universal bank acting as an underwriter conducts due diligence on a cust omer with whom it has had a lending or other relationship. Alternatively, Winton (1999) argues that diseconomies may arise from coordination and administrative cost s when firms offer a broad range of products. Winton further states that these products are often outside the senior managements area of competence. Howeve r, studies of the European Unions universal banking system may not be good predictors of the United States consolidated financial services industr y. Berger, Demsetz and Strahan (1999) state that commercial banking and underwriting in the banking-oriented continental Europe of the past bears li ttle resemblance to commercial banking and underw riting activities in market-oriented financial systems as in the United States. The following subhypotheses explain and define the p roduct market spillover hypothesis. Cash /Cash Flow Lewellen (1971) and Travlos (1987) state th at the merger of two firms that do not possess perfectly positively correlated cash flows reduces the default risk of the new firm, therefore, increasing the value of the combined firm above the sum of the values of the individual firms. Along the same line, Stein (1997) suggests that the gains to conglomerates that transfer assets from one division to another stem from the conglomerates ability to finance positive net present value projects of divisions with growth opportunities and low cash flow. However, there is an internal capital market theory that suggests that the allocation of assets within diversified firms may be suboptimal. Scharfstein and Stein (1997) s uggest that executives over-allocate capital to rent-seeking divisional managers, therefore destroying firm value. The finding of Lamont and Polk (2001) also suggest wastef ul spending or crosssubsidization that reduces cash flow. Similarly, banks may acquire target fi nancial services companies to subsidize the other products and services the bank offers. Thus, the operating 23

PAGE 31

income earned from the target companies are utilized to finance the operation of other divisions. Hence, I hypothesize that the returns to the acqui rer are positively related to the change in the cash holding and cash flows of the consolidated firm. Financial Leverage Palepu (1986) finds that the probabili ty of becoming an acquisition target decreases with the companys debt level. Lo w debt levels are viewed by the acquirers as the targets management inability to maximize firm value. Thus, upon acquisition the acquirer can increase debt levels and ob tain additional assets which in turn may generate extra value. Palepu also finds th at the increase in debt capacity from the acquisition of a low debt target reduces the risk of the acquirers default. In the highly competitive financial services industry, th e acquirers access to additional funding can lead to a competitive advantage over capital-constrained firms. Another explanation for the acquisition of targets with under-utilized debt capacity comes from Lewellen (1971). Lewelle n argues that dive rsified firms have larger tax shields from interest deductions in addition to debt capacity. Majd and Myers (1987) assert that conglomerate firm s pay less in taxes than their divisions would pay separately because of the tax c odes asymmetric treatment of gains and losses. The conglomerate division that expe riences losses has a higher probability of receiving tax benefits from the losses than if it was an independent firm. Therefore, I hypothesize that the returns of the acquirer are positivel y related to the financial leverage of the joint firm resulting from the debt capacity and tax shield benefits of a target. Excess Sales The efficiency hypothesis suggests that asse ts flow to their most efficient use. A new management team replaces under-p erforming incumbent management and manages the acquired assets more efficiently. If this is the case, then the acquiring firms net sales should rise post-acquisi tion. However, I am pr oposing an alternative 24

PAGE 32

explanation of any gains in the post-acqui sition net sales. If the acquiring firm can realize informational and marketing econom ies, cross-marketing and developing products and services will increase sales. Thus the product-market spillover from one division into the other will have a drastic impact on sales. Hence, I hypothesize that the acquirers stock market returns are positiv ely related to the change in sales of the acquiring firm. Given that increased sale s can be a manifestation of increased efficiency, I use excess sales as a proxy to test the spill-over hypothesis. Excess sales is the residual sale change after accounting for the eff ect of efficiency on sales change. Data Description and Empirical Methodology Data Description To examine the wealth effects of FSMA, I collected all U.S. affiliated mergers and acquisitions for all fi nancial firms from November 12, 1999 to December 31, 2002 provided by the Worldwide M&A section of the SDC platinum database. I obtain information on i) the identities of the firms involved in the mergers or acquisitions, ii) the status of the transacti on, iii) the nation of target firms for U.S. acquirers iv) the primary four digit SIC c odes for both acquirers and targets, and v) the number of SIC codes that the acquirers and targets participate in. All acquirers must have three years of returns and accounting data for three years after the acquisition to be included in the database. There are 404 completed financial industry M&As where U.S. firms were acquirers of U.S. target post the ratification of the FSMA. As in Mamun, Hassan and Maroney (2005), I use the acquirers CUSIP as well as the SIC classification from COMP USTAT to identify the commercial banks (SIC 6021 and 6022), brokerage firms (3 digit SIC code 620) and insurance firms (3 digit SIC codes 631, 632, 633) to supplement the SDC platinum data with balance sheet information. The return information for this study comes from the Center for Research in Security Prices (CRSP) tapes. I require these firms to have no missing 25

PAGE 33

trading day returns for at least three consecu tive years after the merger or acquisition, from November 12, 1999 to December 2005. The selection process reduced the sample to 353 U.S. financial institution acquisitions. Among these 353 transactions, 273 were conducted by commercial banks acquiring other commercial banks, 58 commercial banks merging with brokerage firms, and 22 commercial bank and insurance company mergers. There are a total of 194 dis tinct acquirers under study. In addition to returns, I also require a measure of bank efficiency. There are several ways to measure bank efficiency. These include estimated efficiencies (Berger and Mester, 2003) or linear program ming efficiency measures (Wheelock and Wilson, 1999 Alam, 2001). Berger (1993) and Berger, DeYoung, Genay and Udell (2007) utilize the most widely cited ba nking profit efficiency measure using a distribution free random error method. To implement the distribution-free random error method, I estimate the profit functi on using data of 2214 ba nks with continuous and complete annual data for the seve n-year period from 1999 through 2005. Using the results of these estimations, I calculate the profit efficiency for every financial institution with the distribution-free method, which distinguishes efficiency differences from random error by averagin g the profit function residuals over time13. Specifically, the estimate of efficiency for each firm in a data set is determined as the difference between the average residual of the control sample a nd the residual of the firm in the sample. I use this profit effi ciency measure to test the efficiency hypothesis, because it is a more comprehe nsive measure that includes both cost and income variables. The variables included in the profit efficiency estimation model are net income, net sales, cost of sales, earn ing before interest and taxes (EBIT) and nonoperating expenses. As an alternative measure of efficiency, I look to the manuf acturing literature. Harris and Robinson (2002) defi ne operational efficiency as the ratio of operational expenses of financial institutions to that of the assets of financial institutions. This 13 See Berger (1993) for a complete description of the distribution free random error efficiency measure. The general procedure for estimating effi ciency using the distribu tion free method is to estimate input coefficients and random error term to calculate efficiency for each observation in the sample. 26

PAGE 34

efficiency measure is defined in the broadest possible terms to include any effects that increase the consolidating firms exis ting shareholder value. I use operational efficiency as an ancillary measure to di agnose whether financial institutions returns are rooted in cost control ( operational efficiency ) or both cost and sales (distributionfree profit efficiency) and to compare my results to thos e of the previous banking efficiency literature (Berger, DeYoung, Genay and Udell (2007)). However, while presenting both sets of results, I place gr eater weight on the analyses using the distribution-free profit efficiency measure. Methodology This paper attempts to explain the returns to financial services acquirers following the ratification of the FSMA using two competing hypotheses: efficiency hypothesis and product-market spillover hyp othesis. I begin the analysis by examining the impact of efficiency on acqui rers return. In order to provide some preliminary insight into whether the returns of financial institutions are the result of superior efficiency, and to evaluate the economic significa nce of this relationship, I estimate equations of the form: itit tite Efficiency turn 3, *13,Re (1) where returns is the three-year aggregate return of bank i in the period immediately following its investment in an acquisiti on, Efficiency is a measure of bank i s efficiency utilizing the distribution-free pr ofit or operational efficiency variable. A statistically significant positive relationship between Return and Efficiency would be consistent with the efficiency hypothe sis and the findings of Berger, DeYoung, Genay and Udell (2007). 27

PAGE 35

I then use a multivariate regression model (MVRM) suggested by Yildrim, Kwag and Collins (2006) to examine the three-year returns of financial institutions post the passing of the FSMA. With the MVRM the effect of the acquisition on systematic industry risk a nd security returns can be simultaneously captured by adding a variable to account for the indus try diversity varia tion. I add industry diversity variables to account for the va riation in returns resulting from offering different industry services. The model proposed is: ie ngExpenses NonOperati turnRisk s CostofSale s ExcessSale everage FinancialL CashFlow BankMerger Bank Efficiency turntt tt tt tt tt tt tt tit 3, 83, 73, 6 3, 53, 43, 3 2 3, 10 3,Re / ) ( Re (2) where Return is the three-year aggregate return of bank i in the period immediately following its investment in an acquisition, Efficiency is a measure of bank i s efficiency utilizing the distribution-free profit or operational efficiency variable discussed earlier. Bank/Non-bankMerger accounts for the diversity of the acquisition where the variables takes values of 0 for bank/ bank mergers and 1 for bank/ investment firm acquisitions and ba nk/ insurance company mergers. ReturnRisk is the monthly standard devi ation of returns over 36 months multiplied by the square root of 36 resulting in the three-year risk in returns. The Spillover terms account for the three-year cha nges in the resulting firms financial characteristics, such as cash, cash fl ow, financial leverage, and excess sales standardized by the acquirers size. Cash cash flow and financial leverage are reported in Compustat. ExcessSales is defined as the residual change after accounting for the effect of efficiency on change in net sales. After implementing the following regression: 28

PAGE 36

i i ie ofAcquirer Efficiency Sales 10 (3) where Efficiency is a measure of acquiring bank i s efficiency utilizing the distribution-free profit or ope rational efficiency variable I then use the output of this cross-sectional regression (equation 3) in the following model: i iofAcquirer Efficiency Sales sExcessSale* (4) that is actual sales change less predicted sales change, where the estimation of ( hat) is common across all acquirers. Finally, I use abnormal returns (AR) proposed by Asquith and Mullins (1983) to analyze the acquirers announcement da y response. Announcement day returns are computed as jt ti it iRR AR, 1, (5) where Rt,i denotes the date t stock return of the acquiring firm i, Rt,j is the date t return of the equally weighted index of the financial service industry companies or the return of the equally weighted market model. Empirical Analysis Descriptive Statistics for financial institutions Table 1 provides the summary statistics of the main variables for the sample of 353 financial industry acquisi tions. On average, the tran saction value for financial 29

PAGE 37

30 industry mergers is $477 million. Averagi ng across all bank-year observations, the average net asset of the acquirers is ap proximately $64 billion, while the targets average size is smaller, $3.12 billion. The mean financial statistics for the acquirers and targets are reported in column 1. It appears that the acquiring firms are in good financial standing with a mean net sales of $1.3 billion and return on assets (ROA) of 1%. The average financial leverage is $30 b illion. Finally, there is, on average, a positive annual cash flow for the acquiring firms of $154 million. The sample preliminary statistics and characteristics ar e similar to those reported in Berger and Mester (2003) and Akhigbe and Whyte (2001).

PAGE 38

All BankBrokerageInsuranceDifferenceDifference Difference (1) (2) (3)(4) (2)-(3) (2)-(4) (3)-(4) Transaction Value 477.10363.50985.66401.06 -622.16 -37.55 584.60 Number of Acquisitions 353 273 5822 Number of Acquirers 194 145 3613 Acquiror Assets 63896.2417530.0852090.41200774.77-34560.34***-183244.69**-148684.36* Financial Leverage 30425.8113169.2638788.6147018.21-25619.35-33848.95*-8229.60 Operation Expense 313.16245.66579.052577.81 -333.39***-2332.15***-1998.75 Net Sales 1333.111080.235358.6830.07 -4278.45***1050.15 5328.61* Cash 1120.911107.459501.7215.27 -8394.26***1092.19 9486.45* Cash Flow 154.3529.01175.8314.35 -146.82***14.65*161.48* Return on Assets 1.01 1.041.031.33 0.00 -0.30*-0.30 Return on Investment 9.42 9.969.6714.88 0.29 -4.92**-5.21 Cost of Sales 993.41593.071749.732785.04 -1156.66***-2191.97*-1035.31 Non-Operating Expense212.53139.59364.5026.84 -224.92*112.75*337.66 Target Assets 3129.772399.0910993.8860070.00-8594.79***-57670.92 -49076.13Table 1This table presents the summary statistics for the sample of financial services industry acquisitions. The data are for all U.S affiliated mergers and acquisitions for all financial firms from November 12, 1999 to December 31, 2002 provided by the Worldwide M&A section of the SDC platinum d atabase. There are 353 completed financial industry M&As where U.S. firms were acquirers of U.S. target post the ratification of the FSMA.The transact ion value is the total purchase p rice the acquiring bank pays for other financial companies. Acquirer assets are the total assets of the acquirer during time of the acquisition. Acquirer financial leverage is the book debt to total assets during the time of the acquisition. The operation expense is the cost of goods sold for operations. Net sales is defined as the gross sales less returns, discounts and allowances. The acquirer cash is measured as the total book value of cash of the ac quirer during the time of the consolidation. The acquirer cash flow is measured as the change in cash availability of the acquirer the year prior to the acqu isition. The return of assets (ROA) is calculated as the ratio of net income to total assets during the time of the Description of Financial Industry Transactionsconsolidation. Return on investment (ROI) is calculated by dividing net profits less taxes by total assets. The Cost of Sales ( COS) is the cost of goods sold plus any expenses incurred in the selling and delivery of the product or service including the purchase of raw material and manufact ured finished products. The nonoperating income expense is the expense incurred in performance of activities not directly related to the main business of the firm, such as the maintenance of buildings and equipment. Target assets are the total assets of the target during time of the acquisition. Columns 5, 6, and 7 summarizes the differences between the types of acquisition for the various acquirer and target characteristics. ***, ** and denote significance at the 1%, 5% a nd 10% levels respectively. Values are reported in millions. 31

PAGE 39

The mergers between two commercial banks are the largest sub-sample, having 145 mergers with the ac quirer having an average $17. 5 billion in assets and the target having $2.4 billi on in assets. The smallest set of acquisitions are the commercial bank/ insurance company, with 13 acquisitions and a mean of $200 billion of acquirer assets a nd target assets of roughly $60 billion. The assets of the average commercial bank that acquirers ot her commercial banks are approximately $34 billion lower than the acquirers assets of the commercial ba nk/ brokerage firm mergers and about $183 billion smaller than the commercial bank/ insurance company mergers. The financial leverage of the acquirer in commercial bank/ commercial banks mergers does not differ appr eciably from the financial leverage of the brokerage firm acquisitions. The cost of sales are significantly larger for the commercial bank mergers than the brokerage firm or insurance company acquisitions. There is a significant difference (at the 0.01 level) between the mean net sales for commercial bank acquirers and acquirers in commercial bank/ brokerage firm mergers. Financial Services Industry M&A Transaction Value and Premium Table 2 presents the distribution of the transaction value by years and acquisition type. There is a s light decrease in the number of transactions from 2000 to 2002 for the entire acquisition sample. The st eadily decreasing number of financial industry acquisitions suggests that the industry has consolidated. However the number of commercial bank/ br okerage firm merger increased in the final year of the sample. This suggests that firms are recogn izing the benefit of the commercial bank/ brokerage firm combination resulting in a trend of more combinations between commercial banks and brokerage firms. I will more rigorously address this issue below. 32

PAGE 40

Transaction Value Premiu m Entire Sample Entire Sample Section ANmeanstd deviationminmax Nmeanstd deviationminmax 199921389.6131642.4800.24915925.201 210.2000.2740.0040.363 2000138804.3743685.2890.40033554.5791090.1670.2310.0140.312 2001112364.6151764.8602.45013132.151 810.1580.1950.0280.356 200282102.356331.0951.2502870.000 640.1880.1020.0080.324 All Years353477.0982175.2860.24933554.5792750.1720.1940.0040.363 Section B Bank / Bank Mergers Bank / Bank Mergers 199910381.8241746.2060.24915925.201 100.1950.0910.0040.315 2000105509.4202320.3470.40021084.873 840.1660.0680.0390.285 200195396.5521877.9693.20013132.151 710.1520.0720.0420.337 20026367.553598.4901.7002870.000 540.1890.0820.0080.324 All Years273363.5001748.0230.24921084.8732190.1680.0740.0040.337 Section C Bank/Brokerage Firm Merg ers Bank/Brokerage Firm Mergers 19999479.293822.1575.0003335.633 90.2070.1010.0050.363 2000241970.9566918.1712.20033554.579 200.1760.2840.1260.312 200110146.199320.6744.7501010.000 70.1990.0940.0280.356 200215272.63543.2491.250507.508 90.1940.1260.0310.304 All Years58985.6573056.7411.25033554.579 450.1890.1860.0050.363 Section D Bank/ Insurance Company Mergers Bank/ Insurance Company Mergers 199926.9103.8894.1609.660 20.1970.1510.0210.235 20009910.2321340.1291.4002449.297 50.1580.0510.0140.176 2001781.35099.8442.450193.600 30.2060.0420.1740.308 2002411.96310.4412.50026.500 10.103n.a.0.1030.103 All Years22401.055582.2551.4002449.297 110.1730.0620.0140.308 Transaction Value Differences Premium Differences Section E All Years -622.156 -0.021** Section F All Years -37.555 -0.005* Section G All Years 584.602 0.016 Bank/ Bank Mergers vs. Bank/ Insurance Company Mergers Bank/ Bank Mergers vs. Bank/ Insurance Company Mergers Bank/ Brokerage Firm Mergers vs. Bank/ Insurance Company Mergers Bank/ Brokerage Firm Mergers vs. Bank/ Insurance Company Mergers Bank/ Bank Mergers vs. Bank/ Brokerage Firm Mergers Bank/ Bank Mergers vs. Bank/ Brokerage Firm MergersTable 2Financial Services Industry M&A Transaction Value and PremiumThis table reoprts the descriptive statistics for the sample of financial services industry acquisitions by year from November 12, 1999 to December 31, 2002. There are 353 completed financial industry M&As where U.S. firms were acquirers of U.S. target post the ratification of the FSM A.The transaction value is the total purchase price the acquiring bank pays for other financial companies. Premiun is defined as the price paid for the target (transaction value) minus accounting book value of target's equity, this quantity, divided by the transaction value. Section A, B, C and D reports the de scriptive statistics for A) the entire sample, B) commerical bank mergers with other commerical banks, C) commerical bank mergers of brokerage firms, and D) co mmerical bank mergers with insurance companies. Sections E through G reports the tran saction value and premium differences for the various M&A subs amples. ***, ** and denote significance at the 1%, 5% and 10% levels respectively. 33

PAGE 41

The merger premium is also described in Table 2. I define premium as the price paid for the target (i.e., transacti on value) accounting book value of targets equity, this quantity, divided by the transaction value. Section A of Table 2 shows that on a year-by-year basis the acquisiti on premium ranges from 15.8% to 20% for the entire sample. The premiums for the commercial bank/ commercial bank mergers are similar with a low of 15.2% and a high of approxima tely 19.5%. The range of premiums over the sample period for comm ercial bank/brokerage firm mergers is slightly higher with a low of 17.6% and a high of 20.7%. However, the average range of the premium paid in mergers between commercial banks and insurance companies is significantly lower with a minimum of 10.3% and a maximum of 20.6%. For all sub-samples, acquirers paid the highest pr emium in 1999, the last year of the decadelong bull market when hubris would have most likely influenced these deals (Rau and Vermaelen (1998)). Subsequently, there is a sharp drop off in the premium paid. Note that the fall in the market and, hence, decline in the market value of equity of the target cannot completely explain the drop in premium because acquirers could continue to pay the same pr emium over the reduced value of the equity of the target. In addition, Section F of Table 2 reports that acquiri ng banks pay a significantly larger premium to acquirer brokerage firms than to purchase other commercial banks. These findings suggests that by paying a s ubstantial premium acqui rers are expecting to recognize gains potentially in the form of increased efficiency or their ability to cross-market their various products and services. The size of the transaction is likely to be correlated with the financial standing of the acquiring firm. That is, larger, more efficient successful banks are more likely to purchase competing banks or alternat ive financial service companies whereas smaller banks will become targets. Tabl e 3 reports the correlations between the acquirers characteristics and target total assets at the time of the acquisition. The correlation between acquire r total assets and ta rget total assets is 34%. This suggests that the two variables may play different roles in the determination of the premium paid in the acquisitions. However, the corr elations between acquirer total assets and 34

PAGE 42

35 financial leverage, acquirer net sales and ac quirer cash flow are all larger than 97%. This suggests that the acquirers characte ristics explain sim ilar aspects of the acquirers performance and transaction value, which would result in multicollinearity in multivariate regression analysis if not taken into account.

PAGE 43

AcquirorAcquiror AcquirorAcquirorAcquirorAcquirorAcquiror Target TotalAcquiror Total Financial OperationAcquiror NetAcquirorCash Return onReturn onCost ofNon Operating AssetsAssetsLeverageExpensesSalesCashFlowAssetsInvestmentSalesIncome Expense Target Total1 0.33950.38840.30860.32450.62130.76070.1435-0.04560.2747-0.3343 Assets (<.0001)(<.0001)(<.0001)(<.0001)(<.0001)(<.0001)(0.0127)(0.4305)(<.0001)(<.0001) 353 353 347347347350350335332348 305 Acquiror Total 1 0.97340.96310.98360.97100.98800.0897-0.10510.8344-0.9253 Assets (<.0001)(<.0001)(<.0001)(<.0001)(<.0001)(0.4257)(0.3506)(<.0001)(<.0001) 353 341342343341342332329342 305 Acquiror 10.91620.95610.95160.96570.0923-0.10390.8592-0.9363 Financial (<.0001)(<.0001)(<.0001)(<.0001)(0.4127)(0.3558)(<.0001)(<.0001) Leverage 347339341345342331327330 302 Acquiror 10.97310.72390.96820.1187-0.08710.7748-0.8273 Operation (<.0001)(<.0001)(<.0001)(0.1210)0.2558(<.0001)(<.0001) Expenses 347338344345331328331 301 Acquiror Net 10.78160.97740.1103-0.12210.8376-0.9199 Sales (<.0001)(<.0001)(0.1535)(0.1136)(<.0001)(<.0001) 347338337326315332 302 Acquiror 10.81730.0260-0.06460.7905-0.8764 Cash (<.0001)(0.6062)(0.1999)(<.0001)(<.0001) 350350326320332 302 Acquiror 10.2272-0.08060.7585-0.8239 Cash (0.0155)(0.3980)(<.0001)(<.0001) Flow 350329321336 300 Acquiror 10.82870.0574-0.8357 Return on (<.0001)(<.0001)(<.0001) Assets 335328331 301 Acquiror 10.0620-0.8431 Return on (<.0001)(<.0001) Investment 332322 297 Acquiror 1 -0.9582 Cost of (<.0001) Sales 348 305 Acquiror 1 Non Operating Income Expense 305Table 3Correlation analysis of Financial Se rvice Industry Acquisitions from 1999-2002This table presents the correlations for the Financial Service Industry M&As varialbes of interest from November 1999 to Decemb er 2002. There are 353 completed financial industry M&As where U.S. firms were acquirers of U.S. target post the ratification of the FSMA. Target assets are the total assets of the target during time of the acquisition. .Acquirer total assets are the total assets of the acquirer during time of the acquisition. Acquirer financial leverage is the book debt to total assets during the time of the acquisition. The operation expense is the cost of goods sold for operations. Net sales is defined as the gross sales less returns, discounts and allowances. The acquirer cash is measured as th e total book value of cash of the acquirer during the time of the consolidation. The acquirer cash flow is measured as the change is cash availability of the acquirer the year prior to the acqu isition. The return of assets (ROA) is calculated as the ratio of net income to total assets during the time of the consolidation. Return on investment (ROI) is calculated by dividing net profits l ess taxes by total assets. The Cost of Sales (COS) is the cost of goods sold plus any expenses incurred in the selling and delivery of the product or service including the purchase of raw material an d manufactured finished products. The non-operating income expense is the expense incurred in performance of activities not direc tly related to the main business of the firm, such as the maintenance of buildings and equipment. P-values are in parentheses. 36

PAGE 44

Financial Services Industry M&A Transaction Premium Regression In this section, I examine the cross-s ectional variation in premiums for U.S. financial industry mergers an d acquisitions. I use the merger premium to provide some insight into how managers of acqui ring firms perceive the opportunity provided by the FSMA to expand into non-banking activities relative to expanding their banking product via the acquisi tion of other banks. If ba nk managers regard highly the ability to diversify, then the premiu m for doing so should increase as we move from bank/bank to bank/brokerage firm to bank/insurance mergers. That is, this hypothesis proposes a positive and increa sing relationship between the merger premium and the diversity of the deal. Th is is because for a dollar of expected future revenues arising from the merger th e present value is higher the least correlated are the cash flow streams of the acquirer and the target, which re duces the discount rate. An alternative hypothesis is th at although managers may welcome the opportunity to diversify they may not be willing to pay the same premium to acquire non-core banking firms like brokerage and insu rance firms, as they would to acquire other banks. This is because as they move away from their core competencies the level of asymmetric information increases with the diversity of the deal. Thus, the discount rate to be applied to a dollar of incremental revenue s due to the deal increases. Therefore, this hypothesis predicts a positive premium that is declining in the diversity of the deal. It should be noted that managements perception of the value of the diversity of the merger as expressed by the relative sizes of the premiums need not correspond with the markets perception (as reflected in the announcement effect) of benefits of the merger. I examine the latter below. Table 4 presents results for the regression analysis. On average, the target firms capitalization is positively and significantly related to the acquisition premium. In addition, the premium can largely be at tributed to the acquirers size. This is consistent with the finding by Moeller et al. (2004) th at transacti on value is significantly related to the size of the acquiring firm. The acquirers financial characteristics also explain the cross sect ional variation in premium. The financial 37

PAGE 45

38 leverage of the acquirer is negatively linked to acquisition premium. This is consistent with the literature that finds that the acquiring firms inability to finance the deal adversely affects the deals va lue (Palia, 1993, Moel ler et al. 2004). The relationship between majority of the acquirers characteristics and the premium suggest that banks in good financial sta nding pay high transaction premiums to acquire competing banks, products, or services. Table 4 shows that bank/ brokerage firm mergers have higher premiums than bank/bank and that bank/insurance company mergers also have a greater premium than bank/bank mergers. These findings establish a positive and increasing relationship between the merger premium and the diversity of the acquisition. As in Benston et al. (1995) this study reports that banks bid more for merger partners that offer the potential for varying cash flows as a result of earnings diversification.

