Market Concentration, Vertical Integration and Bancassurance:
Consolidation and the “Insurance Middleman”
ARIA Annual Meeting 2006, Washington, D.C.
By
J. David Cummins a
Mary A. Weiss b
Xiaoying Xie c
August 4, 2006
________________________________________________________________________
a
Insurance and Risk Management Department, the Wharton School, University of Pennsylvania,
3620 Locust Walk, Philadelphia, PA19104
Email address: [email protected]
b
Department of Risk, Insurance, and Healthcare Management, The Fox School of Business,
Temple University, 1301 Cecil B. Moore Ave., Philadelphia, PA 19122
Email address: [email protected]
c
Department of Finance, College of Business and Economics, California State University at
Fullerton, 2600 E. Nutwood Ave., Fullerton, CA 92831
Email address: [email protected]
©2006 This paper is preliminary and confidential and may not be quoted or cited without the
written permission of the authors.
Market Concentration, Vertical Integration and Bancassurance:
Consolidation and the “Insurance Middleman”
1. Introduction
The U.S. agent/broker system of insurance distribution has experienced dramatic
changes since the mid-1990s.
Merger and acquisition (M&A) activities were
unprecedented during this period, and these activities have not been halted by the
investigations of insurance brokers that have surfaced since October 2004. M&A deals
involving insurance agents/brokers are primarily dominated by three types of transactions:
M&As between insurance agents/brokers, vertical integration between insurance carriers
and insurance agents/brokers, and bancassurance, in which banks acquire insurance
agents/brokers.1
What drives M&A activities in the U.S. insurance agent/broker industry? One
factor is changes in financial regulation. In 1985, the Office of the Comptroller and
Currency (OCC) granted national banks the right to sell fixed-rate annuities. In 1990, the
right to sell variable-rate annuities was granted. In 1995, the Supreme Court upheld the
OCC’s decisions, and ruled that annuities are not insurance products, but investment
products. In 1996, the Supreme Court further ruled that nationally chartered banks may
sell insurance. These rulings form a contestable market for insurance carriers and large
insurance agents/brokers (Carow 2001). Bank entry into insurance markets led to an
increase in acquisitions of insurance distributors by commercial banks, bank holding
companies, thrifts, and credit unions. In 1999, the Gramm-Leach-Bliley Act removed the
1
Note that the term “vertical integration” as used here does not necessarily imply an exclusive agency
distribution system. It is possible for an insurance carrier to acquire an independent agency or broker firm,
and allow the acquired firm to continue as an independent entity.
2
remaining regulatory barriers between banks and insurers, facilitating the further
development of bancassurance in the U.S.
Changes in the insurance market have also influenced M&A trends among
insurance intermediaries. The intermediary market is restructuring in response to
competition from bank-owned agencies/brokerages and the sensitivity of intermediary
profits to the underwriting cycle. During the soft market of 1997 to 1999, the number of
M&As in the agent/broker market (especially agent/broker to agent/broker transactions)
increased dramatically. Many small agencies/brokerages have been “crowded” out of the
market or been acquired by large public agencies/brokerages, leading to a more
concentrated agent/broker market.2
Meanwhile, the functioning of the agent/broker industry is being transformed.
Agencies/brokerages have adapted to the technology changes in the 1990s by investing in
new technologies and providing sophisticated services to insurers and reinsurers. As a
result, the role of agent/broker has evolved from simply matching clients and insurers to
one of service provider, which in turn is dramatically changing the insurance industry
itself (Swiss Re, Sigma, 2004; Cummins and Doherty, 2005). Insurance companies,
facing competition from banks and a more powerful middleman market, are seeking ways
to maintain their profit margins. The acquisition of agencies/brokerages therefore
provides a fast and efficient way for these carriers to compete with banks and to compete
against the more powerful bargaining power of large intermediaries.
Despite these changes in the insurance intermediary industry, no direct evidence
has been provided on how these changes affect the economic value of banks, insurers,
2
For example, by 2004, Marsh and Aon, two leading commercial insurance brokers, accounted for over
one half of the industry’s revenues. See Swiss Re, Sigma, 2004.
3
and public agencies/brokerages. The purpose of this study is to directly examine the
wealth gains of U.S. agent/broker M&A transactions.3 Given the middleman status of
insurance agents/brokers, brokers and agents involved in M&As provide valuable
information for testing the effects of focusing transactions vs. diversifying transactions on
firm value. Among the diversifying transactions, we can further distinguish among
vertical integration, bancassurance, and purely diversifying transactions.4
We ask three questions in the paper: First, does the market concentration resulting
from M&As create value for the large public agents/brokers? Second, does the market
have a positive reaction to bancassurance, especially to banks’ expansion into insurance
business through the acquisition of insurance agents/brokers? Third, does the market
react similarly to vertical integration and bancassurance in a contestable market
environment? That is, does the market believe the economic rent will be the same for
banks and insurance carriers if both of them sell insurance though a middleman?
This paper contributes to the literature in the following ways. First, this is the first
paper examining the wealth gains of M&As of insurance agents/brokers by using an upto-date sample. The sample period is 1990-2005, and includes all of the important events
regarding financial integration and changes in the agent/broker industry (including initial
reaction to the Spitzer investigation). Second, it directly compares the wealth effects of
vertical integration and bancassurance in the context of a contestable market, thereby
providing a direct test of the impact of financial integration. Third, it isolates the bank-
3
In this paper, “agent/broker” refers to all insurance agency, brokerage, and insurance service firms.
In this paper, vertical integration is defined as M&A transactions between insurance carriers and
insurance agency, broker or service firms. Bancassurance is defined as M&A transactions between banks
and insurance agency, broker or service firms. Pure diversifying transactions are defined as M&A
transactions between insurance agency, broker or service firms and firms other than insurance carriers,
banks, other financial service firms and business service firms.
4
4
agent/broker transactions from the bank-insurance carrier transactions, providing a direct
test of wealth effects associated with the development of bancassurance in the United
States.
We are not aware of any prior papers specifically analyzing the market value
effects of M&As involving U.S. insurance brokers or agents. Carow (2001a) conducted
an event study to determine the market value effect on banks and insurers of rulings that
allowed banks to enter the annuity and insurance business. Although he found that
insurers in general lost value as a result of these events, insurance companies associated
with an agent/broker distribution system earned positive abnormal returns. Carow (2001b)
studied the Citicorp-Travelers Group merger, and found that large banks and life insurers
benefited from the event. This raises a question concerning the role that the insurance
distribution system plays in the process of financial integration and in competition
between insurance carriers and banks. Cummins and Weiss (2004) included insurance
agency transactions in their analysis of European insurance M&As but did not separately
analyze transactions involving agencies in comparison with those involving insurance
companies. Fields, Fraser and Kolari (2005) study M&As between banks and insurance
companies in Europe and the United States. They find a wealth gain for bidders in these
activities in general, but their study does not distinguish between transactions involving
insurance agents/brokers and those involving insurance companies. Thus, our study adds
value to the existing literature by specifically focusing on M&As involving insurance
intermediaries rather than mergers of insurance companies or banks.
The structure of the paper is as follows. In the second section we will present
background information on M&As involving insurance agents/brokers, the development
5
of bancassurance in the United States, and develop the hypotheses tested in this paper. In
the third section we will discuss the data collection process and methodology. The paper
then proceeds to the wealth gain analysis for different types of M&A transactions.
Section 5 will present the cross-sectional regression analysis results, and section 6
provides a concluding discussion of the results of this research.
2. Background, literature review and hypotheses
2.1 Agents/Broker industry in the United States
In the United States, insurance products are sold mainly through agents, brokers
or direct channels. There are thousands of insurance agencies and brokerage firms that
are active in the market.5 The majority of these firms are small in size. Approximately 30
percent of agents and brokers are self-employed. An agent has to maintain certain volume
and production requirements to qualify to represent an insurance carrier. For example,
smaller mutual insurers may require agents to produce $25,000 to $50,000 worth of sales
per year. Large national insurers generally require agents to place $500,000 to $1 million
worth of business in annual sales.6
The agency/broker distribution system developed over more than a century and
faced no dramatic challenges until the mid 1970s. Ever since then, the system has been
challenged by the emergence of direct selling methods, the rapid development of new
technologies, rising insurance costs, increased governmental regulation on insurers’ profit
5
Independent Insurance Agents & Broker of America (IIABA) for example represents a network of more
than 300,000 agents, brokers and their employees nationally.
6
Source: Answer.com, 2005, Insurance agents, brokers, and service,
http://www.answers.com/topic/insurance-agents-brokers-and-service.
6
margins, 7 and a changing regulatory environment (e.g., removal of barriers to banks’
entry into the distribution and underwriting of insurance business). Insurance companies
have been forced to find ways to cut expenses in order to maintain their profit margins,
including expenses related to agents, brokers and services. As a result, the profit margin
of agents/brokers has been reduced due to lower commissions,8 increased operating costs
and escalating competition from new entrants.
Insurance agents/brokers have responded to these challenges primarily in two
ways:
reinventing themselves as important service providers to insurers and their
corporate clients and through mergers and acquisitions.
These two options are not
mutually exclusive. Both strategies are explained more fully below.
As indicated by Sigma (2004), in the past two decades, the role of agents/brokers
has gone far beyond being a market matcher, to being a service provider to insurance
companies and customers. Insurance agents/brokers may serve as risk analysts for their
customers, help them choose appropriate coverages and coverage provisions (e.g.,
deductibles and limits), set up self-insurance plans, establish captive solutions when
appropriate, and even provide employee benefits solutions. Meanwhile, they also provide
services to insurers. In addition to placing business with insurance companies,
agents/brokers serve as risk screeners, matching the right business with the appropriate
insurer.
This is important for insurers that rely on the risk screening ability of
agents/brokers to expand and grow.
7
For example, Proposition 103 in California in 1999 required insurers to reduce automobile insurance rates
by 20 percent, and return money to policyholders.
8
For example, the customary commission for independent agents dropped from 13.4 percent in 1993 to
11.33 percent in 2003. See Cummins and Xie (2005).
7
The agent/broker also has to find ways to maintain a relatively stable profit level
in a market with frequent hard and soft market shifts. Profits from distributing insurance
products are easier to earn if a hard market occurs9 and tend to decline dramatically in a
soft market. In the latter case, firms need to rely on fee income from providing additional
services in order to maintain their profit. Provision of services has also improved through
technology. Most agents/brokers have adopted internet technology to expand their
customer reach and to offer faster and more convenient service to their clients.
Mergers and acquisitions provide an alternative method for agents/brokers to deal
with the challenges they face, especially new competition from banks and other financial
firms. Large agents/brokers expand and increase their market power through acquiring
numerous domestic, and/or international independent agents, managing general agents,
and wholesale brokerage businesses. They have expanded both geographically and across
business lines. A large number of insurance agents/brokers have begun to diversify across
business lines by offering life, property/casualty, and/or health and disability insurance.
The restructuring in the agent/broker industry has also affected the strategy of
insurers. Insurance companies responded to the increased competition, excessive
underwriting capacity and lower profitability in the 1990s in a similar way to
agents/brokers. Insurers, in both the property-liability and life-health insurance industries
have experienced an unprecedented wave of mergers and acquisitions in order to focus on
their core business, thereby keeping up with competition and maintaining profit margins.
