Chapter 10
Life Office Risks and Risk Management
10.1 The Interested Parties
The general concept of life office risk was introduced in Chapter 6,
where the risk was identified as arising from three main sources: insurance
risk, investment risk, and business risk. It is the purpose of this chapter
to look at these sources of risk in more detail, but it is first necessary to
define the overall nature of the life office’s risks.
A logical starting point is to consider the various interests of the
three main parties with whom a life office is concerned, namely the
policyholders, the shareholders (where appropriate), and the regulators.
10.1.1 Policyholders’ Interests
In Section 3.6.4 we described how life insurance companies will seek
to maximize returns, at least for their unit-linked and with-profits
policyholders, subject to the constraint of taking sufficient precautions
to maintain company solvency. This aim can be taken, perhaps, as a possi-
ble definition of success for the company; hence, life office risk can be
considered as the risk that it will fail to meet these aims. The risks are
therefore that:
1. The office will become insolvent, and hence unable to fulfill its
contractual obligations.
2. In some sense, the returns (or ‘‘value for money’’) provided for its
policyholders will be inadequate.
A policyholder’s interpretation of (2) will, of course, depend upon
the type of policyholder concerned. A holder of a nonprofit policy will
have already considered his or her contract as providing acceptable
value for money when agreeing to enter into the contract; hence, such
a policyholder only remains concerned with risk (1). However, offices
303
© 2005 by Taylor & Francis Group, LLC
304 Modern Actuarial Theory and Practice
must continue to ensure that nonprofit policyholders are not over-
charged for the risks they present, otherwise the public will not purchase
these contracts and will essentially be denied this important financial
service.
A holder of a unit-linked policy will wish to ensure that the charges levied
under the contract are fair, and also that the investment managers of the
office secure a ‘‘reasonable’’ return on the assets in which the policyholder’s
unit fund is invested. What is ‘‘reasonable’’ in this context will depend very
much on the nature of the unit fund or funds chosen by the policyholder.
This choice will have been made on the basis of information provided by
the insurance company, whether from the past performance of the funds or
from the stated policy regarding their future management. In either case,
the policyholder will have chosen the fund or funds whose expected profile
of returns matches his or her own risk preferences (see Section 3.8). Hence,
Downloaded by [University of Minnesota - Twin Cities Libraries] at
the duties of the life office are simply to ensure that the distributions of
returns from each fund are in accordance with its stated (or implied)
investment policy, and, subject to this condition, to seek to maximize the
returns obtained from each fund.
With-profits policyholders have the most complex requirements, as
they essentially contribute to the life office’s capital and hence demand
a fair return on that capital (see Section 8.6.1). This interest is obviously
shared with the shareholders where the office is proprietary. As described
in Section 8.2.3, a with-profits policyholder’s reasonable expectation
of returns can be considered to be some kind of smoothed asset
share, the extent depending on the method of profit distribution concer-
ned. There are a number of component elements in this expectation,
relating to:
1. the smoothing policy adopted
2. the charge made for the policy guarantees
3. the investment of the accumulating asset share
All three of these elements depend very much upon the office’s past
behavior and stated intentions of future behavior in these respects.
Furthermore, the office will have to invest its policyholders’ asset shares
in a way that is consistent with the expected smoothing policy and
the policy guarantees, i.e., in a way that would not be expected to threaten
the continuation of (1) and (2). It should be noted that, in many offices, the
with-profits policyholders’ asset shares will include an appropriate
share of the office’s profits from its other (nonparticipating) business,
consistent with the use of the with-profits’ capital in financing these
activities. There may also be contributions to asset shares from the
returns earned from the estate, reflecting policyholders’ expectations
of a share in these returns (in the case of a mutual company, this would
be a 100% share).
© 2005 by Taylor & Francis Group, LLC
Life Office Risks and Risk Management 305
10.1.2 Shareholders’ Interests
The shareholders’ interests are similar to those of the with-profits policy-
holders, except of course that they are concerned with their own share of
the office’s profits or losses in relation to the capital they have provided.
Like the policyholders, they will also expect some return from the estate,
depending upon how much of the estate can be considered as owned by
the shareholders; see Smaller et al. (1996). In proprietary with-profits offices,
the transfer of profit to the shareholders is closely linked to the distribution
of policyholders’ profits (see Section 8.7), so that meeting the interests of
the with-profits policyholders should normally also satisfy the expectations
of the shareholders.