PAGE 46

model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel xmodel xi Variable CoefficienctCoefficienctCoefficienctCoefficienctCoefficienctCoefficienctCoefficienctCoefficienctCoefficienctCoefficienc tCoefficienct Intercept 1.8502.0401.5401.8392.1131.9342.0031.6491.2041.5931.953 (1.81)**(2.34)***(1.37)*(1.74)**(2.53)***(2.00)**(2.12)**(1.67)**(1.29)*(1.40)*(2.03)** Bank/ Brokerage Firm Merger 0.6750.5920.8470.7210.6130.8820.5730.6810.6100.7010.685 (1.82)**(1.76)**(2.03)**(1.91)**(1.81)**(2.21)**(1.74)**(1.86)**(1.78)**(1.89)**(1.86)** Bank/ Insurance Company Merger0.1220.1090.1310.1260.1130.1580.0990.1190.1110.1230.117 (1.41)*(1.30)*(1.52)*(1.49)*(1.33)*(1.76)**(1.29)*(1.40)*(1.31)*(1.47)*(1.39)* Target Assets 1.5051.7611.4221.4801.5761.648 (1.76)**(1.93)*(2.02)***(2.17)**(2.85)***(2.27)** Acquirer Assets 0.9890.991 1.0011.0241.0491.0881.057 (2.40)***(2.84)*** (3.00)***(3.29)***(3.43)***(3.51)***(3.49)*** Acquirer Financial Leverage -0.843 -0.747-1.011 -0.924 -1.044-0.786-0.995 (-2.54)*** (-2.05)**(-2.72)** (-2.66)*** (-2.87)***(-2.09)**(-2.81)*** Acquirer Net Sales 1.4344.131 3.073 -2.847 (1.56)*(2.81)*** (2.43)*** (-1.89)** Acquirer Cash 9.2508.68110.783 (1.59)*(1.50)**(1.74)** Acquirer Cash Flow 20.965 19.38318.459 14.557 0.87 0.660.58 0.50 Acquirer Return on Equity -12.860 -17.221-10.349 -0.54 -0.96-0.43 Acquirer Cost of Sales -17.645-15.243-18.676 -19.544 -16.675 (-1.49)*(-1.75)*(-1.39)** (-1.37)** (-1.64)** Non-Operating Income Expense -25.776-19.754-20.654 -0.48-0.31-0.37 Year Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign. Number of Observations Used 275275230264264275230230264264230 F Test 52.02046.12430.13716.02028.93216.39225.11721.24442.59418.00318.284 P-value 0.00410.00380.00390.00550.00400.00580.00400.00320.00320.00610.0069 Adjusted R2 0.8320.7910.5840.4940.6010.4590.5980.5810.8500.4800.607Table 4Financial Service Industry Premium Regression AnalysisThis table presents the results of the regression model used to explain the cross-sectional variation in Financial Service Indu stry M&As premium.This table presents the results of the multivariante regression model, using the 275 Financial Service Industry acquisition premiums from November 1999 to December 2002 where the a cquirer is a U.S. commerical bank and the target is either 1) a commerical bank, 2) brokerage firm or 3) an insurance company. The dependent variable is premiums paid for financial service in dustry companies. The key indepedent variable is "Industry Complexity" that takes the value of one for acquisition envolving either brokerage firms or insurance companies and zero for me rgers between commerical banks. Other independent varialbes are: logarithm of assets for the acquirer assets during time of the acquisition; acquirer financial leverage is the book debt to tot al assets during the time of the acquisition; the operation expense is the cost of goods sold for operations; net sales is defined as the gross sales less returns, discounts and allowances; the acquirer cash is measured as the total book value of cash of the acquirer during the time of the consolidation; the acquirer cash flow is measured as the change is cash availability of the acquirer the year prior to the acqu isition ; the return of assets (ROA) is calculated as the ratio of net income to total assets during the time of the consolidation; return on investment (ROI) is calculated by dividing net profits less taxes by total assets; the Cost of Sales (COS) is the cost of goods sold plus any expenses incurred in the selling and delivery of the product or service including the purchase of raw material and manufactured finished products; the non-operating income expense is the expense incurred in performance of activities not directly related to the mai n business of the firm, such as the maintenance of buildings and equipment; target assets are the logarithm of assets for the target during time of the acquisition; and year is a dichotomous v ariable representing the year of the acquisition. The t-statistics are in parentheses. ***, ** and denote significance at the 1%, 5% and 10% levels respectively. 39

PAGE 47

Financial Services Industry Announcement-Day Returns This section examines the market response of the financial services industry mergers and acquisitions. Table 5 reports the abnormal returns on the announcement date of the financial services industry acquisitions. I report separate results for bank/bank, bank/brokerage, and bank/insurance firm mergers. I also evaluate the two days surrounding the acquisition date. In order to compare the announcement-day returns of acquiring financial services comp anies and appropriate benchmarks and to determine whether diversifying acquisitions add value for shareholders, I provide a comparison between the acquisition sample and those financial companies that did not experience mergers during the sample period as well as a comparison with the market returns. The results in Table 5 (panel A) show that on the date of acquisition announcement the entire sample experiences no statistically signifi cant response. This finding is consistent with those reported in Jensen and Ruback (1984). However, for both diversifying mergers, between comme rcial banks and brokerage firms and between commercial banks and insurance companies, there is a statistically significant increase in retu rns. Examining the preand post-announcement daily returns indicates that for mergers between commercial banks and brokerage firms, there is a significant increase in abnormal returns on the day of and day after the acquisition announcement of 10.2% and 1.9% respectively. In addition, commercial bank mergers with insurance companies have a positive abnormal market response of 5.9% when compared to the market return (Table 5, panel B). These results are consistent with the hypothesis that di versifying financial services industry acquisitions increase shareholder value. Thes e results are consistent with those of Berger et al. (1999) that finds that join t production within both banking and insurance companies is more efficient for some firms. 40

PAGE 48

denote significance at the 1%, 5% and 10% levels respectively. A: Financial Services Industry Benchmarks NDay -1Day 0Day +1 All Financial Services Mergers 3530.0040.0530.032 (0.50)(1.04)(0.78) Commercial Bank/ Commercial Bank Mergers2730.0030.0490.029 (0.39)(0.89)(0.71) Commerical Bank/ Brokerage Firm Mergers 580.0060.1020.019 (1.01)(2.03)**(1.63)* Commercial Bank/ Insurance Company Mergers220.0020.0930.008 (0.42)(1.86)**(1.35)* B. Market Benchmarks NDay -1Day 0Day +1 All Financial Services Mergers 3530.00120.0410.0042 (0.49)(1.31)*(0.98) Commercial Bank/ Commercial Bank Mergers2730.00090.0330.0037 (0.38)(1.28)*(0.88) Commerical Bank/ Brokerage Firm Mergers 580.00060.0840.0036 (1.03)(1.73)**(1.07) Commercial Bank/ Insurance Company Mergers220.00420.0590.0082 (0.99)(1.67)**(1.19)Table 5Financial Services Industry M&A Announcement Day Returns This table reports the number of financial services industry mergers and acquisitions announcements and the average announcement day returns between November 12, 1999 and December 31, 2002. Announcement dates for the financial services industry acquisitions are obtained from the SDC Platimum database. Announcement day returns are computed as ARi = i, t-1(Ri,t Rt,,j) where Ri,t denotes the date t stock return of the acquiring firm i, R t,,j is the date t return of the equally weighted index of the financial service industry (Panel A) or the return of the equally weighted market model (Panel B).Each panel reoprts the abnormal returns for all financial services industry mergers, commercial bank acquisition of other commercial banks, the mergers between brokerage firms and commercial banks and the mergers of commercial banks with insurance companies. The t-statistics is reported in parentheses below the abnormal returns for the two day period surrounding the announcement date. Day 0 is the announcement date reported in SDC Platimum and Day -1 and Day +1 are the day prior to and after the announcement day, respectfully. The t-statistics are in parentheses. ***, ** and Financial Services Acquirers Efficiency Measurement and Correlations The efficiency measures of the ac quiring financial institutions are not perfectly correlated. The correlation betw een the distribution-free profit efficiency and operational efficiency pre and postacquisition are .61 and .69 respectively 41

PAGE 49

(Table 6). This suggests that the two m easures account for different aspects of banking efficiency. As discussed earlier, the distribution-free profit efficiency measure includes more inputs than the othe r efficiency measures, which may help explain the relatively low correlation betw een the distribution-free profit efficiency and operational efficiency measures. The mean estimates of 0.77 for distribution-free profit efficiency and 0.22 for operation e fficiency are simila r to those found by Berger and Hannan (1998) and Berger et al. (2007) of 0.70 and 0.77, respectively. Standard Operational Variable Sample SizeMeanMediandeviation Efficiency A. Pre-Acquisition Acquirer Profit Efficiency3530.7820.7690.320 0.61 (0.00) Operational Efficiency3280.2280.2200.113 1.00 Target Profit Efficiency3310.6020.5870.221 0.73 (0.00) Operational Efficiency3380.1400.1250.073 1.00 Difference (Acquirer-Target) Profit Efficiency325 0.18** Operational Efficiency328 0.088** B. Post-Acquisition Acquirer Profit Efficiency3530.7740.7620.307 0.69 (0.00) Operational Efficiency3280.2160.2110.107 1.00 Profit This table reoprts the descriptive statistics and correlations for the efficiency measures used to analyze the FSMA post-consol idation three-year returns. The profit efficiency measure was created using a distribution free random error method. To implement the distribution-free random error method, I estimate the profit function using data of 2214 financial institutions with continuous and complete annual data for the seven year period from 1999 through 2005. Using the results of these estimations, I calculate the profit efficiency for every financial institution with the distribution-free method, which distinguishes efficiency differences from r andom error by averaging the profit function residuals over time. Operational efficiency is defined as the ratio of operation expense s of financial institution to that of the assets of financial institution.Table 6Correlations of Efficiency MeasuresCorrelation (p-value) 1.00 1.00 Efficiency 1.00 42

PAGE 50

43 Univariate Efficiency Regression Results This section examines the impact the acquirers efficiency has on the postacquisition three-year return. Efficiency is often credited for the gains acquirers receive post acquisition although questions remain about the effects of intra-division supplementation for conglomerates (Lamont and Polk (2001)). The regression estimates in Table 7 confirm the results in the extant literature (e.g. Benston, 1989; Vennet, 2002 and Berger, DeYoung, Genay a nd Udell, 2007) that the three-year postconsolidation returns to a financial institution can be explained by these firms having high levels of efficiency. Both measures of efficiency have a positive and statistically significant impact of the firms three-year post-acquisition return for the entire sample. As expected, the significance of the distribution-free profit efficiency measure is greater than the operational effici ency measure. This is expected because the distribution-free profit efficiency meas ure is a more comprehensive measure that includes both cost and income variables. As for the different types of financial services industry mergers, the three-y ear returns for mergers between commercial banks are significantly expl ained by both efficiency measures. This result is consistent with the findings of Berger, DeYoung, Genay and Udell, (2007). Both the profit efficiency and operational efficiency measure are statistically significant for the commercial bank/commercial bank mergers. However, both efficiency measures explain more about the commercial bank/in surance company mergers than bank/bank mergers as witnessed by the increase in r-squares. Finally, the profit efficiency measure is the only efficiency measure that is statistically significant for the commercial bank/brokerage fi rm and commercial bank/insurance company mergers. These results establish that banking efficiency impacts not only commercial bank returns, but also the returns of the comb ination firms of comme rcial bank/brokerage firms and commercial bank/insurance companies.

PAGE 51

Section A Commercial Bank/ Commercial Bank/ Commercial Bank/ Commercial Bank Brokerage Firm Insurance Company Variable Coefficientt-statisticCoefficientt-statisticCoefficientt-statisticCoefficientt-statistic Intercept -0.28294-12.49***-0.26941-11.12***-0.41198-7.89***-0.42549-6.93*** Profit Efficiency 0.6509613.93***0.6336612.87***0.8315112.45***0.8437711.39*** Sample Size F-Value Adjusted R-Square Section B Coefficientt-statisticCoefficientt-statisticCoefficientt-statisticCoefficientt-statistic Intercept -0.20186-9.29***-0.20568-8.91***-0.18272-2.54***-0.18107-2.27** Operational Efficiency0.33506551.94**0.33775791.97**0.47807270.49 0.11330.12Table 7Univariate Analysis of Financial Service Inudstry M&As with Efficiency MeasuresFinancial Service Industry All This table presents the results of the univariate regression model used to explain the cross-sectional variation in the post-FS MA three-year returns of Financial Service Industry acquirers. This table presents the results of the regression model, using the 353 Financial Service Industry acquisitions from November 1999 to December 2002 where the acquirer is a U.S. commerical bank and the target is either 1) a commerical bank, 2) b rokerage firm or 3) an insurance company. The dependent varialbe is the post-FSMA three-year return for the consolidated financial firm. Section A rep orts the results for the Profit efficiency measure. The profit efficiency measure was created using a distribution free random error method. To implemen t the distribution-free random error method, I estimate the profit function using data of 2214 financial institutions with continuous and complete annu al data for the seven year period from 1999 through 2005. Using the results of these estimations, I calculate the profit efficiency for every financial institution w ith the distribution-free method, which distinguish es efficiency differences from random error by averaging the profit function residuals over time. Section B reports the results for the Operation efficiency measure. Operation efficiency is defin ed as the ratio of operation expenses of financial institution to that of the assets of financial institution. ***, ** and denote significance at the 1%, 5% and 10% levels respectively. 353 5.72 0.4230 0.4004 4.78 273 58 12.97 0.6895 0.7066 15.51 22 44 Sample Size F-Value Adjusted R-Square 0.0328 0.0054 0.24 51 257 3.66 0.0107 0.0098 3.75 328 20 0.01

PAGE 52

Financial Services Industry Pro duct Market Spillover Measurement The three-year post-consolidation ch ange in the acquirers financial characteristics standardized by the acquirers to tal assets are utilized to measure spillover. Table 8 offers summary statistics of the main variables used to define product market spillover. Averaging across all acquirer-year observations, the ratio of excess sales to total assets has significantly increased over the three-year period, by 5.7 percentage points. Similarly, the change in the operating income to total assets ratio is substantial as indicated by the statistically significant incr ease of 6.6 percentage points over the sample period. The acquiring financial in stitutions are also recognizing efficiency gains. The cost of sales is statistically significantly lower three years followi ng the acquisition. The results for the three-year post-acquisition chan ge in cash flow are significantly larger, increasing by 6.6 percentage points. These re sults are the first indication that financial services acquirers are receivi ng benefits resulting from product market spillover. 45

PAGE 53

Standardized Variables n t t+3 Panel A: Spillover Hypothesis Variables Cash 3460.04670.05510.0085 Cash Flow 3460.01490.08070.0659** Financial Leverage 3470.89520.8290-0.0663* Excess Sales 3200.06590.12290.0570*** Cost of Sales 3210.05490.0375-0.0174* Non-Operating Income Exp e 3050.00910.04210.0330** Panel B: Control Variables Operation Income 3280.03840.10510.0667*** Gross Profit 3270.06350.17480.1113*** difference Table 8Financial Service Industry Post-Consolidation ChangesThis tables presents a summary of explanatory variables scaled by total assets for the acquiring U.S. commerical banks that purchase U.S. targets that are either 1) a commerical bank, 2) brokerage firm or 3) an insurance company. Panel A reports the spillover hypothesis explanatory variables and Panel B reports the values of other control variables. This table provides estimates of the mean difference between the acquirers characteristics weighted by total assets. ***, ** and denote significance at the 1%, 5% and 10% levels respectively. Financial Services Industry Product Mark et Spillover Multivariate Regression Results The objective is to assess the relationshi p between financial institution spillover and the variation in the firms return while accounting for the firms efficiency. In other words, it is important to control for the po ssibility that acquiring banks with efficient policies and practices may be able to transfer these efficiencies to the target firm. These results are reported in Tables 9 and 10. After controlling for the firms efficiency and acquisition diversity, there is a spillover in financial institutions. The relationship between the change in excess sales and returns is economically si gnificant and meaningful. The results also indicate that 46

PAGE 54

47 there is a positive and statistically signifi cant relationship between cash/cash flow and returns. These findings suggest that financial institutions that diversify their product offerings increase shareholder returns resulting from an increase in excess sales. This then leads to an increase in cash and cash flow which in turn increases firm returns. The sample also has a negative and statistically significant coefficient on financial leverage spillover (-0.703, t = -2.38), confirming that financially sound institutions experience strong valuation effects followi ng acquisition. Thus, banks w ith stronger capital ratios are better positioned to benefit from additi onal products and services that accompany financial services acquisiti ons. These spillover results combined with the strong significance of the efficiency measure indica tes that, while the complimentary products and services supplement the firms income, its the firms efficiency that dictates majority of the return. This is evident by the comparison of the univariate efficiency regression model and the multivariate regression model. The univariate efficiency regression model in Table 7 report an adju sted r-square of .42 compared to the multivariate regression model adjusted r-square ranging between .73 and .84. Thus, at least half of the explanatory pow er of the model is due to the efficiency variables. This is consistent with Berger and Mester (2003) who find that profit productivity improved during the 1991-1997 period. However, the spillove r results contradict Berger and Mester (2003) finding that cost productivity declin es annually. The efficiency and spillover findings are consistent with the hypotheses that over the three-year time period, the acquiring financial institutions have contin ued effective and efficient practices while benefiting from the spillover of information fr om the acquired institution. In addition, the coefficients of the interaction terms (efficiency*bank/brokerage and efficiency*bank/insurance) are both positive an d significant. This implies that efficiency has a greater effect on returns when the bank merges with either brokerage firms or insurance companies than when it merges with another bank.