The larger size of insurers post M&As have put them in a better position to do business
with large agents/brokers. Some insurers also started to acquire agents/brokers to
9
However, in a hard market, because the price of insurance is higher, agents/brokers might lose some
customers who seek alternative, less expensive ways to transfer their risks (e.g., through captives).
8
internalize distribution costs (although this was more common in the property-liability
than in the life-health insurance industry). Increasing competition from banks and other
financial firms is speeding up this process further.
2.2. Development of bancassurance in the Unites States
“Bancassurance” first appeared in France after 1980. It originally referred to the
sale of insurance products through banks’ distribution channels. Bancassurance,
especially the distribution of life insurance products by banks, is well established in most
developed countries in Europe. For example, in France, bancassurance accounted for 35
percent of life insurance premiums and 7 percent of property insurance premiums in 2000;
in Spain, bancassurance represented over 65 percent of life insurance premium income in
2001 (Durand, 2003).
The development of bancassurance is closely related to the
regulatory climate of a country, helping to explain differences in its importance across
different countries.
Bancassurance in the U.S. took a much longer time to develop than in Europe.
After the big economic crash in 1929, the federal government passed a series of acts such
as the “Glass – Steagall Act” in 1933 and the “Bank Holding Company Act” in 1956 to
prohibit any commercial bank from owning brokerages or insurers in order to reduce
systematic financial risks. It took nearly half of a century before banks established their
right to sell/underwrite annuity and insurance products. This occurred in stages over time
(as summarized in Carow (2001). As early as April 1985, the OCC permitted national
banks to sell variable annuities. In February 1990, national banks were granted the right
to sell fixed annuity products by the OCC. Then on January 18, 1995, the U.S. Supreme
Court upheld the OCC ruling that annuities are not insurance products, but rather
9
investment products, and could be sold by banks. In August 1986, the OCC, based on
Section 92 of the National Banking Act of 1916, ruled that if a national bank or its branch
was located in a small town (with population of 5,000 or less), then it might sell
insurance to its existing and potential customers anywhere in the U.S. and a 1993 ruling
by the U.S. Supreme Court implied the OCC’s decision to allow banks to sell annuities
was valid. Then in March 1996, the U.S. Supreme Court ruled that nationally chartered
banks might sell insurance products.10 In 1998, the merger of Citi-Corp and Travelers
group started to erode the barrier between banks and insurance underwriting. In 1999,
with the passage of Graham-Leach-Bliley (GLB), the barriers between banking and
provision of other financial services, including insurance services, was eliminated.
Banks can be involved in insurance in two ways: distributing and underwriting
insurance. And the corresponding risks are very different. Banks tend to view insurance
distribution as a less risky business segment with the potential to generate stable and
predictable returns, while they view insurance underwriting as a more risky segment with
low margins. As a result, the development of bancassurance in the U.S. is marked by
mergers and acquisitions (M&As) of insurance agencies and brokers by banks.
There are several reasons why banks may find the agency/broker marketplace
attractive. The first reason is to diversify revenues. By entering insurance and expanding
non-interest-income (i.e., fee income), banks can diversify some of their net interest rate
margin risk without sacrificing profitability. The second reason is economies of scope.
Banks may be able to sell insurance at a lower cost than the independent, small
agencies/brokerages by targeting its existing, sizable customer base; taking advantage of
their significant brand awareness within their geographic regions (thereby lowering
10
For a more detailed description, see Carow (2001).
10
advertising costs); and using their existing employees as an insurance sales force. Cost
economies of scope might exist also with respect to transaction costs (e.g., direct billing
and collection of premium payments).
The third reason is to expand and protect their
existing customer base. Through acquiring local insurance agencies, banks might expand
their customer base to policyholders that are not the current customers of the bank. In
addition, by diversifying their business into insurance, banks might fully exploit their
brand name and provide “one-stop banking” for its existing customers.
The latter
increases the number of products per customer and reduces the possibility of losing
customers to competitors.
The entrance of banks into insurance might provide deeper penetration into the
insurance market, especially the middle-income market. Traditionally agents/brokers did
not serve this segment well due to lack of information and low margins on products sold
to these customers. Banks, in contrast, can easily reach the middle-income market
because they comprise the majority of bank customers, and customer information is
available to the bank at a lower cost.
A bank, when selling insurance, also competes directly with agents/brokers for
customers, especially commercial customers. This is because, first, in the U.S., the
personal insurance market is a standardized, competitive market with low profit margins,
while the commercial insurance market is an area requiring more expertise with
correspondingly higher profit margins. Second, because of the one stop shopping
advantage, banks can provide a package of services to its commercial customers,
including insurance purchases. Moreover, the brand name of the bank in both the local
and regional area might provide the bank an advantage over local, smaller
11
agencies/brokerages. As a bank develops its expertise in distributing insurance products
over time, it might become a threat even to the large, established, brand name
agencies/brokers.
Sales of insurance by banks might also have an impact on the insurance company.
Direct
competition
agencies/brokerages
from
and
banks
accelerates
consequently,
insurers
concentration
are
facing
among
existing
more
powerful
agencies/brokerages. Also, the bank itself might become an insurance competitor,
especially banks that underwrite and then distribute their products through direct selling
or the independent agency system. The competition from banks became potentially
stronger since the passage of the Gramm-Leach-Bliley Act on October 22, 1999.
Usually, banks place the insurance they have produced in several ways. They may
place the business with a wholly controlled insurance subsidiary; but, in most cases, they
sign a definitive agreement with some insurers and effectively operate as brokers. In the
former case, banks become a competitor of direct writers. In the latter case, banks will
mostly place their business with insurers that mainly use the brokerage distribution
system. In a mature insurance market, this might lead to a smaller market share for
insurers using the independent agency or direct selling systems.
2.3. Literature review and hypotheses
Studies exist concerning diversification of banks into financial services, especially
insurance, but none of these focus specifically on the interactions of the agency/broker
industry with the insurance industry and bancassurance. Rose and Smith (1995) studied
BHCs’ expansion into the limited number of insurance lines allowed during the period
1974 to 1990 and found banks experienced positive abnormal returns, especially after
12
1982. Carow (2001a) conducted an event study to determine the market value effect on
banks and insurers of rulings that allowed banks to enter the annuity and insurance
business. He found that insurers in general lost value as a result of these events; banks
neither lost nor gained value. Carow (2001b) studied the Citi-Travelers group merger,
and found that large banks and life insurers benefited from the event. Lown et al (2000)
conducted a simulation to determine whether potential post-GLB diversification might
improve the risk-return trade-off faced by financial companies. They found that simulated
mergers between bank holding companies (BHCs) and life insurers produced firms that
were less risky (and no less profitable), while simulated mergers between BHCs and
property-casualty insurers or securities firms reduced BHCs’ risks only slightly. Fields,
Fraser and Kolari (2005) studied M&As between banks and insurance companies in
Europe and the United States. They found both bank and insurer acquirers experienced
wealth gains, but the study does not differentiate transactions involving insurance
agents/brokers from those involving insurance companies. Since the former represents
the major form of bancassurance in the U.S., it is important to study agent/broker M&As
separately against the broader background of the insurance agent/broker industry.
Further, it is important to distinguish among target agency/broker sectors (i.e., life-health,
property-casualty) to identify where synergies exist.
Banks’ entry into the insurance market, especially insurance distribution can be
viewed as an “industry shock” to the agent/broker industry (Mitchell and Muherin 1996).
This shock, along with other changes in the industry (such as the soft market environment
and availability of more advanced and expensive distribution technology) has been at
least partly responsible for the consolidation taking place between agencies/brokers. The
13
consolidation wave has created a more concentrated agency/broker market than ever
before. If a concentrated market leads to more market power for the acquiring
agencies/brokers, we might observe positive abnormal stock returns for agencies/brokers
involved in M&As with other agencies/brokers. On the other hand, if the purpose of
M&As is to allow acquirers/targets to be better able to compete effectively with banks so
that agencies/brokerages might maintain their profit margins, then we should not observe
any change in stock returns of these firms.
Therefore, the first hypothesis answers the following question: does the market
concentration resulting from M&As in the insurance agent/broker industry create value
for the large public agents/brokers? Overall, we believe that the agency/broker industry
is competitive. In a competitive industry abnormal returns are not possible. This leads to
hypothesis 1:
Null Hypothesis 1: The average abnormal stock returns of the acquiring agents/brokers
are insignificantly different from zero around the M&A event date.
Although allegations and some evidence of anti-competitive behavior in the
agency/broker industry surfaced in 2004, these allegations were primarily focused on a
few mega-brokers. For the bulk of the agent/broker industry, there is no evidence of anticompetitive behavior. Nevertheless, it may be worthwhile to examine the stock returns
of these mega-brokers as they pursued their M&A activity, and this issue is taken up later.
Also, since agents/brokers in both the property-liability and life-health insurance
industries faced similar shocks (albeit to different extents) we believe that Null
hypothesis 1 will hold up regardless of the target agents/brokers’ industry sector.
14
The second hypothesis in the paper answers the question: does the market have a
positive reaction to bancassurance, especially to banks’ expansion into insurance business
through the acquisition of insurance agents/brokers? As suggested by contestable market
theory (Stigler 1971), the entry of banks in insurance operations will enhance competition
in the insurance market, reducing any existing economic rents for insurers. However,
whether a bank can earn an excess return depends on the initial costs of the bank’s entry
into insurance, the cost advantages and disadvantages of banks relative to existing
insurers/agencies/brokers, and the synergies existing between banking and insurance.
Each of these considerations is discussed below.
The initial entry costs into the insurance industry are likely to be small for banks.
Banks had invested heavily already in customer databases, data processing equipment,
and their branches before their entry into insurance. Additional insurance-related costs
are likely to center on building customer relationships and training/hiring employees to
cross-sell insurance products. Through acquiring insurance agencies/brokerages, banks
might quickly build relationships and obtain essential human resources for distributing
insurance.
Banks might have a cost advantage in selling insurance products if cost scope
economies exist between banking and insurance distribution. In addition, if banks’
expansion into insurance helps to diversify banks’ interest rate risk, revenue scope
economies may be realized. If any additional revenue associated with diversification
outweighs the costs of diversification, then positive abnormal returns may occur when
banks acquire agencies/brokerages. Moreover, even if banks are not successful in
generating economic rents by taking away existing business from insurance
15
agencies/brokers, banks might still be able to benefit from M&A transactions if they lead
to deeper insurance market penetration (e.g., by selling insurance products to the
relatively untapped middle market).
This leads to hypothesis 2, which is stated in its null form:
Null Hypothesis 2: The average abnormal stock returns of banks are insignificantly
different from zero around the M&A event date.
Based on the discussion above, we believe that null hypothesis 2 is not likely to be
upheld.
Recall that, beginning in 1999 with the passage of GLB, banks obtained the
additional right to underwrite insurance. This increased flexibility is likely to have
enhanced entry into insurance for banks. Therefore, we hypothesize that the market
should react more positively to M&As of banks and insurance agencies/brokerages after
the passage of GLB. This hypothesis, in its null form, can be stated as
Null Hypothesis 2.1: The average abnormal stock returns of banks involved in M&As
after the GLB Act are insignificantly different from those involved in M&As before the
GLB Act, and are insignificantly different from zero around the M&A event date.