Downloaded by [University of Minnesota - Twin Cities Libraries] at
10.1.3 Regulators’ Interests
As we mentioned in Chapter 9, the aim of the regulators is to ensure that the
policyholders’ reasonable expectations are met. Hence, again satisfying the
policyholders’ interests should, at least in theory, also satisfy the regulators.
However, the regulations ensure that this is not left entirely to chance:
companies have to satisfy the specified regulations in order to continue in
existence as a going concern.
10.2 The Nature of the Risks
Bearing in mind the foregoing, we can summarize the risks that life offices
have to manage into three main elements:
1. The risk that the office will become insolvent (‘‘actual’’ or ‘‘true’’
insolvency).
2. The risk that the office will not be able to meet the statutory solvency
requirements (‘‘statutory’’ insolvency).
3. The risk that the returns to the policyholders and/or the shareholders
will be ‘‘inadequate.’’
Actual insolvency (risk 1) occurs when the actual amount of a company’s
committed (i.e., certain) liabilities exceed the amount of assets at any point
in time. While trading would then cease immediately, it is unlikely that the
existing policyholders would receive very much compensation from the
dispersion of the company’s assets relative to the expected value of their
former contractual benefits. The shareholders would certainly lose all of
their equity.
Hence, risk (1) essentially represents the most severe outcome of risk (3);
the return to shareholders and policyholders could not be more inadequate
than this.
© 2005 by Taylor & Francis Group, LLC
306 Modern Actuarial Theory and Practice
However, it is important to identify true insolvency as a separate risk,
because it marks a point at which something dramatic happens: the office
ceases to honor its contractual liabilities to its policyholders. Prior to that
point, full contractual benefits may still have been payable on claims
(financed, for example, by cash flow and/or by depletion of assets). The
returns to policyholders, therefore, fall dramatically at the point of insol-
vency; the returns to shareholders will show a similar effect but it may be a
little less abrupt, depending upon whether or not dividends were being
paid up to the bitter end.
The above discussion of true insolvency is, of course, almost completely
hypothetical. This is because, in practice, true insolvency would never be
allowed to happen; statutory insolvency would always happen first, as
a result of which the fund would normally be wound up or transferred
to another insurer. In any case, the policyholders and shareholders
Downloaded by [University of Minnesota - Twin Cities Libraries] at
will receive some value for their respective entitlements to the assets
(which will not be zero), depending on the terms of termination agreed
with the regulators, although it might leave policyholders with an entitle-
ment to smaller benefits than originally contracted under their policies.
Hence, risk (2) marks another occasion in which a sudden fall in policy-
holders’ returns occurs, i.e., it has a similar effect to actual insolvency
but with a much reduced degree of severity. The shareholders, however,
will almost certainly lose the whole of their equity on statutory insolvency,
as all of the available assets will be used to provide compensation to the
policyholders.
Management needs to concentrate on managing the risks in the order
(1) to (3). A completely solvent office would, therefore, be concerned
with managing risk (3) (although not taking any action which would
unduly threaten risk (2)), while an office with a small statutory solvency
margin would be mostly concerned with managing risk (2). Should
statutory solvency be irrevocably breached, then the concern would focus,
immediately, upon risk (1).
Risk (1) will be prevented if, overall, the income to the office generated by
each tranche of business is at least equal to the outgo incurred, by the time
that the tranche is terminated. The true insolvency risk, therefore, depends
upon the constituents of the office’s income and outgo, i.e.:
for income — capital provision, premiums, investment income, and asset
appreciation
for outgo — claims, expenses, tax, and distributions of profit
Risk (2) depends additionally on the generation of statutory valuation
strain during the policy term. For example, one tranche of nonprofit policies
would produce statutory insolvency at duration 0þ (i.e., immediately after
the first premium has been paid) if there was insufficient capital provided;
see Section 9.3.1. The most important additional intrinsic factor affecting
© 2005 by Taylor & Francis Group, LLC
Life Office Risks and Risk Management 307
risk (2) is product design, because of its importance in determining the
statutory profit signature.
