The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Working Paper Series
Junk Bond Holdings, Premium Tax
Offsets, and Risk Exposure at
Life Insurance Companies
E lija h B re w e r III a n d T h o m a s H. M o n d s c h e a n
• W R
:
JUM
R '
P r -
FEDERAL h .
e.
BANK OF CHICAGO
W o rk in g P a p e rs S e rie s
Is s u e s in F in a n c ia l R e g u la tio n
R e s e a rc h D e p a rtm e n t
F e d e ra l R e s e rv e B a n k o f C h ic a g o
A p ril 1 9 9 3 (W P -9 3 -3 )
1
Junk Bond Holdings, Prem ium Tax Offsets, and Risk Exposure at
L ife Insurance Companies
Elijah Brewer HI
Senior Economist
Federal Reserve Bank of Chicago
and
Thomas H. Mondschean
Assistant Professor of Economics
DePaul University
Chicago, Illinois
April 1993
Acknowledgements
We thank Herbert Baer, George Benston, Larry Mote, Jonathan Neuberger, and an anonymous
referee for helpful comments. The research assistance of Andrew A. Bergad, Craig L. Knight,
George Rodriguez, and Peter Schneider is greatly appreciated. All views expressed here are
those of the authors and are not necessarily those of the Federal Reserve Bank of Chicago or the
Federal Reserve System.
2
Junk Bond Holdings, Prem ium Tax Offsets, and Risk Exposure at
L ife Insurance Companies
Abstract
Life insurance company (LIC) risk exposure has increased in the 1980s while capital ratios
have been declining. Although state guaranty funds exist to handle losses to policyholders in the
event of an LIC failure, these funds can create incentives for excessive risk taking as did the
federal deposit insurance system for savings and loan associations. This paper also examines the
relationship between stock market risk and LIC risk exposure. Stock market risk is found to be
positively related to financial leverage as well as to differences in asset mix in a sample of 44
LICs, confirming that market data can help identify institutions with greater risk exposure.
3
Junk Bond Holdings, Prem ium Tax Offsets, and Risk Exposure at
L ife Insurance Companies
The failures of several large life insurance companies, such as Mutual Benefit and Executive
Life in 1991, have raised the possibility of pervasive failures in yet another category of financial
intermediaries that would require government intervention and taxpayer expense.1 Given the
cost of the savings and loan (S&L) bailout to federal taxpayers and the importance of life
insurance to millions of policyholders, an evaluation of the current risk exposure and regulatory
structure of the U.S. life insurance industry would be useful for several reasons. First, the life
insurance industry is a major supplier of funds to capital markets, so insolvency problems could
affect credit availability. Second, because state guaranty funds provide protection for
policyholders, the incentive for LICs to take additional risks might be greater than it would be in
the absence of these guarantees, so it is important to evaluate whether the current guaranty
system affects LIC risk taking.2 Third, since guaranty fund assessments can be credited against
an LICs premium taxes in most states, insurance failures could reduce expected tax revenues to
state governments.
This paper is divided into six sections. Section two provides background information on
recent industry performance and assesses its current risk exposure. Section three analyzes the
characteristics of several LICs that failed in 1991. Section four examines the regulatory
environment and the role of state guaranty funds. Section five reports empirical results on the
stock market's assessment of life insurance riskiness. Section six concludes with some policy
implications.
H. BACKGROUND
In the process of offering risk protection to their customers, LICs expose themselves to a
number of risks. Mortality and morbidity risk are related to the probability of a policyholder
dying, conditional on age, illness, and other variables. In addition to these insurance risks, a life
insurance company faces risks similar to those of other financial intermediaries, including
4
interest rate risk, liquidity risk arising from policyholders' right to borrow against their policies
or to cash them in for their surrender value, and credit risk.
Historically, life insurance companies have played an important role in the bond and
mortgage markets.3 Within the bond market, LICs are major buyers of private placement debt,
which are securities issued in the U.S. but not registered with the SEC.4 Life insurers are also
very active in the commercial real estate segment of the mortgage market, which provides a
market for loans on nonresidential properties, such as office buildings and manufacturing plants.
Life insurance companies, commercial banks, and S&Ls are the major suppliers of credit to this
market.5
Lending in the private placement and commercial real estate markets requires substantial
amounts of information gathering in the form of credit evaluation and monitoring of borrowers'
managements through covenant enforcement. Recent studies of the private placement and
commercial real estate markets have indicated that the loans made by LICs in these markets
generally have terms that are less uniform than other investments, such as publicly traded
corporate bonds. As a result, private placements and mortgages are less liquid. Yields are
higher to reflect information gathering costs and greater default risk.
According to the Federal Reserve Flow of Funds Accounts, LICs represented approximately
11.6 percent of total assets held by financial intermediaries in the U.S. over the 1986-1990
period. Table 1 contains balance sheet data for the industry for selected years from 1970 to
1991. As of the end of 1991, life companies held over $1.5 trillion in assets. Government
securities as a percentage of total assets rose sharply in the early 1980s. Holdings of corporate
bonds grew rapidly in the latter half of the 1980s, especially between 1986 and 1988 when they
rose from 36.5 to 41.2 percent of total assets, due in part to the growth in corporate debt
securities not of investment grade ("junk" bonds). Direct mortgage loans declined from over
one-third to less than one-fifth of total assets from 1970 to 1991. These portfolio changes reflect
the movement toward greater securitization of financial instruments as well as rapid growth in
the amount of corporate debt outstanding. While these securities have enhanced the liquidity of
5
LIC portfolios, they may also have exposed LICs to prepayment risk (in the case of mortgagebacked securities) and credit risk (in the case of junk bonds).6
Turning to the liability side of the balance sheet, one can observe the growing importance of
pension and annuity business relative to traditional life insurance. Policy reserves for life
insurance in force fell from 55.7 percent of total assets in 1970 to 24.0 percent in 1991, while
reserves to cover annuity payments rose from 23.5 to 57.6 percent between 1970 and 1991.
However, regulatory capital as a fraction of total industry assets declined from 9.4 percent in
1970 to 8.0 percent in 1991. Regulatory capital as a percent of general account assets (total
assets less separate account assets) also fell from 9.7 percent in 1970 to 8.5 percent in 1990.7 In
1991, the life insurance industry increased its capital-general account asset ratio to 9.3 percent,
signalling an improved ability of firms to absorb losses without becoming insolvent [see Benston
(1992), Kane (1992), and Kaufman (1992)].
Effects of inflation and high interest rates in the 1970s and early 1980s forced changes in both
LIC investment strategy and the types of insurance products offered. To stem outflows and
attract additional funds, LICs developed new insurance products, such as universal and variable
life, which differed from traditional whole life policies in that the size of the death benefit and/or
the annual premium could change to reflect investment performance over the duration of the
policy. Another product was the guaranteed investment contract (GIC), which promised a fixed
return for a specified period. At year-end 1991, the share of industry general account assets
financed by GICs was about 8 percent.