We believe that null hypothesis 2.1 is not likely to be supported and that the passage of
GLB created value for banks acquiring agencies/brokerages after 1999.
Given the existing experience with bancassurance in Europe (Falautano and
Marsiglia 2003) and the studies projecting synergies for bancassurance around the CitiTravelers merger and GLB Act, we provide further hypotheses related to the target
agency/broker sector. First, we argue that a bank’s acquisition of an existing life-health
insurance agency/brokerage should produce positive abnormal stock returns because of
16
the bank’s ability to better penetrate the middle-class market and the bank’s ability to sell
at lower costs. Second, we predict that a bank’s acquisition of an existing propertycasualty insurance agency/brokerage will have an ambiguous effect on the bank’s stock
returns. This is because there is more limited room for growth in the highly competitive
personal property-liability lines. Although one might argue that banks have the trust of
their commercial customers, it is unclear whether distributing commercial propertyliability products can produce higher than normal returns. In addition, commercial
property-liability insurance distribution is more service intensive, and it is unclear
whether banks have developed the specific skill set necessary to provide these services.
However, if banks have developed expertise comparable with agencies/brokerages and
their risk/return performance improves from the fee income generated from selling
property-liability insurance, abnormal returns might reasonably occur.
Null Hypothesis 2.2: The average abnormal stock returns of banks that acquire lifehealth insurance agencies/brokerages are insignificantly different from those that
acquire other types of agencies/brokerages, and are insignificantly different from zero
around the M&A event date.
We have no clear expectation about the results of this hypothesis test.
The third question asked in this paper is, does the market react similarly to
vertical integration and bancassurance in a contestable market environment? By acquiring
an agency/brokerage, an insurer and a bank both obtain the ability to sell insurance
through a middleman. The difference is that insurers, in most cases, also underwrite the
business, while banks, in most cases in the U.S., place the business with an insurer
willing to underwrite the business. Insurers, whose primary function it is to underwrite
17
risks, are generally interested in underwriting a very large volume of business. Banks’
entry into insurance provides both opportunities and challenges for insurers. For insurers
that traditionally use the brokerage distribution system, bancassurance might mean more
business for them; however, this does not necessarily mean higher profits since banks
usually have more bargaining power than local insurance agents/brokers. For insurers that
use a direct selling system and/or independent agents, the rise of bancassurance might
result in losing part of the market to banks (or to insurers that have business relationships
with banks). For all types of insurers, the increased competition resulting from the
accelerated number of M&As involving insurance agencies/brokerages might result in
lower profit margins.
In this operating environment, vertical integration between insurers and
agents/brokers (mainly insurers’ acquisition of agencies/brokerages) is a reasonable
response of insurers to the shock affecting the agent/broker industry. By internalizing
agent/broker costs, insurers may gain better control of operating expenses and be better
able to stabilize their business than if no integration between insurers and
agencies/brokerages occurred. If the shock is the primary reason for adopting a vertical
integration strategy, then we should not observe an excess return for insurers that acquire
agencies/brokerages. However, if vertical integration is a better operating strategy than
other types of integration regardless of whether an industry shock exists or not, we should
expect such transactions to bring positive effects to the insurers.
Null Hypothesis 3: The average abnormal stock returns of the insurers that adopt a
vertical integration strategy are insignificantly different from zero around the M&A event
date.
18
We have no priors on the result of this hypothesis test.
3. Data and methodology
3.1 Data selection
This paper analyzes mergers and acquisitions in the U.S. insurance agent/broker
industry from 1990 to 2005. This sample period includes most of the important events in
the development of bancassurance and the changes in the agent/broker industry in the
U.S., such as the important rulings from OCC and the Supreme Court, the merger of CitiTravelers, the passage of GLB, and the Spitzer’s lawsuits against large brokers.
The merger and acquisition data are primarily from two sources: SDC Platinum
(covering the period 1990-1996) maintained by Thomson Financial and the SNL
DataSource-Merger and Acquisition database for the insurance broker industry
maintained by SNL Financial LC (covering the period 1997-2005). We require that at
least one of the partners --acquirer or target-- be an insurance agency or broker.11 The
reason for using two data sources is that SDC Platinum’s SIC code information is
sometimes inaccurate. SNL Datasource provides more complete and accurate information
than SDC; however, it only reports on transactions since 1997. Therefore we combine the
transactions from the two data sources12 and verify the information for every transaction
from the corresponding Factiva reports for the transactions.
11
The overwhelming majority of transactions involve insurance agents/brokers as targets. There are only
six cases where the target in these transactions is not an insurance agent/broker and stock data are available.
For these observations, three involved property-liability insurers as targets, one involved a bank as a target
and one involved a life insurer as a target.
12
At the time of this work, the 2005 transactions from SNL were not available yet (except for the aggregate
industry statistics used in Table 1). The paper will be updated to include additional transactions from the
SNL DataSource.
19
Table 1 presents the aggregate M&A transactions in the U.S. agent/broker
industry during the period 1990-2005. In total, 2172 transactions were reported during
our sample period, with an aggregated reported value of $22.7 billion.13 There is a time
trend present with respect to these transactions. The pace of M&A activities accelerated
since 1998, coinciding with the removal of barriers between bank and insurers, and
continued until 2004. The number of transactions slowed down a bit in 2005, perhaps
because of Spitzer’s investigation of insurance broker operations beginning in October
2004. The reported transaction values were largest for the year 1996 through 1998.
To study the wealth gains of the parties involved in the M&A transactions, we
require that the transactions used in our analysis be completed and that they resulted in a
change of control in the targets. Since we focus on analyzing the stock market reaction to
the events, we further require the acquirer in the transaction to be a listed company.14
With these requirements, 1040 transactions were identified for potential analysis.15 Table
2 provides the number of transactions for each acquirer-target category included in the
sample. It indicates that most transactions involve agencies/brokerages purchasing other
agencies/brokerages or banks acquiring agencies/brokerages.
The majority of
agencies/brokerages acquired are property-casualty insurance agencies/brokerages. (Data
indicating the agent/broker insurance sector were obtained from SNL DataSource and
from Factiva reports.)
13
Approximately two-thirds of the transactions reported in the data sources did not disclose the transaction
values and the terms.
14
We also looked at the public targets; however, the overwhelmingly majority of targets are private firms,
and only a few are publicly traded with stock data available. As a result, we do not report on target wealth
gains in this paper.
15
These 1,040 transactions include all public acquirers that potentially have stock data available. In later
analyses, some of these transactions were dropped because stock data were not available. Stock data were
unavailable because the acquirers were traded OTC, because the acquirers’ stock data were not available
before the event date, or because the acquirers’ stock data were not available until after the event date.
20
Table 3 reports the time trend for each type of transaction. Purchases of
agencies/brokerages by other agencies/brokerages (hereafter referred to as focusing
transactions) were more frequent than any other type of transaction in the early 1990s,
and the number of these transactions increased dramatically since 1997. Purchases of
agencies/brokerages by banks (hereafter referred to as bancassurance transactions) were
relatively rare in the early 1990s, and started to grow from the mid 1990s, especially after
1998. The pace slowed down a bit since 2004, probably due to fierce competition for
quality agencies from public brokers (Wepler et al, 2004). It is interesting to note that the
changes in the agent/broker industry attracted more property-casualty insurers into
M&As than life-health insurers. Yet, the number of M&As between insurers and
insurance agencies/brokerages (hereafter referred to as vertical integration) is much
smaller than for focusing and bancassurance transactions. For both property-casualty and
life-health insurers, transactions were concentrated in 1997 to 1999, but there were much
fewer transactions involving life-health insurers. In addition to banks and insurers, there
are other types of financial services firms that showed an interest in entering the
insurance agent/broker industry, including health services firms, business services firms,
and investment services firms. They became more involved in insurance agent/broker
transactions since the mid 1990s.
For each merger and acquisition, we also collected the following data. Additional
transaction information such as announcement date, completion date, payment method,
transaction value, and percentage of stock acquired were obtained from SDC Platinum
and SNL DataSource for the relevant years. Stock market data were obtained from
Center for Research in Security Prices (CRSP). Acquirers’ financial data were obtained
21
from COMPUSTAT.
Since banks and insurers have different characteristics, we
collected additional information for them. For acquiring banks, we tried to identify
whether the bank already owned an insurance subsidiary or insurance agent/broker
subsidiary before the M&A transaction. The relevant information is from the Factiva
report for the transaction and from the Directory of Corporate Affiliations, when
available. For insurers, distribution system information and rating information before the
M&A transaction were obtained from Best's Key Rating Guide.
Table 4 presents summary statistics for all acquirers that had stock information
available to do the analysis. In total, 980 public acquirers are present in our analysis.
Among them, 445 are agencies/brokerages, 361 are banks, 79 are property-casualty
insurers, 29 are life-health insurers, and 66 are other types of services firms. Among the
bancassurance transactions, 126 transactions involved national banks, and 235
transactions involved state banks. About 282 banks owned an insurance subsidiary or
insurance agent/broker subsidiary before the M&A transaction. The mean bank size
before the transaction was $7,007 million. For the total 108 vertical integration
transactions, 86 insurers had distribution system information available. Most of them
used the independent agent distribution system (52 property-casualty insurers and 13 lifehealth insurers), and a few were direct writers. The majority of insurers involved in these
transactions had a B (very good) rating or higher, with only 8 having vulnerable ratings.
The average size of equity for property-casualty insurers prior to an M&A transaction
was $1,799 million, and the average size was $3,947 million for life-health insurers. The
average equity size for the agent/broker acquirers was $1,917.
22
Sources of financing for the M&A transactions varied. Cash was the primary
financing source for 111 transactions; 122 transactions were financed by stock, and 101
were financed by a mixture of stock and cash and other methods. In 646 of the 980
transactions, the financing method was not disclosed. Tobin’s Q was also calculated for
each acquirer in our sample, and the average Tobin’s Q value for all acquirers was 1.65.
The Tobin’s Q value of public agent/broker acquirers was 2.2, while it was 1.11, 1.27,
1.26, and 2.57 for banks, property-casualty insurers, life-health insurers, and other
acquirers, respectively. An ANOVA analysis shows that the Tobin’s Q values for public
agent/broker acquirers and other acquirers are significantly higher than those for bank,
property-casualty insurers and life-health insurers in the sample, and Tobin’s Q value for
the latter three categories are not significantly different from each other.
3.2. Event study methodology
Because the acquirers in the analysis are from different industries, showing
significant differences in size and market to book values, we adopted the Fama-French
time-series model to estimate wealth changes for the acquirers at the time of
announcement of acquisitions (Fama and French, 1993). In addition to the market index,
two more factors, a high-minus-low market-to-book ratio factor, and a small-minus-big
market capitalization factor are added to the model to estimate acquirers’ excess stock
returns.16 The model specification is as follows,
R jt = α j + β j Rmt + s j SMBt + h j HMLt + ε jt
(1)
where
16
The market model (MacKinlay 1997) and market adjusted model were estimated also as a robustness test.
The results are consistent with that of the three factor Fama-French model with the latter model showing
more significant results.