Risk (3) is the risk of not meeting policyholders’ (and shareholders’)
reasonable expectations of making returns. The aim of life office manage-
ment is, therefore, that these reasonable expectations are at least met, and
in most cases this will manifest itself in attempts to increase profitability and
to ensure a fair distribution of those profits. Attempts to maintain or increase
policyholders’ and shareholders’ returns by overdistribution (i.e., by
distributing consistently more profit than earned) will ultimately increase
risks (2) and (1). On the other hand, persistent retention of profit (i.e., over
and above what would be considered a reasonable charge for capital or
for prudent management of the estate) is unfair to the policyholders and
shareholders and contrary to expectation.
Of course, part of this management process is to ensure that the
Downloaded by [University of Minnesota - Twin Cities Libraries] at
expectations of its policyholders are, in fact, reasonable. Hence, expectations
should not be influenced by:
figures of past policy payouts in isolation of the conditions which led to
those payouts
monetary projections of future policy payouts, unless properly qualified by
the assumptions underlying the projections
Policyholders’ expectations are likely to be significantly influenced
by the advice given to them during the sale of their policies; the behav-
ior of the sales intermediary is of paramount importance in shaping
policyholders’ expectations, reasonable or otherwise. Hence, an important
element in controlling the risk is the control of the behavior of the
intermediaries.
10.3 Risk Control
We are now going to consider the various contributing elements to the life
office’s overall risk and the ways in which these risks can be controlled. In
order to do this, it is necessary, or at least very helpful, to have a benchmark
against which experience can be compared. For conventional nonprofit
contracts, the obvious benchmark is the office’s premium basis: the
assumptions that were made in setting the prices at which the policies
were, and are being, sold. Should all these assumptions be exactly met in
practice, then the office will earn the levels of profit it expected when the
policies were issued and, provided the policies were appropriately priced,
the office should meet all three of its risk objectives. Should the experience
be worse, in aggregate, than that allowed for in the premium basis, then the
amount of profit will be lower than anticipated. This will, ultimately,
translate into lower than expected returns to the participating policyholders
© 2005 by Taylor & Francis Group, LLC
308 Modern Actuarial Theory and Practice
and/or the shareholders (depending upon which parties share in the profits
from the nonparticipating business). The risk must, therefore, not only be
considered in terms of whether the experience is worse than the chosen
benchmark, but also by how much it is worse. Hence, the significance of the
risk depends upon both its incidence and its intensity.
All other types of contract provide benefits that depend in some
way upon future investment returns. Hence, the appropriate benchmark
for the investment risk under these contracts will relate directly to what
the corresponding policyholders consider to be adequate in the prevailing
conditions: this was discussed in Section 3.6 and Section 3.8. The other
sources of risk for these contracts, i.e., the insurance and business risk
elements, can refer to the pricing basis benchmark described above. For
example, under a unit-linked contract the company runs the (investment)
risk that the returns it achieves on its linked assets will be inadequate for
Downloaded by [University of Minnesota - Twin Cities Libraries] at
its policyholders; it also runs the risks that its expenses and mortality
experience will be worse than that assumed in calculating its charges.
The control of life office risk should, therefore, be considered with
respect to each of its three main sources described in Chapter 6, namely
investment risk, insurance risk, and business risk. Life office investment
risk was described in Chapter 3; we consider insurance risk and business
risk below.
10.4 Insurance Risk
This is the risk that the cost of contractual claims will be higher than that
assumed in the premium basis. This risk has two elements: the amount of
claim and the incidence of claim. The incidence of claim is determined by
the occurrence of contingencies specified in the policy conditions, such
as death, survival, and sickness. The premiums are usually calculated on
the basis of the expected behavior of these random events, according to
some statistical model. The incidence of claim can, therefore, differ from
that assumed due to:
error in the parameterization of the model, e.g., that the assumed expected
mortality rates in an assurance premium basis are lower than the actual
expected mortality rates
random variation around the actual expected incidence of claim
error in the choice of model structure
These elements have been discussed by Daykin et al. (1994), Cairns (1995)
and by Hooker et al. (1996). They are sometimes referred to as parameter
error, process error, and specification error respectively and are not confined to
life insurance risk: they apply equally to the use of any model in actuarial
management (see Section 11.7 and Section 14.3 for examples).
© 2005 by Taylor & Francis Group, LLC
Life Office Risks and Risk Management 309
Most long-term policy claim amounts are determined by precise, fully
specified rules. This contrasts with the majority of general insurance (short-
term business) policies, which are contracts of indemnity (see Chapter 12),
under which the benefit is determined by the event itself, i.e., by the amount
of loss incurred by the event. Claim amounts under long-term insurance
business are, therefore, fundamentally predictable, and this feature greatly
assists in restricting the extent of the insurance risk.