Because the interest rate credited on universal life policies and other interest sensitive
products affected the demand for these instruments, LICs had an incentive to offer high rates
during the early years of these policies to attract new customers and forestall policy lapses and
surrenders by existing customers. Wright (1991) contends that competition with other
companies and pressure from insurance agents and brokers also kept rates high. The decline in
market interest rates in the mid- and late 1980s, however, squeezed profit margins for many
LICs. In order to maintain the high returns being paid on GICs and other liabilities, Wright also
6
contends that many insurance companies decided to increase interest income by either taking on
riskier real estate loans or reducing the quality of their corporate bond portfolios.7
Table 2 examines the financial characteristics of LICs, categorized by their book value net
worth-total asset ratios in 1990. Almost seventy percent of the industry's assets were held by
LICs with capital and surplus less than 6 percent of assets (low capital LICs). Low capital LICs
held more mortgage loans and junk bonds than companies with capital ratios greater than 6
percent (high capital LICs). At the same time, low capital LICs held a smaller proportion of
their assets in equity investments. GICs are a relatively more important funding source for low
capital LICs than for high capital companies. Return on equity is generally lower for low capital
LICs than for higher capital companies.
The effects of losses from inflation and disintermediation coupled with the decision of many
companies to increase risk exposure in order to maintain high returns on interest sensitive
liabilities have eroded the book value of industry capital over the past two decades. Moreover,
because under historical cost accounting losses on bonds, mortgage loans, and real estate
investments do not have to be realized unless they are sold or written down according to
generally accepted accounting principles, the accounting Value of net worth can overstate the
economic value of capital available to withstand future losses.
HI. CHARACTERISTICS OF RECENT LIFE INSURANCE INSOLVENCIES
One consequence of greater LIC risk exposure in recent years has been a rising number of
"financially impaired companies" (FICs). FICs are firms that were subjected to any one of the
following actions by state insurance departments: involuntary liquidation, receivership,
conservatorship, cease-and-desist orders, suspension, license revocation, administrative orders,
supervision, or any other action which restricted companies' freedom to conduct business [A. M.
Best (1992)]. Figure 1 shows that from 1976 to 1986, an average of 10 companies were
financially impaired each year. Since 1986, the number of FICs has averaged 36 per year, with
62 troubled firms in 1991. Moreover, the average size of FICs rose substantially in 1991. In
1991, FICs held about 3.2 percent of the industry's total assets. By contrast, for the 1984-1990
7
period, the average annual ratio of FICs' assets to industry assets was approximately 0.10
percent. To examine the 1991 experience in greater detail, characteristics of seven large l f
ie
insurance companies that failed are compared with those for the industry as a whole in Table 3.
The two Executive Life companies that failed were subsidiaries of First Executive
Corporation. Executive Life of California was seized by California regulators in April 1991.
The company expanded very rapidly throughout i s history, with total assets of $29 million in
t
1975, $606 million in 1980, $5.6 billion in 1985, and a peak of $13.1 billion in 1989. An
increase in policy lapses and surrenders and a writedown of i s bond portfolio caused total assets
t
to decline to $10.2 billion in 1990. The principal cause of failure was a large exposure to junk
bonds, which represented 62.7 percent of the company’ total assets at year-end 1990. As Table
s
3 indicates, the exposure was so large that a decline of only 8.3 percent in the value of their junk
bond portfolio was sufficient to wipe out completely both i s mandatory security valuation
t
reserve (which are reserves set aside, based on the risk classes of the bonds held by the insurance
company, to cover possible losses on bond holdings) and i s capital and surplus.
t
The situation at Executive Life of N e w York was quite similar. The company had total assets
of $42.6 million in 1975, $230 million in 1980, $2.4 billion in 1985, and $3.9 billion in 1989. In
1990, total assets f l 18.7 percent due to policyholder runs and asset writedowns. The company
el
was placed in receivership in April 1991. As with Executive Life of California, a large exposure
to junk bonds (64 percent of total assets at the end of 1990) was the principal cause of i s demise.
t
Both units of Executive Life possessed a bond rating of A (the second highest rating) from A. M.
Best at the end of 1988.
Fidelity Bankers of Virginia and First Capital of California, subsidiaries of First Capital
Corporation, failed for similar reasons as did First Executive. Fidelity Bankers, with a low book
capital ratio and 36.9 percent of i s assets in junk bonds in 1990, could withstand only an 11.7
t
percent decline in i s portfolio before exhausting i s capital and mandatory security valuation
t
t
reserve. The company was placed in receivership in M a y 1991. A. M. Best had given the
company i s highest rating of A + at the end of 1989, and downgraded i to B + at the end of 1990.
t
t
8
First Capital was also placed in receivership in M a y 1991 due to investment losses on i s bond
t
portfolio and increasing levels of policyholder surrenders. The company had grown from $221.4
million in 1980 to $4 billion in 1990. Approximately 40 percent of general account assets were
invested in junk bonds. Given i s low book capitalization (2.66 percent in 1990), the company
t
had l t l protection from future losses on i s assets. Nevertheless, i was rated A- in 1989 and B
ite
t
t
at the end of 1990.
Guarantee Security Life of Florida had $6.6 million in assets in 1980. Starting in 1982, i
t
concentrated on writing individual annuity contracts, and i assets grew rapidly. B y the end of
t
1990, i held $686.4 million in total assets with 60.7 percent invested in junk bonds, despite the
t
fact that i had not purchased any additional junk bonds since the beginning of 1988. Due to
t
insufficient information, A. M. Best did not assign a rating for the company.
The failures of Monarch Life of Massachusetts and Mutual Benefit of N e w Jersey do not f t
i
into the same pattern as the other five companies. Both companies' growth rates throughout the
1980s were well below the industry average, and holdings of junk bonds relative to capital were
comparable to those for the industry as a whole. Monarch had 10.9 percent of i s assets in junk
t
bonds, while Mutual Benefit's proportion stood at 3.1 percent. The problem with Mutual Benefit
was a large exposure to commercial real estate loans, primarily in the northeastern United States.
The ratio of capital to mortgage loans at the end of 1990 was 9.41 percent, implying that a
market value decline of
10 percent in i commercial real estate loans would be sufficient to wipe
t
out the firm's capital. Following a liquidity run by policyholders, Mutual Benefit was placed in
receivership by N e w Jersey regulators on July 15,1991. I received a rating of A from A. M.
t
Best as recently as the end of 1990.
Monarch Life of Massachusetts managed a relatively large amount of separate account assets
($3.6 billion in 1990) in addition to i s general account assets. The life company was seized by
t
regulators in M a y 1991 apparently to protect i from problems at the parent company, Monarch
t
Capital Corporation. According to Kopcke and Randall (1991), after the stock market crash of
1987, Monarch Capital used bank debt to invest heavily in real estate development and venture
9
capital deals in the N e w England region. The parent company experienced losses on these
investments, triggering a default on the bank loans. Although Monarch Life had a capital-asset
ratio well above the industry average, i s commercial real estate portfolio was vulnerable to the
t
N e w England recession. Monarch Life possessed an A + rating in 1988 and an A rating at the
end of 1989.