23
R jt :
the return on jth stock on day t;
Rmt :
the CRSP equally weighted stock index of returns including
dividends as a proxy of market index;
SMBt :
the average return on small market-capitalization portfolios minus
the
average return on three large market-capitalization portfolios;
HMLt :
minus the
the average return on two high book-to-market equity portfolios
average return on two low book-to-market equity portfolios;
α j , β j , s and h j :
j
OLS parameters for stock j, representing the idiosyncratic return,
beta
coefficient of stock j, sensitivity of stock j to the difference
between small and large stock returns, sensitivity of stock j to the
difference between value and growth stock returns, respectively;
ε jt :
residual for stock j on day t,
It is assumed ε jt is i.i.d. through time t and is jointly multivariate normal across stocks,
with E( ε jt ) = 0, VAR ( ε jt ) = σ ε2j . The parameters α j , β j , s j and h j are estimated for
each stock j in our sample using pre-event returns for a 150 trading day period (starting
from 11 days before the announcement date).
The abnormal return on day t in the event window (e.g., if t is within the window
τ 1 to τ 2 ) for stock j is then defined as:
AR jt = R jt − αˆ j − βˆ j Rmt − s j SMBt − h j HMLt
(2)
The most important value used in this study is cumulative abnormal returns
(CAR), which is calculated as,
τ2
CAR j (τ 1 ,τ 2 ) = ∑ AR jt
(3)
t =τ1
24
To test the overall impact of the events, we aggregate abnormal returns both
through time and across securities. The value we generally report is the mean CAR ( CAR )
across all N securities over the event window:
CAR (τ 1 ,τ 2 ) =
1
N
N
∑ CAR (τ ,τ
j =1
j
1
2
)
(4)
The results are estimated by Eventus, and three test statistics are reported in this
paper, the calendar time t test, the portfolio time-series t test and the generalized sign Z
test. Details concerning the test statistics can be found in Cowan (2005).
4. Wealth gains analysis
4.1. Overall wealth effects of acquirers
Table 5 presents the CARs over various event windows for acquirers from the
Fama-French model estimation. The CARs by acquirer industry sector are reported also.
Since the most significant results appear in the (0, +1) window, our analysis will focus on
this two-day window.
We find that acquirers during the sample period 1990-2005 on average earned a
positive 0.18 percent abnormal return around the announcement date in the (0, +1)
window, and this result is significant at the 5 percent level under the portfolio time-series
t test . Given the relatively large size of acquirers (average $ 4.3 billion), this result is
very important in terms of economic value.
Examining wealth gains by acquirer industry type allows us to determine whether
wealth gains vary among focusing transactions, bancassurance and vertical integration.
The public agent/broker acquirers (focusing transactions) on average earned 0.02 percent
abnormal returns, but this result is not statistically significant. This provides support for
25
our Null hypothesis 1, which states that acquisition activities by the public
agencies/brokerages are primarily aimed at dealing with an industry shock (i.e., a more
challenging operating environment and competition from banks). M&As for them are an
important means to remain competitive and maintain adequate shareholder earnings.
Bank acquirers, during the sample period, on average earned a 0.31 percent
abnormal return, which is statistically significant under both the portfolio time-series t
test and generalized sign Z test. This result is also economically significant because the
average size of bank acquirers is large (about $7 billion). This finding leads us to reject
our Null hypothesis 2, which states that under the contestable market theory banks will
not earn abnormal returns just from expanding into insurance. Possible explanations for
this are: First, cost and revenue scope economies may exist between banking and
distribution of insurance. A bank’s existing customer base and customer information may
provide banks a comparative advantage when entering the insurance business (Todd and
Murray, 1998), and lower post-integration costs may arise when a bank acquires an
insurance agent/broker. Second, selling insurance may provide banks with a relatively
stable stream of non-interest income, which helps diversify the risks of banks (revenue
scope economies). The diversification might be more important for state banks than for
national banks, since the risk of national banks is already well diversified over a national
geographic area and over a large group of customers. Third, as noted in Carow (2001a),
the permission to sell insurance effectively grants an option to banks. Exercising the
option might deliver a signal to the market that the bank can benefit from such a
transaction, which in turn results in a positive reaction from the market.
26
Insurer acquirers (vertical integration transactions) in our sample on average earn
0.21 percent in abnormal returns, but this result is not statistically significant. This
provides support for our Null hypothesis 3. By separating insurers into property-casualty
insurers and life-health insurers we are able to investigate the difference between the two
insurance sectors. Property-casualty insurers on average earn 0.70 percent abnormal
returns, which is significant at the 10 percent level under the portfolio time-series t test.
Life-health insurers on average earn a negative 1.15 percent in abnormal returns, which is
also significant at the 10 percent level under the portfolio time-series t test. The
difference between the two insurance sectors might reflect investors’ expectations that
banks entering the life insurance business will benefit more than when life insurers
distribute products. The other service industries on average earn an insignificant positive
abnormal return.
To test the wealth gain differences among focusing, bancassurance and vertical
integration transactions, we performed a one way ANOVA analysis. The results are
shown in Table 6. The results indicate that there are no significant wealth gain
differences among focusing, bancassurance and vertical integration transactions in the
property-casualty industry, but that life-health insurers earn a significantly lower
abnormal return than the other types of acquirers. Again, this might provide evidence that
banks are fierce competitor for life-health insurers.
4.2. The impact of Graham-Leach-Bliley Act on acquirer wealth effect
The passage of the GLB Act in 1999 provided banks the legal right to enter all
aspects of insurance business. Although bancassurance activities started years before the
GLB Act, we expect that the act would raise the market’s expectations for bancassurance
27
transactions because after 1999 banks obtained the additional option to underwrite
insurance. In order to address the wealth effect difference before and after GLB, we
perform a sub-period analysis, and the results are shown in Table 7.
For all acquirers in general, the average pre-GLB wealth gain was not statistically
significant, while the post-GLB acquisitions on average earned a significant 0.18 percent
abnormal return under both the portfolio time-series t test and the generalized sign Z test.
The public agency/broker acquirers showed no significant wealth effect before and after
the GLB Act for the (0,+1) window, but positive and significant abnormal returns were
found for the windows (-3,+3), (-5,+5), (-10,+10) and (0, +10). This is puzzling since the
GLB Act did not change the business possibilities for agencies/brokerages. Perhaps this
result suggests that after 1999 investors believed that market concentration had reached a
point where market power might accrue to some brokers allowing them to earn excess
returns.
Therefore, another analysis was conducted in which the abnormal stock returns
from transactions involving Marsh were isolated and compared to the abnormal returns
from all other focusing transactions both pre and post the GLB Act. A similar analysis
was conducted in which the abnormal stock returns associated with transactions by Marsh
and Aon combined, and Marsh, Aon, and Willis combined were isolated from all other
focusing transactions both pre and post GLB. (Marsh, Aon, and Willis were ranked
number 1, 2, and 3, respectively, in terms of revenue size over the sample period.) The
results (not shown) indicate that there were no significant abnormal stock returns
associated with transactions by Marsh, Marsh and Aon combined, or Marsh, Aon and
Willis combined prior to GLB. After GLB, however, significant positive abnormal stock
28
returns were realized by Marsh of 0.032 percent. Also, relative to all other focusing
transactions, Marsh, and Marsh and Aon combined, earned positive and significant
abnormal stock returns of 0.039 and 0.016 percent, respectively. Thus, after 1999, the
market reacted very favorably to growth in the largest public brokers; however it is
impossible to say whether this was due to expectation of future economic rents to be
realized from exercise of market power by these brokers. A caveat of this analysis is that
it is based on a small number of observations.
The average wealth gain for bancassurance activities was 0.27 percent before the
GLB Act, which is not significant. The transactions after GLB on average earned 0.33
percent abnormal returns, which is significant at the 5 percent level under the portfolio
time-series t test and at the 10 percent level under the generalized sign Z test. This result
rejects our Null hypothesis 2.1 that bank acquirers will show no significant difference in
wealth gains before and after the passage of GLB. An explanation for the difference
might be that the market believes the synergies between distributing insurance and/or the
potential opportunity of underwriting insurance will enhance bank value.
The abnormal returns from vertical integration activities on average showed no
significant difference before and after GLB overall, but differences exist if propertycasualty insurers are separated from life-health insurers. There were no significant wealth
gain differences for property-casualty insurers before and after the GLB Act, indicating
that the passage of GLB might have had less impact on this segment of the industry.
Perhaps this is because the primary pressures on this industry are focused on factors such
as underwriting cycles, distribution system reform, and increased concentration in the
agent/broker market. Life-health insurers experienced a difference in wealth gains before
29
and after the act. Before GLB, life-health insurers earned a negative 0.30 percent
abnormal return from their M&A activities, which is statistically insignificant. After GLB,
life-health insurers on average earned a negative 2.35 percent abnormal return, which is
significant at the 5 percent level under the portfolio time-series t test. Again, the possible
explanation for this is that the passage of GLB is expected to lead to a fiercer competitive
operating environment for life-health insurers, and this situation is not likely to be
resolved by adopting a vertical integration strategy.
4.3 Wealth gains and target agent/broker insurance sector
The analysis in this paper is concerned with possible synergies between insurance
agents/brokers and insurance carriers and banks. But the general literature argues that
property-casualty and life-health insurance are very different businesses and this might
lead to differing synergies in M&A transactions with banks and insurers. Therefore,
abnormal returns are examined separately for life-health agents/brokers and propertycasualty agents/brokers.
For the target agencies/brokerages in our sample, 361 were property-casualty
agencies/brokerages, 159 were life-health agencies/brokerages, 179 were multiline
agencies/brokerages, and in 281 cases the insurance sector of the target agency/brokerage
was unknown. The wealth effects of various combinations of acquirers and targets by
insurance sector are presented in Table 8.
Public agencies/brokerages that acquired property-casualty agencies/brokerages
and multiline agencies/brokerages, on average, experienced no wealth gains, while
acquisition of life-health agencies/brokerages produced a slight positive (0.21 percent)
abnormal return (significant at the 10 percent level under the generalized sign Z test). An
30
explanation for this result might be that acquisition of a life-health agency/brokerage is
an acquisition of human resources and these transactions tend to bring valuable, longterm customers to the acquirers. On the other hand, most property-casualty insurance
contracts are short-term contracts. Also, affiliating with large brokers might change the
operating style of the target property-casualty agencies/brokerages resulting in a loss of
customers at the target agency/brokerage. Another explanation is that large public
agencies/brokerages are more likely to overpay for property-casualty agencies/brokerages
because they are under more earnings pressure than banks just to remain competitive and
purchasing
a
property-casualty
agency/brokerage
would
allow
large
public
agencies/brokerages to be more competitive.
In the bancassurance case, acquisitions of property-casualty agencies/brokerages,
on average, resulted in a 0.62 percent abnormal return to banks, which is significant at the
1 percent level under all three tests. Acquisitions of life-health agencies/brokerages, on
average, provided bank acquirers with a 0.76 percent abnormal return, which is
significant at the 5 percent level under the portfolio time-series t test and generalized sign
Z test. However, acquisitions of multiline agencies/brokerages were associated with no
wealth gains. This finding rejects our Null hypothesis 2.2, which states that no significant
wealth gain differences exist between bank’s acquisition of property-casualty
agencies/brokerages and life-health agencies/brokerages. It is no surprise that banks gain
by acquiring life-health agencies/brokerages, but somewhat surprising that banks also
gain by acquiring property-casualty agencies/brokerages given the lower expected
synergies between banking and property-casualty insurance business. The wealth gain for
such transactions might primarily arise from the diversification of revenues for banks.