Nevertheless, claim amounts still contribute significantly to the insurance
risk of life offices, due to the potential concentration of large sums assured
payable on the death of single individuals. If we have nx independent
policyholders aged x, each with probability qx of dying before age x þ 1, and
for which sx, i is the sum assured payable on death to the ith such
policyholder, then it can be shown that the variance of the random amount
of total claim Cx under the nx policies over the year of age is
Downloaded by [University of Minnesota - Twin Cities Libraries] at
X
nx
VarðCx Þ ¼ qx 1 qx ðsx, i Þ2
i¼1
Hence, for the same total sum assured at risk and for the same number
of policies insured, the variance of Cx will be greater the more concentrated
the risk is in individual lives, even though the expected death cost will
be the same. This is equivalent to saying that the greater the variability
in the sum assured between individual policies, the greater will be the
variability in claim cost and, hence, the higher the probability of insolvency
(all else being equal).
10.4.1 Underwriting and Risk Classification
The main way in which life offices control their claim incidence rates
(or ‘‘frequencies’’) under their life and sickness insurance contracts is
by initial underwriting. Underwriting is the means by which life offices
judge whether an applicant should be accepted for insurance, and if so to
establish the terms of acceptance. The aim is to classify the applicants
into the following general risk groups:
1. insurable at the office’s standard rate of premium for the contract
applied for
2. insurable, but with special conditions attached
3. possibly insurable at a later date
4. uninsurable
For marketing reasons, the premium rate assumptions should be such
that the majority of applicants would be acceptable at the standard
© 2005 by Taylor & Francis Group, LLC
310 Modern Actuarial Theory and Practice
rate of premium. These standard rates will be specific to the age of the
policyholder (and term of the contract, if appropriate), and in some cases
may be dependent upon other risk characteristics, e.g., upon the smoking
habits of the applicant. (See also Section 10.4.13.)
A minority of applicants should fall into the second category, which
is referred to as the ‘‘impaired lives’’ category. The special terms may
include a debt on the policy (a deduction from the normal level of
benefit) which might reduce over the term of the policy, charging of a higher
premium, or exclusion of certain causes of death from the terms of
the insurance. The terms are decided on an individual basis by the life
office’s underwriters in order to reflect their best assessment of the
particular mortality risk presented by the applicant. Examples of impair-
ments leading to special terms include known health conditions which
could affect future mortality (e.g., heart disease, diabetes, obesity), various
Downloaded by [University of Minnesota - Twin Cities Libraries] at
‘‘risky’’ aspects of lifestyle (e.g., smoking, alcohol consumption, drug
use, sexual promiscuity, dangerous sports) and family history (e.g., a
high incidence of certain diseases among close blood relatives can
sometimes affect the terms of acceptance). There is now also the possibility
of genetic test results providing information about mortality risk;
see Section [Link].
A small number of applicants (group (3) above) may be currently
uninsurable because of uncertainty regarding the level of risk involved.
Examples would include applicants who are recovering from an opera-
tion or currently receiving treatment for some serious condition. The life
office may defer consideration of these applicants until some time in the
future when the prognosis becomes clearer.
There will be a final group of applicants whose condition is such
that their expected claim experience cannot be predicted with sufficient
accuracy to enable them to be insured. This would include people with
certain congenital diseases, and most people with terminal illnesses, such
as cancer and AIDS. This group, which should form a very small minority
of applicants, would be declined insurance.
The overall aim of the underwriting process is for the average morta-
lity experience under each group of applicants to be reflected by the pre-
miums or charges paid. In particular, underwriting seeks to avoid adverse
selection, e.g., to prevent individuals who present a significantly higher
than average mortality risk from being charged the standard (and hence
inadequate) rate of premium, or from securing an exceptionally high level
of insurance cover (see Section 10.4.2).
[Link] Genetic Testing
A developing concern in the area of risk classification and underwriting
is the development of genetic testing. This is where individuals can be
screened for the presence or absence of specific errors in their genetic
© 2005 by Taylor & Francis Group, LLC
Life Office Risks and Risk Management 311
constitution, which can (so far only in rare cases) be a relevant predictor
of mortality risk. The most quoted example is the case of Huntington’s
chorea, where an early (pre-senescent) death can be accurately predicted by
the presence of a particular genetic error, which is identifiable by a genetic
test (see Barnaby (1997), Wilkie (1997), and Le Grys (1997)). However,
the possibility of using genetic test results routinely to make accurate
risk assessments of major causes of death, such as heart disease and cancer,
still seem a long way off, if not impossible to achieve, especially when
the significant influence of environmental factors is taken into account.