The failures of these seven companies can be divided into two groups. The fi s five failed
rt
because of a high exposure to junk bonds relative to their capital. They a l exhibited extremely
l
rapid asset growth. The last two companies failed due to problems in commercial real estate
investments in the Northeast. With the exception of Monarch, a l had capital-asset ratios below
l
the industry average at the end of 1990. Given these and other LIC insolvencies, an important
policy question i : who bears the cost of a l f insurance insolvency? To help answer this
s
ie
question, we now discuss the role of government regulation of l f insurance.
ie
IV. IN SO LV EN CY REG U LA TIO N O F L IFE INSURANCE CO M PA N IES
State insurance departments, not the federal government, regulate insurance companies. The
motivations for regulating l f insurance companies are similar to those for regulating financial
ie
services industries. Life companies have better information about the market value of their net
worth than do policyholders. Although many LICs have changed the nature of their insurance
products from insurance policies to investment contracts, traditional long term insurance
products s i l make up a significant proportion of l f companies' l a i i i s In the absence of
tl
ie
iblte.
regulation, l f companies would have an incentive to increase risk taking after the writing of an
ie
insurance policy. In the event of an LIC failure, there i also the possibility of contagion effect i
s
f
policyholders at other LICs lose confidence in their own companies' ability to meet their
obligations and exercise their surrender options.8
Companies that wish to write insurance in an individual state must receive permission from
the state insurance commissioner. Regulators enforce rate setting, asset restrictions, and other
policies established by state legislation. Insurance companies are required to f annual reports
ile
to state insurance commissioners with detailed balance sheet and income statement information.
10
The state insurance departments perform periodic examinations (usually once every 3 years) of
the companies operating within their borders. Most states also tax premiums in part to finance
the cost of regulation.
To protect policyholders and to manage insolvencies, a l 50 states (including the District of
l
Columbia) have set up guaranty funds. Prior to 1970, only one state had a guaranty system to
cover the obligations of l f and health insurance companies. In 1970, the National Association
ie
of Insurance Commissioners (NAIC) adopted a "model" guaranty system for subsequent
consideration by individual state legislatures. Within one year, nine states adopted legislation
based on or similar to the N A I C model. The guaranty systems are designed to satisfy benefit
claims of policyholders and annuitants in the event that, after liquidation, an insolvent company
does not have enough assets. These funds are financed by ex post assessments on the surviving
insurance firms which operate in the individual s ate. The size of the assessment for each
t
company i based on the proportion of the total premium income they generate. In 39 states, the
s
assessment can be offset against the company's state taxes, thereby shifting the cost of failure
directly onto state taxpayers. In other states, LICs are allowed to impose a premium surcharge to
cover the cost of the assessment.
For most states, coverage under guaranty funds i $300,000 in death benefits, $100,000 in
s
cash or withdrawal value for l f insurance, $ 100,000 in present value of annuity benefits, and
ie
$100,000 in health benefits. As shown in Table 4, some states cover a l insurance policies
l
written by an insolvent firm located in the s a others cover residents only. Some states cover
t te;
unallocated annuities such as GICs purchased by companies to fund pension plans up to a certain
amount, usually $5 million.
Because of differences in state guaranty funds and in how insolvencies are managed, who
actually bears the cost of an insurance failure varies across states. Surviving insurance
companies initially pay the assessment and claim i as an expense on their federal corporate
t
income tax return, reducing their federal income taxes. As the companies receive the tax credits
in subsequent years, these credits become taxable income. Because of the time value of money,
11
the federal government bears part of the cost of an insolvency since i does not fully recover the
t
present value of the tax decrease granted in the assessment year. The majority of the cost,
however, is paid by state taxpayers in the form of a loss in tax revenues. Barrese and Nelson
(1992) found that for 1990 life/health guaranty fund assessments, 73.6 percent was paid for by
state taxpayers, 8.9 percent by federal taxpayers, and 17.5 percent by the equity holders of the
surviving firms.
The way in which state guaranty funds are financed raises several policy concerns. First, the
LIC does not have to make any ex ante payment to receive the guarantee. Second, the
assessments are based on the ex post cost of the failure and have no relationship to current or
future LIC risk exposure. Third, companies in states with premium tax offsets have l t l
ite
incentive to monitor each other because over 80 percent of the assessment will be recouped
through lower taxes. Finally, insurance guaranty funds can weaken market discipline by
policyholders. Without insurance guaranty funds, policyholders would have an incentive to buy
insurance products from safe insurers.
As we learned from the S & L c
risis, underpriced deposit insurance creates a moral hazard
problem. Institutions with low net worth have every incentive to gamble for resurrection by
investing in riskier assets. If the investments make good, they keep a l the gains. If the
l
investments turn sour, the deposit insurer bears the cost of failure. The method of financing
insurance guaranty funds i worse in terms of the incentives i creates than federal deposit
s
t
insurance because assessments are ex post and can be credited on state taxes. Thus, a poorly
capitalized LIC has an incentive to write policies (even charging lower premiums to attract
business from healthy institutions), invest in risky assets, and hope for the best. If the
investments turn bad, the guaranty fund steps in and covers the costs of failure.9
Given the incentives that guaranty funds create for risk taking, i i important that state
t s
regulators vigilantly monitor the institutions to ensure they do not take excessive risks given
their level of capital. Since i may be difficult to measure risk exposure using accounting data,
t
we show in the next section how to use stock market data to examine the risk exposure of LICs.
12
V. T H E EFFE C T S O F ASSET M IX AND LE V ER A G E ON ST O C K M A R K ET R ISK
A. Theoretical C onsiderations
D o changes in asset mix at LICs significantly affect their riskiness? W e address this question
by examining the relation between the volatility of LIC stock returns and various asset groups.
The fi s step in the development of the model, following Black and Scholes (1973) and Galai
rt
and Masulis (1976), is to relate the volatility of the market return on LIC equity, o^y, to the
volatility of the return on an LICs assets, a A :
()
1
where (dMV/9A)(A/MV) i the elasticity of market value of equity with respect to the value of
s
the assets of a representative LIC. Equation (1) indicates that the volatility of LIC equity returns
i a function of: the volatility of the asset returns, ct^; the change in market value capital with
s
respect to the change in total assets, 3MV/3A; and the leverage ratio, A/MV.
Because we cannot observe a l the right hand side variables in equation (1), a simplified
l
specification, following Christie (1982), can be written as:
()
2
where Oj t i the equity return volatility (o^y) of the ith LIC in period t LEVj ti the ratio of
s
,
s
total assets to market value of capital of the ith LIC in period t ej ti an error term, and the
,
s
coefficients s q and sj are parameters to be estimated. Since greater leverage increases LIC
riskiness, we predict sj >
0.