PAGE 55

Variable model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel xmodel ximodel xii Intercept 0.23810.26770.25820.30180.28640.21570.25240.20230.19930.20040.19030.2176 (10.71)***(9.93)***(11.02)***(11.23)***(9.90)***(10.83)*** (8.98)***(8.73)***(7.43)***(7.28)***(7.11)***(9.06)*** Profit Efficiency 0.64370.65320.64930.66710.67410.68460.65780.65930.69940.69970.68170.6832 (10.38)***(11.03)***(10.77)***(11.21)***(12.03)***(12.54)***(1 2.31)***(11.91)***(13.01)***(13.14)***(12.32)***(12.23)*** Bank/ Brokerage 1.04391.10721.08141.01501.10041.03481.0121.12591.12651.04470.97061.0072 (1.79)**(1.96)*(2.04)**(1.46)*(2.01)**(1.67)**(1.61)*(2.15)**(2.23)**(1.54)*(1.34)*(1.82)** Bank/ Insurance 0.98300.84340.73490.53581.00340.99030.54060.89321.00810.98140.48310.8711 (1.32)*(1.56)*(1.47)*(1.29)*(1.66)**(1.36)*(1.31)*(1.29)*(1.92)**(1.50)*(1.26)*(1.68)** Financial Leverage Spillover -0.7032-0.7103-0.6996-0.7207 -0.6675-0.6549-0.6254-0.6614 (-2.38)***(-2.58)***(-2.44)***(-2.76)*** (-2.08)**(-1.54)*(-1.36)*(-1.77)** Excess Sales Spillover 0.03390.03590.03560.03110.03270.03210.03790.0307 (1.32)*(1.28)*(1.35)*(1.31)*(1.26)*(1.29)*(1.54)*(1.25)* Cash Flow Spillover 0.07320.07690.0710 0.07130.07260.0761 (1.62)*(1.68)**(1.59)* (1.63)*(1.29)*(1.44)* Cash Spillover 0.0903 0.09790.09560.09840.0918 (1.40)* (1.61)*(1.60)*(1.54)*(1.52)* Cost of Sales Spillover -0.2907 -0.3150-0.2725 -0.3452 -0.3278 -0.2882 (-0.64) (-0.69)(-0.70) (-0.62) (-0.68) (-0.73) Non-Operating Expense Spillover -0.0631 -0.0601 -0.0625-0.0626 -0.0665 (-0.07) (-0.07) (-0.05)(-0.06) (-0.10) Return Risk 0.16270.1954 0.1906 0.17590.19290.1892 (0.74)(0.96) (0.82) (0.78)(0.91)(0.88) Efficiency* Bank/ Brokerage Interaction0.66930.6779 0.69380.71120.6833 0.73870.7003 (12.19)***(12.62)*** (12.78)***(12.91)***(12.95)*** (14.38)***(12.71)*** Efficiency* Bank/ Insurance Interaction 0.64520.6509 0.66030.7089 0.6918 (11.86)***(12.04)*** (12.17)***(12.38)*** (12.54)*** Joint Significance between Efficiency 0.67430.69830.6 5830.66190.70020.72330.69310.69310.71190.78120.79760.7079 and Interaction (12.39)***(12.87)***(12.15)***(12.38)***(12.66)* **(13.09)***(13.12)***(12.43)***(12.77)***(14.57)***(12.88)***( 12.84)*** Year Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign. Number of Observations Used 320305310302305320305305321317305314 F Test 43.4242.8154.9255.1127.8334.8237.3234.1937.6233.6737.9832.09 P-Value 0.00020.00090.00050.00020.00210.00300.00370.00330.00410.00350.00490.0032 Adjusted R2 0.80060.79200.83290.84080.73960.76230.78310.77510.78960.78150.79100.7665 financial service industry mergers.The t-statistics are in parentheses. ***, ** and denote significance at the 1%, 5% and 10% levels, respectively.Table 9This table presents the re sults of the multivariate regression model used to explain the cross-sectional variation in the postFSMA three-year returns of Financial Service Industry acquirers. This table presents the results of the regression model, using the 353 Fi nancial Service Industry acquisitions from November 1999 to Decem ber 2002 where the acquirer is a U.S. commerical bank and the target is either 1) a commerical bank, 2) brokerage firm or 3) an insuran ce company. The dependent varialbe is the post-FSMA three-year r eturn for the consolidated financial firm. The key indepedent variable is Profit efficiency. The profit efficiency measure was created using a distribution free random error method. To implement the di stribution-free random error method, I estimate the profit function using data of 2214 financial institutions w ith continuous Using the resu lts of these estimations, I calcu late the profit efficiency for every financial institution with the dist ribution-free method, which distinguishes efficiency differences from random error by averaging the profit function residuals over time. Bank/Brokerage is a dichotomous variable that takes the value of one for commerical bank and brokerage firm mergers and zero otherwise. Bank/Insurance is a dichotomous variab le that takes the value of one for commerical bank and insurance company mergers and zero otherwise. Other post-consolidation acqui rer independent variables are: logarithm of assets for the a cquirers' assets; financial leverage is the book debt to total assets; the operation expense is the cost of goods so ld for operations; excess sales is actual sa les change less predic ted sales changeis. Efficiency, measured as the banks effi ciency utilizin g the distribution-free profit or operational efficiency variable to pred ict sales changes; cash is measured as the total book value of cash of the consoldate d firm; the cash flow is m easured as the change is cas h availability of the acquirer the year prior to the acquisition; the Cost of Sales (COS) is th e cost of goods sold plus any expenses incurre d in the selling and delivery of the product or service including the purchase of raw material and manufactured finished products; the non-operating income expens e is the expense incu rred in performance of activities not directly related to the main business of the firm, such as the maintenance of b uildings and equipment; retu rn risk is the standard deviation of returns for 36 months mu ltiplied by the square root of 36 resu lting in the three-year risk in returns; and year is a dichotomous variable represe n Multivariate Analysis of Financial Service Inudstry M&As with the Profit Efficiency Measure 48

PAGE 56

Utilizing the operational efficiency meas ure to account for the acquirers postconsolidation efficiency, the Table 10 result s on spillover confirm the earlier results. There is spillover from the target firm to the acquiring bank. There is a significant positive association between the change in excess sales and the firms returns while controlling for firm risk and accounting for operation efficiency. The Excess Sales spillover and the Cash/ Cash Flow spillover ar e slightly larger than previously reported when efficiency is measured using operationa l efficiency. This is probably due to the inclusiveness of the profit efficiency m easure and the limitations of the operation efficiency measure. The efficiency measur es (profit and operation efficiencies) are statistically significant (Tables 9 and 10). Thes e findings suggest that the acquiring firms efficiency help explains the three-year returns. This is consistent with the results of Rhoades (1998) that finds modest cost efficiency gains using data from the early 1990s.14 However, the increase in statis tical significance for the spillover variables in Table 10 suggests that product market sp illover explains the future returns of acquiring banks. 14 Note that Berger (1998) reports conflicting results finding very little improvement in cost efficiency for M&As of either large or small banks using data from 1991 to 1997. 49

PAGE 57

Variable model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel xmodel ximodel xii Intercept 0.50220.41920.49310.42030. 42160.412890.5192380.52310.43080.479280.39990.4002 (11.43)***(11.38)***(11.65)***(11.10)***(11.83)***(11.21)*** (10.32)***(10.94)***(9.92)***(10.32)***(11.01)***(10.44)*** Operational Efficiency 0.22730.29950.2442 0.25550.20020.22350.22850.24690.22240.23840.24450.2827 (1.61)*(1.66)**(1.72)**(1.58)**(1.69)**(1.36)*(1.44)*(1.59)*(1.41)*(1.37)*(1.67)**(1.75)** Bank/ Brokerage 1.26541.31101.21481.2098 1.23051.26751.35101.30371.19861.20471.30021.3094 (2.08)**(2.37)**(1.82)**(1.74)**(1.89)**(2.05) **(2.21)**(2.39)**(1.54)*(1.59)*(2.48)**(2.51)** Bank/ Insurance 1.00251.00111.00251.0042 1.00451.00331.00231.00351.00231.00241.00351.0028 (1.79)**(1.65)**(1.43)*(1.85)**(1.81)**(1.98)** (2.03)**(1.93)**(1.41)*(1.43)*(1.95)**(1.83)** Financial Leverage Spillover -1.4399-1.4538-1.2843-1.2930 -1.6287-1.5843-1.6390-1.6303 (-2.20)***(-2.24)**(2.13)**(2.14)** (-2.51)***(-2.47)**(-2.30)***(2.53)*** 6 Excess Sales Spillover 0.06240.06290.07070.06300.06920.07200.06730.0693 (1.53)**(1.56)*(1.76)**(1.51)*(1.63)*(1.88)**(1.57)*(1.65)** Cash Flow Spillover 0.09210.10290.0903 0.10100.09130.0912 (1.99)**(2.25)**(1.97)** (2.19)*(2.01)**(1.89)** Cash Spillover 0.1030 0.10350.10500.10210.1007 (1.68)** (1.74)**(1.85)**(1.58)*(1.38)* Cost of Sales Spillover -0.6294 -0.6527-0.6385 -0.6103 -0.7024 -0.6584 (-1.02) (-1.01)(-0.82) (-0.69) (-0.71) (-0.51) Non-Operating Expense Spillover -0.0112 0.0111 0.0142-0.0193 -0.0164 (-0.63) (-0.89) (-0.59)(-0.48) (-0.32) Return Risk 0.407430.5109238 0.5402931 0.46928370.5029810.4198241 (0.97)(1.10) (1.21) (1.03)(1.14)(1.15) Efficiency* Bank/ Brokerage Interaction1.49241.5299 1.42051.51041.5584 1.33921.4982 (2.78)***(2.88)*** (2.36)***(2.80)**(3.11)*** (1.68)**(2.79)*** Efficiency* Bank/ Insurance Interaction 1.28371.3003948 1.3022931.2908 1.3211 (1.67)**(2.00)** (2.10)**(1.89)* (1.91)** Joint Significance between Efficiency 1.78291 .84401.52981.58901.72981.82981.98281.59011.55111.59821.79321.6035 and Interaction (3.02)***(3.24)***(1.88)**(2.89)***(2.99)***(3 .15)***(3.34)***(2.90)***(2.76)***(1.88)**(3.13)***(2.80)*** Year Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign.Insign. Number of Observations Used 320305310302305320305305321317305314 F Test 24.1826.0627.5422.7119.6917.3923.3713.7721.3212.1923.9113.20 PValue 0.00060.00110.00090.00020. 00150.00230.00150.00300.00080.00380.00160.0029 This table presents the results of the multivariate regression model used to explain the cross-sectional variation in the postFSMA three-year returns of Financial Service Industry acquirers. This table presents the results of the regression model, using the 353 Financial Service Industry acquisitions from November 1999 to Decem ber 2002 where the acquirer is a U.S. commerical bank and the target is either 1) a commerical bank, 2) brokerage firm or 3) an insurance company. The dependent varialbe is the post-FSMA three-year r eturn for the consolidated financial firm. The key indepedent variable is Operation efficiency. Operation efficiency is defined as the ratio of operation expenses of financial institution to that of th e assets of financial institution. Ban Adjusted R2 .0.50390.51070.52030.49930. 48220.46380.50020.43870.49750.39280.50260.4173 k/ Bro k erage i s a di c h otomous var i a bl e t h at ta k es t h e va l ue o f one f or commer i ca l b an k an d b ro k erage fi rm mergers an d zero ot h erw i se. Ban k/ Insurance i s a di c h otomous var i a bl e t h at ta k es t h e va l ue of one for commerical bank and insurance company mergers and zero otherwise. Other post-consolidation acquirer independent var iables are: logarithm of assets for the acquirers' assets; financial leverage is the book debt to total assets; the operation expense is the cost of goods sold for operations; excess sales is actu al sales change less predicted sales changeis. Efficiency, measured as the banks efficiency utilizing the distribution-free profit or operational efficiency variable to predict sales changes; cash is measured as the total book value of cash of the consoldated firm; the cash flow is measured as the change is cash availability of the acquirer the year prior to the acquisition; the Cost of Sales (COS) is the c ost of goods sold plus any expenses incurred in the selling and delivery of the product or service including the purchase of raw material and manufactured finished products; the non-operating income expense is the expense incurred in performance of activities not directly related to the main business of the firm, such as the maintenance of buildings and equipment; return risk is the standard deviation of returns for 36 months multiplied by the square root of 36 resulting in the three-year risk in returns; and year is a dichotomous variable r epresenting.the year of the acquisition. Interaction variables are also included to examine the relationship of efficiency to the various types of financial service industry mergers.The t-statistics are in pa rentheses. ***, ** and denote significance at the 1%, 5% and 10% levels, respectively.Table 10Multivariate Analysis of Financial Service Inudstry M&As with the Operation Efficiency Measure 50

PAGE 58

Conclusion In this paper, I examine the long-term effects of the Financial Services Modernization Act of 1999 which allowed th e combination of banking, brokerage and insurance services under one financial conglom erate. This paper examines the different types of financial combinations and the size of the transaction valu e for the varying types of mergers. This study attemp ts to explain the cross secti onal acquisition pr emium for the varying types of financial service mergers. I employ a multivariate regression analysis to study the merger transaction value and explai n the long-term returns of the financial combinations. As in the extant manufacturing literature, the banking industry mergers and acquisitions transaction value is related to th e size of the target. The results also indicate that the acquiring firms financial characteri stics including the firms size, leverage and operating cost determine the acquisition premiu m. This suggests that not only does the target value dictate the amount pa id for the assets but also the target firms ability to be integrated and compliment acquiring firm financial standing. In addition, the large transaction values w ithin the financial industry merg ers/ acquisitions has lead to significant premiums paid for the targets. The premiums paid for the firms within the financial service industry vary from acquisition type to acquisition type. The result s show that on average the premium paid is approximately 17% above target value. Th e merger premiums for commercial bank/ commercial bank and commercial bank/ brokerage firm mergers are st atistically larger than those for commercial bank/ insurance company mergers. This may be due to the ease with which the acquiring commercial bank can cross sell its products and services with other banks and brokerage firms. The difference in premiums can also be the function of the difference in the size of a ssets and clientele between commercial bank, brokerage firms and insurance companies. Since insurance companies are significantly larger than other types of financial serv ice companies there may be greater spillover benefit in cross selling products and services. 51

PAGE 59

This paper also examines the announcem ent day market response for acquisitions within the financial service industry to specifically de termine whether diversifying acquisitions add value for shareholders. This study shows that on the date of acquisition announcement the entire sample experiences no statistically sign ificant response. However, both diversifying mergers between commercial banks and brokerage firms and commercial banks with insurance companies repo rt a statistically significant increase in returns. Examining the preand postannounc ement daily returns in dicates that for the commercial bank mergers with brokerage firm sample, there is a significant increase in abnormal returns on the day of and day afte r the acquisition announcement of 10.2% and 1.9% respectively. In addition, commercial bank mergers with insurance companies have a positive abnormal market response of 5.9% wh en compare to the market return. All of the results are consistent with the theory th at diversifying financia l services industry acquisitions increase shareholder value. The large premiums reflect expectations that the acquiring firms will be able to cross-sell their products and utilize inform ation from different divisions to develop superior services or for the acquirer to put in place efficient polic ies and practices to receive superior returns. U tilizing the distribution-free random error method, I find evidence that acquiring financial firms are more efficient than target firms and that these efficiencies dictate long-term performance of the firms. I find that not only profit efficiency but also operation efficiency determin es future returns. There is also evidence that when banks engage in diversifying merger s with brokerage firms the effects of profit efficiency on post-consolida tion returns are larger than those of commercial bank/ commercial bank mergers. A competing view holds that financial acquirers are able to cross-market products and services resulting from spillover between di visions. Looking at a variety of spillover characteristics, the results indicate that ac quiring financial firms take advantage of the diversity. The results indicate that there is a positive and statistically significant relationship between the three-year change ex cess sales and returns. This suggests that financial institutions that diversify their products and services are able to increase 52

PAGE 60

shareholder wealth by selling more products. Spillover can also be se en by the change in cash flow for the post-consolidat ed financial conglomerate. On average, if the financial conglomerate can increase cash flow that will lead to an increase in shareholder returns. The ratification of the 1999 Financial Services Modern ization Act authorized commercial banks, brokerage firms and insurance companies to combine their businesses and thus significantly broadened their produ ct and service offerings and improve their ability to compete with industry rivals. Estim ation of the multivariate regression models show some evidence of support for both the efficiency and spillover hypotheses. Thus, the proprietary customer information gathered by one financial intermediary is efficiently and effectively disseminated to complimentary divisions of the financial conglomerate. 53

PAGE 61

Essay 2 Down but Not Out Mutual Fund Man ager Turnover within Fund Families Introduction Ever since Berle and Means (1932) first established that there is a separation between ownership and control and Jensen and Meckling (1976) recognized that this disconnect between managers and sharehol ders causes agency issues, financial economists have discussed ways to eliminate or at least minimize these agency concerns. The financial literature has advanced two fundamental theories about how to address agency problems and influence manager behavi or. First, financial economists suggest that the board of directors design compensation sc hemes to provide managers with effective incentives to maximize shareholder value (p ay-performance). Secondly, the market for corporate control imposes some constraints on the managers actions. These two approaches are designed to align the mana gers behavior with shareholders wealth maximization. Despite the interest in this area, there has yet to be a study that examines the effects of both the pay-pe rformance and the market for corporate control theories simultaneously. The uniqueness of the mutual fund industry and the mutual fund manager contracts allows us to reex amine these agency issues. Within the mutual fund industry, this i ssue is significant given the importance of management in the implementation of the funds investment strategy, the sizable assets under their control, and the poten tial impact it has on the overall succes s and profitability of the fund complex. The issue is also cri tical in terms of the different corporate governance mechanisms and principal-agent problems that exist between investors, shareholders, and management. This is because investors that entrust funds to managers cannot participate in exerci sing corporate control in the same manner in which shareholders can exercise their collective will on company boards. Accordingly, while internal control mechanisms of investment management organizatio ns are likely to be related to corporate governance practices experienced by industrial organizations, the 54

PAGE 62

literature has not devoted si gnificant attention to organi zational structure that is associated with changes in mutual f und management of investment firms. Thus far, the mutual fund literature ha s investigated how mutual fund manager turnover is affected by past performan ce and manager age. Examining the relation between mutual fund managerial replacemen t and prior performance, Khorana (1996) finds evidence of an inverse relation be tween the probability of fund manager replacement and past performance. Extending th e work of Khorana (1996), Chevalier and Ellison (1999) investigate the link betw een mutual fund managers age and the probability of manager replacement. They find that younger managers are more likely to be replaced if the funds systematic and unsys tematic risks deviate from the investment objectives average risk level. Khorana (1996) also documents that the magnitude of underperformance that investment advisors are willing to accept before replacing a manager is positively related to the volatility of the underlying assets being managed by fund managers. Khorana states that his findi ngs are consistent with well-functioning internal and external market mech anisms for mutual fund managers.15 However, is it possible that previous resear ch, by ignoring the specific orga nizational form and, more specifically, the management structure of the fund family, might have significantly overstated the sensitivity of managerial replacement to past performance? My first objective in this paper is to ex tend the Khorana (1996) and Chevalier and Ellison (1999) results to a setting that acc ounts for both the fund family organizational structure and the individual characteristics of fund managers. While previous literature helps us understand the replacement-performance relationship of mutual fund managers, we know little about how the managerial struct ure of the mutual fund family influences the sponsors willingness to replace underperforming managers. 15 One weakest of the mutual fund managerial tu rnover literature is th at it is difficult to distinguish between turnover due to promotion and turnov er due to demotion caused by underperformance. In a working paper, Hu, Hall and Harvey (2000) separates manager changes into promotions and demotions. Thei r evidence suggests that the probability that a manager is likely to be fired or demoted is negatively correlated with the fund's current and past performance and the promotion probabi lity is positively related with the fund's current and past performance. 55

PAGE 63

In addition, this study identifies the importance of the management structure within the mutual fund industry. Within so me fund families, a portfolio manager works autonomously managing only one fund. At othe r fund families, an individual portfolio manager is responsible for two or more mutu al funds within the same sector, related sectors or with complementary investment objectives. For the sample period, 22% of mutual funds accounting for 25% of the funds under management are now in the multiple fund management structure. Fund sponsor s make manager tur nover decisions by comparing the cost of firing UFM versus MFM16 and the benefits of having the MFM structure. The incremental cost of repl acing a unitary fund manager includes the employee search cost and hiring of a new f und manager and the potential cost of losing loyal investment customers of the replaced manager. Under the MFM structure, these costs arent necessarily a concern for fund s ponsors since the replaced managers remain with the fund family. In addition, the MFM structure lowers the individual cost of operating each fund. If fund sponsors are le ss likely to end the services of a UFM manager, because it is more costly to the sp onsor, then this is a clear indication of a conflict of interests because for the same level of underperformance investors would benefit more if the pay for performan ce relationship (proposed by Khorana (1996)) worked effectively for the costlier funds/f und management system. However, without considering the specific organizational form and, more specifically, the management structure of the fund family previous rese arch might have signi ficantly overstated the sensitivity of managerial replacement to past performance. I show that, in addition to prior performance and managerial experien ce, the number of individual funds managed by a fund manager increases the probability of manager replacement. UFM are -2.77% less likely to be replaced than a MFM, even though both managers are underperforming. This suggests that fund sponsors tend to repl ace underperformers only when it is cheap because replacing a UFM is more expensive than taking one fund from a manager that 16 In the case of an MFM, we regard a manager as having been replaced if he is relieved of his duties related to one or more funds of the two or more that he manages, even if he continues to be in charge of other funds. 56

PAGE 64

operates multiple funds. This presents an obvious conflict of inte rests between fund investors and fund management. The second objective of this study, as in Khorana (2001), is to examine whether funds that experienced manager replacemen t underperform funds where the manager maintains responsibility and, if so, by how much and for how long prior to replacement. However, unlike Khorana (2001) this st udy takes into account the fund family management structure prior to replacement. Consistent with Khoran a (2001), I find that new fund managers exhibit dramatic performance improvement in the post-replacement period. This finding suggests that the previo us manages were replaced due to poor past performance. Potential explanations for poor past performance is that fund managers have too many funds or fund objectives to mana ge to be effective and/ or diminished fund management abilities. I also find th at unitary fund managers significantly underperform their objective and risk adjusted peers (1.8%, 2.8% re spectively), which is a greater underperformance than multiple fund managers (1.2%, 2.6% respectively). It appears that fund sponsors are more tolerant of unitary fund manage rs underperformance than that of multiple fund managers. Contra ry to Khorana (1996, 2001), these findings suggest weaker internal control mechanism than previously thought. The final objective is to extend the wo rk Chevalier and Ellison (1999a,b) and Gallagher (2003), who examine performan ce related to investment manager characteristics, including experience, institut ional asset size, and investment management characteristics. As in Chevalier and E llison (1999a,b) and Gallagher (2003) this study documents an inverse relationship between manager tenure and the probability of replacement. After simultaneously accounting for manager tenure and past performance, I find a statistically significant negative relationship with th e probability of a manager being replaced and the combination of ma nager tenure and past performance. This finding suggests that sponsors are reluctant to fire poor performers if they are experienced fund managers. This may be beca use even with underperformance, relative to peers, fund managers have an establishe d relationship with inve stors and firing the 57

PAGE 65

manager can signal problems to potential invest ors and result in an outflow of funds from current investors. This would lead to a reduction in sponsor income. This study represents the first signi ficant and rigorous examination of the relationship between performance, manager ch aracteristics and fund family management structure. In this paper, I present eviden ce that mutual fund replacement is not only contingent on previous performance and ma nager tenure but also on the number of individual funds managed by the fund manager. In addition, I document the importance of the management structure w ithin the mutual fund industry. While previous literature helps us understand the replacement-performance relationship of mutual fund managers, we know little about how the organizational fo rm, specifically the managerial structure, of the mutual fund family influences the sens itivity of replacement to past performance. The remainder of the paper is organized as follows. Section II discusses the related literature and develops the hypotheses tested. Section III describes the data and methodology used for analysis. Section IV provi des a sample description and preliminary statistics of the repla ced fund managers. Section IV also presents the empirical results of the study. I conclude this paper with a su mmary of my findings in Section V. Related Literature/ Hypotheses Development Agency Issues Financial economists have found that agen cy problems or conflict of interests between shareholders (principal s) and managers (agents) co me from two sources. First, managers and shareholders have different go als and preferences. Secondly, managers and shareholders have imperfect information as to each others knowledge, actions and preferences. Berle and Means (1932) notes th at this separation provides managers with the ability to act in their own self-interest ra ther than in the interest of shareholders without corporate governance mechanisms. Shle ifer and Vishney (1997) also finds that managers use their discretion to benefit themselves personally in a variety of ways such as empire building (Jensen, 1972), failure to distribute excess cash when the firm does not have profitable investment opportunities (Jensen, 1986), and manager entrenchment 58

PAGE 66

(Murphy, Shleifer and Vishney, 1989). Within the mutual fund industry, the interaction between investors and fund management repr esents a principalagent relationship. Investors delegate assets to professional fund managers with the expectation that performance will be commensurate with the funds investment objective. However, while performance is important to the fund family, the primary goal for a fund complex is to maximize the total assets under management, as revenue is generated as a percentage of fund assets under management. Although perf ormance and fund size are interrelated (Gruber (1996)), the first objective for a fund manager is to maximize total assets under management. There are two distinct theories about how to effectively deal with these agency problems. In general terms, these theories can be viewed as internal and external control mechanisms (Fama, 1980). The design of th e compensation contracts by the board of directors is considered an internal contro l mechanism while the market for corporate control is an external control mechanism. There has b een extensive res earch conducted on compensation contracts and agency issues.17 Several papers find that compensation contracts seem to reflect managerial rent-seeking rather than the proper incentives to align manager actions with shareholder interest (Blanchard, Lopez-de-Silanes and Shleifer, 1994; Yermack, 1995; Bertrand a nd Mullainathan, 2001; Bebchuk, Fried and Walker, 2003). However, Jensen and Murphy (1990) asserts that optimal contracting arrangements require large am ounts of compensation for execu tives to provide managers with powerful incentives to enhance shareholder value. This suggests the use of equitybased compensation contracts to make pay mo re sensitive to performance. The mutual fund industry utilizes the suggested equity-bas ed compensation contracts by aligning the management fee (a stated rate) with the value of the funds ne t assets. Thus, the managers compensation increases only as the funds net assets grow. The second control mechanism, the market for corporate control, is such an important issue that the Journal of Financial Economics published a special issue on the topic in 1983. Jensen and Ruback (1984) is a survey paper of these papers. They view the 17 See Murphy (1999) and Core, Guay and Lacker (2003) for surveys on optimal contracting models and agency issues. 59