31
The insignificant wealth gain for acquisitions of multiline agencies/brokerages might be
due to potentially high post-acquisition integration costs from such transactions.
Most transactions of property-casualty insurers involve acquisitions of propertycasualty agencies/brokerages, indicating that these insurers are seeking benefits from
vertical integration. The wealth gain from these transactions, on average, is positive
(0.965 abnormal return) and is significant at the 10 percent level under the portfolio timeseries t test. No significant wealth gains exist for other types of transactions for propertycasualty insurer acquirers.
The majority of transactions for life-health insurers involve acquisitions of lifehealth agencies/brokerages, indicating that life-health insurers are also seeking vertical
integration. However, these life-health insurers are not very successful when measured in
terms of abnormal stock returns. Life-health insurers earn negative abnormal returns from
these transactions.
5. Regression analysis
Table 9 presents the results of regression analysis on the acquiring firms’ returns.
The advantage of performing regression analysis is that it allows us to control for several
features of the transactions simultaneously. The dependent variable in the regression
models is the acquirer’s CAR for the event window (0, +1). A Huber/White/Sandwich
variance estimator and controls for potential cluster problems for acquirers that appear
more than once in the sample are used.
Model 1 includes basic control variables for type of acquirer (bank, insurer, and
agency/brokerage) and type of transaction (focusing, bancassurance, vertical integration)
by agent/broker sector (property-casualty, life-health).
Model 2 includes additional
32
control variables for acquiring firm characteristics. The additional variables in Model 2
are distribution system of the insurer, equity size of acquirer, Tobin’s Q value and
dummy variables for whether the bank acquirer had an insurance subsidiary and for
whether the acquirer was involved in multiple purchases of agencies/brokerages during
the year. (Tobin’s Q value is used as a proxy for the firm’s performance and risk level.)
All regression models include time dummies.
The results in Model 1 indicate that no significant differences in wealth gains for
banks, agencies/brokers and insurers relative to acquirers in other service industries
occurred. That is, the coefficients for the dummy variables for bank, property-casualty,
life-health and agency/broker acquirers are not significantly different from zero. (The
omitted category for the type of transaction categorization is acquirers in other service
industries.)
However, positive and significant abnormal returns are associated with
banks’ purchases of life-health agencies/brokerages since the coefficient for purchases of
a life-health agency/brokerage is positive and significant.
Model 2, the more complete regression model, indicates that banks appear to gain
more than other types of acquirers from acquisitions of insurance agencies/brokerages,
confirming that the market believes bancassurance has some advantage over other types
of transactions and can potentially enhance banks’ value. That is, the bank dummy
variable and the interactions of the bank dummy variable with life health
agencies/brokerages and with property-casualty agencies/brokerages are all positive and
significant.
However, banks that acquire life-health insurance agencies/brokerages
earned smaller positive abnormal returns than those acquiring property-casualty
agencies/brokerages. This potentially raises a question about whether distributing
33
property-casualty insurance products enables a bank to better diversify its business than
distributing life-health insurance products.
Contrary to general expectations, the possibility of increased market power of
public agencies/brokers associated with M&A activity, in general, does not seem to
reward the shareholders of these firms with significantly higher abnormal returns. That is
the coefficients of all of the focusing variables in Model 2 are insignificant.
Vertical integration transactions apparently also do not enhance an insurer’s
market value because the coefficients associated with both life-health and propertycasualty insurers are not significant. A distribution system variable is included in the
regression to test whether the distribution system used by an insurer before the M&A
transaction is associated with wealth gains. Insurers that use independent agency and
direct selling systems should face more competition from banks than insurers using
brokerage because banks that distribute insurance are more likely to rely on brokerage
contracts with insurers (Carow 2001a). Our regression results do not provide support for
this argument.
The impact on brokers, insurers and banks from M&A activity might vary by firm
size. For larger acquirers, the target is more likely to represent a smaller fraction of their
operations, and the economic gain from the transaction would be spread over a larger
equity base. Hence the abnormal stock return would be expected to be smaller in this
case. Otherwise, since larger brokers and insurers are more established in the market,
they might be impacted less by changes and competition from new market players (i.e.,
banks). Banks, on the other hand, might have a size disadvantage. A large bank might
already be adequately diversified because it is more likely to extend over a larger
34
geographic area and offer a large portfolio of banking products. So the marginal benefits
from further diversification might be small compared with a smaller bank. In addition, a
large bank may be more likely to take insurance distribution “in-house” and set up its
own in-house operations than a smaller bank. This would lead to higher post-integration
costs than continuing to operate the acquired agency/broker as a separate entity.
To determine whether size is related to acquirers’ wealth gains, Model 2 includes
four size variables in the regression: bank size, agency/broker size, property-casualty
insurer size and life-health insurer size. Size is measured as equity value before the M&A
transaction. Only bank size is significant, and it has a negative sign, implying that large
banks may benefit less from further diversification into insurance. However, this could
simply be a size effect. Since most of the targets are small firms, a similar wealth gain
might generate higher returns for smaller banks than for large banks.
6. Conclusion and Discussion
This paper investigates the consolidation wave in the U.S. agent/broker market
over the past 15 years. By looking at M&A transactions, we document the changing
structure of the agent/broker industry to better understand how different players in the
insurance marketplace are evolving. The period of this study encompasses important
regulatory and market changes including the entrance of banks into insurance and
investigations into the brokerage industry regarding anti-competitive practices.
We
compare the abnormal stock returns resulting from focusing transactions, vertical
integration and bancassurance in this study. A distinction is made between life-health
and property-casualty agency/brokerage.
35
The results show that consolidation in the agency/broker industry has tremendous
wealth effects for shareholders, especially for shareholders of banks that pursue a
bancassurance strategy. Conversely, vertical integration between property-casualty
insurance carriers and insurance distributors does not provide significant wealth gains to
shareholders, indicating perhaps that vertical integration is a defensive strategy in
response to a more concentrated agent/broker market and competition from banks. In
addition, any market power achieved through M&As did not provide large public brokers
significant wealth gains in the short run except for the largest brokers after 1999.
Despite the natural affinity between the investment products of life-health
insurance
and
banks,
the
majority
of
transactions
involving
agencies/brokerages concern property-casualty agencies/brokerages.
banks
and
The latter
transactions also generate higher abnormal returns than M&As of banks with life-health
agencies/brokerages. Possible reasons for this are that it might be easier for a bank to set
up a life-health insurance agency or sell life-health insurance in-house than propertycasualty insurance. Then an M&A would be a quick and convenient way for banks to
enter the property-casualty insurance business. Second, distributing property-casualty
insurance products may better help diversify banks’ revenues with non-interest (fee)
income, and improve banks’ performance measures.
Finally abnormal returns associated with bancassurance are less dramatic for large
banks, perhaps indicating that scope diseconomies outweigh scope economies for large
banks or that a size effect exists. Future research is needed to examine whether banks
engaged in bancassurance show significant performance improvement and whether this
36
improvement is different for banks that distribute relatively more property-casualty
insurance than life-health insurance.
37
References
Carow, Kenneth A, 2001a, “The Wealth Effects of Allowing Bank Entry into the
Insurance Industry,” Journal of Risk and Insurance, 68(1), 129-150.
Carow, Kenneth A, 2001b, “Citicorp-Traveler’s Group Merger: Challenging Barriers
Between Bank and Insurance,” Journal of Banking and Insurance, 25, 11553-1571.
Cowan, Arnold R., 2005, “Eventus 8.0: Software for Event Studies and CRSP Data
Retrieval,” Cowan Research, L.C..
Cummins, J. David, 1977, "Economies of Scale in Independent Insurance Agencies,"
Journal of Risk and Insurance, 44(4), 539-553.
Cummins, J. David and Neil A. Doherty, 2005, “The Economics of Insurance
Intermediaries,” Working Paper, the Wharton School, University of Pennsylvania.
Cummins, J. David and Mary A. Weiss, 2004, “Consolidation in the European Insurance
Industry: Do Mergers and Acquisitions Create Value for Shareholders?” BrookingsWharton Papers on Financial Services:
Falautano, Isabella and Emanuele Marsiglia, 2003, “Integrated Distribution of Insurance
and Financial Services and Value Creation: Challenges ahead,” The Geneva papers on
Risk and Insurance, Vol. 28, No. 3, 481-494.
Fama, Eugene F. and Kenneth R. French, 1993, “Commons Risk Factors in the Returns
on Stocks and Bonds,” Journal of Financial Economics, 33(1), 3-56.
Fields, L. Paige, Donald R. Fraser, and James W. Kolari, 2005, “What’s Different About
Bancassurance? Evidence of Wealth Gains to Banks and Insurance Companies,”
Working Paper, Texas A&M University.
Lown, Cara S., Carol L. Osler, Philip E. Strahan, and Amir Sufi, 2000, “The Changing
Landscape of the Financial Services Industry: What lies ahead?” FRBNY Economic
Policy Review, 39-55.
Mitchell, Mark L., and J. Harold Mulherin, 1996. The Impact of Industry Shocks on
Takeover and Restructuring Activity. Journal of Financial Economics 41, 193-229
Romain Durand, 2003, “Bancassurance across the Global: Meets with very mixed
response,” SCOR technical Newsletters, No. 10.
Rose, Lawrence C., and Dean G. Smith, 1995, “Expansion into Insurance Product-lines
and Bank Shareholder Returns,” Journal of Financial and Strategic Decisions, V8, N0. 2,
13-25.
38
Stigler, G., 1971, “The Theory of Economic Regulation,” Bell Journal of Economics and
Management Science, Vol. 2, No. 1, 3-21.
Swiss Re, 2004, “Commercial Insurance and Reinsurance Brokerage-Love thy
Middleman,” Sigma, No. 2.
Todd, J., and M. Murry, 1988, “banks in Insurance: Increase or Reduce Competition?”
Journal of Insurance Regulation, 6(4): 518-537.
Wepler, John M., Thomas R. Linn and Patrick T. Linnet (2004), “Banks in Insurance: A
Five-Year Retrospective since the Passage of GLB,” Bank Director Magazine, 3rd
Quarter.
39
Table 1
M&As of Insurance Broker, All Transactions, 1990-2005
Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Count
67
63
53
71
78
104
103
96
159
240
175
192
202
195
239
135
Reported Value ($ millions)
1079.0
1371.2
1633.5
1241.1
1488.6
1163.3
2809.4
2597.4
2587.7
336.2
361.9
1149.7
1280.6
1221.9
1418.7
953.5
Data Source: SDC Platinum (1990-1996), SNL DataSource (1997-2005). The number in the table includes
both public and private acquirers/ targets.