Christie (1982) indicates that the volatility of equity returns i affected by other variables
s
besides leverage. For example, if an LIC holds a portfolio of assets with differing degrees of
risk, then changes in asset mix can either increase or decrease the volatility of equity returns.
The precise behavior of Cj t will depend on the variance/covariance structure of the returns on
the various asset categories. The asset categories analyzed in this study include: junk bonds
(JUNK), mortgage loans (MORT), real estate direct investments (DIRECT), equity holdings
13
(STOCK), government and investment grade corporate bonds (OBOND), and other assets
(OASSET) which include cash and policy loans.10 Changes in the relative investment in these
different assets can affect the volatility of LIC equity returns.
The volatility of equity returns i also affected by how much of the cost of managing
s
insolvent LICs i borne by stockholders. For example, if guaranty fund assessments can be
s
credited against state premium taxes, then taxpayers and not shareholders pay the majority of
these costs. Thus, changes in asset risk will be reflected more fully in stock return volatility for
firms that operate in states with no premium tax offsets than for firms that do [see Brickley and
James (1986) for a test of this hypothesis for S&Ls].
To examine the impact of financial leverage and asset mix variables on stock return volatility,
we estimate:
oi,t =s+Y
. 0
t2
=
s W +sLV/,/+sJN.r+s O T + D R C .
n . l E. 2.UKi, 3M R . s I E Tu
.
4
0,t
t
i,t
+ seS T O C K .i,t + s , O B O N D .i,t + ei,t ,
.
5
6
()
3
where a l asset variables are divided by total assets, and W ti a time d u m m y variable that i
l
s
s
equal to one for year t (t=2,...,T) and zero otherwise. W e included time d u m m y variables in the
empirical specification to control for possible correlation across time. W e excluded one asset
variable (OASSET) to avoid perfect multicollinearity; hence, one should interpret the
coefficients on the individual asset variables as the impact of switching funds from cash plus
policy loans into the particular asset category.
Estimation of equation (3) for a cross-sectional time series sample of l f companies can
ie
provide a test of the relation between asset mix variables and LIC risk as reflected in the
volatility of LIC equity returns. This measure of LIC risk, however, i imperfect because the
s
criterion of volatility of equity returns could cause an LIC to be judged more risky than another
LIC even if the former has more capital in every state of the world. A n alternative i a risk
s
measure employed by Boyd and Graham (1986), Wall (1987), and Brewer (1989). This
measure, referred to as Z-score, represents the probability that losses (negative profits) will
14
exceed equity. For each LIC, z values, defined as [(-l-n^/CTg], where | E i the mean weekly
i s
stock return in a given period and a E i i s standard deviation, are computed. Assuming the
s t
return on equity is a normal random variable, i can be shown that z is i s standard normal
t
t
variate, representing how far, in standard deviations, the rate of return would have to f below
all
i s expected value for the LIC to f i . To be consistent with other studies that have used this
t
al
measure of risk, we use the negative of z and denote i as Z-score. The advantage of this risk
t
measure i that Z-score (Zj t takes account of each LIC's equity and expected rate of return. A
s
)
version of equation (3) was estimated by using Zj t as the dependent variable. Since higher Zscores indicate lower risk, the coefficients in this equation are expected to have signs opposite to
those in the equation in which the risk measure c j ti the dependent variable.
t s
A third approach to obtaining a measure of risk i to estimate the probability of failure using
s
the square of the Zj tvalues for each LIC. I can be shown by Chebychev’ theorem that for any
t
s
symmetric distribution with a finite variance, the probability of failure, PROB, will be such that:
PROB< ---- .
(1+n/
(4)
W e employ the equality in equation ( along with estimates of a E and ( E,to obtain
4),
i
estimates of the probability of failure for each LIC. Equation (3) was estimated with the risk
measure P R O B jtrather than Cj tas the dependent variable to provide a check on the robustness
of our results attained by using the fi s two risk measures.
rt
B. D ata Sources an d E stim ation P rocedure
The data used in this paper are for 44 life insurance corporations whose stocks were traded on
the N e w York Stock Exchange, American Stock Exchange, or over the counter and who filed
annual reports of condition for each year from 1986 to 1990. Stock market data are from
Interactive Data Services, Inc. For multiple LIC holding companies, the assets of individual LIC
subsidiaries are consolidated using reports of condition to construct the balance sheet variables
discussed below.
15
To obtain our measures of risk, we use weekly stock market data. For each year in the
sample period, estimates of the standard deviation of the weekly returns on an LIC's equity were
computed using data covering the twelve month period of each year. The second risk measure,
Zj t i obtained by dividing one plus the average rate of return on equity by the standard
,s
deviation of the rate of return. Finally, the risk measure P R O B jtis estimated by calculating the
reciprocal of the square of the Zj tvalues. The market value of equity i calculated by
s
multiplying the number of shares outstanding at the end of each year by the price of the LIC’
s
equity at the end of the year.
The asset-capital ratio (LEV) i calculated as the ratio of total book value of invested assets to
s
the market value of capital. The variables JUNK, M O R T , DIRECT, O B O N D , and S T O C K
represent book values of junk bonds, mortgage loans, real estate direct investments, other bonds,
and corporate stock holdings, respectively.11 All asset variables were divided by the book value
of total assets. Equation (3) was estimated for a pooled cross-section, time series sample of LICs
from 1986 through 1990. Time d u m m y variables, D U M 8 6 - D U M 8 9 , are specified in each
version of equation (3) to control for the effects of risk of changes in time-specific factors that
are not captured by our other independent variables.12
Selected financial characteristics of the 44 stock LICs at the end of 1990 are presented in
Table 2. The sampled companies held about 22 percent of industry assets and had a book value
capital-asset ratio of 8.2 percent. They held a greater proportion of their assets in mortgage loans
and other bonds and financed a greater proportion of their assets with GICs than the industry.
C. Empirical Results
Results from estimating equation (3) using ordinary least squares are reported in Table 5
.
The estimated values of the parameters represent their cross-sectional average values. In Table
5, the f r t set of equations shows the results of tests in which the risk measure Cj ti the
is
s
dependent variable.13 The second set of equations presents the results of tests with Zj tas the
dependent variable. Finally, the last three equations show the results of tests with the risk
measure PROBj t as the dependent variable.
16
The results with Gj t as the dependent variable indicate a significant positive relationship
between LIC risk and the asset-capital ratio, verifying that greater stock return volatility i
s
associated with increased financial leverage. The coefficients on J U N K and O B O N D are
positive and statistically significant. In equation (2) of Table 5, M O R T , DIRECT, and S T O C K
are not significantly correlated with Gj t The coefficients on the time d u m m y variables indicate
.
that LIC risk rose in 1987, f in 1988, was unchanged in 1989, and rose in 1990.
ell
Equations (4) and (5) present results of tests with the risk measure Z j tas the dependent
variable. Because Zj ti an indicator of the probability of failure, i i a different measure of risk
s
t s
than Gj t These results suggest a negative correlation between Z lt and J U N K and O B O N D .