PAGE 67

market for corporate control as a market in which alternative managerial teams compete for the rights to manage corporate resources. Thus, in theory, the market for corporate control influences both the managerial labor market and managerial behavior. However, Bebchuk, Coates and Subramanian (2002) find th at a hostile bidder must be prepared to pay a substantial premium in order to acqui re a target firm, providing the target management with a golden parachute and weakening the disciplinary force of the market for corporate control.18 Fund family manager turnover allows us to examine both the replacement-performance and market for corporate control mechanisms simultaneously. Past Performance Within the mutual fund industry, where pa y and performance ar e directly linked, the managerial labor market has played a ma jor role in enhancing shareholder wealth. Khorana (1996) examines the relation between the replacement of mu tual fund managers and their prior performance. He finds an inverse relation betw een the probability of managerial replacement and fund performance, us ing the growth rate in the funds assets and portfolio returns. Similarly, Ding and Wermers (2005) document a positive crosssectional relation between performance and replacement. Khorana (2001) goes on to examine the impact of mutual fund ma nager replacement on subsequent fund performance. He documents significant im provements in post-replacement performance relative to the past performance of the fund, suggesting that the market for corporate control benefits shareholder wealth. Hence, I hypothesize that past performance will have an inverse relation with fund family mana ger turnover. However, after accounting for fund manager characteristics and fund family responsibilities past performance will have a decreased effect than previously reported in Khorana (2001). 18 Bebchuk, Coates and Subramanian (2002) find that during the second half of the 1990s, the average premium in hostile acquisitions was 40 percent. 60

PAGE 68

Management Structure Mutual fund sponsors and mutual fund investors have di fferent goals and preference for their fund managers. The fund sponsors require managers earn high management fees while maintaining low cost s for the fund(s) they manage. To achieve these desired goals, fund sponsor increase total profits by maintaining the level of fund performance or inflows and decreasing the individual cost of operating each fund. In addition, fund sponsors can optimize manage ment fees for a given level of fund performance. On the other hand, mutual fund i nvestors prefer managers to obtain superior fund returns and charge minimum fund expe nses. Investors ra ther fund managers maximize fund returns by focusing on a single funds performance and efficiently pursuing these maximized funds returns. The difference in goals and preferences results in conflict of interests betw een sponsors and investors and can manifest in the fund management structure. At some fund families, a portfolio manager works autonomously managing only one fund. At other fund families, an individu al portfolio manager is responsible for two or more mutual funds within the same sector, related se ctors or with complementary investment objectives. For example, in 2001 Fidelity Funds manager John Carlson managed the Fidelity Emerging Market fund, Fi delity Strategic Income fund and Fidelity International Bond fund simultaneously. Similarly, Charles Melhouse managed the Fortis Capital fund, Fortis Fiduciar y fund and Fortis Growth and Income fund in 2002 for Fortis Funds Inc. However, both Fidelity Inc. and Fortis Funds Inc. implement the UFM structure as well. In 2002, Stephen Poling si ngularly managed Fortis Growth Fund and Jason Weiner solely managed Fidelity Contrafund. I contend that manager turnover policy is affected by the management structur e. That is, fund sponsors are less likely to end the services of a manager that manages a si ngle fund, because it is more costly to the sponsor, then this is a clear i ndication of a conflict of interests because for the same level of underperformance investors would benef it more if the pay for performance relationship (proposed by Khorana (1996)) work ed effectively for the costlier funds/fund 61

PAGE 69

management system. Fund sponsors are less li kely to fire a UFM than a MFM with similar underperform because the cost of searching and employing a unitary fund manager and the threat of the investors wi thdrawing funds is greater. The increased replacement costs with a decrease in fund management fee result in lower fund sponsors profits. Thus, sponsors may be slower to repl ace UFM than MFM giving rise to a conflict of interests between investors that expect the superior fund performance and management decisions regardless of the management structure. Shleifer and Vishney (1989) argue that managers engage in diversification acquisitions to make themselves indispensable to the firm. They note that when the acquired assets or subsidiary ceases to pr ovide further entrenchment benefits, the manager initiates divestures. Empirical evidence has shown that divested divisions do better as stand-alone entities than as part of a larger conglomerate (Myerson (1982), Harris, Kriebel, and Raviv (1982) and Hubbard and Pahlia (1999)). However, John and Ofek (1995) find that the typical divested divi sion is performing as well as the industry at the time of the divesture, suggesting that th e divested managers are benefiting from the good fortune of the industry and not their management ability. Similarly, Massa (2003) shows that the degree of product differentia tion negatively affects fund performance and positively affects fund proliferation. Furthe r, Nanda, Wang, and Zheng (2004) find that families that are more concentrated perfor m better. After accounting for the management structure given a certain le vel of underperformance, I hypot hesize that unitary fund managers (multiple fund managers) are less (more) likely to experience manager turnover. Expense Ratio and Management Fee The expense ratio is an important metr ic when comparing funds, because money paid for expenses is money that is not invested and earns no profit. Khorana (2001) suggests that in a competitive market, expense ratios should decline over time where investors become more pricesensitive, investment management firms increase in size and improve their economies of scale, and new entrants commence operations. Santini 62

PAGE 70

and Aber (1998), shows exactly the opposite is true: As fund size increases, fund expenses tend to rise rather than fall They find that fund complexes less likely to compete on expenses because people don't seem to care. In addition, high expense ratios are not proportional to better management. We rmers (2000) finds that high-expense funds underperform index funds, which are minima lly managed and have very low expense ratios. However, fund managers still earn a management fee regardless of the funds overall performance. Management fee is th e largest component of expense ratio. The management fee is the portion of the expense ratio that the fund mana ger receives for his/ her advising and stock selections. As stated earlier, the fund sponsors primary goal is for fund managers to earn high management fees while maintaining low costs for the fund(s) they manage. Thus, suppose a fund sponsor is faced with two similarly underperforming funds, but which provide different levels of management fees per dollar of managed assets. Given that the sponsor is compensate d on the basis of assets under management and not on performance, it seems unlikely that the sponsor would more quickly replace the high-fee fund. This is es pecially the case given that manager replacement may lead to redemption. This would be a cl ear conflict of interests because it is in investors interest if more expeditious action is taken against the underperforming high-fee fund. Hence, I hypothesize that the probability of replacement is lower with high expense/ management fee funds for any given level of performance. I also hypothesize that the probability of replacement is lower for high expens e, UFM funds for any given level of underperformance. This is because replacing a high-fee fund manager could lead to greater redemptions and a loss of fee income while having to incur greater costs of replacing a manager who manages a single fund. Fund Size The sensitivity of investor inflows to fund performance is well documented (Ippolito, 1992; Sirri and Tufano, 1992; and Chevalier and Ellison, 1997). Similarly, Gruber (1996) finds evidence th at sophisticated i nvestors are able to recognize superior 63

PAGE 71

management, witnessed by the fact that the fl ow of new money into and out of mutual funds follows the predictors of future performance. Fund families recognize the importance and the benefits of having popul ar, well-performing f unds. Analyzing the determinants of mutual fund starts, Khorana and Servaes (1999) identify several factors that induce fund families to set up new funds, such as economies of scale and scope, the overall level of funds invested, and the family s prior performance. Fund families market not only the superior performance of their ma nagers but also their funds in general to increase investor inflows and thus increas e total net assets managed and management fees. Elton, Gruber and Busse (2004) find that investors buy funds with higher marketing costs than the best-performing funds. Fund fam ilies also market the performance of their star funds to increase fund family inflows. Massa (1998) shows a positive spillover to other family funds from having a star fund. Nanda, Wang, and Zheng (2004) also finds a positive spillover effect on the inflows of other family funds resulting from having a star performing fund without the negative effect from a poor performing fund. Guedj and Papastaikoudi (2004) reports that this star performance is more prevalent for large fund families than their smaller peers. Thus, larg er fund families receive benefits from having star managers and funds due to the spillover into other family funds. Since managers are evaluated on past-performance and assets un der management, it stands to reason that the fund size, in total net assets, and funds age will have an inverse relation with the probability of manager turnover.19 Manager Tenure and Reputation When an investor buys a managed equity mutual fund, she is buying a managers expertise in picking stocks. When investor s evaluate funds most investors track the historical (typically the previ ous 3, 5 or 10 year) performan ce of the fund rather than the performance of the manager in place at the time of superior performance. It is important 19 Note that improved performance and increase in age, while related are not synonymous. For instance, if the fund attracts new investors afte r a bout of heavy advertising, it could experience an increase in its size while experiencing lower returns (net) if the adver tisement is paid for form (increased) 12b-1 fees. 64

PAGE 72

for investors to look for an investment manager who has not only supervised the fund for substantial length of time, but also who has be en in charge of the fund when it produced its best results. The longevity of a manage r shows that the manager can produce in both the bull and bear markets and that the mana ger is not the recipi ent of luck-based returnsreturns associated with profit incr eases that are entirely generated by external factors (such as changes in oil prices) rather than by the managers expertise. A manager who follows a consistent trad ing strategy and who delivers consistent returns over a relatively long period of time benefits investors by decreasing the volat ility of investors returns (Busse 1999). Thus the managers tenure leads to a reputation effect that the fund family can benefit from. Diamond (1989) states that reputation is impor tant when there is a diverse pool of observationa lly equivalent firms. Rosson and Brooks (2004) states that reputation can be seen as th e collective judgment of outsiders about an organizations actions and achievements. Fombrun (1996) pos its that when positive, this reputational capital is viewed as an asset that beco mes a competitive advantage to the company. Hence, I hypothesize that the ma nagers experience and tenure is inversely related to the likelihood of manager replacement within th e fund family. Additi onally, I hypothesize that for any level of recent poor performance, longer tenure reduces the probability of replacement. Style Drift/ Tracking Error A portfolio managers selecti on of securities should be c onsistent with the mutual funds investment objective, which is stated in the funds prospectus. A mutual funds (stated) investment objective is established when the fund is created and can be changed only with a majority vote of the funds shar eholders. However, Busse (2001) reports that managers increase risk levels or style drif t to increase return performance following a period of poor performance. Thus, an increase in style drift provide s some indication of manager incompetence. However, Brown and Ha rlow (2006) find that funds with greater style drift performs be tter than their peers during recessions or in down markets. Hence, I 65

PAGE 73

hypothesize that the probability of a manager being repl aced increases with the managers increase in style drift. Fund Styles/ Competency To select the most suitable mutual fund an investor must be able to differentiate clearly amongst the numerous investment objectives that fund families offer and understand the basic strategies by which the fund manager seeks to achieve the stated objective. Each investment ob jective requires th e fund manager have specific knowledge, expertise and level of competency. This ma y require a manager, who manages multiple funds, to have experience in a variety of fund styles. There are several investment objectives, each targeted to an investor w ith a specific risk tolerance and time horizon. For example, the growth objective can be di vided into aggressive growth, established growth, growth and income, large-cap grow th, micro-cap growth, mid-cap growth, and small-cap/small company growth funds. Funds w ith assets of different characteristics require different management skills (Deli (2002)). For instance, stock funds require greater competence than bond funds. Due to th e variety of fund styles, I hypothesize that the number of objectives managed is inversel y related to the like lihood of a manager being replaced. That is managers who are offered multiple objectives to manage have greater skill and become entrenched ma king them less likely to be fired. Data and Descriptive Statistics Sample Selection Procedure I examine the returns and ch aracteristics of replaced fund managers over the 1997 to 2001 period. This database is construc ted from two sources. First, I obtain the information on the month and year in whic h the current manager commenced overseeing the operations of the fund and thus the month and year in which the previous manager was replaced from the Morningsta r Principia database. I am also able to track the number of funds and objectives each fund manager operates by the manager characteristics 66

PAGE 74

provided in the Morningstar database. In addition, I receive th e annual fund style, turnover ratio, expense ratio, fund size (in to tal net assets-TNA), capital gains overhang, fund age (in years), 12b-1 fees, fund family affiliation and fund returns20 from the Morningstar database. Second, using fund na mes, fund family affiliation and other fund information, I supplement the Morningstar data base with the Center for Research and Securities Prices (CRSP) database which provides monthly returns and investment objectives. I utilize this information to calcula te the twelve month tracking error and style drift variables (Ammann and Zimmermann (2001) and Brown and Harlow (2006). To compute the tracking error, I fo llow Ammann and Zimmermann (2001), and use the square root of the non-central sec ond moment of deviation according to the following equation, ))1/() (((2 1 nRR TEtbench n t ti i (6) where Ri,t denotes the return of the tracking fund in time t, Rbench,t the return of the predetermined benchmark portfolio in period t, and n is the sample size. To calculate the tracking error and style dr ift variables, I first classify each fund according to the Morningstar investment style grid. I then selected a benchmark for each fund based on the above classi fication. Following Brown and Harlow (2006), I selected the Russell group of style benchmarks, whic h are available on line from the Frank Russell Company. I regress each funds retu rns over the last 12 months before the replacement of the fund manager on the benchmark returns and take 1-R2 as the measure of style drift.21 20 Morningstar and the Center for Research and Securitie s Prices (CRSP) list fund returns net of expenses and taxes. 21 There are several broadly similar approaches to estimating style drift. Brown and Harlow (2006) use the standard deviation of differences in returns relative to a benchmark that reflects the investment style of the fund and 1R2 from a regression of the fund returns on the benchmark. Chan et al. (2002) take the absolute difference in the factor loadings from a regression of a funds returns on the Fama-French factors over consecutive sub-periods. Amman and Zimmerman (2001) take the standard deviation of the residuals from 67

PAGE 75

Consistent with Gruber (1996), I define fund net flow as the growth in the fund assets net of growth in existing assets22: 1, 1, ,/])1([% ti titi tiTNA TNAR TNA wth NetFundGro (7) where TNAi,t denotes fund i s total net assets at the end of month t and Ri,t is return of fund i over month t To get the final manager replacement sample, I first exclude funds without manager tenure, turnover ratio, expense ratio, fund total net asset, capital gains overhang, fund age, 12b-1 fees, fund family affiliati on and fund returns data. Second, for calculations, I include the weighted average of all classes of fund shares in the final sample. Third, I exclude funds that list multip le fund managers for an individual fund (team managed funds). I also exclude funds having fewer than two years of monthly returns. After this sample selection procedure, I am left with 891 fund manager replacements in the final sample, wh ich consist of 188 unitary managed fund replacements and 703 replaced funds with the multiple fund manager structure23. The control sample is drawn from a ll funds from 1997 to 2001 that did not undergo a managerial change. To be include d in the control sample, Morningstar or CRSP must report the funds turnover ratio, ex pense ratio, fund tota l net asset, capital gains overhang, fund age, 12b-1 fees, manager tenure, fund family affiliation and fund returns data. All funds that are team managed are excluded from the control sample. In this study, the control sample is utilized as reference funds to cal culate the cumulative abnormal returns (CARs), risk-adjusted retu rns (RARs), and objective-adjusted returns (OARs) and for the logistic regressions. The reference funds are matched by both the a regression of the funds returns on the returns of its benchmarks. Brown and Harlow (2006) find that the results are not sensitive to the approach taken. 22 The Sirri and Tufano (1998) measure for asset flows wa s also utilized with similar results. This is defined as (TNA,t TNAi,t-1) x (1+Ri,t-1)/TNAi,t-1 where TNAi,t is the total net asset for fund i at time t ; and Ri,t-1 the raw return at time t-1 23 Both the Morningstar and the CRSP databases cover dead funds as well as active funds, therefore, survivorship bias is not a concern for this study. 68

PAGE 76

69 stated CRSP objective and Morningstar inve stment style. Table 1 provides annual summary statistics for the control sample. Th e control sample consists of 8477 funds that did not have a managerial replacement dur ing the sample period. Of the 8477 control sample funds, 1866 have unitary fund managers and the remaining 6611 control sample fund managers operate multiple funds simu ltaneously. Panel A through H of Table 2 reports the results of the univariate fund-specific characteristics for the control sample. Description of full sample Table 1 summarizes the frequency with which fund managers are replaced in a year for the sample period 1997 to 2001. For each sample year, I report the total number of fund manager replacements as well as the cumulative number of replacement for the sample period. The largest number of manager replacements occurred in the final year of the sample period, 2001, with 198 replacements and 2000 had the least number of replacements, 140.

PAGE 77

Panel A: Replacement Sample Cumulative Cumulative Unitary Fund ManagerMulti-Fund Manager YearFrequencyPercentFrequencyPercentageReplacement Replacement 199717719.8653177 19.865 30 147 199819621.9978373 41.863 42 154 199918020.202553 62.065 44 136 200014015.7127693 77.778 31 109 200119822.2222891 100 41 157 Total891100 188 703 Panel B: Control Sample Total ReplacementUnitary ReplacementMulti-fund Replacement YearTotalUnitary Multifund Percentage Percentage Percentage 19971420367 1053 12.465 8.174 13.960 19981719390 1329 11.402 10.769 11.588 19991686372 1314 10.676 11.828 10.350 20001792374 1418 7.813 8.289 7.687 20011860363 1497 10.645 11.295 10.488 Total847718666611 53.001 50.355 54.072 Control SampleThistablesummarizesthemanagerreplacementandcontrolsamples,whichwerecreatedbymatchingtheMorningstarfundmanager databasewiththeCenterforResearchandSecuritiesPricesdatabase.Managerreplacementsampleconsistofdistributioninformationo f 891mutualfundswithstart-updatesbetweenJanuary1997toDecember2001.Amanagementchangeisdefinedasanychangeinthefund's portfoliomanager.Managerialreplacementsarepresentedbyreplacementyear.Eachreplacementisfurtherdividedintomanagement structureaccordingtothenumberoffundssimultaneouslymanaged.TheUnitaryFundManageroperatesasinglefundwhilethemulti-fund manageroperatesmultiplefundssimultaneously.Thecontrolsampleconsistsof8477fundsthatdonotundergoamanagerialreplacement forthegivenperiod.Eachcontrolsamplefundisfurtherdividedintomanagementstructureaccordingtothenumberoffunds simultaneously managed.Table 11: Managerial Replacement Distribution 70

PAGE 78

I decompose the sample of 891 fund ma nager replacements based on the number of funds managed simultaneously over the sa mple period. Fund managers that operate one fund are placed in the uni tary fund management sample (UFM) and those managers that operate multiple funds simultaneously ar e placed in the multiple funds management sample (MFM). This sample decomposition yields 703 fund managers in the multiple funds management sample and 188 funds a nd fund managers in the unitary fund management sample. Table 2, Panel A through H, summarizes statistics for variables used in the analysis for each sample year as well as over the entire sample period. For each sample year, I report the total number of funds as well as the average size (measured by total net assets), net fund growth, manager tenure, fund age, expense ratio, turnover ratio, 12b-1 fees, and capital gains overha ng. I compare each of the 891 replacement sample with an objective and style matched samp le of mutual funds that did not have any managerial turnover (control sample). 71

PAGE 79

N199719981999200020011997-2001 Panel A Total Replacement Sample 891314.0250399.5877400.6508524.5375515.2213430.8045 Fund Size UFM Replacement Sample188274.2100391.8349349.6533469.3833476.7853392.3734 MFM Replacement Sample 703324.6725401.6610414.2888539.2872525.5000441.0819 Total Control Sample 8477504.3041558.4486618.1343641.5718626.5382589.7994 UFM Control Sample 1866917.46161210.58381253.04191435.74921292.11431221.7902 MFM Control Sample 6611387.6875374.3789438.9273417.4097438.6748411.4157 Panel B Total Replacement Sample 8917.556110.10228.570112.00279.53949.5541 Capital Gains OverhangUFM Replacement Sample18813.295016.70009.071415.892014.286013.8489 MFM Replacement Sample 7036.02148.33778.436010.96268.27008.4056 Total Control Sample 84777.563811.16549.14309.920510.75069.7086 UFM Control Sample 186612.591815.437012.923113.778115.347014.0154 MFM Control Sample 66116.14469.95978.07618.83179.45328.4930 Thistablesummarizesvariousfundcharacteristicsforthe5sampleyearsandforthewholesampleperiod(1997to2001).Foreachsampleyear,Ireport thetotalnumberof fundsaswellasthesummarystatisticsforthemanagerialreplacementandcontrolsamples.Statisticsforthewholesampleperiodareaveragesovera llfund-years.UFM representsfundsthathavemanagersthatoperateasoleunitarymanagerwhileMFMrepresentsthefundsthataremanagedbymanagersthatoperatemulti plefunds simultaneously.FundSizeisthetotalfundnetassetsinbillions.CapitalGainsOverhangisthenetunrealizedappreciation(ordepreciation)duringtheperiodreportedbythe Morningstardatabase.NetFundGrowthisthechangeinthefundassetsnetofgrowthinexistingassets.Managertenureisthenumberofyearsaportfoli oasoverseena particularfund.Fundageisthenumberofyearsthefundasbeeninoperation.TurnoverRatioistotalpurchasesandsalesdividedbyfund'saveragenet assetvalue.Expense Ratio is the mutual fund's total annual operating expenses (including operational fees, distribution fees, and other expenses) stated as a percentage of the fund's average net a s ManagementFeeisthefeethefundcomplexreceivesfo r managingshareholdersassets,expressedinbillions.Managementfeeiscalculatedasthetotalassetsmanagedper fundtimesthefund'sexpenseratio.Thefinalcolumnsummarizesthemeandifferencesbetweenthemanagementstructures;unitaryfundmanagementan dmultiplefund management. ***, ** and denote significance at the 1%, 5% and 10% levels respectively. Table 12: Descriptive Statistics -48.7085** 810.3745*** 5.4433* 5.5224* 1997-2001 Year(s) Management Structure Difference (UFM-MFM) 72

PAGE 80

Panel C Total Replacement Sample 891-13.762425.354229.663227.7719-15.021810.8010 Net Fund Growth UFM Replacement Sample188-9.143644.799731.609444.8689-11.612920.1043 MFM Replacement Sample 703-14.997620.154029.142723.1998-15.93348.3131 Total Control Sample 8477-1.296934.496227.292931.7334-12.438615.9574 UFM Control Sample 1866-2.487343.731030.356542.3107-8.491521.0839 MFM Control Sample 6611-0.960931.889626.428228.7479-13.552614.5104 Panel D Total Replacement Sample 8913.71005.78984.15105.30876.11415.0147 Manager Tenure UFM Replacement Sample1882.88894.64293.36364.10714.62183.9249 MFM Replacement Sample 7033.92966.09664.36155.63006.51325.3062 Total Control Sample 84774.19914.18504.55694.91054.91924.5542 UFM Control Sample 18662.95313.63834.17485.34905.28614.2803 MFM Control Sample 66114.55084.33934.66484.78684.81564.6315 Panel E Total Replacement Sample 8918.95219.790910.129710.28829.13519.6592 Fund Age UFM Replacement Sample1887.82788.97489.52419.21948.76258.8617 MFM Replacement Sample 7039.252810.009210.291710.57409.23479.8725 Total Control Sample 84779.74109.99899.83529.45869.67169.7411 UFM Control Sample 18669.57979.21139.67208.93289.06449.2921 MFM Control Sample 66119.786610.22129.88139.60709.84309.8678 -1.0108 Table 12: Descriptive Statistics (Continued) 11.7912* 6.5735* -1.3813* -0.3512 -0.5757 ** 73