Table 2
Public Acquirers: Sample Size by Industry Sector, 1990-2005
Target Insurance Sector
Acquirer Industry Sector
Life-Health
PropertyInsurer Casualty Insurer Other
Agency/Broker Bank
Property-Casualty
Agency/Broker
189
Life-Health Agency/Broker
Total
125
3
60
9
386
64
39
22
9
32
166
Multiline Broker
74
109
1
3
5
192
Unknown
149
95
3
23
26
296
476
368
29
95
72
1040
Total
Data Source: SDC Platinum (1990-1996, 2005), SNL DataSource (1997-2004). Target Insurance Sector is
from SNL DataSource and Factiva Report. "Other" acquirer sector includes firms such as managed care
firms, claim adjusting firms, and other business service firms. "Multiline" target insurance sector indicates the
target has both property-casualty and life-health business. "Unknown" target insurance sector means the
insurance sector was not disclosed and includes managed care, general managing agency, and
independent agency.
40
Table 3
Public Acquirers: Sample Size by Year and Industry Sector, 1990-2005
Year/Acquire
Target agency/broker
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Total
Year/Acquire
Target agency/broker
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Total
P-L
4
3
5
5
5
6
7
11
16
10
13
17
24
24
26
13
189
L-H
3
2
1
1
1
5
6
5
4
4
6
6
8
12
64
Agency/Broker
Multiline
Unknown
18
1
15
1
12
1
9
1
10
5
6
2
4
2
5
4
12
1
16
5
10
21
10
10
8
9
4
9
5
2
5
74
149
P-L
1
L-H
Property-Casualty Insurer
Multiline
Unknown
1
1
1
1
2
4
9
11
16
3
3
4
1
2
1
60
1
2
2
1
1
1
1
2
5
3
5
1
3
1
2
1
1
9
3
1
23
Total
25
21
19
16
16
17
14
23
38
32
32
52
48
43
48
32
476
Total
1
1
1
1
2
3
7
15
17
23
4
6
7
2
2
3
95
P-L
L-H
Bank
Multiline
Unknown
Total
1
3
1
5
7
2
6
17
12
15
10
6
6
4
95
1
1
3
2
6
11
14
37
48
57
49
57
52
22
8
368
1
1
2
3
17
13
17
12
21
27
10
2
125
P-L
1
1
4
3
2
3
9
12
2
1
1
39
L-H
1
1
1
5
11
16
25
13
14
17
5
1
109
Other
Multiline
Unknown
1
3
2
2
3
Total
1
1
1
3
1
2
2
6
13
9
2
8
9
9
5
5
26
72
1
1
2
1
1
1
1
1
2
1
2
3
4
3
2
3
6
5
1
9
32
2
2
2
3
7
2
P-L
1
1
1
3
L-H
1
1
1
1
1
1
4
3
3
1
1
2
1
1
22
Life-Health Insurer
Multiline
Unknown
1
1
1
1
1
3
Total
1
1
1
2
1
2
5
4
4
2
1
1
2
1
1
29
Data Source: SDC Platinum (1990-1996, 2005), SNL DataSource (1997-2004). Target Insurance Sector is from SNL DataSource and Factiva Report. "Other" acquirer sector includes
firms such as managed care firms, claim adjusting firms, and other business service firms. "Unknown" target insurance means target insurance sector not disclosed, includes managed
care, general managing agency, and independent agency.
41
Table 4
Descriptive Statistics
Number of
Observations
Banks with Cumulative Abnormal Returns (CARs)
Available
National Banks (SIC=6021)
State Banks
1
Bank with Insurance Subsidiaries
Total Banks with CARs Available
126
235
282
361
Insurers with CARs Available
Property-Casualty Insurer
Life-Health Insurer
Total insurers with CARs Available
79
29
108
Agencies/Brokers with CARs Available
445
Other Types of Acquirers with CARs Available
Distribution Systems of Insurance Acquirers
Brokerage
66
2
Property-Casualty
Insurer
Life-Health Insurer
Property-Casualty
Insurer
Life-Health Insurer
Property-Casualty
Insurer
Life-Health Insurer
Agency
Direct
Total no. of insurers with distribution system data
10
6
52
13
5
1
86
3
Rating of Insurance Acquirers
Superior
Excellent
Very Good
Vulnerable
Total no. of insurers with rating data
30
39
10
8
87
Purchasing Method
Cash
Stock
Mixed
Unknown
Total
Total market value of equity of acquirers ($ millions)
Agency/Broker
Bank
Life-Health Insurer
Property-Casualty Insurer
Other
All Observations
111
122
101
646
980
N
Mean
Std
Dev.
445
361
29
79
66
980
1917
7007
3947
1799
8605
4294
3234
16750
7122
2791
39923
14927
4
5
Tobin’s Q
Agency/Broker
409
2.20
0.82
Bank
344
1.11
0.13
Life-Health Insurer
28
1.27
0.58
Property-Casualty Insurer
76
1.26
0.46
Other
63
2.57
1.23
All Observations
857
1.65
0.80
1
2
3
Data is obtained from Factiva Report and Directory of Corporate Affiliations, various years; and :Data is from Best's
4
5
Key Rating Guide; Data is calculated from CRSP, one month-end equity value prior to announcement date; Q is
calculated from COMPUSTAT data
42
Acquirer
All Acquirers
Table 5
Wealth Gains of Acquirers, All Transactions and Transactions by Acquirer Industry, 1990-2005
Days
N
Mean CAR
Median
Positive:
Calendar
Portfolio
Generalized
(0,0)
980
0.04%
-0.05%
466:514
0.277
0.614
-0.567
(-1,+1)
980
0.10%
-0.06%
481:499
0.68
0.786
0.392
(-2,+2)
980
0.09%
-0.02%
488:492
0.262
0.551
0.839
(-3,+3)
980
0.10%
0.05%
497:483)
0.596
0.508
1.414$
(-5,+5)
980
0.14%
-0.07%
484:496
0.92
0.582
0.583
(-10,+10)
980
0.28%
-0.06%
487:493
0.667
0.834
0.775
(0,+1)
980
0.18%
0.01%
492:488
1.484
1.733*
1.095
(0,+2)
980
0.19%
0.00%
490:490
1.179
1.502$
0.967
(0,+3)
980
0.15%
0.05%
499:481)
1.091
1.022
1.542$
(0,+5)
980
0.06%
-0.05%
484:496
0.631
0.325
0.583
(0,+10)
980
-0.01%
0.05%
494:486
0.146
-0.05
1.223
Agency/Broker vs. Agency/Broker (Focus)
(0,0)
445
-0.03%
-0.10%
207:238
-0.125
-0.324
-0.678
(-1,+1)
445
0.02%
-0.15%
209:236
0.388
0.118
-0.488
(-2,+2)
445
0.01%
-0.11%
218:227
-0.118
0.063
0.366
(-3,+3)
445
0.18%
0.09%
226:219
0.611
0.72
1.125
(-5,+5)
445
0.41%
0.12%
227:218
1.307
1.281
1.22
(-10,+10)
445
0.33%
0.24%
228:217)
0.686
0.757
1.315$
(0,+1)
445
0.02%
-0.06%
218:227
0.645
0.168
0.366
(0,+2)
445
0.08%
0.09%
227:218
0.54
0.473
1.22
(0,+3)
445
0.18%
0.05%
228:217)
1.097
0.935
1.315$
(0,+5)
445
0.19%
-0.04%
221:224
1.042
0.815
0.65
(0,+10)
445
0.06%
0.34%
235:210>
0.668
0.199
1.979*
Bank Vs. Agency/Broker
(0,0)
361
0.04%
-0.04%
173:188
-0.111
0.406
-0.417
(-1,+1)
361
0.27%
0.06%
184:177
0.897
1.604$
0.742
(-2,+2)
361
0.23%
-0.10%
178:183
0.478
1.091
0.11
(-3,+3)
361
0.22%
0.08%
185:176
0.67
0.882
0.847
(-5,+5)
361
0.18%
-0.06%
177:184
0.723
0.579
0.005
(-10,+10)
361
0.29%
-0.10%
177:184
0.74
0.66
0.005
(0,+1)
361
0.31%
0.08%
190:171)
1.258
2.329**
1.373$
(0,+2)
361
0.19%
0.05%
183:178
0.384
1.162
0.636
(0,+3)
361
0.12%
0.16%
191:170)
0.375
0.642
1.479$
(0,+5)
361
0.02%
0.01%
181:180
0.192
0.075
0.426
(0,+10)
361
-0.09%
-0.11%
177:184
-0.052
-0.278
0.005
Insurer Vs. Agency/Broker
(0,0)
108
-0.24%
-0.01%
54:54
-0.682
-0.939
0.393
(-1,+1)
108
0.35%
0.27%
59:49)
0.348
0.797
1.356$
(-2,+2)
108
0.20%
0.18%
57:51
0.211
0.353
0.971
(-3,+3)
108
-0.05%
0.10%
55:53
0.2
-0.074
0.586
(-5,+5)
108
-0.17%
-0.45%
53:55
-0.344
-0.2
0.201
(-10,+10)
108
0.92%
-0.63%
52:56
-0.055
0.799
0.008
(0,+1)
108
0.21%
-0.14%
50:58
0.067
0.584
-0.377
(0,+2)
108
0.23%
-0.23%
48:60
0.311
0.525
-0.762
(0,+3)
108
-0.26%
-0.67%
48:60
-0.302
-0.516
-0.762
(0,+5)
108
-0.48%
-0.26%
52:56
-0.71
-0.786
0.008
(0,+10)
108
0.09%
0.00%
54:54
-0.591
0.112
0.393
P-L insurer vs. Agency/Broker
(0,0)
79
-0.13%
-0.05%
39:40
-0.349
-0.432
0.207
(-1,+1)
79
0.78%
0.51%
43:36
0.752
1.436$
1.108
(-2,+2)
79
0.65%
0.17%
41:38
0.443
0.932
0.658
(-3,+3)
79
0.62%
0.18%
41:38
0.584
0.752
0.658
(-5,+5)
79
0.43%
0.09%
40:39
-0.041
0.415
0.432
(-10,+10)
79
0.82%
-0.57%
39:40
-0.075
0.57
0.207
(0,+1)
79
0.70%
-0.09%
38:41
0.751
1.592$
-0.018
(0,+2)
79
0.84%
-0.20%
35:44
1.079
1.550$
-0.693
(0,+3)
79
0.47%
-0.29%
38:41
0.359
0.752
-0.018
(0,+5)
79
-0.01%
0.11%
40:39
-0.563
-0.017
0.432
(0,+10)
79
-0.54%
0.35%
41:38
-0.801
-0.517
0.658
L-H insurer vs. Agency/Broker
(0,0)
29
-0.51%
0.06%
15:14
-0.703
-1.012
0.417
(-1,+1)
29
-0.83%
0.20%
16:13
-0.467
-0.956
0.789
(-2,+2)
29
-1.04%
0.23%
16:13
-0.419
-0.927
0.789
(-3,+3)
29
-1.88%
-0.03%
14:15
-0.542
-1.413$
0.045
(-5,+5)
29
-1.79%
-0.58%
13:16
-0.596
-1.075
-0.327
(-10,+10)
29
1.19%
-1.73%
13:16
-0.036
0.518
-0.327
(0,+1)
29
-1.15%
-0.39%
12:17
-0.759
-1.616$
-0.698
(0,+2)
29
-1.44%
-0.51%
13:16
-0.819
-1.655*
-0.327
(0,+3)
29
-2.24%
-1.12%
10:19(
-0.844
-2.235*
-1.442$
(0,+5)
29
-1.76%
-0.82%
12:17
-0.463
-1.430$
-0.698
(0,+10)
29
1.81%
-0.52%
13:16
0.177
1.085
-0.327
Other vs. Agency/Broker
(0,0)
66
1.04%
-0.06%
32:34
0.782
2.648**
0.046
(-1,+1)
66
-0.68%
-0.61%
29:37
-1.087
-0.996
-0.693
(-2,+2)
66
-0.36%
0.25%
35:31
-0.78
-0.408
0.785
(-3,+3)
66
-0.92%
-0.37%
31:35
-1.036
-0.885
-0.201
(-5,+5)
66
-1.40%
-1.70%
27:39
-1.244
-1.068
-1.186
(-10,+10)
66
-1.20%
-1.33%
30:36
-0.912
-0.667
-0.447
(0,+1)
66
0.43%
0.13%
34:32
-0.152
0.768
0.539
(0,+2)
66
0.85%
-0.22%
32:34
0.21
1.245
0.046
(0,+3)
66
0.75%
-0.33%
32:34
0.101
0.946
0.046
(0,+5)
66
0.26%
-0.47%
30:36
-0.037
0.27
-0.447
(0,+10)
66
-0.27%
-0.96%
28:38
-0.526
-0.21
-0.94
Note: Fama-French Time-Series Model, Equally Weighted Index; The symbols $,*,**, and *** denote statistical significance at the 0.10,
0.05, 0.01 and 0.001 levels, respectively, using a 2-tail test. The symbols (,< or ),> etc. correspond to $,* and show the significance and
direction of the generalized sign test.