.
Since higher values of Zj t signal a lower probability of failure, the implications are the same as
the Gj t t s : higher risk i associated with greater LIC holdings of junk and other bonds. The
et
s
coefficients of M O R T and S T O C K are positive and are now statistically significant from zero.
Lower risk i associated with greater LIC holdings of mortgage loans and stocks. The time
s
d u m m y variables exhibit the same qualitative patterns as those found in the Gj tequations.
W h e n PROBj t i used as the dependent variable, in equations (7) and ( ) of Table 5, the
s
8
results are somewhat weaker than those of Gj t In both equations, LEV, JUNK, and the time
.
d u m m y variables have the same qualitative effects as those found in equations (1) and ( ) of the
2
table. However, there i no evidence of an association between LIC risk and O B O N D .
s
The evidence provided with respect to the impact of asset mix variables on LIC risk i mixed.
s
Regression results using two measures of risk indicate a significant positive relationship between
J U N K and O B O N D and risk, but the results using the third measure of risk show only a positive
relationship between J U N K and risk. These results are consistent with the notion thatjunk bond
investments are risk increasing. Holding leverage constant, there is l t l evidence of a consistent
ite
positive relationship between LIC risk and M O R T , DIRECT, or STOCK.
W e also examined whether states' policy on granting tax credits for guaranty fund
assessments affected the relationship between asset mix variables and LIC risk. Since
shareholders would bear the risk of unexpected assessments in states without tax offsets, we
17
expect changes in asset mix to have a greater impact on market risk for companies that do most
of their business in these states. W e test this hypothesis by interacting the asset mix variables
with a d u m m y variable, D U M , that equals one for LICs with 20 percent or more of their
premium income from states without premium tax offset and zero otherwise.14 The results are
presented in equations (3), (6) and (9) of Table 5. In the stock return volatility and P R O B
,
equations, only the coefficient of (JUNK)(DUM) i significantly different from zero. In the
s
equation using Zj t as the measure of risk, none of the interactive terms i significant. The stock
s
return volatility and P R O B of LICs with 20 percent or more of their premium income from states
without premium tax offsets tend to be more positively related to junk bond holdings than the
return volatility and P R O B of other LICs. Many of the recent LIC failures have been associated
with high junk bond exposure and occurred in states without premium tax offsets.
Thus, firms with large holdings of junk bonds appear to have greater stock return volatility.
Consistent with the findings of Brickley and James (1986), we find that access to government
subsidies (e.g. premium tax offsets) tends to affect firm common stock returns. Because firms in
no premium tax offset states are not insulated from the financial impact of an LIC failure,
changes in junk bond holdings will be reflected more fully in changes in stock return volatility
than firms that have access to premium tax offsets. An F-test was used to determine i the
f
estimated coefficients for LICs in states with no premium tax offset were significantly different
from those of LICs in states with a premium tax offset. In the aj tand P R O B j tequations, the
null hypothesis that a l asset mix coefficients are equal for the two groups can be rejected at the
l
10 percent level (F5 199 = 2.12 and F 5 199 = 12. ,respectively).
02
VI. CO NCLUSIO N
Recent failures of several large insurance companies have raised questions about the overall
risk exposure of the industry. Our analysis leads us to a number of conclusions. First, the
industry's overall risk exposure appears to have increased in the 1980s. Second, LICs with lower
capital ratios have higher concentrations of junk bonds and commercial real estate than do well-
18
capitalized LICs. Third, our empirical results show that stock market risk is positively related to
increases in both financial leverage and holdings of riskier assets such as junk bonds.
The case studies of seven large 1991 failures raise concerns about the incentive effects of the
guaranty funds on risk taking. Five of the failures resulted from rapid asset growth, low capital,
and excessive investment in junk bonds. Protecting policyholders from the effects of insurance
insolvencies may be worthwhile, but i may lead to increased risk taking. Many guaranty funds
t
are financed mostly by state taxpayers through premium tax offsets or by policyholders through
premium surcharges. Of the seven large 1991 failures discussed in this study, five were in states
that permitted l t l , if any, premium tax offset.
ite
The empirical results show that the lack of premium tax offsets dampens the relationship
between high junk bond exposure and stock market risk. W e believe that this i occurring
s
because premium tax offsets shift the risk of unexpected ex post assessments from shareholders
of risky firms to taxpayers because LICs do not bear the full cost of these assessments.
Guaranty funds essentially provide insurance for policyholders in the event of an insolvency.
LICs can benefit from these guarantees because their products become more attractive to their
customers. Thus, LICs should pay for the access to these guarantees. Currently, LIC premiums
are a fixed proportion of premium income and are assessed ex post, independent of asset risk.
As in the S & L crisis, this policy could lead to greater risk taking by LICs. To reduce potential
moral hazard problems, ex ante risk based premiums should be collected from a l participating
l
companies. Life companies should not be allowed to receive a tax credit for these premiums;
insurance i an expense of doing business. By making guaranty fund premiums risk based,
s
control over LIC risk taking may be improved. Finally, the amount of capital an LIC holds
should be made dependent on the overall riskiness of i s operations.
t
19
REFERENCES
Balestra, Pietro, and M. Nerlove, "Pooling Cross-Section and Time-Series Data in Estimation
of a Dynamic Model: the Demand for Natural Gas," Econometrica, July 1966,585-612.
Barrese, James, and Jack M. Nelson, "Distributing the Cost of Protecting Life-Health
Insurance Consumers," unpublished paper, The College of Insurance, April 1992.
Barth, James R., The Great Savings and Loan Debacle, Washington, D.C.: The AEI Press,
1991.
Barth, James R., Philip F. Bartholomew, and Carol Labich, "Moral Hazard and the Thrift
Crisis: An Analysis of 1988 Resolutions," Proceedings of a Conference on Bank Structure and
Competition, Federal Reserve Bank of Chicago, 1989.
Benston, George J , "The Purpose of Capital for Institutions with Government-Insured
.
Deposits," Journal of Financial Services Research, April 1992, 369-384.
A. M. Best Company, Inc., Best's Insolvency Study: Life/Health Insurers 1976-1991, June
1992.
Black, Fischer, and Myron Scholes, "The Pricing of Options and Corporate Liabilities,"
Journal ofPolitical Economy, May/June 1973, 637-659.
Boyd, John H., and S. L. Graham, "Risk, Regulation, and Bank Holding Company Expansion
into Nonbanking," Quarterly Review, Federal Reserve Bank of Minneapolis, Spring 1986, 2-17.
Brewer, Elijah HI., "Relationship between Bank Holding Company Risk and Nonbank
Activity," Journal of Economics and Business, 41 (1989), 337-353.