PAGE 81

Panel F Total Replacement Sample 891107.304387.0276100.013692.185997.749496.8561 Turnover Ratio UFM Replacement Sample188110.241498.6842112.232698.7778103.5828104.7037 MFM Replacement Sample 703106.518883.910396.745990.423196.189394.7575 Total Control Sample 847790.867392.571791.533393.234497.547193.1508 UFM Control Sample 186697.153499.139492.7976105.0851108.6239100.5599 MFM Control Sample 661189.093190.717991.176589.889594.420691.0595 Panel G Total Replacement Sample 8911.13281.14261.14141.10310.97271.0985 Expense Ratio UFM Replacement Sample1881.20471.27801.17161.12161.07501.1702 MFM Replacement Sample 7031.11361.10641.13331.09810.94531.0794 Total Control Sample 84771.13261.14311.17171.18541.18581.1637 UFM Control Sample 18661.27601.22941.28491.30881.27231.2743 MFM Control Sample 66111.09211.11881.13971.15061.16131.1325 Panel H Total Replacement Sample 8910.15950.17720.19030.13130.16210.1641 12b-1 fees UFM Replacement Sample1880.07330.08840.09110.06170.07280.0775 MFM Replacement Sample 7030.18260.20100.21680.14990.18600.1872 Total Control Sample 84770.15900.17430.18390.14450.14270.1609 UFM Control Sample 18660.09260.10730.11300.07020.07090.0908 MFM Control Sample 66110.17770.19320.20390.16540.16300.1806 Panel I Total Replacement Sample 8913.71714.22634.33184.57534.41874.2538 Management FeesUFM Replacement Sample1884.1837125.0228645.1612295.3873285.2036624.9918 MFM Replacement Sample 7033.5923414.0132844.1100324.3581124.2087734.0565 Total Control Sample 84775.84386.36777.46437.89607.57967.0303 UFM Control Sample 186611.67248613.97103616.07921818.42399116.27686415.2847 MFM Control Sample 66114.1985684.2215855.0326974.9244025.1247434.7004 9.9462 9.5004 0.0908 Table 12: Descriptive Statistics (Continued) 10.584320*** 0.1418 -0.1098 -0.0898 0.935251* 74

PAGE 82

Several notable features emerge from the descriptive statistics in Table 2. For instance, the size of the average MFM replacemen t fund is consistently larger than that of the average UFM replacement f und and both sets are smaller than the average size of the control sample (Panel A, Table 2). This suggests that multiple fund managers are managing similar amounts of assets as the uni tary fund manager just spread across more funds. The capital gains overhang is constantly larger for th e unitary management sample than for the multiple management sample, i ndicating that, on average, UFM funds might have done better for existing investors in term s of capital gains. On the other hand, it suggests that multiple management structur e may be preferred by (potential) new investors who desire to avoid the tax liability of previously built-up capital gains. Not surprisingly, the turnover rati o is larger for the unitary management structure who may sell more frequently to get rid of the cap ital gains overhang. As expected, the fund growth of the replacement sample, regard less of the management structure, is consistently lower than that of the industry average. This finding is consistent with the previous literature on the relationship between performance and manager replacement (see, e.g. Khorana (1996)). It is also interesting to note th at of the replaced sample, MFM had lower fund growth in each year and over the full sample than UFM funds. It appears that the multiple fund management structure benefits from economies of scale resulting in a lower expense ratio (by about 10 basis point s) than the unitary management structure. In addition, the 12b-1 fees are lower for the multiple management structure, suggesting a cost benefit to multip le fund management. One implication of this is that, if I find that fund sponsors are less likely to end the servi ces of a manager that manages a single fund, possibly because it is more costly to the sponsor, then this is a clear indication of a conflict of interests b ecause for the same level of underperformance investors would benefit more if the internal control mechanisms (pay for performance) worked effectively for the costlie r funds/fund management system. The average managerial tenure for the un itary management replacements for the entire sample period is slightly shorter at 3. 92 years than for the managerial tenure of the 75

PAGE 83

multiple fund management structure, 5.31 year s. This finding highlights the importance of managerial experience to operate multip le funds simultaneously. Furthermore, the average fund age across all funds for the sample with multiple fund management structure is 9.87 years, which is statistica lly significantly older than for the unitary management structure, 8.86 years. Thus the MFM sample has more managerial experience than the UFM sample and operate older funds (Panel D and E). Table 3 provides the descriptive statistics on both the multiple fund management and the multiple objective sub-samples. Pa nel A of Table 3 shows the mean (median) number of funds operated simultaneously by a manager that was replaced, where replacement means the manager was relieved of his responsibilities fo r at least one fund. The average number of funds operated simu ltaneously for the replacement sample (4.903 funds) is slightly larger than the control sample (4.305 funds). However, the number of objectives managed simultaneously by the replacement sample (2.039 objectives) is smaller than the control sample (4.301 objecti ves), as reported in Panels C and D of Table 3. Taken together, these statistics indi cate that not only are individual managers being asked to manage multiple funds simulta neously, but they are also being asked to manage funds with different objectives. De pending on how different these objectives are this practice could dampen their performance. 76

PAGE 84

Panel A. Multiple Fund Management (MFM) replacement Sample Year(s)NMeanMedianStd DevMinimumMaximum 19971474.267 4 1.457 26 19981544.953 5 1.621 27 19991365.149 5 1.868 27 20001094.873 5 1.492 26 20011575.259 5 2.016 26 All Years7034.903 5 1.601 27 Panel B. Multiple Fund Management (MFM) Control Sample Year(s)NMeanMedianStd DevMinimumMaximum 199710533.951 41.01925 199813294.012 41.78326 199913144.639 51.89427 200014184.863 51.95426 200114973.994 41.02625 All Years66114.305 51.82127 Panel C. Multiple Objective Management (MOM) replacement Sample Year(s)NMeanMedianStd DevMinimumMaximum 1997962.037 20.88726 1998892.012 20.824525 1999912.024 20.67724 2000982.005 20.661923 20011042.110 20.875326 All Years4782.039 20.783926 Panel D. Multiple Objective Management (MOM) Control Sample Year(s)NMeanMedianStd DevMinimumMaximum 19979953.951 31.03425 199811434.012 31.08726 199910964.639 41.10227 200011874.863 41.13626 200111653.994 31.04725 A ll Years55864.301 31.10027 Table 13: Multiple Funds and Multiple Objectives preliminary statistics Thistablereportsthemeanandmediannumberoffundsandobjectivesofasampleofmutual fundsexperiencingmanagerialturnoverandthecontrolsamplebetween1999and2001. MultipleFundManagementstructure(MFM)representsfundsthathavemanagersthatoperate multiplefundssimultaneously.MultipleObjectivesManagementstructure(MOM)represents fundsthathavemanagersthatoperatemultipleobjectivessimultaneously.Thestandard deviationaswellastheminimumandmaximumnumberoffunds/objectivesoperatedarealso presented for each sample year. 77

PAGE 85

I also decompose the replacement sample by objective and style. Table 4 displays the distribution of 891 fund manager replacements from 891 funds across fund objectives and styles over the sample period. The equity funds belong to one of nine Morningstar equity style categories, which group funds on the basis of the ma rket capitalization and growth potential of their portfolios24. As expected, the majority of the replacements involves equity objectives/sty le funds. As noted in Brown and Goetzmann (1997), the dispersion in styles among the funds from th e same objective category is quite high, which is consistent with the existing evidence (Grinblatt and Titman (1989,1993), Grinblatt, Titman and Wermers (1995), Dani el, Grinblatt, Titman and Wermers (1997), and Wermers (2000)) on misclass ification of funds in the objective categories. For instance, the aggressive growth, the long-term growth and international equity funds have at least one fund in each of the nine Morningstar equity-style categories. Similar levels of dispersion across styles are also observed in the 323 bond fund replacements sample. The high quality bond objective has the most dispersion with a fund in eight of the nine fixedincome style categories. Only in the Single State Municipa l Bond objective is there 70% of the funds concentrated in two style categories. 24 See the appendix in Goriaev (200 3) The relative impact of different classification schemes on mutual fund flows for the definition of the Morningstar styles. 78

PAGE 86

Panel A: Managerial Replacement Equity Funds Objective/ Style Large ValueLarge BlendLarge GrowthMedium ValueMedium BlendMedium GrowthSmall ValueSmall BlendSmall GrowthTotal Aggressive Growth (Ag) 14 3 7 3 7 13 10 7 1983 Balance (BL) 8 8 4 2 4 6 0 11 043 Global Equity (GE) 2 6 6 0 2 5 2 4 027 Growth and Income (GI) 21 17 3 2 2 0 0 0 045 International Equity (IE) 17 14 16 14 7 9 6 5 795 Income (IN) 9 3 4 5 0 0 0 0 021 Long Term Growth (LG) 14 24 14 13 12 9 8 6 8108 Precious Metals (PM) 0 0 0 0 5 0 4 3 618 Sector Fund (SF) 13 17 16 23 9 8 5 0 091 Total Return (TR) 4 4 3 5 7 0 3 0 026 Utility Fund (UT) 8 0 0 3 0 0 0 0 011 Total 110 96 73 70 55 50 38 36 40568 Panel B: Managerial Replacement Bond Funds Objective/ Style High-ShortHigh-IntermediateHigh-LongMedium-Shor tMedium-IntermediateMedium-LongLow-ShortLow-IntermediateLow-Long Total High Quality Bond (BQ) 18 10 5 7 8 12 0 6 066 High Yield Bond (BY) 0 0 0 0 0 4 5 7 016 Global Bond (GB) 0 9 0 0 6 0 0 4 019 Ginnie Mae Bond (GM) 6 7 5 0 0 0 0 0 018 Government Security Bond (GS)13 16 6 7 0 0 0 0 042 High Quality Municipal Bond (MQ)9 10 17 2 4 3 0 0 045 Single State Municipal Bond (MS)0 39 48 0 6 11 5 0 0109 High Yield Municipal Bond (MY) 0 0 0 0 6 2 0 0 08 Total 46 91 81 16 30 32 10 17 0323 Thistablereportsthenumberoffundmanagerreplacementobservationswitha givenstatedCenterforResearchandSecuritiesPrices(CRSP)objectiv eandMorningstarinvestmentstyleoverthe periodJanuary1997toDecember2001.TheCenterforResearchandSecuritiesP ricesdatabasereportselevenequityandeightbondobjectivecategori es.TheMorningstardatabasereportsnine equitystyleandninefixedincomestylecategories.PanelAreportsthedistributionforthemanagerialturnoverequityfundsamplewhilePanelBrep ortsthedistributionforthemanagerialturnover bond fund sample. There are 891 managerial replacements including 568 equity fund replacements and 323 bond fund replacements. Table 14: Managerial Replacement Sample Distribution 79

PAGE 87

Methodology I measure abnormal returns for a replacement event-fund as the difference in returns between the replacement event-fund and the equal-weighted fund style category to which the fund belongs. For example, th e style category-adjust ed return for fund i during month t is: ]1)1([]1)1([, to tiR R RAR (8) where Ri,t is the return for fund i in month t and Ro,t is the equal-weighted return of all funds in fund i s category in month t The average category-adjusted return during month t is calculated as ti tRAR N RAR,1 (9) where N equals the number of funds that experience a manager replacement event. Finally, the cumulative category-adjusted return over k event months is simply the sum of RAR t t kttRAR CRAR, (10) As demonstrated in Table 4, funds w ithin the same category have different investment objectives and e xposed to different risk f actors. Thus, I construct a performance measure that uses the equal-weig hted average of all funds with the same investment objective as the benchmark, OAR. The use of the objective-adjusted performance measure is consistent with the argument put forth by Morck, Shleifer, and Vishny (1989) that firms make their managerial replacement decisions based on the industry benchmarks. The advantag e of this benchmark is that it better controls for risk than the broader style category-based benchmark. However, both calculations measure fund performance relative to othe r managers in the peer group. Estimating the managerial-turnover relationshi p, I control for the determinants of replacement previously identified in the litera ture, such as past performance, size, age, fees, fund flows, and manager tenure (see, e.g., Khorana, 2001, Chevalier and Ellison, 1997, Sirri and Tufano, 1998, and Nanda, Wa ng, and Zheng, 2000). As in Khorana 80

PAGE 88

(1996), I use the objective and category-adjusted returns as separate performance measures in the following regression: itiOAR placement P 1 0) (Re (11) itiRAR placement P 1 0) (Re where OAR and RAR are the objectiveand category-adjusted fund returns, respectively. Interaction terms are also included to examine the relations hip between abnormal returns and the probability of fund mana gers being replaced, when I account for management structure, manager tenure and total management fees. The interaction variables examine the relati onship between replacement a nd management structure, manager tenure and management fees for a given level of underperformance. With these three interaction terms, I am able to fu rther explore how well the internal governance mechanisms work for fund managers. As in Ai and Norton (2003), I include the following marginal interaction term effect estimation to understa nd the economic impact of the interaction terms. ) (*)](1)[( arg e Performanc mance iUFMPerfor dUFMxFxF inalEffect M (12) where F(x) is the average implied probability of management replacement computed for each observation using the logit coefficien ts. Fund Performance is defined by the objective adjusted return (OAR) in year t and the risk adjusted return (RAR) in year t Unitary Fund Manager is a dichotomous variable that takes the value of one if a replaced manager operates a unitary fund and zero if that manager operates multiple funds simultaneously. 81

PAGE 89

82 Empirical Results Performance-Replacement Relationship The relation between fund manager tur nover and past performance has been established and well documented (see, e. g. Khorana, 1996, 2001, and Chevalier and Ellison, 1997). However, without considering the specific organizational form and, more specifically, the management structure of the fund family previous research might have significantly overstated the sensitivity of manage rial replacement to past performance. In this section, I further analy ze the relationship between manage rial turnover and fund past performance with respect to both the obj ective and style category by including the management structure of the fund family.

PAGE 90

-2 to -1-1 to 0to 00 to +0 to +1+1 to +2 -2 to + to +2 Panel A: Risk Abnormal ReturnFull Sample N 870 870 870870870870870 870 CAR -0.0321-0.0203-0.01580.02 750.03020.0265-0.01760.0294 t-statistic -0.2024-1.7834**-1.5395*1.2930*1.3950*1.3476*-1.6295*1.5018* Panel B: Risk Abnormal ReturnEquity Fund Sample N 552 552 552552552552552 552 CAR 0.0860-0.0116-0.02440.06 780.07370.0833-0.09170.1556 t-statistic 0.3390-1.6382*-2.2718**2. 2512**1.6627**2.3148**-1.6502**2.3349*** Panel C: Risk Abnormal ReturnBond Fund Sample N 318 318 318318318318318 318 CAR -0.0236-0.0210-0.00800.01 630.01050.0339-0.03160.0576 t-statistic -0.5513-0.7722-0.7918 0.68431.19240.9253-0.92490.6981 Panel D: Objective Abnormal ReturnFull Sample N 870 870 870870870870870 870 CAR -0.0477-0.0382-0.02530.02 350.02940.0294-0.02460.0274 t-statistic -0.8353-1.4833*-1.5550*1. 7206**1.9400**1.5921*-1.2870*1.3295* Panel E: Objective Abnormal ReturnEquity Fund Sample N 552 552 552552552552552 552 CAR -0.0273-0.0364-0.04010.02 210.02710.0409-0.03550.0282 t-statistic -1.0389-1.6182*-1.7161**1 .3446*1.8799**1.5014*-1.4140*1.5427* Panel F: Objective Abnormal ReturnBond Fund Sample N 318 318 318318318318318 318 CAR -0.0245-0.0079-0.02080.01 380.00020.0014-0.00170.0127 t-statistic -0.3315-0.1858-0.84672.3946*1.3856*1.4701*-1.11461.5664* Years with respect to Managerial Turnover Thistablepresentsthemeanperformanceofactivelymanagedfundsthatexperiencedmanagerialreplacementintheperiod1997to2001.Thepurposeof therisk-adjustedand objective-adjustedmatchedsampleapproachistocomparefundsthatexperiencedreplacementwiththosethatdidnotforthegivenperiod.Thetabler eportstheabnormalreturn valuesofthereplacedmanagersampleineachyearincludinga6monthwindowaroundthereplacementdate.Year0referstotheyearinwhichreplacement occurred.Thelast twocolumnsofthetablereporttheaveragedifferencesbetweenthepre-andpost-replacementcategory-adjustedreturnacrossfunds,usinga-2toyearand+to+2year event window, respectively. The symbols ***, **, and denot e significance at the 1%, 5% and 10% levels, respectively. Table 15: Performance in the years Preand PostManager Turnover: Full Sample 83

PAGE 91

I examine the preand postreplacement changes in objectiveand style-adjusted performance. As in Khorana (2001), the impact of manageri al turnover on fund performance is examined based on the change s in performance measures during four subperiods surrounding the event date: year -2 co rresponds to the second year or 13 to 24 months prior to replace ment year, year -1 corresponds to the first year or 1 to 12 months prior to replacement year, so on and so fort h. The overall results in Table 5, Panel A indicate a monotonic decrease, wh ich is statistically signifi cant different from zero, in fund performance for the repl acement sample in the prereplacement period, followed by a statistically significant increase in perf ormance in the post-replacement period. Based on the style category performance estimates managers exhibit significantly negative abnormal returns of 2.4 percent in the six months preceding managerial replacement. In Panel B, abnormal underperformance of funds with replaced managers is statistically significant for the equity fund replacement sa mple. This finding suggests that replaced managers perform significantly worse than those in the style category control group. 84

PAGE 92

-2 to -1-1 to 0to 00 to +0 to +1+1 to +2 -2 to + to +2 Panel A: Risk Abnormal ReturnFull Sample N 716 716 716 716 716 716 716 716 CAR 0.0824-0.0191-0.02590.03650.02410.0062-0.01240.0342 t-statistic 0.4940-1.5652*-1.8131**1.2829*1.6004*1.1262*-1.2981*1.4319* Panel B: Risk Abnormal ReturnEquity Fund Sample N 461 461 461 461 461 461 461 461 CAR 0.0989-0.0167-0.04560.06670.07890.0264-0.02410.0624 t-statistic 0.4750-1.3459*-2.4559***1.6893**1.3348*1.1973-1.7642*1.6523** Panel C: Risk Abnormal ReturnBond Fund Sample N 255 255 255 255 255 255 255 255 CAR 0.0359-0.0263-0.0091-0.0188-0.0350-0.0244-0.0225-0.0286 t-statistic 0.5660-1.0515-0.5512-0.4141-1.0774-0.7080-0.8778-0.8744 Panel D: Objective Abnormal ReturnFull Sample N 716 716 716 716 716 716 716 716 CAR -0.0393-0.0348-0.02950.02570.02680.0130-0.02960.0218 t-statistic -1.0573-1.3821*-1.6251*1.6450**1.6426*1.6376*-1.2899*1.7860** Panel E: Objective Abnormal ReturnEquity Fund Sample N 461 461 461 461 461 461 461 461 CAR -0.0455-0.0413-0.04640.02360.03640.0465-0.03640.0354 t-statistic -0.7957-1.6867**-1.4331*1.2885*1.9153**1.8543**-1.7325**1.7512** Panel F: Objective Abnormal ReturnBond Fund Sample N 255 255 255 255 255 255 255 255 CAR -0.0323-0.0103-0.02630.01840.01040.0185-0.03240.0154 t-statistic -0.3965-0.4325-0.98532.4432***1.4870*1.5673*-1.23281.7643** ThistablepresentsthemeanperformanceofMultipleFundManagement(MFM)samplethatexperiencedmanagerialreplacementintheperiod1997to2001 .Thepurposeofthe risk-adjustedandobjective-adjustedmatchedsampleapproachistocomparefundsthatexperiencedreplacementwiththosethatdidnotforthegiven period.Thetablereportsthe abnormalreturnvaluesofthereplacedmanagersampleineachyearincludinga6monthwindowaroundthereplacementdate.Year0referstotheyearinwh ichreplacement occurred.Thelasttwocolumnsofthetablereporttheaveragedifferencesbetweenthepre-andpost-replacementcategory-adjustedreturnacrossfu nds,usinga-2to-yearand + to +2 year event window, respectively. The symbols ***, **, and denote significance at the 1%, 5% and 10% levels, respecti vely. Years with respect to Managerial Turnover Table 16: Performance in the years Preand PostManager Turnover: Multiple Fund Management (MFM) Sample 85

PAGE 93

-2 to -1-1 to 0to 00 to +0 to +1+1 to +2 -2 to + to +2 Panel A: Risk Abnormal ReturnFull Sample N 154 154 154 154 154 154 154 154 CAR -0.0588-0.0354-0.02980.04490.03390.0299-0.01880.0311 t-statistic -0.8643-1.9474**-1.2846*1.3172*1.2873*1.2997*-1.7346**1.6239* Panel B: Risk Abnormal ReturnEquity Fund Sample N9 19 19 19 19 19 19 19 1 CAR 0.0954-0.0167-0.04440.07760.07860.1006-0.09570.1613 t-statistic 0.2783-1.7988**-2.0432**2.3855***1.8935**2.5255***-1.3468*2.4858*** Panel C: Risk Abnormal ReturnBond Fund Sample N6 36 36 36 36 36 36 36 3 CAR -0.0319-0.0286-0.01150.01940.01370.0359-0.04270.0710 t-statistic -1.1345-0.5325-1.00451.10351.7286**1.2149-1.02451.2286 Panel D: Objective Abnormal ReturnFull Sample N 154 154 154 154 154 154 154 154 CAR -0.0588-0.0529-0.02980.02940.03120.0443-0.02880.0295 t-statistic -0.6136-1.5466*-1.3658*1.9013**2.2654**1.4735*-1.28651*1.2853* Panel E: Objective Abnormal ReturnEquity Fund Sample N9 19 19 19 19 19 19 19 1 CAR -0.0358-0.0429-0.05780.03000.03090.0492-0.05650.0323 t-statistic -1.3455*-1.5863*-2.0424**1.5786*1.6935**1.4015*-1.3065*1.5133* Panel F: Objective Abnormal ReturnBond Fund Sample N6 36 36 36 36 36 36 36 3 CAR -0.0254-0.0139-0.03150.02370.01230.0235-0.01390.0286 t-statistic -0.2755-0.1468-0.74332.1843**1.3655*1.43458*-1.08581.5063* ThistablepresentsthemeanperformanceofUnitaryFundManagement(UFM)samplethatexperiencedmanagerialreplacementintheperiod1997to2001. Thepurposeofthe risk-adjustedandobjective-adjustedmatchedsampleapproachistocomparefundsthatexperiencedreplacementwiththosethatdidnotforthegiven period.Thetablereportsthe abnormalreturnvaluesofthereplacedmanagersampleineachyearincludinga6monthwindowaroundthereplacementdate.Year0referstotheyearinwh ichreplacement occurred.Thelasttwocolumnsofthetablereporttheaveragedifferencesbetweenthepre-andpost-replacementcategory-adjustedreturnacrossfu nds,usinga-2to-yearand + to +2 year event window, respectively. The symbols ***, **, and denote significance at the 1%, 5% and 10% levels, respecti vely. Years with respect to Managerial Turnover Table 17: Performance in the years Preand Pos tManager Turnover: Unitary Fund Management Sample 86

PAGE 94

There are similar patterns when we disaggr egate the sample of replaced managers. For the multiple fund management replacements sample Table 6, Panel A, reports that manager replacement is preceded by poor retu rns and that these returns improve during the period following the replacement. Specifi cally, during year -1, replacement event funds underperform their category averages by 1.9 percentage points. However, this underperformance turns into overperformance as early as six months following the managerial replacement. Table 6, Panel D, indicates that there is a 1.6 percentage points underperformance between the replacement event funds and the control funds for the Objective Abnormal Returns (OAR). As Table 7 indicates, I obtain si milar results for the changes in the objective-adjusted return for the unitary management equity fund sample. Finally, there is a statistically and econo mically significant change in performance between the [-2,-] and [+,+2] event windows that is robust across both performance measures for both the unitary and multiple management structures. The average increase in abnormal performance is 3.4%, based on the MFM style category model, and 3.1%, based on the objective-adjusted return UFM samp le. Thus, consistent with the findings in Khorana (2001), the event study statistics presented indicate a strong relationship between managerial turnover and past performa nce. In the next section, I implement a univariate regression model followed by a comprehensive multivariate model to further explore managerial turnover. It appears that MFMs have a s horter underperformance period before replacement. According to Khorana (1996), sponsors seem to tolerate underperformance of UFMs longer before acting. Panel A of Tables 6 and 7 indicate that MFMs have lower cumulative RARs in the -2,window (-0.0124 vs. -0.0188) and experience negative returns for only one year before replacement, whereas UFMs experience losses for two years before replacement. This tolerance is not in the interest of investors, but may benefit sponsors as they can defer the higher costs involved in replacing a UFM manager. This is even of more importance to investors because it appears that UFM funds have a greater sp eed of recovery after a replacement, as evidenced by the larger aver age returns in the [0, +] wi ndow 4.49 percentage points 87

