·
What is a life insurance policy?
o In
the United States a life insurance policy is a contract between the policy
holder and the insurance carrier (i.e. an insurer), where the insurer promises
to pay a designated beneficiary a sum of money (the "benefits") upon
the death of the insured person. In return, the policy holder agrees to pay a stipulated
amount (the "premium").
·
What is a senior life settlement?
o A senior
life settlement is a financial transaction in which the owner of a life
insurance policy sells a policy to a third party for more (than its cash value that
it would have received if the insurance policy had been surrendered to the
insurer) and less than its death benefit.
As upheld by the United States Supreme Court the policy is an asset of
the owner and may be traded. An
insured must typically be 65 or older to obtain interest from buyers. The price paid is typically calculated
by judging the relative health of the individual against future premiums
payments and the eventual death benefit to be paid.
·
What is LE?
o Life
expectancey (“LE”) is not a
prediction of how long an individual will live, but rather is the average survival time (i.e. the
50/50 point) amongst a particular risk cohort.
·
What is a viatical settlement; how is it different
from a senior life settlement?
o A
viatical is the sale of an insurance policy where the insured has a very short
medically underwritten LE – typically 24 months or less (e.g. the insured is
very sick, such as with AIDS, terminal cancer, etc.). In an effort to prevent economic abuse of the terminally
ill, viaticals are regulated separately from senior life settlements. A policy sale which would otherwise be
a senior life settlement becomes a viatical settlement if the LE is short
enough.
· Who
provides LE’s for the senior life settlement industry?
o Life
Expectancy Providers (LEPs, LE Providers) are specialized independent companies
that issue life expectancy reports (LERs) that estimate the life expectancy
(LE) of an individual (typically the insured individual on whose life a life
insurance policy involved in a life settlement is based).
o LEPs are typically made up of actuaries and medical
underwriters who utilize actuarial models based on published or proprietary
mortality (life) tables and medical underwriting based on various
debits/credits for various morbidity characteristics similar to the medical
underwriting performed by life insurance company underwriters and reinsurance
underwriters.
o Examples of LEP companies are AVS, 21st, Fasano,
etc.
o Most LEPs have factored in the experience data
underlying the 2008 VBT, as well as their own experience data and other
factors, as a basis for their mortality tables. This resulted in a significant
lengthening of average LEs in the fourth quarter of 2008 for some LEPs.
All major LEPs have continued the practice of developing and using
proprietary and confidential mortality tables based on extensive medical
research and mortality experience.
· What is
/ where did the VBT table come from?
·
Until recently, the most
commonly used mortality table was the 2001 Valuation Basic Table (VBT)
published by the Society of Actuaries based on data supplied by contributing
life insurance carriers. In 2008, the Society of Actuaries published a new table,
the 2008 VBT, that is based on 695,000 lives representing $7.4 trillion in
death benefits which is almost 3 times more lives than the former 2001
VBT.
·
Among other factors, the
standard VBT tables group individuals based on age, gender and tobacco use.
· What
is “Duration 0”, or “current q”, or “not back-dating the q”? What does it mean
with respect to the 2008 VBT?
·
The VBT tables show the rate
of expected mortality within a group starting at an attained age. Each year’s expected mortality is
referred to as q (q(x) is the probability that someone aged x will die before
reaching age (x+1).
·
An individual has a series of
q’s from the VBT table beginning at an observed age; the q numbers represent
the expected mortality based on the probability of having survived to the
observed age. This means that a 72
year old today has a different set of q numbers than a 70 year old today
looking two years in the future – this is because there is a probability that
the 70 year old may die prior to reaching age 72.
·
A “current q” assumes that
the individual’s observed age is his current age; this presents a conservative
approach. Other terminology used
when referring to the current q includes “Duration 0” and “not backing the q”.
·
In a “back dated q”, an individual
who is now 72 would be presented using a 70 year old’s q’s two years into the
table. As an example, using
current q’s a 72 year old has a
87.86% probability of reaching age 80.
Using the q’s for a 70 year old a time x+2 (ie they are now 72), the
probability of that individual reaching age 80 is only 85.91%. By back-dating the q we have increased
the probability that the insured will die. Any time the probability of survival is decreased (i.e. the
probability of dying is increased), the actual expected IRR results will be higher,
the costs to carry the policy will be lower and other negative factors come
into play.