Brewer, Elijah HI, and Thomas H. Mondschean, "An Empirical Test of the Incentive Effects
of Deposit Insurance: the Case of Junk Bonds at Savings and Loan Associations," Journal of
Money, Credit, and Banking, forthcoming 1993.
Brickley, James A., and Christopher M. James, "Access to Deposit Insurance, Insolvency
Rules, and the Stock Returns of Financial Institutions," Journal ofFinancial Economics 16 (July
1986), 345-371.
20
Brumbaugh, R. Dan, J . Thrifts Under Siege, Cambridge, M A : Ballinger Publishing Co.,
r,
1988.
Cabanilla, Nathaniel, "Analyzing Developments in the Life Insurance Industry— Commercial
Mortgage Investments," Proceedings of a Conference on Bank Structure and Competition,
Federal Reserve Bank of Chicago, M a y 1992, 878-905.
Carey, Mark, S. Prowse, J Rea, and G. F. Udell, "The Private Placement Market:
.
Intermediation, Life Insurance Companies, and a Credit Crunch," Proceedings of a Conference
on Bank Structure and Competition, Federal Reserve Bank of Chicago, M a y 1992, 843-877.
Christie, Andrew A., "The Stochastic Behavior of C o m m o n Stock Variance," Journal of
Financial Economics, December 1982,407-432.
Curry, Timothy, and Mark Warshawsky, "Life Insurance Companies in a Changing
Environment," Federal Reserve Bulletin, July, 1986,449-460.
Fenn, George, and Rebel Cole, "Announcements of Asset-Quality Problems and Contagion
Effects in the Life Insurance Industry," Proceedings of a Conference on Bank Structure and
Competition, Federal Reserve Bank of Chicago, M a y 1992, 818-842.
Galai, Dan, and Ronald W. Masulis, "The Option Pricing Model and the Risk Factor of
Stock," Journal of Financial Economics, January/March 1976, 53-81.
General Accounting Office, "Insurer Failures: Life/Health Insurer Insolvencies and
Limitations of State Guaranty Funds,." Washington: Government Printing Office, March 1992.
Harrington, Scott E., "Should the Feds Regulate Insurance Company Solvency," Regulation,
Spring 1991,53-61.
IDS Financial Services, Inc., "Will the U.S. Life Insurance Industry Keep i s Promises," IDS
t
position paper, March 1990.
Kane, Edward J , "Corporate Capital and Government Guarantees," Journal of Financial
.
Services Research, April 1992,357-368.
Kane, Edward J , The S & L Insurance Mess, Washington, D.C.: Urban Institute Press, 1989.
.
21
Kaufman, George G., "Capital in Banking: Past, Present and Future," Journal ofFinancial
Services Research, April 1992,385-402.
Kopcke, Richard W., "The Capitalization and Portfolio Risk of Insurance Companies," New
England Economic Review, Federal Reserve Bank of Boston, July/August 1992,43-57.
Kopcke, Richard W., and Richard E. Randall, "Insurance Companies as Financial
Intermediaries: Risk and Return," in The Financial Condition and Regulation ofInsurance
Companies, Conference Series No. 35, Federal Reserve Bank of Boston, June 1991, 19-52.
Lang, Larry H., and Rene M. Stulz, "Contagion and Competitive Intra-industry Effects of
Bankruptcy Announcements," unpublished paper, N e w York University, January 1992.
Saunders, R.A., Life Insurance Company Financial Statements: Keys to Successful Reporting,
Chicago, Illinois: Teach’
em, Inc, 1986.
Wall, Larry D., "Has Bank Holding Companies' Diversification Affected their Risk of
Failure?" Journal ofEconomics and Business, 39 (1987), 313-326.
Wright, Kenneth M., "The Structure, Conduct, and Regulation of the Life Insurance
Industry," in The Financial Condition and Regulation ofInsurance Companies, Conference
Series No. 35, Federal Reserve Bank of Boston, June 1991,73-96.
22
FOOTNOTES
lrhe term life insurance companies (LICs) i used throughout to refer to firms that are classified
r
s
as l and/or life-health insurance companies.
ife
2For a discussion of incentive effects of deposit insurance, see Kane (1989), Barth (1992), Barth,
Bartholomew, and Labich (1989), Brumbaugh (1988), Brewer and Mondschean (1993), among
others.
3
According to Kopcke and Randall (1991), LICs held about one-half of the outstanding
corporate bonds during the 1960s. This share has declined to one-third in a recent five year
period. Over the past 30 years, LICs have held approximately 30 percent of al commercial
l
mortgages, but their shares of residential mortgage loans have declined.
4See Cary, Prowse, Rea, and Udell (1992) for a discussion of l f companies' participation in the
ie
private placement market.
5Cabanilla (1992) discusses the role of l f companies in the commercial real estate market.
ie
6Life insurance companies are not required to separate corporate and mortgage-backed securities
in the summary tables of balance sheet data filed with state insurance commissioners, so i i not
t s
possible for us to separate these two classes of debt in the tables.
7Separate accounts are defined as groups of assets in which the investment risk i borne by the
s
policyholder, and the insurer's guarantee i limited to mortality and expense charges [see
s
Saunders (1986)].
8See Lang and Stulz (1992) for an excellent discussion of the contagion effects of bankruptcy
announcements. Fenn and Cole (1992) analyzes the impact of policyholder behavior on the
market value of insurance companies in the event of an insolvency.
9 Harrington (1991) makes this point for property/casualty companies, which also benefit from
state guaranty fund coverage.
10O A S S E T also includes a small but unknown amount of collateralized loans and other
investment assets.
23
11With the exception of corporate stock holdings, all asset categories are reported at book values.
Corporate stock holdings which include preferred stocks and common stocks are recorded at
market values. However, Saunders (1986) indicated that preferred stocks on which dividends
have been paid for the past three years are carried at cost, while others are carried at market
value. All common stocks are reported at market values.
12For a discussion of the existence of "other effects" see Balestra and Nerlove (1966).
13Equation (2) was also estimated with each asset mix variable entering separately. The results
indicate a positive and significant association between risk and J U N K and risk and O B O N D ,
while they indicate a negative and significant association between risk and M O R T and risk and
STOCK. The results indicate a marginally significant correlation between risk and DIRECT.
^Alternatively, the d u m m y variable was defined to divide the sample into two groups using
10
and 30 percent of premium income from states without premium tax offset. The results of these
tests were qualitatively similar to those reported.
TABLE 1
Balance Sheet of Life Insurance Companies, Selected Years, 1970-1991
(Percent)
1970
1980
1990
1991
5.3
15.0
13.1
1.9
17.4
15.6
3.1
6.9
3.5
3.3
Corporate securities
Bonds
Stock*3*
42.7
35.3
7.4
47.4
37.5
9.9
50.5
41.4
9.1
50.8
40.2
Mortgage loans
35.9
27.4
19.2
17.1
Real estate
3.0
3.1
3.1
3.0
Policy loans
7.8
8 .6
4.4
4.3
Other assets0
5.3
6 .6
7.8
7.4
$207.2
$479.2
$1408.2
$1551.2
Assets3
Government securities
U.S.