PAGE 95

for UFMs vs. 3.65 for MFMs. Overall, the evidence is suggesting that since UFMs experience larger losses for longer periods befo re replacement but recover faster, if fund sponsors act in the interest of investors then we should observe that the probability of replacement is higher for UFMs for a given level of underperformance than for MFMs. However, this may not be the case because, as reported in Table 2, UFMs tend to have higher asset growth rates and higher fees than MFMs, suggesting that sponsors can benefit more from keeping them intact. Univariate Logistic Analysis I examine the managerial replacement decision using a univa riate regression model on the multiple management structures as well as the determinants of replacement previously identified in the literature: past pe rformance (Khorana, 1996, 2001), fees, fund size and fund age (Chevalier and Ellison, 1997), and manager tenure (Nanda, Wang, and Zheng, 2000). I perform a logistic regression on a dichotomous variable equal to one if the fund undergoes managerial turnover and zero if the incumbent ma nager continues to operate the fund. The logistic regressions control for clustering along two dimensions (fund complex and year), as described in Cameron, Gelbach and Miller (2006) and Petersen (2007). 88

PAGE 96

constant Explanatory Variables Psuedo R Observations Risk-Adjusted Return 0.0309 -0.0288 0.0259 9172 (0.0096) (<.0001)*** Objective-Adjusted Return 0.0204 -0.0521 0.0560 9172 (0.0007) (0.0022)*** Unitary Fund Manager 0.0482 -0.0125 0.0090 9368 (0.0062) (0.0070)*** Multiple Objectives Managed -0.2415 0.0658 0.0008 9368 (0.0496) (0.3799) Logisticregressionestimatesofmanagerialreplacementfor891managersare reportedoverthe1997to2001period.Managerreplacementisthedichotomous dependentvariableequaltooneforthereplacementsampleandzeroforthe controlsamplethathavenomanagerialturnover.Theobservationsareinfundyears.FundPerformanceisdefinedbytheobjectiveadjustedreturn(OAR)inyear t andtheriskadjustedreturn(RAR)inyear t .UnitaryFundManagerisa dichotomousvariablethattakesthevalueofoneifareplacedmanageroperatesa unitaryfundandzeroifthatmanageroperatesmultiplefundssimultaneously. MultipleObjectivesManagedisadichotomousvariablethattakesthevalueofone ifareplacedmanageroperatesmultipleobjectivessimultaneouslyandzeroifthat manageroperatesfund(s)withoneobjective.Trackingerrorisconstructedby takingthestandarddeviationoftheresidualsfromaregressionofthefundsreturn on the returns of its benchmark. To calculate Style Drift, I regress each fundsreturnsovertheyearpriortoreplacementonthebenchmarkreturnsand take1-R2asthemeasureofstyledrift.CapitalGainsOverhangisthenet unrealizedappreciation(ordepreciation)duringtheperiodreportedbythe Morningstardatabase.Managertenureisthenumberofyearsaportfoliomanager hasoverseenaparticularfund.FundSizeisthenaturallogoftotalfundnetassets. Fundageisthenaturallogofthefundsage.NetFundGrowthisthechangeinthe fundassetsnetofgrowthinexistingassets.ManagementFeeisthefeethefund complexreceivesformanagingshareholdersassets,expressedinmillions. Managementfeeiscalculatedasthetotalassetsmanagedperfundtimesthe fund'sexpenseratio.TurnoverRatioistotalpurchasesandsalesdividedbyfund's averagenetassetvalue.ExpenseRatioisthemutualfund'stotalannualoperating expenses(includingoperationalfees,distributionfees,andotherexpenses)stated asapercentageofthefund'saveragenetassets.12b-1feeischargebymutual funds for advertising, promotion, distributions, marketing expenses, and often comm inparentheses.Thesymbols***,**,and*denotesignificanceatthe1%,5%and 10% levels, respectively. Table 18: Mutual Fund Manager Replacement Univariate Regressions 89

PAGE 97

(Continued) constant Explanatory Variables Psuedo R Observations Management Fee -0.3951 0.0028 0.0003 9015 (<.0001) (0.3264) Expense Ratio 1.0384 -0.6123 0.0052 9015 (0.0078) (0.0200)** Fund Size 0.0548 -0.0009 0.0002 9364 (0.0068) (0.3285) Manager Tenure 1.8952 -1.1400 0.4522 9246 (<.0001) (<.0001)*** Style Drift -0.9847 0.5863 0.0063 8985 (0.0015) (0.0079)*** Tracking Error -0.7974 0.1026 0.0140 9172 (0.0006) (0.0052)*** Capital Gains Overhang 0.8737 -0.0132 0.0059 8892 (0.0063) (0.0089)*** Fund Age -0.4642 0.1636 0.0074 9285 (0.0004) (0.0077)*** Net Fund Growth -1.3295 0.0051 0.0004 9364 (<.0001) (0.1387) Turnover Ratio 0.0431 -0.0018 0.0004 8942 (<.0001) (0.1934) 12b-1 Fee 1.3843 -1.2433 0.0038 8939 (0.0320) (0.0493)** Table 18: Mutual Fund Manager Replacement Univariate Regressions 90

PAGE 98

Consistent with previous literature, managerial replacement is inversely related to the past performance of a fund (Khorana 199 6, 2001). Table 8 shows that both past performance regressions (RAR and OAR) in dicate the presence of a significantly negative relation between the probability of managerial tu rnover and past performance (p-value = 0.0001 and 0.0022, respectively) The tracking error has a positive and statistically significant relation with the repl acement of a fund manager. Brown, Harlow and Starks (1996) suggests that underperformi ng fund managers tend to have more erratic trading behavior seeking to improve their year-end performance. This positive relation between tracking error and managerial replacement can be explained by the fact that managers are compensated for their ability to outperform the benc hmarks they track. Consistent with the findings in Nanda, Wang, and Zheng (2000), probability of replacement has an inverse relation with ma nager tenure and expl ains a significantly large amount of the replacement decision, pseudo R2 = .45. As hypothesized earlier, the unitary fund management structure is negatively related to the probability of a manager bei ng replaced. This finding suggests that unitary fund managers are less likely to be repl aced than their multiple fund manager counterparts. Finally, the fees received by the management complex, expense ratio and 12b-1 advertising fees, expressed as a percenta ge of total assets, have a negative and statistically significantly influence on the managerial tur nover. This finding suggests that fund complexes are hesitant to replace manage rs than earn a significant amount of revenue for the company. The results in Table 8 indicate that there is no relation between the number of diverse objectives managed an d manager replacement. Overall, these results provide the first indication that, like th e literature suggests, there are a variety of criteria that have influence on the mana gerial replacement decision. Amongst these criteria is the management structure of th e fund complex, measured by the number funds simultaneously operated. 91

PAGE 99

92 Logistic Regressions To examine if fund management structure affects the performancereplacement relationship in a manner inconsistent with well-functioning internal control mechanism, I implement a multivariate regressi on in which I use the entire replacement sample to examine jointly the previously iden tified variables that influence managerial turnover. Specifically, I examine the re lation between unitary fund management structures and the managerial replacem ent decision after controlling for fund characteristics such as size, age, expense ratio, turnover ratio, advert ising fees and growth and other previously identified variables that affect the replacement decision. In Table 9, I report results of the logi stic regressions. Similar to Khorana (1996), I find a significantly negative relation between the probability of manager replacement and the previous year fund performance [model (i), (i ii), (iv) and (x)]. These results were obtained using the style category risk-adjusted return measure of managerial performance and are robust to using the objective-adjust ed abnormal return (OAR ) [model (v), (vii), (viii) and (ix)]. Consistent with the findi ngs of Nanda, Wang, and Zheng (2000), models (ii, iii, iv, viii) confirm the inverse relati on between manager tenure and the probability of a manger replacement.

PAGE 100

Explanatory Variables model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel x Intercept 0.05012.30223.21990.25312.87773.00103.10683.47600.58270.6392 (0.6136)(<.0001)***(<.0001)***(0.0840)*(0.0002)***(<.0001)***(<.0001)***(<.0001)***(0.0190)**(0.0003)*** Risk-Adjusted Return -0.0299 -0.0246-0.0298 -0.0253 (<.0001)*** (0.0017)***(<.0001)*** (<.0001)*** Objective-Adjusted Return -0.0061 -0.0087-0.0092-0.0079 (0.0028)*** (0.0009)***(0.0838)*(0.0273)** Unitary Fund Manager -0.2360-0.3564-0.2794-0.2253-0.2539 -0.2308-0.2675-0.0736 (0.0327)**(0.0072)***(0.0026)***(0.0011)***(0 .0036)*** (0.0039)***(0.0027)***(0.0065)*** Multiple Objectives Manager -0.1163 -0.1421-0.1389 (0.3252) (0.5662)(0.7138) Management Fee 0.00270.00110.0002 0.0022 0.0026 (0.9593)(0.3318)(0.6375) (0.6047) (0.5823) Expense Ratio -0.6024 -0.2549-0.6540-0.3493-0.5220-0.0185 (0.0211)** (0.4659)(0.0010)***(0.3035)(0.0392)**(0.9117) Fund Size -0.0004-0.0069-0.0008-0.0001-0.0007-0.0013-0.0063 (0.9217)(0.8769)(0.5537)(0.2951)(0.5571)(0.2689)(0.8916) Manager Tenure -1.2130-1.3443 -1.3847-1.2437-1.4032-1.3813 (<.0001)***(<.0001)*** (<.0001)*** (<.0001)***(<.0001)***(<.0001)*** Style Drift 0.24890.62960.32410.3675 (0.5890)(0.1676)(0.2640)(0.2754) Tracking Error 0.0158 (0.8118) Managertenureisthenumberofyearsaportfoliomanagerhasoverseenaparticularfund.TocalculateStyleDrift,Iregresseachfundsreturnsovert heyear priortoreplacementonthebenchmarkreturnsandtake1-Rasthemeasureofstyledrift.Trackingerrorisconstructedbytakingthestandarddeviati onofthe residualsfromaregressionofthefundsreturnonthereturnsofitsbenchmarks.CapitalGainsOverhangisthenetunrealizedappreciation(ordepr eciation) duringtheperiodreportedbytheMorningstardatabase.Fundageisthenaturallogofthefundsage.NetFundGrowthisthechangeinthefundassetsnet of growthinexistingassets.TurnoverRatioistotalpurchasesandsalesdividedbyfund'saveragenetassetvalue.12b-1feeischargebymutualfundsfo r advertising,promotion,distributions,marketingexpenses,andoftencommissions.AlsoincludedinTable9isaninteractionterm,abnormalretur nwithUFM, measuringtheabnormalreturnoftheunitaryfundmanagerinthepre-replacementperiod.Iincludethemarginalinteractiontermeffectestimationt o understand the economic impact of the interaction terms. In Table 9, I also report the results of the joint significance of Logisticregressionestimatesofmanagerialreplacementfor891managersarereportedoverthe1997to2001period.Managerreplacementisthe dichotomousdependentvariableequaltooneforthereplacementsampleandzeroforthecontrolsamplethathavenomanagerialturnover.Theobservat ions areinfund-years.FundPerformanceisdefinedbytheobjectiveadjustedreturn(OAR)inyear t andtheriskadjustedreturn(RAR)inyear t .UnitaryFund Managerisadichotomousvariablethattakesthevalueofoneifareplacedmanageroperatesaunitaryfundandzeroifthatmanageroperatesmultiplefu nds simultaneously.MultipleObjectivesManagedisadichotomousvariablethattakesthevalueofoneifareplacedmanageroperatesmultipleobjective s simultaneouslyandzeroifthatmanageroperatesfund(s)withoneobjective.ManagementFeeisthefeethefundcomplexreceivesformanagingshareho lders assets,expressedinmillions.ExpenseRatioisthemutualfund'stotalannualoperatingexpenses(includingoperationalfees,distributionfees, andother expenses) stated as a percentage of the fund's average net assets. Fund Size is the natural log of total fund net assets. the interaction va riable and abnormal return. The p-valu es of the regression coefficients are in parentheses. The symbols ***, **, and denote significance at the 1%, 5% and 10% levels, respectively. Table 19: Multivariate Regression results for all Mutual Fund Manager Replacements: Unitary Fund Management Specific 93

PAGE 101

Explanatory Variables model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel x Capital Gains Overhang -0.0016-0.0109-0.0124-0.0017-0.0072-0.0107-0.2489-0.0071-0.0027 (0.5739)(0.0078)***(0.0120)**(0.5537)(0.2731)(0.0090)***(0.4560)(0.2167)(0.4815) Fund Age 0.33010.10650.0907 (0.0081)***(0.1126)(0.5171) Net Fund Growth 0.0015 0.0003 0.00290.00080.0017 (0.1496) (0.7548) (0.7119)(0.3601)(0.1399) Turnover Ratio -0.0015 -0.0015-0.0018 -0.0027 (0.1858) (0.3281)(0.1402) (0.1932) 12b-1 Fee -0.4801-0.1031-1.2093 -1.2764-0.3883-1.1706-0.9637-0.2823 (0.1790)(0.8301)(0.0529)* (0.0968)*(0.4262)(0.1131)(0.1080)(0.4933) Return*Unitary Fund Manager-0.0043 -0.0057-0.0062-0.0046 -0.0045-0.0056-0.0031 (0.0472)** (0.0358)**(0.0284)**(0.0501)* (0.0491)**(0.0357)**(0.0437)** Marginal Interaction Term Effect-0.0191 -0.0346-0.0190-0.0277 -0.0245-0.0316-0.0066 Return and Interaction Variable-0.0163 -0.0158-0.0156-0.0025 -0.0047-0.0047-0.0025 Joint Significance (0.0002)*** (0.0015)***(0.0004) ***(0.0005)*** (0.0008)***(0.0018)***(0.0027)*** Observations 9127863484057927875381718733839281938204 Psuedo R 0.28500.48470.59070.29800.51140.50870.51280.58160.31790.1363 Table 19: Multivariate Regression results for all Mutual Fund Manager Replacements: Unitary Fund Management Specific (Continued ) 94

PAGE 102

In addition, I find that the fund management stru cture has a statistically significant relation with manage rial turnover. The evidence i ndicates that fund managers in UFM fund have a lower probabi lity of being replaced than managers of MFM funds. I also include an interaction term, abnormal return with UFM, measuring the abnormal return of the unitary fund manager in the pr e-replacement period. This is an important test of the internal governance mechanis m whereby fund sponsors evaluate managers based on performance and management structur e. The interaction coefficient indicates that for a given level of performance, manage rs of funds with unita ry management have a lower probability of being replaced than managers of MFM funds. The marginal interaction term effect reported in Table 9 i ndicates the probability of being replaced will be different for UFM and MFM at different levels of underperformance. Thus, the marginal interaction term ef fect of -0.0277 (model v) m eans that underperforming UFM are -2.77% less likely to be replaced than underperforming MFM counterparts. I also report the results of the joint significance of the interaction variable and abnormal return. The importance of the joint test is to see if the significant negative sign on return disappears or falls once we account fo r management structure. The statistically significant and negative coefficient of the join t significance variable confirms that even when unitary fund managers underperform the probability of getting replaced is lower than for multiple fund managers. Thus, th e performance-replacement relationship is stronger for multiple fund managers than that of their unitary fund counterparts. This implies that fund complexes are more likely to replace underperformers when it is cheap because replacing a unitary fund mana ger is more expensive than taking one fund from a manager that manages multiple funds. The above results remain unchanged when past performance and manager tenure are considered jointly (model iii, vii and viii), or when fund characteristic control variables are included (model v and vi). The magnitude and the statistical significance of estimated unitary fund management coeffici ent are robust to changes in the model specification. The explanatory power of the unitary fund management plus other variables is significant across all models with relatively high R2s. These findings suggest 95

PAGE 103

96 that there may be some economies of scale associated with the multiple fund management structure. However, once perf ormance is compromised the fund complex replaces the manager. However, the multip le objective management variable has no explanatory power with respect to th e managerial replacement decision.

PAGE 104

Explanatory Variables model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel x Intercept 0.41290.28350.84480.46650.90280.33970.78400.89711.02400.0287 (0.0002)***(0.0070)***(0.0039)***(0.0104)***(0.0577)*-0.1596(0.0144)**(<.0001)***(0.0014)***(0.9207) Risk-Adjusted Return -0.0342 -0.0318-0.0357 -0.0295 (<.0001)*** (<.0001)***(<.0001)*** (<.0001)*** Objective-Adjusted Return -0.0072 -0.0101-0.0089-0.0114 (0.0792)* (0.0114)**(0.0232)**(0.0005)*** Unitary Fund Manager -0.0106-0.0283-0.0178 -0.0134-0.0128 -0.0213 (0.04895)**(0.0851)*(0.0326)** (0.0465)**(0.0465)** (0.0228)** Multiple Objectives Manager -0.0045 (0.9635) Management Fee 0.00340.0041 0.0023 (0.5485)(0.4612) (0.9135) Expense Ratio -0.3505 -0.2243-0.1859-0.1470 -0.1033 (0.0619)* (0.3674)(0.0206)**(0.4379) (0.5709) Fund Size -0.0023-0.0047-0.0025-0.0029-0.0062-0.0039-0.0031 (0.7813)(0.9236)(0.6583)(0.5637)(0.9923)(0.4668)(0.5662) Manager Tenure Dummy -1.9100-1.8817-2.2794-2.1437-2.365-2.9509-3.4167-3.4065-3.3916-2.8326 (<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)*** Style Drift 0.13440.09010.0869 valueofoneifareplacedmanagerhasatenureinthetophalfoftheindustryandzeroifthatmanagerhasatenureinthebottomhalfoftheindustry.Tocal culate StyleDrift,Iregresseachfundsreturnsovertheyearpriortoreplacementonthebenchmarkreturnsandtake1-Rasthemeasureofstyledrift.Track ingerroris constructedbytakingthestandarddeviationoftheresidualsfromaregressionofthefundsreturnonthereturnsofitsbenchmarks.CapitalGainsO verhangisthe netunrealizedappreciation(ordepreciation)duringtheperiodreportedbytheMorningstardatabase.Fundageisthenaturallogofthefundsage.N etFundGrowth isthechangeinthefundassetsnetofgrowthinexistingassets.TurnoverRatioistotalpurchasesandsalesdividedbyfund'saveragenetassetvalue. 12b-1feeis chargebymutualfundsforadvertising,promotion,distributions,marketingexpenses,andoftencommissions.AlsoincludedinTable10isanintera ctionterm, abnormalreturnwithhighmanagertenure,measuringtheabnormalreturnofthemanagerswithhightenureinthepre-replacementperiod.Iincludethe marginal interaction term effect estimation to understand the economic impact of the interaction terms. In Table 10, I also report Logisticregressionestimatesofmanagerialreplacementfor891managersarereportedoverthe1997to2001period.Managerreplacementisthedicho tomous dependentvariableequaltooneforthereplacementsampleandzeroforthecontrolsamplethathavenomanagerialturnover.Theobservationsareinfu nd-years. FundPerformanceisdefinedbytheobjectiveadjustedreturn(OAR)inyear t andtheriskadjustedreturn(RAR)inyear t .UnitaryFundManagerisadichotomous variablethattakesthevalueofoneifareplacedmanageroperatesaunitaryfundandzeroifthatmanageroperatesmultiplefundssimultaneously.Multiple ObjectivesManagedisadichotomousvariablethattakesthevalueofoneifareplacedmanageroperatesmultipleobjectivessimultaneouslyandzeroi fthat manageroperatesfund(s)withoneobjective.ManagementFeeisthefeethefundcomplexreceivesformanagingshareholdersassets,expressedinmill ions. ExpenseRatioisthemutualfund'stotalannualoperatingexpenses(includingoperationalfees,distributionfees,andotherexpenses)statedasap ercentageofthe fund's average net assets. Fund Size is the natural log of total fund net assets. High Manager Tenure is a dichotomous variable that takes the the results of the joint significance of the interaction variable and abnormal return. The p-values of the regression coefficie nts are in parentheses. The symbols ***, **, and denote significance at the 1%, 5% and 10% levels, respectively. Table 20: Multivariate Regression result s for all Mutual Fund Manager Replaceme nts: High Management Tenure Specific 97 (0.6823)(0.7780)(0.7800)

PAGE 105

Explanatory Variables model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel x Tracking Error 0.0261 0.0681 (0.5568) (0.1006) Capital Gains Overhang -0.0053-0.0050-0.0058-0.0049-0.0046-0.0051-0.0035-0.0037 (0.9858)(0.0682)*(0.0973)*(0.1563)(0.3397)(0.0753)*(0.4213)(0.3864) Fund Age 0.19390.20470.03110.18460.1249 (0.0232)**(0.0145)**(0.7384)(0.0113)**(0.1440) Net Fund Growth -0.0014-0.0014-0.0006 -0.0015-0.0012-0.0013 (0.0856)*(0.0746)*(0.4022) (0.0934)*(0.1467)(0.1087) Turnover Ratio -0.0017 -0.0039-0.0012 -0.0066 -0.0009 (0.8321) (0.7209)(0.0785)* (0.3880) (0.3395) 12b-1 Fee -0.5217 -0.2568 -0.0104-0.5899-0.1340 -0.3049 (0.1654) (0.5769) (0.9849)(0.1262)(0.7679) (0.4802) Return*High Manager Tenure-0.0148 -0.0198-0.0088-0.0043 -0.0092-0.0085-0.0082-0.0288 (0.0591)* (0.0552)*(0.0782)*(0.0980)* (0.6096)(0.0644)*(0.0656)*(0.0419)** Marginal Interaction Term Effect-0.0799 -0.1423-0.1773-0.1482 -0.2807-0.2818-0.2769-0.1979 Return and Interaction Variable-0.0324 -0.0308-0.0333-0.0067 -0.0090-0.0079-0.0105-0.0294 Joint Significance (<.0001)*** (<.0001)***(<.0001)***(0.0846)* (0.0187)**(0.0353)**(0.0007)***(<.0001)*** Observations 7933793382178193859879698197821381868456 Psuedo R 0.16770.13910.21650.20970.21770.21960.25350.25480.25050.2376 Table 20: Multivariate Regression results for all Mutual Fund Manager Replacements: High Management Tenure Specific (Continued) 98

PAGE 106

I interpret the findings that managers of UFM have a lower probability of being replaced, for any given level of performance, than managers of MFM funds as evidence of a conflict of interests between investors and fund sponsors. This, as discussed before, is because fund sponsors reluctance to terminate single-fund managers is driven by costsavings consideration of the sponsor. The prel iminary evidence (Table 2) indicates that UFM funds have higher asset growth rates, which is beneficial to the sponsors. In contrast, these funds have higher expense ratios, which makes th eir governance even more important to investors because higher ex penses reduces investors terminal wealth while benefiting fund sponsors whose manage ment fees are included in the funds expense ratio. Therefore, taken together, th e evidence does not support the claim of wellfunctioning internal mechan isms for mutual fund managers at least not without qualifications. To further explore the in ternal governance mechanisms for fund managers, I examine the joint significance of manager te nure and management fees with abnormal returns. In Table 10, I conduc t a logistic regression model in which the probability of managerial turnover is explai ned by a dichotomous manage r tenure variable. Manager tenure takes the value of one if the replaced managers tenure is greater than the median (Table 2, Panel D), and zero otherwise. Consistent with the findings in Chevalier and Ellison (1999a,b) and Gallagher (2003), the manager tenure dummy variable is inversely related to the probability of managerial replacement. The interaction term, abnormal return and manager tenure, measures the infl uence of abnormal returns for managers with longer tenures on the probability of replacemen t. The evidence indicates that for any given level of performance the probability of getting replaced is lower than that of less experienced managers. Thus, according to the ma rginal effect term, an experienced fund manager will be 14.8% less likely to be repl aced than an inexperienced manager, even though both of them are underperforming. The evidence for tenure does not really make a strong case because if a tenured manager with a history of good performance hits a rough spot there is good reason to hope he is going to become a high performance later, so 99

PAGE 107

100 sponsors may tolerate low performance. Thes e findings suggest weak and limited internal governance mechanisms for fund management. Finally, I examine the inte rnal governance of mutu al funds with the joint significance of abnormal return and the high total fees binary variable. Table 11 documents these results. For all models, the joint significance of abnormal return*management fee variable is statistical ly insignificant. This finding suggests some level of governance concerning the revenue to the fund complex.