43
Table 6
Difference in Wealth Gains by Type of Transaction, CAR Window (0,+1)
Parameter
Estimate
Standard Error
t Value Pr > |t|
-0.29%
0.003
-1.120
0.261
Compare Focus & Bancassurance
1.17%
0.007
1.670
0.096
Compare Focus & Life-Health insurer
-0.68%
0.004
-1.520
0.128
Compare Focus & Property-Casualty insurer
-0.41%
0.005
-0.840
0.402
Compare Focus & Other
1.46%
0.007
2.070
0.039
Compare Bancassurance & Life-Health insurer
-0.39%
0.005
-0.860
0.393
Compare Bancassurance & Property-Casualty insurer
-0.11%
0.005
-0.230
0.817
Compare Bancassurance & Other
-1.85%
0.008
-2.330
0.020
Compare Life-Health insurer & Property-Casualty insurer
-1.58%
0.008
-1.930
0.054
Compare Life-Health insurer & Other
0.28%
0.006
0.450
0.652
Compare Property-Casualty insurer & Other
Note: One-Way ANOVA Analysis. "Focus" is defined as M&As between insurance agency, broker or service firms. Bancassurance is defined as
M&A transactions between banks and insurance agency, broker or service firms. "Other" is defined as M&A transactions between managed care
firms, claim adjusting firms, other business service firms and insurance agency, broker or service firms.
44
Table 7
Wealth Gains of Acquirers, All Transactions and Transactions by Acquirer Industry, Pre-GLB and Post GLB
Before GLB Act (PreGLB)
Portfolio
Mean Median Positive: Calendar Time- Generalized
CAR
CAR Negative Time t series t
Sign Z
After GLB Act (PostGLB)
Portfolio
Mean Median Positive: Calendar Time- Generalized
CAR
CAR Negative Time t series t
Sign Z
Acquirer
Days
N
N
All Acquirers
(0,0)
346 0.12% -0.03% 169:177
0.897
0.956
0.514
634 0.01% -0.09% 297:337
-0.437
0.063
-1.085
(-1,+1)
346 0.11% -0.12% 160:186
0.565
0.536
-0.455
634 0.09% 0.05% 321:313
0.404
0.609
0.821
(-2,+2)
346 0.05% -0.30% 163:183
0.031
0.183
-0.132
634 0.11% 0.07% 325:309
0.324
0.576
1.139
(-3,+3)
346 -0.11% -0.13% 170:176 -0.533
-0.355
0.622
634 0.21% 0.16% 327:307)
1.222
0.931
1.298$
(-5,+5)
346 -0.23% -0.67% 153:193 -0.707
-0.577
-1.209
634 0.34% 0.39% 331:303)
1.661$
1.199
1.616$
(-10,+10)
346 -0.39% -0.69% 159:187 -1.449
-0.692
-0.563
634 0.64% 0.43% 328:306)
1.785$
1.615$
1.378$
(0,+1)
346 0.18% -0.24% 158:188
0.91
1.062
-0.67
634 0.18% 0.11% 334:300>
1.186
1.435$
1.854*
(0,+2)
346 0.25% -0.18% 167:179
0.966
1.164
0.299
634 0.16% 0.09% 323:311
0.706
1.056
0.98
(0,+3)
346 0.10% -0.17% 166:180
0.542
0.413
0.191
634 0.17% 0.17% 333:301>
0.973
1.01
1.775*
(0,+5)
346 0.00% -0.34% 154:192
0.325
-0.005
-1.101
634 0.09% 0.15% 330:304)
0.542
0.426
1.536$
(0,+10)
346 -0.51% -0.43% 163:183 -1.283
-1.256
-0.132
634 0.26% 0.36% 331:303)
0.954
0.899
1.616$
Agency/Broker vs. Agency/Broker (Focus)
(0,0)
183 0.07% -0.01% 90:93
0.53
0.44
0.426
262 -0.10% -0.19% 117:145
-0.732
-0.87
-1.24
(-1,+1)
183 0.11% -0.18% 79:104
0.235
0.385
-1.202
262 -0.05% -0.05% 130:132
0.309
-0.219
0.367
(-2,+2)
183 -0.19% -0.37% 82:101
-1.086
-0.504
-0.758
262 0.16% 0.13% 136:126
0.74
0.58
1.109
(-3,+3)
183 -0.13% -0.29% 87:96
-1.114
-0.282
-0.018
262 0.40% 0.44% 139:123)
1.685$
1.25
1.480$
(-5,+5)
183 -0.17% -0.88% 80:103
-1.359
-0.296
-1.054
262 0.81% 0.99% 147:115>> 2.780**
2.021*
2.469**
(-10,+10)
183 -0.78% -0.76% 82:101
-1.996*
-1.008
-0.758
262 1.11% 1.51% 146:116>> 2.436*
2.012*
2.345**
(0,+1)
183 0.15% -0.20% 85:98
0.626
0.624
-0.314
262 -0.07% 0.06% 133:129
0.289
-0.385
0.738
(0,+2)
183 0.11% 0.11%
95:88
0.297
0.375
1.166
262 0.06% 0.06% 132:130
0.462
0.271
0.614
(0,+3)
183 0.17% -0.17% 88:95
0.376
0.491
0.13
262 0.19% 0.29% 140:122)
1.132
0.782
1.603$
(0,+5)
183 0.11% -0.23% 86:97
-0.015
0.26
-0.166
262 0.25% 0.06% 135:127
1.422
0.846
0.985
(0,+10)
183 -0.55% -0.35% 86:97
-1.106
-0.986
-0.166
262 0.49% 0.85% 149:113>> 1.825$
1.235
2.716**
Bank Vs. Agency/Broker
(0,0)
75 -0.13% -0.13% 32:43
-1.012
-0.569
-0.736
286 0.08% -0.02% 141:145
0.536
0.791
-0.092
(-1,+1)
75 0.35% -0.06% 36:39
0.235
0.917
0.19
286 0.24% 0.15% 148:138
0.914
1.344$
0.736
(-2,+2)
75 0.51% -0.57% 34:41
0.276
1.026
-0.273
286 0.16% 0.02% 144:142
0.389
0.687
0.263
(-3,+3)
75 0.40% -0.11% 36:39
0.358
0.672
0.19
286 0.18% 0.14% 149:137
0.567
0.644
0.854
(-5,+5)
75 0.00% -0.69% 33:42
0.278
0.006
-0.505
286 0.23% 0.03% 144:142
0.668
0.668
0.263
(-10,+10)
75 -0.82% -0.85% 34:41
-0.857
-0.799
-0.273
286 0.58% 0.02% 143:143
1.274
1.216
0.145
(0,+1)
75 0.27% -0.23% 35:40
-0.129
0.869
-0.042
286 0.33% 0.11% 155:131)
1.617
2.213*
1.564$
(0,+2)
75 0.32% -0.27% 33:42
-0.131
0.821
-0.505
286 0.16% 0.16% 150:136
0.608
0.885
0.973
(0,+3)
75 0.15% 0.14%
39:36
-0.135
0.333
0.884
286 0.12% 0.16% 152:134
0.535
0.557
1.209
(0,+5)
75 -0.51% -0.84% 29:46(
-0.948
-0.928
-1.430$
286 0.16% 0.22% 152:134
0.73
0.61
1.209
(0,+10)
75 -1.24% -1.14% 33:42
-1.985$ -1.675*
-0.505
286 0.21% 0.04% 144:142
0.955
0.621
0.263
Insurer Vs. Agency/Broker
(0,0)
60 0.12% 0.18% 34:26)
0.096
0.441
1.367$
48 -0.68% -0.39%
20:28
-1.209
-1.419$
-0.938
(-1,+1)
60 0.50% 0.18%
32:28
0.44
1.091
0.85
48 0.15% 0.81%
27:21
-0.037
0.182
1.084
(-2,+2)
60 0.77% 0.20%
33:27
0.907
1.302$
1.108
48 -0.52% 0.06%
24:24
-0.934
-0.491
0.218
(-3,+3)
60 0.52% 0.56%
33:27
0.633
0.738
1.108
48 -0.76% -0.21%
22:26
-0.452
-0.602
-0.36
(-5,+5)
60 0.33% -0.45% 29:31
0.258
0.378
0.075
48 -0.79% -0.25%
24:24
-0.685
-0.5
0.218
(-10,+10)
60 1.90% 0.15%
30:30
0.833
1.557$
0.333
48 -0.31% -1.09%
22:26
-0.528
-0.141
-0.36
(0,+1)
60 0.28% -0.33% 27:33
0.353
0.741
-0.442
48 0.12% -0.01%
23:25
-0.413
0.175
-0.071
(0,+2)
60 0.58% -0.27% 26:34
0.799
1.27
-0.701
48 -0.22% -0.18%
22:26
-0.548
-0.265
-0.36
(0,+3)
60 0.26% -0.59% 27:33
0.538
0.484
-0.442
48 -0.90% -0.67%
21:27
-1.124
-0.948
-0.649
(0,+5)
60 0.55% -0.26% 28:32
0.734
0.839
-0.184
48 -1.77% -0.36%
24:24
-1.326
-1.515$
0.218
(0,+10)
60 1.04% 0.52%
33:27
0.756
1.184
1.108
48 -1.10% -0.51%
21:27
-1.004
-0.694
-0.649
P-L insurer vs. Agency/Broker
(0,0)
43 0.03% 0.05%
23:20
-0.248
0.104
0.684
36 -0.34% -0.20%
16:20
-0.259
-0.585
-0.441
(-1,+1)
43 0.76% -0.10% 21:22
0.223
1.359$
0.074
36 0.79% 1.03%
22:14)
1.156
0.799
1.561$
(-2,+2)
43 1.20% -0.02% 21:22
0.664
1.656*
0.074
36 0.00% 0.81%
20:16
-0.326
-0.002
0.893
(-3,+3)
43 1.19% 0.18%
23:20
0.49
1.391$
0.684
36 -0.06% 0.04%
18:18
0.312
-0.039
0.226
(-5,+5)
43 0.79% 0.42%
22:21
0.165
0.735
0.379
36 0.00% -0.25%
18:18
-0.22
0
0.226
(-10,+10)
43 2.62% 2.11%
24:19
0.881
1.766*
0.989
36 -1.34% -1.22%
15:21
-0.577
-0.509
-0.775
(0,+1)
43 0.51% -0.32% 19:24
0.207
1.107
-0.537
36 0.94% 0.28%
19:17
1.309
1.155
0.56
(0,+2)
43 1.05% -0.27% 18:25
0.821
1.865*
-0.842
36 0.59% -0.14%
17:19
0.742
0.594
-0.107
(0,+3)
43 0.81% -0.06% 21:22
0.484
1.243
0.074
36 0.07% -0.78%
17:19
-0.079
0.061
-0.107
(0,+5)
43 0.53% -0.27% 20:23
0.144
0.672
-0.231
36 -0.67% 0.90%
20:16
-0.739
-0.473
0.893
(0,+10)
43 1.31% 0.63% 25:18)
0.544
1.222
1.295$
36 -2.75% -0.51%
16:20
-1.084
-1.441$
-0.441
L-H insurer vs. Agency/Broker
(0,0)
17 0.33% 0.28%
11:6)