Despite its current rarity, there is considerable debate concerning
the ethics of using genetic test results for insurance risk classification
purposes (see Barnaby (1997) for a brief overview). Practice in different
countries is certainly not uniform; in some territories, known test results
may have to be disclosed in insurance applications, whereas in others
Downloaded by [University of Minnesota - Twin Cities Libraries] at
the use of any genetic test information in underwriting is not allowed. To
our knowledge, nowhere is it permissible practice for insurance companies
to require applicants to undergo a genetic test.
[Link] The Underwriting Process
The process of underwriting is described by Diacon and Carter (1992),
Luffrum (1989), and by Black and Skipper (2000). A brief outline will be
given here. Further reading on underwriting can also be found in Fisher
and Young (1965), and in Leigh (1990).
Proposers for long-term insurance usually have to complete a kind of
medical questionnaire, normally on the proposal form. On the strength
of their responses to the questions, the insurer may seek to obtain further
information from the proposer’s medical practitioner, and/or commission
an independent medical examination. While each additional stage in this
sequence provides the underwriters with an increasingly reliable means
of assessment of the risk presented by the proposer, they necessarily
incur additional expense. Hence, these additional stages (medical report
and medical examination) are only followed when the risk appears to be
potentially higher than usual. This would occur when:
The proposer indicates some unfavorable or potentially unfavorable
aspect of his or her health or lifestyle in the initial proposal.
The level of benefit applied for is higher than some preset limit for the age
of the applicant (reflecting the importance of higher than average
benefit levels on the variability of the claim cost, as described earlier).
In this way, it is hoped to employ the more detailed underwriting proce-
dures in the most cost-effective way, i.e., only in those areas in which the
more accurate classification of the risk is most likely to lead to significant
savings in claim costs in the future.
© 2005 by Taylor & Francis Group, LLC
312 Modern Actuarial Theory and Practice
An office must, therefore, decide upon its desired compromise between
underwriting cost and the accuracy of its risk classification, both of
which will also affect the premium rate that it can charge. There may
be marketing as well as risk control considerations which will affect this
decision, as, for example, it would be clearly very expensive and inefficient
to have a very competitive premium basis if the company’s market
consists of relatively high risks, such that the majority of its applicants
might have to be accepted on ‘‘special terms.’’
[Link] Preferred Lives
In some territories (particularly the U.S.) life insurance premium rates
differ according to a much wider range of rating factors, such as income
level, occupation, geographical region, family history, and the height-
Downloaded by [University of Minnesota - Twin Cities Libraries] at
to-weight ratio (Le Grys, 1997). The variation in premium rates is broadly
50% to standard rates. The advantages are that good risks get cheaper
rates (and insurance companies can attract more low-risk business), but
the disadvantages are that bad risks can feel unfairly discriminated
against, particularly where the cause of their higher risk is largely or
entirely outside of their control (e.g., due to social class background or to
family history). In the U.K., for example, insurers have made little headway
in this direction, other than charging different rates for males and females,
and for smokers and nonsmokers.
[Link] Underwriting Annuity Risks
Underwriting is not necessary for applicants for annuities, as they are
‘‘self-selected’’: people who perceive themselves to be in poor health are
unlikely to purchase annuities, so that the average initial mortality
of annuitants would be expected to be lower than that of the general
population. The office must, therefore, seek to ensure that the mortality
rates assumed in its annuity premium rates are at least as low as the
average mortality that will be experienced by the annuitants, in order to
avoid losses. It should be noted that population mortality rates in most
developed countries have generally improved (i.e., reduced) over recent
years, so that it is particularly important to incorporate anticipated future
reductions in mortality rates in the annuity premium basis.
[Link] Impaired Life Annuities
A recent and increasingly popular development is the issuing of
impaired-life annuities. Here, applicants for annuities who have poor
health or have a predictably raised mortality risk (such as smokers) may
be offered more favorable annuity conversion rates, reflecting their shorter
life expectancy. This has proved more popular in the U.K., for example, than
© 2005 by Taylor & Francis Group, LLC
Life Office Risks and Risk Management 313
its assurance equivalent (the preferred lives concept described above).