PAGE 108

Explanatory Variables model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel x Intercept 0.14611.28762.59610.30640.29192.80741.99980.28900.29920.3205 (0.3480)(<.0001)***(<.0001)***(0.0942)*(0.1092)(<.0001)***(<.0001)***(0.2508)(0.3195)(0.3854) Risk-Adjusted Return -0.0299 -0.0351-0.0305 -0.0295 (<.0001)*** (0.0018)***(<.0001)*** (0.0003)*** Objective-Adjusted Return -0.0086-0.0132-0.0084 -0.0096 (0.0139)**(0.0175)**(0.0258)** (0.0284)** Unitary Fund Manager -0.0866-0.0813-0.3701 -0.1357 -0.0913 (0.0580)*(0.0618)*(0.0179)** (0.0419)** (0.0639)* Multiple Objectives Manager -0.3548 -0.2570 (0.2643) (0.1002) Total Expenses & Fees Dummy-0.1596-0.1571-0.1437-0.0617-0.0201-0.0807-0.0197-0.1197-0.2776-0.1845 (0.2422)(0.3794)(0.5716)(0.6943)(0.8892)(0.8547)(0.9432)(0.0480)**(0.2439)(0.3154) Fund Size -0.0002-0.0030-0.0030-0.0050 -0.0078-0.0021 (0.1672)(0.4999)(0.5100)(0.6787) (0.8665)(0.6768) Manager Tenure -1.1857-1.2905 -1.3539-1.3438 (<.0001)***(<.0001)*** (<.0001)***(<.0001)*** Style Drift 0.75180.3338 0.7149 (0.0899)*(0.2394) (0.0914)* Tracking Error 0.0094 0.01560.0095 replacedmanagerhastotalexpensesinthetophalfoftheindustryandzeroifthatmanagerhastotalexpensesinthebottomhalfoftheindustry.FundSi zeisthe naturallogoftotalfundnetassets.Managertenureisthenumberofyearsaportfoliomanagerhasoverseenaparticularfund.TocalculateStyleDrift,Iregresseach fundsreturnsovertheyearpriortoreplacementonthebenchmarkreturnsandtake1-Rasthemeasureofstyledrift.Trackingerrorisconstructedby takingthe standarddeviationoftheresidualsfromaregressionofthefundsreturnonthereturnsofitsbenchmarks.CapitalGainsOverhangisthenetunreali zedappreciation (ordepreciation)duringtheperiodreportedbytheMorningstardatabase.Fundageisthenaturallogofthefundsage.NetFundGrowthisthechangein thefund assetsnetofgrowthinexistingassets.TurnoverRatioistotalpurchasesandsalesdividedbyfund'saveragenetassetvalue.12b-1feeischargebymu tualfundsfor advertising,promotion,distributions,marketingexpenses,andoftencommissions.AlsoincludedinTable11isaninteractionterm,abnormalretu rnwithtotalexpenses (expense ratio, management and 12b-1 fees), measuring the abnormal return of the managers with Logisticregressionestimatesofmanagerialreplacementfor891managersarereportedoverthe1997to2001period.Managerreplacementisthedicho tomous dependentvariableequaltooneforthereplacementsampleandzeroforthecontrolsamplethathavenomanagerialturnover.Theobservationsareinfund-years. FundPerformanceisdefinedbytheobjectiveadjustedreturn(OAR)inyear t andtheriskadjustedreturn(RAR)inyear t .UnitaryFundManagerisadichotomous variablethattakesthevalueofoneifareplacedmanageroperatesaunitaryfundandzeroifthatmanageroperatesmultiplefundssimultaneously.Mul tipleObjectives Managedisadichotomousvariablethattakesthevalueofoneifareplacedmanageroperatesmultipleobjectivessimultaneouslyandzeroifthatmanageroperates fund(s)withoneobjective.ManagementFeeisthefeethefundcomplexreceivesformanagingshareholdersassets,expressedinm illions.ExpenseRatioisthemutual fund'stotalannualoperatingexpenses(includingoperationalfees,distributionfees,andotherexpenses)statedasapercentageofthefund'saveragenetassets.Total Expenses is a dichotomous variable that takes the value of one if a highmanagementfeesinthepre-replacementperiod.Iincludethemarginalinteractiontermeffectestimationtounderstandtheeconomicimpactoft heinteraction terms.InTable11,Ialsoreporttheresultsofthejointsignificanceoftheinteractionvariableandabnormalreturn.Thep-valuesoftheregression coefficientsarein parentheses. The symbols ***, **, and denote significance at the 1%, 5% and 10% levels, respectively. Table 21: Multivariate Regression results for all Mutual Fund Manager Replacements: Total Expenses Specific 101 (0.9886) (0.7109)(0.3240)

PAGE 109

Ex Cap Fund Age Net Tu Ret M Ret Jo O Ps planatory Variables model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel x ital Gains Overhang -0.0014-0.0121-0.0123 -0.0062-0.0057-0.0032 (0.6149)(0.0040)***(0.0151)** (0.3365)(0.3298)(0.3947) 0.1056 0.31470.06960.05960.07560.26260.0618 (0.1139) (0.0118)**(0.3675)(0.4374)(0.5855)(0.03664)**(0.4376) Fund Growth -0.0019-0.0012-0.0011-0.0026-0.0120-0.0019-0.0056 (0.0953)*(0.0629)*(0.0861)*(0.7776)(0.2378)(0.0325)**(0.4266) rnover Ratio -0.0021 -0.0015 -0.0996 (0.0435)** (0.2957) (0.3275) urn*Total Expnses & Fees-0.0029 -0.0195-0.0030 -0.0155-0.0099-0.0024-0.0059-0.0061 (0.8405) (0.4412)(0.8555) (0.2271)(0.3548)(0.7460)(0.7401)(0.6931) arginal Interaction Term Effect-0.0045 -0.0162-0.0051 -0.0081-0.0024-0.0028-0.0082-0.0066 urn and Interaction Variable-0.0207 -0.0223-0.0236 -0.0034-0.0071-0.0052-0.0059-0.0064 int Significance (<.0001)*** (0.0071)***(<.0001)*** (0.0472)**(0.0885)*(0.0570)*(0.0741)*(0.0733)* bservations 7927813884048169816987538392819385908183 uedo R 0.02860.49610.57920.03190.02970.50920.58280.02060.04170.0638 Table 21: Multivariate Regression result s for all Mutual Fund Manage r Replacements: Total Expenses Specific (Continued) 102

PAGE 110

Managerial Turnover from Demotion One of the major issues with in the mutual fund manageri al turnover li terature is the difficulty in distinguishing manager re placement due to promotion and manager replacement due to demotion. Hu, Hall and Harvey (2000) identify management promotions and demotions by cross referencing the Morningstar database with the reports from Lexis Nexis Inc. and define demotion as a manager moving to a smaller size fund or forced out of the mutual fund industry. Since this study focuses on the governance mechanisms within the mutual fund industr y, I am only concerned with managerial replacements due to demotions. Khorana (1996 ) reports an inverse relation between the probability of fund manager repl acement and past performance. Thus, this study defines replacement due to demotion as poor performa nce; the one year negative abnormal return of a mutual fund manager. In Table 12 I conduct robustness test usi ng only those manager replacements that had negative one-year pre-replacement returns. The results in Table 12 also indicate a negative relationship between the probability of managerial changes and fund structure. A comparison of Tables 9, 10 and 11 with Ta ble 12 suggests that fund structure exhibits a stronger inverse relationship with manager demotion than with manager replacement. This evidence suggests that a unitary fund manager is less lik ely to be fired or demoted than their multiple fund manager peers. As stat ed earlier, since it is more costly for a fund sponsor to replace a unitary fund manager than take a fund from a multiple fund manager, the fund sponsor is more hesitant to replac e a unitary fund manage r regardless of the funds performance. This presents a conflict of interests between the investors request for superior returns and the sponsors desire to lowe r costs for the sponsor. As hypothesized above, the coefficients of the abnormal return variables and net fund growth variables are negative in Table 12. There is a stronger association between the performance variables and manager replace ment. For the demotion sample (Table 12) these coefficients are rangi ng from -0.0109 and -0.0718, wh ereas for the replacement sample (Table 9) they range from -0.006 and -0.0299. In contrast to the replacement sample, the expense ratio has consistent and statistically significant explanatory power 103

PAGE 111

104 for the probability of demotion. However, the management fee and total expenses and fees variables are insignifi cant in all models. The coeffi cients on the manager tenure variables are negative and stat istically significant in all seven models. Compared to the models that include the entire replacemen t sample, the demotion replacement sample exhibits lower correlation be tween the probability of replacement and manager tenure. The evidence suggests that there are conflict of interests between fund sponsors and investors due to the management structure a nd that there are governance mechanisms in place to address these c onflicts. However, these governance mechanisms dont completely protect investors from fund sponsor interests.

PAGE 112

Table 22: Multivariate Regression results for Poor Performing Funds Logisticregressionestimatesofmanagerialreplacementfor539managersthathadoneyearofpoorperformance(negativecumulativeabnormalretur ns) priortoreplacement.Managerreplacementisthedichotomousdependentvariableequaltooneforthereplacementsampleandzeroforthecontrolsamp le thathavenomanagerialturnover.Theobservationsareinfund-years.FundPerformanceisdefinedbytheobjectiveadjustedreturn(OAR)inyear t andthe riskadjustedreturn(RAR)inyear t .UnitaryFundManagerisadichotomousvariablethattakesthevalueofoneifareplacedmanageroperatesaunitary fundandzeroifthatmanageroperatesmultiplefundssimultaneously.MultipleObjectivesManagedisadichotomousvariablethattakesthevalueofo neifa replacedmanageroperatesmultipleobjectivessimultaneouslyandzeroifthatmanageroperatesfund(s)withoneobjective.ManagementFeeisthefe ethe fundcomplexreceivesformanagingshareholdersassets,expressedinmillions.ExpenseRatioisthemutualfund'stotalannualoperatingexpenses (including operational fees, distribution fees, and other expenses) stated as a percentage of the fund's average net assets. TotalExpensesisadichotomousvariablethattakesthevalueofoneifareplacedmanagerhastotalexpensesinthetophalfoftheindustryandzeroifth at managerhastotalexpensesinthebottomhalfoftheindustry.FundSizeisthenaturallogoftotalfundnetassets.Managertenureisthenumberofyears a portfoliomanagerhasoverseenaparticularfund.HighManagerTenureisadichotomousvariablethattakesthevalueofoneifareplacedmanagerhasa tenureinthetophalfoftheindustryandzeroifthatmanagerhasatenureinthebottomhalfoftheindustry.TocalculateStyleDrift,Iregresseachfun ds returnsovertheyearpriortoreplacementonthebenchmarkreturnsandtake1-Rasthemeasureofstyledrift.Trackingerrorisconstructedbytaking the standarddeviationoftheresidualsfromaregressionofthefundsreturnonthereturnsofitsbenchmarks.CapitalGainsOverhangisthenetunreali zed appreciation(ordepreciation)duringtheperiodreportedbytheMorningstardatabase.Fundageisthenaturallogofthefundsage.NetFundGrowthi sthe change in the fund assets net of growth in existing assets.Turnover Ratio is total purchases and sales divided by fund'saveragenetassetvalue.12b-1feeischargebymutualfundsforadvertising,promotion,distributions,marketingexpenses,andoftencommis sions. AlsoincludedinTable12isaninteractionterm,abnormalreturnwithUFM,measuringtheabnormalreturnoftheunitaryfundmanagerintheprereplacementperiod;abnormalreturnwithhighmanagertenure,measuringtheabnormalreturnofthemanagerswithhightenureinthepre-replacement period;andabnormalreturnwithtotalexpenses(expenseratio,managementand12b-1fees),measuringtheabnormalreturnofthemanagerswithhigh managementfeesinthepre-replacementperiod.Iincludethemarginalinteractiontermeffectestimationtounderstandtheeconomicimpactofthein teraction terms.InTable12,Ialsoreporttheresultsofthejointsignificanceoftheinteractionvariableandabnormalreturn.Thep-valuesoftheregression coefficients are in parentheses. The symbols ***, **, and denote significance at the 1%, 5% and 10% levels, respectively. Explanatory Variables model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel x Intercept 2.10501.42903.43262.05943.01131.03282.92041.95842.13291.6227 (<.0001)***(0.0025)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(0.0030)***(<.0001)***(0.0031)*** Risk-Adjusted Return -0.0718 -0.0683-0.0699 -0.0703 -0.0645 (<.0001)*** (<.0001)***(<.0001)*** (<.0001)*** (<.0001)*** Objective-Adjusted Return -0.0109-0.0116 -0.0224 -0.0193 (0.0004)***(<.0001)*** (0.0201)** (0.0066)*** Unitary Fund Manager -0.8205-0.9138-0.8864-0.8307-0.8376 -0.8154 -0.7853-0.8091 (0.0004)***(0.0005)***(0.0003)***(<.0001)***(0.0002)*** (0.0006)*** (0.0007)***(0.0007)*** Multiple Objectives Manager -0.4305 -0.3875-0.2950 (0.2065) (0.3384)(0.2986) Management Fee -0.0338-0.0912-0.0117 (0.9023)(0.9108)(0.9157) Expense Ratio -0.5230-0.3995-0.6041 -0.4203 (0.00518)*(0.0428)**(0.0032)** (0.0495)** Total Expenses & Fees -0.5212-0.5526-0.4837 105 ( 0.4274 )( 0.4061 )( 0.3985 )

PAGE 113

Explanatory Variables model imodel iimodel iiimodel ivmodel vmodel vimodel viimodel viiimodel ixmodel x Fund Size -0.0022-0.0054-0.0003-0.0005-0.0007-0.0091-0.0077 (0.8913)(0.7697)(0.6538)(0.5993)(0.8416)(0.8621)(0.7812) Manager Tenure -1.1380-1.0570 -1.1011-1.0534 (<.0001)***(<.0001)*** (<.0001)***(<.0001)*** High Manager Tenure -1.3460-1.6775-1.7317 (<.0001)***(<.0001)***(<.0001)*** Style Drift 0.4733 0.6694 0.5432 (0.3792) (0.2129) (0.2583) Tracking Error 0.0378 0.0478 0.0575 (0.6398) (0.7985) (0.7154) Capital Gains Overhang -0.0058 -0.0260 -0.0732 -0.0813-0.0268-0.3352 (0.4831) (0.3903)** (0.2512) (0.2743)(0.4137)(0.3585) Fund Age 0.2146 0.1832 0.2195 (0.1073) (0.1166) (0.1319) Net Fund Growth -0.0018 -0.0020 -0.0026 -0.0031 -0.0073 (0.0816)* (0.0916)* (0.0744)* (0.0671)* (0.0921)* Turnover Ratio -0.0009 -0.0005-0.0004 -0.0007 (0.2186) (0.3384)(0.4147) (0.3285) 12b-1 Fee -0.2667 -0.6146 -0.7828-0.4257 (0.1121) (0.10507) (0.1205)(0.1006) Return*Unitary Fund Manager-0.0912 -0.0849 -0.0217 (<.0001)*** (<.0001)*** (0.0004)*** Return*High Manager Tenure -0.9710-0.7724-0.8968 (<.0001)***(<.0001)***(<.0001)*** Return*Total Expnses & Fees -0.0366-0.0245-0.0345 (0.0035)***(0.0221)**(0.0013)*** Marginal Interaction Term Effect-0.0760 -0.0866-0.1160-0.1631-0.1542-0.0502-0.0481-0.0460-0.0717 Return and Interaction Variable-0.0072 -0.0057-0.0663-0.0084-0.0097-0.0034-0.0011-0.0024-0.0008 Joint Significance (<.0001)*** (<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)***(<.0001)*** Observations 8796830380747596842278408402806178627873 Psuedo R 0.40070.14270.46450.32760.45360.31330.42060.30480.41350.3106 Table 22: Multivariate Regression results for Poor Performing Funds (Continued) 106

PAGE 114

Conclusion The inverse relationships between mana ger turnover on the one hand and past performance and manager tenure, respectiv ely, on the other hand have been well documented. However, this paper is the first to document that management structure and other fund characteristics affect the probabi lity of managerial turnover in a manner consistent with the existence of a conflict of interests between investors and sponsors. Using a sample of 891 equity and bond fund managerial replacements over the 1997 to 2001 period, I document that unitary fund mana gers have an approximately 2% lower probability of experiencing replacement than their multiple fund management peers. As hypothesized, fund complexes tend to repla ce underperformers only when it is cheap because replacing a unitary fund manager is mo re costly to the fund sponsor than taking one fund from a manager that operates multiple funds. Conversely, the number of funds a manager operates simultaneously has a positiv e relation with the probability of that manager being replaced. Coupled with the past performance, these re sults are consistent with the argument that there are some ec onomies of scale benefits to multiple fund management. However, once the fund performan ce deteriorates the ma nager is released from his duties for that fund faster than the manager who manages a single fund. Despite the large body of research on managerial turnover, previous studies have only examined (or assumed there exists only ) the unitary manageme nt structure. The failure to account for the multiple fund mana gement structure ignores an additional impact fund managers have on the fund comp lex. For instance, fund complexes increase total profits by increasing (or at least maintaining) the level of inflows and decreasing the individual cost of operating each fund. Khor ana (2001) suggests that in a competitive market, management expense ratios should de cline over time where investors become more price-sensitive, investment management firms increase in size and improve their economies of scale. As not ed earlier, fund complexes that deploy the multiple management structure have lo wer expense ratios on average. Using a series of carefully constructed multivariate logistics regressions, I examine the internal governance mechanisms within the mutual fund industry utilizing 107

PAGE 115

interaction terms and joint signi ficance analysis. This study util izes a marginal interaction term effect methodology presented in Ai a nd Norton (2003) to document to marginal change in management structures for a gi ven level of underperformance. I also employ cluster analysis to account for two dimensi ons (fund complex and year) of variation. I document weak and limited in ternal governance mechanisms within the mutual fund industry because for a given amount of underp erformance fund sponsors are significantly likely to replace fund managers who manage a single fund than those who manage multiple funds. These findings suggest that if managers operate a single fund or have a great deal of experience they are even less lik ely to be replaced than previous literature states. In summary, this area of research is si gnificant given the re sponsibility of fund sponsors in managing their investment manage rs, the sizable assets under their control, the significant research effort and resource s dedicated to the res earch of investments management institutions. Thus, the use of th e fund management structure that benefits both the fund complex as well as the investor provides a dditional insights into this dynamic and multifaceted industry. 108

PAGE 116

References Ai, C. and E.C. Norton, 2003, Interac tion terms in logit and probit models, Economics Letter 80, 123-129. Akhavein, J., Berger A., and Humphr ey, D., 1997, The Effects of Bank Megamergers on efficiency and prices : Evidence from th e profit function, Review of Industrial Organization 12, 95-139. Akhigbe A. and A. M. Whyte, 2001, The Markets Assessment of the Financial Services Modernization Act of 1999, The Financial Review 36, 119-138. Akhigbe A. and A. M. Whyte, 2004, Th e Gramm-Leach-Bliley Act of 1999: Risk implications for the financial services industry, Journal of Financial Research Volume 27 Issue 3 Page 435-446. Alam I., 2001, A non-parametric approach for assessing productivity dynamics of large banks, Journal of Money, Credit, and Banking 33,121-139. Allen, L., Jagtiani, J., Peristiani, S., and A. Saunders, 2004, The Role of Bank Advisors in Mergers and Acquisitions, Journal of Money, Credit and Banking 36, No. 2, 197-224. Allen, P. and W. Wilhelm, 1988, The im pact of the 1980 De pository Institutions Deregulation and Monetary Control Act on th e market value and risk: Evidence from the capital markets, Journal of Money, Credit and Banking 20, 364-380. Almazan A., 1996, A model of competition in banking: bank capital vs. expertise, working paper from Massachusetts Institution of Technology. Altunbas, Y., Molyneux, P. and Thornton, J., 1997, Big mergers in EuropeAn analysis of the cost implications, Economica 64:317-329. Amel, D., and Liang, N., 1992, The Relationship Between Entry into Banking Markets and Changes in Legal Restrictions on Entry, The Antitrust Bulletin Fall, pp. 631-49. Ammann, M. and H. Zimmermann, 2001, Tracking error and tactical asset allocation, Financial Analysts Journal 57 (2), 32-43. Amihud, Y. and Lev, B., 1981, Risk reduction and managerial motive for conglomerate mergers, Bell Journal of Economics 12, 605-617. 109

PAGE 117

Apilado, V., Gallo, J. and Lockwood, L., 1993, Expanded securities underwriting: implications for bank risk and return, Journal of Economics and Business 45, 143-158. Asquith, P., 1983. Merger Bids, Unce rtainty, and Stockholder Returns. Journal of Financial Economics 11: pp 51-83. Asquith, P. and D. Mullins, 1983, The Imp act of Initiating Dividend Payments on Shareholders Wealth, The Journal of Business 56, 77-96 Ball, R. and S. Kothari, 1989, Security Returns around Earnings Announcements, The Accounting Review 66, 718-738. Barth, J., Brumbaugh R., and Wilcox, J., 2000, The repeal of Glass-Steagall and the Advent of board banking, Journal of Economic Perspectives 14(2), Spring, 191-204. Bebchuk, L., J. Coates and G. Subraman ian, 2002, The powerful antitakeover force of staggered boards: Theory, evidence, and policy, Stanford law review, 54, 887951. Bebchuk L., J. Fried and D. Walker, 2003, Executive Managerial power and rent extraction in the design of executive compensa tion, University of Chicago Law Review 69, 751-846. Benston, G.J., Hunter, W.C., Wall, L.D ., 1995, Motivations for bank mergers and acquisitions: Enhancing the deposit insurance put option versus earnings diversification, Journal of Money, Credit, and Banking 27, 777-788. Berger, A., 1993, Distribution-Free Estimates of Efficiency in the U.S. Banking Industry and Tests of the Standa rd Distributional Assumptions, Journal of Productivity Analysis 4, 261-292. Berger, A., Cummins, J., Weiss, M., and Zi, H., 1999, Conglomeration versus Strategic Focus: Evidence from the Insurance Industry, Board of Governors of the Federal Reserve System. Berger, A., Demsetz, R., and Strahan, P., 1999, The consolidation of the financial services industry: Causes, consequences, and implications for the future, Journal of Banking and Finance 23, 135-194. Berger, A., DeYoung, R., Genay, H., and Udell G., 2000, Globalization of financial institutions: Evidence from cross-border banking performance, Brookings Papers on Economic Activity, 2: 23-158. 110