0.339
0.531
1.481$
12 -1.69% -0.93%
4:8
-1.372
-2.099*
-1.112
(-1,+1)
17 -0.16% 0.33%
11:6)
0.323
-0.147
1.481$
12 -1.79% -1.78%
5:7
-1.169
-1.277
-0.534
(-2,+2)
17 -0.30% 1.32%
12:5>
0.457
-0.22
1.967*
12 -2.08% -0.81%
4:8
-1.012
-1.154
-1.112
(-3,+3)
17 -1.18% 0.78%
10:7
0.251
-0.723
0.995
12 -2.86% -1.35%
4:8
-1.019
-1.338$
-1.112
(-5,+5)
17 -0.82% -0.58%
7:10
0.154
-0.401
-0.463
12 -3.16% -1.17%
6:6
-0.79
-1.181
0.043
(-10,+10)
17 0.06% -2.18%
6:11
0.024
0.021
-0.95
12 2.79% 2.35%
7:5
-0.053
0.756
0.621
(0,+1)
17 -0.30% -0.34%
8:9
0.212
-0.344
0.023
12 -2.35% -1.85%
4:8
-1.495
-2.057*
-1.112
(0,+2)
17 -0.58% -0.15%
8:9
0.092
-0.546
0.023
12 -2.65% -1.01%
5:7
-1.266
-1.895*
-0.534
(0,+3)
17 -1.13% -1.44%
6:11
0.161
-0.913
-0.95
12 -3.82% -0.53%
4:8
-1.445 -2.369**
-1.112
(0,+5)
17 0.58% -0.26%
8:9
0.929
0.381
0.023
12 -5.07% -1.48%
4:8
-1.328 -2.564**
-1.112
(0,+10)
17 0.36% -0.52%
8:9
0.424
0.176
0.023
12 3.85% -0.65%
5:7
-0.066
1.440$
-0.534
Other vs. Agency/Broker
(0,0)
28 1.04% -0.18% 13:15
1.055
1.670*
-0.081
38 1.05% 0.07%
19:19
0.287
1.901*
0.127
(-1,+1)
28 -1.37% -1.47% 13:15
-0.757
-1.271
-0.081
38 -0.17% -0.41%
16:22
-0.781
-0.182
-0.846
(-2,+2)
28 -1.17% -0.13% 14:14
-0.524
-0.841
0.298
38 0.24% 0.25%
21:17
-0.586
0.192
0.776
(-3,+3)
28 -2.76% -0.31% 14:14
-1.014
-1.679*
0.298
38 0.43% -0.37%
17:21
-0.388
0.295
-0.522
(-5,+5)
28 -2.51% -1.70% 11:17
-0.898
-1.22
-0.838
38 -0.57% -2.11%
16:22
-0.848
-0.313
-0.846
(-10,+10)
28 -1.53% -1.18% 13:15
-0.533
-0.536
-0.081
38 -0.97% -1.33%
17:21
-0.754
-0.383
-0.522
(0,+1)
28 -0.05% -0.54% 11:17
-0.231
-0.058
-0.838
38 0.78% 0.58%
23:15)
0.007
1.001
1.426$
(0,+2)
28 0.21% -0.54% 13:15
-0.123
0.194
-0.081
38 1.32% 0.30%
19:19
0.376
1.384$
0.127
(0,+3)
28 -0.80% -1.24% 12:16
-0.739
-0.64
-0.459
38 1.88% 0.17%
20:18
0.784
1.705*
0.452
(0,+5)
28 -0.53% -1.60% 11:17
-0.163
-0.35
-0.838
38 0.85% 0.05%
19:19
0.089
0.625
0.127
(0,+10)
28 -1.55% -1.61% 11:17
-0.684
-0.753
-0.838
38 0.67% -0.40%
17:21
-0.095
0.365
-0.522
Note: Fama-French Time-Series Model, Equally Weighted Index; The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels,
respectively, using a 2-tail test. The symbols (,< or ),> etc. correspond to $,* and show the significance and direction of the generalized sign test.
45
Table 8
Wealth Gains of Acquirers, Transaction by Acquirer Industry and Target Insurance Sector, Event Window (0,+1), 1990-2005
Acquirer
Target
Property-Casualty Agency/Broker
Life-Health Agency/Broker
All
Multiline Agency/Broker
Unknown
Property-Casualty Agency/Broker
Insurance Agency/Broker Life-Health Agency/Broker
Multiline Agency/Broker
Property-Casualty Agency/Broker
Bank
Life-Health Agency/Broker
Multiline Agency/Broker
Property-Casualty Agency/Broker
Property-Casualty Insurer Life-Health Agency/Broker
Multiline Agency/Broker
Property-Casualty Agency/Broker
Life-Health Insurer
Life-Health Agency/Broker
Multiline Agency/Broker
Property-Casualty Agency/Broker
Other
Life-Health Agency/Broker
Multiline Agency/Broker
N
361
159
179
281
178
60
65
124
39
105
47
8
3
3
22
1
9
30
5
Mean
CAR
0.22%
0.38%
0.08%
0.07%
-0.19%
0.21%
-0.40%
0.62%
0.76%
-0.08%
0.96%
0.17%
0.72%
-0.88%
-1.40%
8.48%
-0.75%
1.59%
7.51%
Median Positive: Calendar
Portfolio
Generalized
CAR Negative Time t Time-series t
Sign Z
0.05% 186:175
1.31
1.486$
1.15
0.42% 92:67>> 0.704
1.378$
2.468**
-0.40% 76:103< -0.041
0.357
-1.679*
-0.04% 138:143 0.712
0.364
0.224
-0.16% 81:97
-0.963
-0.926
-0.656
0.27% 34:26)
1.026
0.538
1.382$
-0.42% 27:38
-0.526
-1.116
-1.08
0.41% 76:48>> 2.645**
2.764**
2.734**
0.56% 25:14>
0.45
1.832*
1.959*
-0.40% 44:61(
-0.81
-0.327
-1.476$
0.00% 24:23
0.768
1.589$
0.316
0.25%
4:4
0.21
0.18
0.104
-0.43%
1:2
0.629
0.288
-0.516
-1.36%
1:2
-1.194
-0.543
-0.448
-0.47% 10:12
-1.341
-1.985*
-0.255
8.48%
1:0
6.901***
1.069
-0.80%
4:5
-0.618
-0.574
-0.418
0.67% 19:11>
1.277
1.840*
1.659*
0.42%
3:2
1.071
3.840***
0.537
Note: Fama-French Time-Series Model, Equally Weighted Index; The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using
a 2-tail test. The symbols (,< or ),> etc. correspond to $,* and show the significance and direction of the generalized sign test. "Other" is defined as M&A transactions between
managed care firms, claim adjusting firms, other business service firms and insurance agency, broker or service firms.
46
Table 9
Regression Analysis on Focusing Transaction, Vertical Integration and Bancassurance
Sample Period 1990-2005, Fixed Time Effects Included
Dependent Variable: CAR (0,+1)
Independent Variable
Broker
1
Model 1
-0.0041
(0.0150)
-0.0023
(0.0156)
-0.0084
(0.0173)
-0.0142
(0.0216)
-0.0032
(0.0029)
-0.0001
(0.0039)
-0.0037
(0.0046)
0.0065
(0.0037)*
0.0077
(0.0062)
0.0079
(0.0077)
-0.0078
(0.0211)
0.0068
(0.0103)
Model 2
0.0037
(0.0172)
2
Bank
0.0300
(0.0153)*
3
0.0237
P-L insurer
(0.0286)
4
L-H insurer
0.0071
(0.0330)
5
-0.0038
Broker acquires P-L insurance broker
(0.0032)
6
-0.0039
Broker acquires L-H insurance broker
(0.0039)
7
Bank with insurance subsidiaries
-0.0002
(0.0041)
8
0.0061
Bank acquires P-L insurance broker
(0.0030)**
9
0.0107
Bank acquires L-H insurance broker/
(0.0062)*
10
P-L insurer acquires P-L insurance broker
0.0053
(0.0062)
11
-0.0093
L-H insurer acquires L-H insurance broker
(0.0217)
12
0.0090
Agency/direct selling distribution system
(0.0108)
13
Size of broker acquirers, log(equity)
0.0004
(0.0018)
13
-0.0031
Size of bank, log(equity)
(0.0010)***
13
-0.0027
Size of P-L insurer, log(equity)
(0.0041)
13
Size of L-H insurer, log(equity)
-0.0008
(0.0038)
14
Acquirers with multiple transaction
0.0069
(0.0035)**
15
-0.0006
Tobin’s Q of acquirer
(0.0020)
Constant
0.0033
-0.0120
(0.0145)
(0.0143)
Observations
950
891
Note: *, **, *** denote significance at the 10, 5 and 1 percent levels, respectively. Robust standard errors are reported in
parentheses.
1
2
3
indicates Broker Dummy (1 if acquirer is broker). indicates Bank dummy (1 if acquirer is bank). indicates P-L insurer dummy (1 if
4
5
acquirer is a P-L insurer). indicates L-H insurer dummy (1 if acquirer is an L-H insurer). indicates Broker acquires P-L insurance
6
7
Agency/Broker. indicates Broker acquires L-H insurance Agency/Broker. indicates Bank with insurance (agency/broker)
8
9
subsidiaries at the time of acquisition. indicates Bank acquires P-L insurance Agency/broker. indicates Bank acquires L-H
10
11
insurance Agency/broker. indicates P-L insurer acquires P-L insurance Agency/broker. indicates L-H insurer acquires L-H
12
13
insurance Agency/broker. indicates An Insurer using agency/direct selling as distribution channel at the time of acquisition.
14
indicates Size of Broker, Bank, P-L insurer, and L-H insurer, measured by log(equity) one month-end prior to acquisition. indicates
15
firms buy more than once within a year. Firms Q value, measured by market equity/book equity.