Nonetheless, some important ethical issues are involved, e.g., if smoking is
used as a rating factor then it could provide people with a financial
incentive to smoke. (For further information on impaired life annuities,
see Ainslie (2000).)
10.4.2 Financial Underwriting
There is some evidence, particularly from North America, that higher
mortality experience tends to be associated with policies having higher
sums assured. It would be rational for policyholders who perceive
themselves to be a high risk to be more likely to choose a high level of
benefit; this is an example of adverse selection.
Downloaded by [University of Minnesota - Twin Cities Libraries] at
There are potential ‘‘moral hazards’’ in life insurance. There may be a
temptation to induce early claims, e.g., by suicide. Particularly suspect
cases are proposals for life insurance on the life of a different person
from the one who will be paying the premiums, referred to as ‘‘life of
another’’ cases. There is also a risk that claims may be fabricated, usually
by claiming deaths to have occurred in some overseas country where
identification and registration of deaths are not rigidly enforced.
Financial underwriting is designed to identify these potential risks,
if possible, in the following ways:
1. In order to ensure that the proposer has a valid insurable interest in
the life of the policyholder that is at least as large as the amount to
be insured. An insurable interest can be considered to exist if the
policyholder would incur financial loss by the insured event, i.e., on
the death of the life assured. While all individuals are assumed to have
unlimited insurable interest in their own or their spouse’s lives, in any
other ‘‘life of another’’ case the insurable interest has to be demon-
strated. Valid ‘‘life of another’’ cases do exist, of course; examples
include a person who may have an insurable interest in his or her
parents; a firm, who may have an insurable interest in particular
individual ‘‘key’’ employees, whose death in employment would be
expected to incur a reduction in profit to the firm (known as keyman
assurance). (See Diacon and Carter (1992) for more information on
insurable interest.)
2. In order to ensure that the need for the insurance (and at the
proposed level) is present, and that the prospect has adequate financial
means to pay for the anticipated level of cover throughout the
term of the policy. This should uncover most cases of moral hazard,
such as intended suicide, and would also tend to reveal prospects
who, however altruistically, perceive their mortality risk to be excep-
tionally high.
© 2005 by Taylor & Francis Group, LLC
314 Modern Actuarial Theory and Practice
Financial underwriting risks can extend over different life offices;
proposers could secure a very high total level of death benefit by
taking out modest-sized policies with a number of different firms.
Therefore, it is customary to ask proposers to state whether they
have applied for policies from other life companies, although compliance
is difficult to verify.
10.4.3 Reinsurance
Reinsurance is essential for reducing the risk caused by the concentration of
sums assured under individual policyholders. The essential principle of
reinsurance is that it is a means by which a single risk, accepted by one life
office, is shared with another life office, or offices.
Downloaded by [University of Minnesota - Twin Cities Libraries] at
The whole insurance and reinsurance process can be summarized as
follows (from Spedding (1989)).
1. A life office (the direct-writing office) issues a contract of assurance with
a member of the public for a sum assured of £A.
2. The direct writing office (the ceding company) may effect a contract of
reinsurance for £B (B A) with a reinsurance company, with benefit
payment contingent on the original risk (e.g., on the death of the
policyholder). It may alternatively effect several contracts, each with
different reinsurance companies, with total amounts reinsured equal
to £B.
3. The reinsurance company may effect a contract (or contracts) of
reinsurance with other insurers (the retrocessionaires) for amounts
totalling £C (C B).
4. Retrocession could continue until the original risk of £A is shared out
among a number of insurance providers, at levels that individually do
not constitute an excessive risk to each office. Clearly, the larger A is,
the greater the number of offices that will generally be involved in
collectively covering the risk.
Thus, reinsurance allows each office to keep the total benefits insured
at any one time for any individual life below a certain maximum
level, thereby restricting total claim amount variability and as a result
helping to reduce the probability of insolvency. If the life office also acts as a
reinsurer, then it will gain a greater diversity of risk in its portfolio
of policyholders, which will also help to reduce variability, e.g., by reducing
the incidence of nonindependent deaths among the lives assured. The
reinsurance process also allows individual offices to accept much larger
risks than they could do otherwise, in the knowledge that the office
itself will only need to retain a certain amount of the total loss under its
own account. This obviously helps the office to sell business, and the
© 2005 by Taylor & Francis Group, LLC