PAGE 118

Berger, A., Hancock, D. and Humphrey D., 1993, Bank efficiency derived from the profit Function, Journal of Banking and Finance 17, 317-347. Berger, A., and Hannan, A., 1998, The Effi ciency Cost of Market Power in the Banking Industry: A Test of the Qui et Life and Related Hypotheses, Review of Economics and Statistics 80 (August): 454. Berger, A., Hanweck, G., and Humphrey D., 1987, Competitive viability in banking: Scale, scope, a nd product mix economies, Journal of Monetary Economies 20, 501-520. Berger A., and Humphrey D., 1992, Megamergers in banking and the use of cost efficiency as an antitrust defense, Antitrust Bulletin 37, 541-600. Berger, A., and. Mester, L., 1997, Insi de the Black Box: What Explains Differences in the Efficiencies of Financial Institutions?, Journal of Banking and Finance 21, no. 7 (July): 895-947. Berger A. and Mester L., 2003, Explai ning the dramatic changes in the performance of U.S. banking: Technological change, deregulation, and dynamic changes in competition, Journal of Financial Intermediation 12. Berger, P., and Ofek, E., 1995, Divers ifications effect on firm value, Journal of Financial Economics 37, 39-65. Berle, A., and A. Means, 1932, The modern corporat ion and private property. New York: Commerce Clearing House. Bertrand M. and S. Mullainathan, 2001, Do people mean what they say? Implications for subjective survey data, Economics and social behavior, 67-72 Black, F., and M. Scholes, 1974, The eff ects of dividend yi eld and dividend policy on common stock prices and returns, Journal of Financial Economics 1, 1-22. Blanchard, O., F. Lopez-de-Silanes and A. Shleifer, 1994, What do firms do with cash windfalls?, Journal of Fi nancial Economics, XXXVI, 337-360. Boyd J., Graham, S., and Hewitt, S., 1993, Bank Holding Company Mergers with Nonbank Financial firms: Effect s on the Risk of Failure, Journal of Banking and Finance 17, 43-63. Boyd J., and Runkle D., 1993, Size a nd Performance of Banking Firms, Journal of Monetary Economics 31, 47-67. 111

PAGE 119

Bradley, M., Desai A., and E. Han Kim. 1983, The Rationale Behind Interfirm Tender Offers: Information or Synergy?, Journal of Financial Economics 11, no. 1-4: pp 183-206. Brewer, E., Fortier, D., and Pavel, C ., 1988, Bank Risk from Nonbank Activities, Federal Reserve Bank of Chicago, Economic Perspectives (July/August) pp. 14. Brook, Y., Hendershott, R. and Lee, D., 1998, The gains from takeover deregulation: Evidence from the end of interstate banking restrictions, Journal of Finance 53, 2185-2204. Brown S,J, and W.N. Goetzmann, 1997, Mu tual fund styles. Journal of Financial Economics 43 (March): 373-399. Brown, K. and W. Harlow, 2006, Staying the course: Performance persistence and the role of investment style consistency in professional asset management, Unpublished Working Paper, University of TexasAustin. Brown, K. C., W. Harlow, and L. T. Starks, 1996, Of tournaments and temptations: An analysis of managerial incentives in the mutual fund industry, Journal of Finance 51, 85-110. Busse, Jeffrey, 2001, Another Look at Mutual Fund Tournaments, Journal of financial and Quantitative Analysis 36, 53-73. Calomiris, Charles W., and Jason Karces ki, 1998, Is the Bank Merger Wave of the 90s Efficient? Lessons fr om Nine Case Studies, in: St ephen Kaplan, ed., Mergers and Productivity (Unive rsity of Chicago Press).) Calem, P., 1994, The Impact of Geographic deregulation on small banks, Business Review, Federal Reserve Bank of Philadelphia, 17-31. Cameron A., J.B. Gelbach and D. Miller, 2006, Robust Inference with Multi-Way clustering, NBER Technical Work ing Paper No. 327 (September). Carow, K., Kane, E., 2002. Event-Study Evidence of the Value of Relaxing Longstanding Regulatory Re straints on Banks, 1970-2000. The Quarterly Review of Economics and Finance 42, 505-527. Chan, K., 1988, On the Contra rian Investment Strategy, Journal of Business 61, 147-163. Chan, Louis K.C., Hsiu-Lang Chen, and Josef Lakonishok, 2002, On Mutual Fund Investment Styles, The Review of Financial Studies 15, 1407-1437. 112

PAGE 120

Chan, Y., Greenbaum, S., and A. Thakor, 1986, Information Reusability, Competition, and Bank Asset Quality. Journal of Banking and Finance 10, 243-253. Chevalier, J., 1999. Why do firms unde rtake diversifying mergers? An examination of the investment policies of merging firms. Unpublished working paper, University of Chicago. Chevalier, J. and G. Ellison, 1997, Risk Taking by Mutual Funds as a Response to Incentives, Journal of Political Economy 105: 1167-1200 Chevalier, J. and G. Ellison, 1999, Career Concerns of Mutual Fund Managers, Quarterly Journal of Economics 114, 389-432. Chevalier, J. and G. Ellison, 1999, Are Some Mutual Fund Managers Better than Others? Cross-Sectional Patterns in Behavior and Performance, Journal of Finance 54, 875-900. Chopra, N., Lakonishok, J. an d J. Ritter, 1992, Measuring abnormal performance: Do stocks overreact?, Journal of Financial Economics 31, 235-268. Clarke, M., 2004, Geographi c Deregulation of Banki ng and Economic Growth, Journal of Money, Credit and Banking 36, No. 5, 929-942. Conrad, J. and G. Kaul, 1993, Long-Term Market Overreaction or Biases in Computed Returns?, The Journal of Finance 48, 39-63. Core, J.E., W. Guay and D.F. Lacker, 2003, Executive equity compensation and incentives: A survey, Economic Policy Review 9, 27-50. Cornett, M. and H. Tehranian, 1989, Stock market reactions to the Depository Institutions Deregulation and Monetary Control Act of 1980, Journal of Banking Finance 13, 81-100. Cornett, M. and H. Tehranian, 1990, An Examination of the Impact of the GarnSt. Germain Depository Institutions Act of 1982 on Commercial Ba nks and Savings and Loans, Journal of Finance 45, 95-111. Covington, A. and J. Burling, 1999, Financial Modernization: The Gramm-LeachBliley," American Bankers Association (November, 1999). Daniel, Kent, Mark Grinblatt, Sher idan Titman, and Russ Wermers, 1997, Measuring Mutual Fund Performance w ith Characteristic-Based Benchmarks, Journal of Finance 52, 1035-1058. 113

PAGE 121

DeBondt, W. and R. Thaler, 1985, Does the stock market overreact?, Journal of Finance 40, 793-808. DeBondt, W. and R. Thaler, 1987, Further evidence on investor overreaction and stock market seasonality, J ournal of Finance 42, 557-581. Deli, D. N., 2002, Mutual Fund Advisory contracts: An empirical investigation, Journal of Finance 57, 109-133. Demsetz R., and Strahan, P., 1997, Diversific ation, size, and risk at bank holding companies, Journal of Money, Credit, and Banking 29, 300-313. DeYoung, R., 1997, A diagnostic test fo r the distribution-free efficiency estimator: An example using U.S. commercial bank data European Journal of Operational Research Volume 98, Issue 2 Pages 243-249. Diamond, D., 1984, Financial intermed iation and delegate monitoring, Review of Economic Studies 51, 393-414. Dodd. P., 1980, Merger proposals, managerial discretion and stockholder wealth, Journal of Financial Economics 8 (June), 105-138. Diamond D., 1989, Reputation Ac quisition in Debt markets, Journal of Political Economy, XCVII, 828-862. Ding B. and R. Wermers, 2005, Mutual fund performance and governance structure: the role of portfolio managers and boards of directors. Unpublished working paper, State University of New York at Albany and University of Maryland. Eckbo, B. Espen. 1983. "Horizontal Mergers, Collusion, and Stockholder Wealth." Journal of Financial Economics 11: pp 241-273. Elton, Edwin J., Martin J. Gruber, and Jeffrey Busse,2004, Are Investors Rational? Choices among Index Funds, Journal of Finance Vol. 59, No. 1, Fama, E., 1980, Agency problems and the theory of the firm, The Journal of Political Economy 88, April, 288-307 Fama, E., 1991, Efficient Capital Markets: II, Journal of Finance 46, 1575-1617. Fombrun C., 1996, Reputation: Realizing value from corporate image, Boston, MA: Harvard Business Scholl Press. 114

PAGE 122

Gale, D., 1993, Informational capacity and financial collapse, in C. Mayer and X. Vives (eds.), Capital Markets and Financial Intermediation, Cambridge University Press, Cambridge. Gallagher D., 2003, Investment manage r characteristics, strategy, top management changes and fund performance, Accounting and Finance 43 (3), 283-309. Graham, J., Lemmon, M. and Wolf, J., 2002, Does corporate diversification destroy value? Journal of Finance 57, 695-720. Greenbaum, S., Kanatas, G., and Venezia, I., 1989, Equilibrium Loan Pricing under the bank-client relationship, Journal of Banking and Finance 13 (May), 221-235. Grinblatt, Mark, and Sheridan Titman, 1989, Mutual Fund Performance: Analysis of Quarterly Portfolio Holdings, Journal of Business 62, 394-415. Grinblatt, Mark, and Sheridan Titman, 1992, The Persistence of Mutual Fund Performance, Journal of Finance 47, 1977-1984. Grinblatt, Mark, and Sheridan Titman, 1993, Performance Measurement without Benchmarks: An Examination of Mutual Fund Returns, Journal of Business 66, 47-68. Grinblatt and Titman (2002, 2nd ed.; 1997, 1st ed.), Financial Markets & Corporate Strategy, Irwin/McGraw-Hill. Grinblatt, Mark, Sheridan Titman, and Russ Wermers, 1995, Momentum Investment Strategies, Portfolio Performance, and Herding: A Study of Mutual Fund Behavior, American Economic Review 85, 1088-1105. Goriaev, A., 2003, The relative impact of different classification schemes on mutual fund flows, Working Paper, New Economics School. Gruber, Martin J., 1996, Another Puzzle: The Grow th in Actively Managed Mutual Funds, The Journal of Finance 51, 783-810. Guedj I. and J. Papastaikoudi, 2004, Can Mutual Fund Families affect the performance of their funds?, Working Paper, University of Texas at Austin. Harris M., C. Kriebel, and A. Raviv, 1982, Asymmetric information, incentives, and intrafirm resource allocation, Management Science 28, 604620. Harris, R. I. D., and C. Robinson (2002) Foreign Ownership and Productivity in the United Kingdom. Estimates for U.K. Manufacturing Using the ARD, Mimeograph 115

PAGE 123

Hendershott, R., D. Lee and J. G. Tom pkins (2002) Winners and Losers as Financial Service Providers Converge: Evidence from the Financial Services Modernization Act of 1999, The Financial Review, 37, pp. 53. Hu, F., Hall, A. and C. Harvey, 2000, Promotion or Demotion? An Empirical Investigation of the Determinants of Top Mutual Fund Manager Change, Working Paper, Duke University. Hubbard G. and D. Pahlia, 1999, A re-e xamination of the conglomerate merger wave in the 1960s: An internal capital market view. Journal of Finance Ippolito, R., 1992. Consumer reaction to m easures of poor quality: Evidence from the mutual fund industry. Journal of Law and Economics 35, 45 70. James, C., 1983, An analysis of intraindustry differences in the effect of regulation: The case of deposit rate ceilings, Journal of Monetary Economics 12, 417432. Jayaratne, J, and Strahan. P., 1996, Th e FinanceGrowth Nexus: Evidence from Bank Branch Deregulation, Quarterly Journal of Economics August, pp. 639-70. Jayaratne, J, and Strahan. P., 1998, Entr y restrictions, Industry Evolution, and dynamic efficiency: Evidence fr om commercial banking, The Journal of Law and Economics 41, 239-273. Jensen, M., 1972, Capital Mark ets: Theory and Evidence, Bell Journal of Economics and Management Science 4, 357-398. Jensen, M., 1986, Agency costs of free cash flow, corporate finance as antitakeover mechanisms: The recent evidence, Journal of Financial Economics 20, 129152. Jensen M. and W. Meckling, 1976, Theo ry of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3, 305-360. Jensen M. and K.J. Murphy, 1990, CEO IncentivesIts not how much you pay, but how, Harvard Busine ss Review 68, 138-149. Jensen, M., and Murphy, K., 1990. Performance Pay and TopManagement Incentives. Journal of Political Economy 98 (April): 225. Jensen, Michael C. and Richard S. R uback, 1984, The market for corporate control: The scientific evidence, Journal of Financial Economics 11, 5-50. 116

PAGE 124

John, K. and Ofek, E., 1995, Asset sales and increase in focus, Journal of Financial Economics 37, 105-126. Kashyap, A., and Stein J., 1995, The imp act of monetary po licy on bank balance sheets, Carnegie-Rochester Series on Public Policy, 42, 151-195. Kellner, S., and G. Frank Mathewson. 1983. Ent ry, Size Distribution, Scale, and Scope Economies in the Li fe Insurance Industry. Journal of Business 56 (January): 25. Khorana, A., (1996): Top management tur nover: An empirical investigation of mutual fund managers. Journal of Financial Economics vol. 40, 403-427 Khorana, A., 2001, Performance change s following top management turnover: Evidence from open-end mutual funds, Journal of Financial and Quantitative Analysis 36, 371393. Khorana, Ajay, and Henri Servaes, 1999, The Determinants of Mutual Fund Starts, The Review of Financial Studies 12, 1043-1074. Kwan, S. and Laderman, E., 1999, On the portfolio eff ects of financial convergenceA review of the literature, Federal Reserve Bank of San Francisco Economic Review No. 2, 18-31. Kwast, M., 1989, The impact of underwri ting and dealing on bank returns and risk, Journal of Banking and Finance 13, 101-125. Laeven, L., and R. Levine, 2007, Is there a diversification di scount in financial conglomerates? Journal of Financial Economics 85, 331-367. Lamont, O. and Polk, C., 2001, The divers ification discount: Cash flows versus returns, Journal of Finance 56, 1693-1721. Lang, L. and Stulz, R., 1994, Tobins Q corporate diversifi cation, and firm performance, Journal of Political Economy 102, 1246-1280. Lewellen W., 1971, A pure financial ra tionale for the conglomerate merger. Journal of Finance 26, 521. Loughran, T. and Vijh, A., 1997, Do long-term shareholders benefit from corporate acquisitions? Journal of Finance 52, 1765. Macey, J.R., 2000, The Business of Banking: Before and After Gramm-LeachBliley, Journal of Corporation Law Vol. 25, No. 4 (Summer), pp. 691. 117

PAGE 125

Majd, S., and Myers, S. 1987. Valuing th e governments tax cl aim on risky assets. Working Paper no. 1553. Cambridge, Mass.: National Bureau of Economic Research. Malatesta, P., 1983, The wealth effect of merger activit y and the objective functions of merging firms, Journal of Financial Economics 11, 155-181. Mamun,A., Hassan, M.K. and Maroney, N. The wealth and risk effects of the GrammLeachBliley Act (GLBA) on the US banking industry, Journal of Business Finance & Accounting 32 (2005), pp. 351. Mandelker, G., 1974. Risk and return the case of merging firms. Journal of Financial Economics 1, p. 303. Martinez, B., 2002. Adoption costs, age of capital and technological substitution. IRES, Universite catholique de Louvain, DP 2002-24. Massa M., 1998, Are there too many mutual funds? Mutual fund families, market segmentation and financial perfor mance, Working Paper, INSEAD Massa M., 2003, How do family strategi es affect fund performance? When performance-maximization is not the only game in town, Journal of Financial Economics 67, 249-304. McLaughlin, S., 1994, The impact of inte rstate banking and branching reform: Evidence from the states, Current Issues in Economics and Finance, Federal Reserve Bank of New York. Mester, Loretta J. 1987, Multiple Market Contact between Savings and Loans: A Note, Journal of Money, Credit, and Banking 19 (November): 538. Mester, Loretta J., 1993, Efficiency in the Savings and Loan Industry, Journal of Banking and Finance 17 (April): 267. Michaely, R., Thaler, R. and Womack L ., 1995, Price Reactions to Dividend Initiations and Omissions: Overreactions or Drift?, Journal of Finance 50, 573-608. Morck, R., Shleifer, A. and Vishny, R., 1990, Do managerial objectives drive bad acquisition? Journal of Finance 45, 31-48. Morck R., A. Shleifer, and R. Vishny, 1989, Alternative Mechan isms of corporate control, American Economic Review 79, 842-852. 118

PAGE 126

Muphy K., A. Shleifer and R. Vishney, 1989, Income Distribution, Market Size and Industrialization, Quarterly Journal of Economics 103, 537-564. Murphy K., 1999), Executive Compensation, in Orley Ashenfelter and David Card, eds., Handbook of labor economics, Vol. III., Amsterdam: North-Holland, 24852563. Myerson R. B., 1982, Optimal Coordination mechanisms in generalized principalagent problems, Journal of Math ematic Economics 10:1, 67-81. Nanda, V., Narayanan, M. P., Warther, V. A., 2000. Liquidity, investment ability, and mutual fund structure. Journal of Financial Economics 57, 417-43. Nanda, V., Wang, Z. J., Zheng, L. 2004. Family values and the star phenomenon. Review of Financial Studies 17, 667-698. Neale, F.R. and P.P. Peterson (2003), T he Effect of the Gramm-Leach-Bliley Act on the Insurance Industry, Working Pa per (Florida State University). Nippani, S., and Green, K. W. 2002. Th e banking industry after the Riegle-Neal Act: Re-structure and overall performance. The Quarterly Review of Economics and Finance. 42(5):901-909. Palepu, K., 1986. Predicting takeover ta rgets: a methodological and empirical analysis. Journal of Accounting and Economics 8, pp. 3. Palia, D., 1993, The manageri al, regulatory, and financ ial determinants of bank merger premiums, Journal of Industrial Economics 41, 91-102. Peristiani, S., 1997, Do Mergers Improve the X-Efficiency and Scale Efficiency of U.S. Banks? Evidence from the 1980s, Journal of Money, Credit & Banking Vol. 29, 326-337. Perold, Andre F., and Robert S. Salomon, 1991, The Right Assets under Management, Financial Analysts Journal 47, 31-39. Petersen M., 2007, Estimating Standard Errors in Finance Panel Data Sets: Comparing approaches, Working Paper, Kell ogg School of Management, Northwestern University. Pilloff, Steven J., and Anthony M. Santomero. 1998, The Value Effects of Bank Mergers and Acquisitions, In Bank Mergers and Acquisitions edited by Yakov Amihud and Geoffrey Miller, 59. Norwell, Ma ss.: Kluwer Academic Publishers. 119

PAGE 127

Rajan, R., 1996. Commercial Bank Entry in to the Securities Business: A Survey of Theories and Evidence. In: Saunders A & Walter I (eds.) Universal Banking: Financial System Design Reconsidered. Irwin, Chicago. Rau, R., and Vermaelen, T., 1998, Glamour, value, and the post-acquisition performance of acquiring firms. Journal of Financial Economics 49, 223. Rhoades, Stephen A. 1993. The Efficien cy Effects of Horizontal (In-Market) Bank Mergers. Journal of Banking and Finance 17 (April): 411. Rhoades, Stephen A., 1998, The Effici ency Effects of Bank Mergers: An Overview of Case Studies of Nine Mergers. Journal of Banking and Finance 22 (March): 273. Savage, M., 1993, Career mobility and class formation: British banking workers and lower middle classes, Building European Society: Occupational change and Social Mobility in Europe, 196-216. Rosson, P. and Brooks, M. R. 2004. M& As and corporate visual identity: An exploratory study. Corporate Reputation Review, 7(2): 181-193. Santini, D., and Aber, J., 1998, Determinan ts of net money flows to the equity mutual fund industry, Journal of Economics and Business Vol. 50, Issue 5, 419-429. Scharfstein D., Stein, J., 1997. The dark side of internal capital markets: Divisional rent-seeking and in efficient investment. Disc ussion Paper No. 5969, NBER, Cambridge, MA. Scharfstein, D., and Stein, J., 2000, D. Sc harfstein and J. Stein The dark side of internal capital markets: divisional rent-seeking and inefficient investment. Journal of Finance 55 (2000), pp. 2537. Servaes, H., 1991, Tobins Q and the gains from takeovers, Journal of Finance 46, 409-419. Servaes, H., 1996, The value of diversif ication during the conglomerate merger wave, Journal of Finance 51, 1201-1225. Shefrin, H. and M. Statman, 1984, Explaining investor preference for cash dividends, Journal of Financial Economics 13, 253-282. 120

PAGE 128

Shleifer A. and R. Vishney, 1997, A survey of corporate governance, Journal of Finance 52, 737-783. Shleifer, A., and Vishny R., 1989, Ma nagement entrenchment: The case of manager-specific investments, Journal of Financial Economics 25, 123-140. Sirri E. and P. Tufano, 1992, The demand for mutual fund services by individual investors, Working paper, Harvard Business School. Sirri, Erik R., and Peter Tufano, 1998, Costly Search and Mutual Fund Flows, Journal of Finance 53, 1589-1622. Stein, J., 1997, Internal capital mark ets and the competition for corporate resources. Journal of Finance 52 (1997), pp. 111. Stein J., 1998, An adverse selection model of bank asset and liability management with implication for the transmission of monetary policy, RAND Journal of Economics 29, 456-486. Stiglitz, J., 1989, Using Tax Policy to Curb Speculative Short-Term Trading, Papers t2, ColumbiaCenter for Futures Markets. Stiroh, K. and Strahan, P. 2003, The competitive dynamics of competition: Evidence from U.S. banking deregulation, Journal of Money, Credit and Banking 35, 801-828. Stulz, R. 1990. Managerial discretion and optimal financing policies. Journal of Financial Economics 26 (July): 3. Thakor A., 1996, Capital requirements, monetary policy, and aggregate bank lending: Theory and Empirical evidence, Journal of Finance 51, 279-324. Travlos, N.G., 1987, Corporate takeover bi ds, methods of payment, and bidding firms stock return, Journal of Finance 42 943-963. Von Thadden, E., 1998, Asymmetric in formation, bank lending and implicit contracts: The winners curse, worki ng paper, Universite de Lausanne. Wall, L., A. Reichert and S. Mohanty (1993), Deregulation and Opportunities for Commercial Bank Di versification, Economic Review, Federal Re serve Bank of Atlanta (September/October), pp. 1. 121

PAGE 129

Whalen, G., 2000, Trends in Organizati onal Form and Thei r Relationship to Performance: The Case of Foreign Secu rities Subsidiaries of U.S. Banking Organizations, Journal of Financial Services Research 16, 181-238. Wheelock, David C., and Paul W. Wilson. 1996. Technical Progress, Inefficiency, and Productivity Change in U.S. Banking, 1984. Working Paper 94021B. Federal Reserve Bank of St. Louis. Wermers, Russ, 2000, Mutual Fund Perf ormance: An Empirical Decomposition into Stock Picking Talent, Style, Transactions Costs, and Expenses, Journal of Finance 55, 1655-1695. Winton, Andrew J. 2000. Dont Put All Y our Eggs in One Basket? Diversification and Specialization in Lending. Working Pape rsFinancial Institutions Center at the Wharton School, 1-43. Yermack D., 1995, Do corporations aw ard CEO stock options effectively?, Journal of Finance and Economics 39, 237-269. Yildrim,H., Kwag S.W., and Collins M., 2006, An Examination of the equity market response to the Gramm-Leach-Bl iley Act Across Commercial Banking, Investment Banking, and Insurance Firms, Journal of Business Finance and Accounting 33, 1629-1649. Zheng, L., 1999, Is money smart? A study of mutual fund investors' fund selection ability, Journal of Finance 54, 901-933. 122

PAGE 130

123 About the Author Lonnie Lashawn Bryant is a Ph.D. candidate in the Depa rtment of Finance at the University of South Florida. His primary area of research is corporate finance, with focus in the financial services industry, mutual funds, and mergers and acquisitions (M&As). He has presented his papers at national ac ademic conferences including the Financial Management Association conference and Sout hern Finance Association conference. He holds a Masters of Business Administration with a concentration in finance (2002) from the University of North Carolina at Chapel Hill, and a Bachelors degree in Electrical Engineering (1998) from the Florida Agricultural and Mechanical University. Lonnie has the distinction of being a Florida Education Fund McKnight Scholar.