State, local, and foreign
Total assets (billions of dollars)
2 .2
1.8
10.6
Liabilities and net worth
Policy reserves
Life insurance, total
Individual
Ordinary
Industrial
Group**
55.7
41.3
24.8
24.0
48.3
5.9
1.5
36.6
2 2 .8
2 2 .2
2 .6
2.1
0.9
0 .8
1.1
Annuities, total
Group
Industrial6
23.5
16.4
7.1
37.8
29.3
8.5
57.8
36.6
2 1 .2
57.6
35.3
22.3
Health insurance
1.7
2.3
2.4
2.5
81.0
81.4
85.0
84.1
Other liabilities^
9.6
10.0
7.5
7.8
Net worthS
9.4
8.5
7.5
8 .0
Total policy reserves
1.1
aThese data include assets in separate accounts.
^Market value.
cIncludes cash, due and deferred premiums, due and accrued investment income, and other items.
^Includes reserves for credit life.
includes reserves for individual annuities and supplementary contracts with and without life contingencies.
^Includes policy dividend accumulation, funds set aside for policy dividends, and other items.
^Includes capital, surplus funds, and mandatory securities valuation reserves. Adapted from Curry and Warshawsky
(1986).
Source: American Council of Life Insurance. Numbers may not add to totals due to rounding.
TABLE 2
Financial Characteristics of Life Insurance Companies as of December 31,1990
(percent of total general account assets)
Net worth
category
Proportion
of industry
Mortgage
loans
Junk
bonds
Other
bonds
Equity
investments
Real
estate
direct
investments
Policy
loans
Other
assets
Guarantee
investment
contracts
ROA
ROEa
00
.
-5.00
-
BVAb < 0
01
.
15.7
25.6
29.5
4.4
7.1
5.6
12.6
0< BVA <3
5.7
21.7
9.5
58.6
1.3
1.1
2.4
5.2
17.1
0.07
2.9
3 < BVA < 6
63.6
24.7
7.7
49.0
3.9
3.0
5.1
6.5
11.1
0.51
11.3
< BVA<9
13.4
2 1 .6
4.7
54.7
3.0
2.2
6.5
7.3
14.6
0 .6 8
9.9
9 < BVA <12
4.5
12.1
4.8
56.7
7.1
1.5
5.0
12.7
6.1
2.17
20.8
12 < BVA <15
3.6
13.3
3.4
63.7
5.6
0.9
5.9
7.2
00
.
2.03
15.9
BVA > 15
9.1
7.9
2.5
56.7
15.7
2 .0
3.8
11.5
0.5
4.54
15.4
2 1 .6
6 .6
51.9
4.9
2.6
5.0
7.4
10.3
1.00
13.2
6.7
12.9
1.64
20.0
6
Industry
44 Publiclv traded life insurance companies
2 1 .6
23.6
6 .0
54.0
4.9
1.8
aNel income as a percent of net worth.
^Book value net worth-total asset ratio.
Source: National Association of Insurance Commissioners (NAIC) Database of Annual Statements.
3.0
TABLE 3
Characteristics of Insolvent Insurance Companies
(Percent)
Growth in total assets
1980-85 1985-90
_________________ I29Q________________________
General
Net worth-total
Ratio to junk bonds of
account assets asset ratio
net wortha
net worth + MSVRb
(millions of dollars)
Executive Life (California)
823.45
81.63
$ 10,167
4.66
7.44
8.29
Executive Life (New York)
922.29
34.49
3,172
5.82
9.10
10.41
34.39 1,683.45
4,069
3.00
8 .1 2
11.72
Fidelity Bankers Life (Virginia)
675.17
135.13
4,035
2 .6 6
6.69
11.26
Guarantee Security Life (Florida) 5,519.37
85.23
686
4.48
7.38
1 1 .0 0
Monarch Life (Massachusetts)
35.17
12.98
851
11.70
107.01
115.82
Mutual Benefit Life (New Jersey)
58.73
41.28
13,006
3.38
108.81
138.89
Industry
66.38
69.20
SI,248,386
7.32
107.21
124.50
First Capital Life (California)
aNet worth includes surplus.
^MSVR refers to mandatory security valuation reserve.
Sources: Best’s Insurance Reports, NAIC database, and the 1991 American Council of Life Insurance Fact Book Update.
TABLE 4
Basic Provisions of State Life/Health Guaranty Funds
State
Coverage
GICs
Effective date
Max annual
assessments
Premium tax
offset
Alabama
0
S
1/1/83
2%
none
Alaska
l
Y
5/16/90
2%
20% for 5 years
Arizona
l
S
8/27/77
2%
20% for 5 years
Arkansas
l
Y
3/9/89
1%
recoup from policy surcharge
California
l
N
1/1/91
1%
none
Colorado
l
N
6/1/91
1%
20% for 3 yrs. 7.5% for 2 yrs. for
life and annuity, for health
recoup from policy surcharge
Connecticut
l
Y
10/1/72
2%
20% for 5 years
Delaware
l
Y
7/23/82
2%
20% for 5 years
District of
Columbia
l
S
7/22/92
2%
10 % for 10
Florida
l
S
10/1/79
1%
. 1 % per year
Georgia
l
Y
7/1/81
2%
20% for 5 years
Hawaii
l
N
7/1/88
2%
20% for 5 years
Idaho
l
N
6/1/77
2%
100% in 1 of the following 5 years
Illinois
l
Y
1/ 1/86
2%
20% for 5 years
Indiana
l
Y
7/1/78
2%
2 0 % per year or
years
recoup from policy surcharge
Iowa
l
Y
7/1/87
2%
20% for 5 years
Kansas
l
N
7/1/82
2%
20% for 5 years
Kentucky
l
N
6/17/78
2%
20% for 5 years
Louisiana
l
N
9/30/91
2%
20% for 5 years
Maine
l
S
7/25/84
2%
recoup from policy surcharge
Maryland
l
S
7/1/71
2%
none
Massachusetts
l
N
4/3/86
2%
10% for 5 years
Michigan
l
Y
5/1/82
2%
amount varies according
to a formula
Minnesota
l
Y
5/27/77
2%
none
Mississippi
l
Y
4/9/85
2%
25% for 2 years
Missouri
l
N
8/13/88
2%
20% for 5 years
T A B L E 4 (c o n t'd )
State
Coverage
GICS
Effective date
Max annual
assessments
Premium tax
offset
Montana
1
S
7/1/74
2%
2 0 % for five years
Nebraska
1
S
8/24/75
2%
20% for 5 years
Nevada
1
N
7/1/73
2%
20% for 5 years
New Hampshire
0
S
6/25/79
4%
none
New Jersey
1
Y
1/1/91
2%
10% for 5 years
New Mexico
0
S
4/9/75
2%
none
New York
1
Y
8/2/85
2%
80% when aggregate assessments
for all insurers exceeds $ 1 0 0 million
North Carolina
1
Y
4/13/74
2%
20% for 5 years
North Dakota
1
Y
7/1/83
2%
20% for 5 years
Ohio
1
Y
9/14/88
2%
20% for 5 years
Oklahoma
1
N
9/1/81
2%
20% for 5 years
Oregon
1
N
9/13/75
2%
20% for 5 years
Pennsylvania
0
S
1/25/79
2%
20% for 5 years
Rhode Island
1
S
6/20/85
3%
10% for 5 years
South Carolina
0
s
7/14/72
4%
20% for 5 years
South Dakota
1
N
7/1/89
2%
20% for 5 years
Tennessee
1
N
7/1/89
2%
10 % for 10
Texas
1
Y
9/27/73
1%
10 % for 10
Utah
1
Y
7/1/86
2%
20% for 5 years
Vermont
0
S
4/27/72
2%
20% for 5 years
Virginia
1
N
7/1/76
2%
.05% for 5 years
Washington
1
Y
5/21/71
2%
20% for 5 years
West Virginia
1
S
6/21/77
2%
none
Wisconsin
1
s
8/22/69
2%
20% for 5 yrs. if can't recoup
through policy rates
Wyoming
1
s
7/1/90
2%
10 % for 10
0=A11 policyholders
S=SILENT N=NO
l=Residents only
Y=YES
Source: National Organization of Life and Health Guaranty Associations.
yrs. or . 1% of premium
written, whichever is less
years
years
TABLE 5
The Impact of Financial Leverage and Asset Mix on Life Insurance Company Risk
(Pooled Cross-Section Time Series Results for 1986-1990)
Equation
Dependent
Variable
Intercept
LEV
JUNK
MORT
DIRECT
STOCK
OBOND
DUM
()
i
a
4.6643
(13.83)***
0.0286
(7.43)***
8.6163
(5.64)***
(2 )
a
2.4269
(1.56)
0.0263
(7.48)***
8.5416
(4.18)***
-2.6081
(-1.50)
8.0069
(1.31)
-3.2042
(-1.17)
a
-0.3131
(-0.14)
0.0145
(3.03)***
9.2055
(3.81)***
-0.0571
(-0 .0 2 )
9.6337
(1.47)
-0.5063
(-0.15)
7.8054
(3.14)***
Z
26.0371
(15.08)***
-0.0605
(-3.07)***
-30.5363
(-3.90)***
-
-
-
Z
35.1637
(4.57)***
-0.0442
(-2.54)***
-26.7556
(-2.65)***
14.1143
(1.64)*
-22.5754
(-0.75)
34.6522
(2.56)**
-22.1328
(-2.61)***
(6 )
Z
44.0501
(3.96)***
-0.0429
(-1.78)*
-35.5925
(-2.91)***
5.1386
(0.45)
-33.2300
(-1 .0 0 )
23.5076
(1.39)
-31.0212
(2.47)**
(7)
PROB
0.2559
(4.53)***
0.0066
(10.27)***
1.4277
(5.57)***
(8 )
PROB
0.1068
(0.38)
0.0065
( 10 . 12)***
1.4025
-0.3156
(3.75)*** (-0.99)
1.2188
(1.09)
(9)
PROB
-0.4497
(-1.2 2 )
(2.82)***
1.1919
(2.96)***
1.2400
(1.14)
-2.4712
(-0.55)
-
(5)
0.0397
(3.20)***
2.9541
(1.18)
(4)
(MORT)(DUM)
4.4254
(2.58)***
(3)
(JUNK)(DUM)
0 .0 0 2 2
-
-
1.6500
(0.43)
-
-
-
-
-
-
-
-10.9742
(-0.87)
-
0.0233
(0.37)
11.9362
(0.52)
-
-
-
-
-0.0906
(-0.18)
0.2943
(0.94)
-
-
-
0.2442
(0.44)
1.088
(2 .6 8 )***
0.5276
(1.26)
0.0160
(7.73)***
The l-slatistics in parentheses are starred if the regression coefficients are significantly different from zero at the 10(*), 5(**), and !(***) percent levels.
-
-0.2774
(-0.37)
TABLE 5 (continued)
The Impact of Financial Leverage and Asset Mix on Life Insurance Company Risk
(Pooled Cross-Section Time Series Results for 1986-1990)
Equation
Dependent
Variable
(DIRECT)(DUM)
a
-
(2 )
a
a
(4)
Z
--
(5)
Z
Z
(7)
PROB
(8 )
PROB
DUM88
DUM89
-
-
-0.9688
(-2.13)**
-0.0023
(-0 .0 1 )
-1.2536
(-2.83)***
-1.2430
(-2.80)***
0.4247
216
-0.5570
(-1.34)
0.2831
(0.70)
-1.1890
(-2.97)***
-1.2601
(-3.14)***
0.5306
216
-0.5021
(-1.23)
0.3046
(0.77)
-1.2144
(-3.105)***
-1.3303
(-3.39)***
0.5540
216
1.7440
(0.75)
-3.2875
(-1.45)
4.7911
(2 . 11 )**
6.8519
(3.01)***
0.2236
216
-0.8760
(-0.43)
-4.9948
(-2.50)**
4.3327
(2.19)**
6.9384
(3.49)***
0.4087
216
6.7868
(0.42)
-0.8393
(-0.41)
-4.7862
(-2.39)**
4.5305
(2.29)**
7.1598
(3.62)***
0.4145
216
-
-1.8037
(-2.37)**
-0.8422
(-1.13)
-2.0029
(-2.70)***
-1.8081
(-2.43)**
0.5141
216
-0.0652
(-0 .8 8 )
-0.1970
(-2.69)***
-0.1815
(-2.47)**
0.5280
216
-0.1320
(-1.95)*
-0.0458
(-0.69)
-0.1946
(-2.99)***
-0.1952
(-2.99)***
0.6275
216
PROB
(9)
DUM87
--
(6 )
DUM8 6
“
(3)
(OBOND)(DUM)
-0.1548
(-2.04)**
a)
(STOCK)(DUM)
-14.1972
(-0.38)
31.7707
(0.17)
-
-4.0852
(-0 .6 6 )
-0.3276
(-0.04)
-
-0.5115
(-0 . 12)
-
0.3663
(0.26)
-2.0927
(-0 .6 6 )
-
-0.5265
(-0.99)
R2
N
FIG U R E
1
N u m b e r of Financially I m p a i r e d Life I n s u r a n c e C o m p a n i e s ( 1 9 7 6 - 9 1 )
p e rc e n t
70 r
60
50
40
30
20
10
0
1976
1978
1980
1982
1984
1986
1988
1990
The numbers in parentheses are the percentages of the life insurance companies that were classified by A. M.
Best Company as financially impaired companies.
SO URC ES: A. M. Best Company (1992) and American Council of Life Insurance.
@FedFRASER