·
Accordingly, parties using
back-dated q’s (e.g. the Duration 0 date, the date the insured first bought
their insurance policy, the date the policy was sold and became a life
settlement, etc.) will project a higher expected IRR then parties that use a
“current q”.
· What
is a mortality factor?
o The VBT projects the mortality of an individual in normal
health using a series of annual mortality numbers. A Mortality Factor (“MF”) is used to decrease or increase
the rate of the projection based upon the better than average or worse than
average (impaired) health of the individual. A MF above 1 (e.g. 1.10 or 110%) indicates an impairment
(the higher the number the greater the impairment) while a number below 1 (e.g.
.90 or 90%) would indicate better than standard health. An individual with an MF of 2.0 would
show a rate of mortality twice that of the base VBT curve.
·
One can choose to
mathematically ignore an insured’s mortality factor (such as provided by the LE
Provider) and apply an overall mortality factor to a policy or pool of policies,
or to mathematically recalculate a mortality factor from a given LE using a
chosen mortality table, such as the 2008 VBT (e.g. referring to the example provided above, using the 2008
VBT @ 100%, @ 110%, @ 90%) or a custom table.
· How
is the purchase price of a life settlement calculated?
·
At the most basic level, a
policy is purchased using a combination of two distinct data points. The first is the insurance policy which
is a legal contract which governs the amounts and timing of premium payments
and death benefit receipts. While
certain variables may change in the future, such as interest rates, as a
contract, the minimum amount of premium payments may be calculated. The second component is the LE which
predicts how and when the insured may die.
·
For the most part the life
settlement industry uses these two data points ((i) premium payment and death
benefit information from the policy, and (ii) LE) combined with the desired IRR
(e.g. 16%) within a Deterministic model to calculate the price paid for a
policy.
·
With a Deterministic model, the
insured is treated as though he (or she) were a large group (“probabilistic
approach”) – here, the model projects incremental mortality based upon the
annual q. Death benefits are collected
in small amounts and future premiums are reduced by the cumulative mortality. Thus if the curve projected 10%
mortality in the first year the model might show 10% of the death benefit
collected and only pay 90% of the premiums for year 2. This generates a stream of expected cash
flows.
·
The Deterministic cash flow
information is combined with the desired IRR to determine the maximum price
that will be paid.
·
Why is there a problem to
use the LE as the point in time when the person is dead?
·
This approach assumes that an
insured dies at a pre-established time – i.e. at the LE. This causes the model to project
100% of premiums until this date and collect 100% of the death benefit on this
date.
·
As discussed above, the LE is
not the point in time when the insured is expected to be dead. Using the Deterministic method
above, at the same LE point in time only 50% of the death benefit would be
included within the calculation.
· What is deterministic and stochastic? Why should a
person use both? What happens if you only use deterministic?
·
Again, with a Deterministic
model, the insured is treated as though he were a large group– here, the model
projects incremental mortality based upon the annual q. Death benefits are collected in small
amounts and future premiums are reduced by the cumulative mortality. Thus if the curve projected 10%
mortality in the first year the model might show 10% of the death benefit
collected and only pay 90% of the premiums for year 2. This generates a stream of expected
cash flows that is predicated on the assumption that a person can be partially
dead / alive at any given point in time.
·
With the Stochastic approach
– the insured is either 100% alive or dead at any given point in time, and the timing
of death for each insured is determined randomly. The model calculates and discounts cash flows based upon premiums
paid until death and death benefits then received. The process is repeated a large number of times and results
are aggregated. This is referred
to as a Monte Carlo analysis.
·
Each of these two methods
present various strengths and weaknesses:
o
Using a Deterministic
approach alone presents a set of cash flows which will not mimic reality. A Deterministic model ignores tail risk,
early deaths and the generally increasing cost of the insurance policy as the
individual ages. This method,
however, is very simple to model and requires very little actuarial knowledge.
o
Using the Stochastic (i.e.
Monte Carlo analysis) approach
supplies a range of outcomes and provides
a much better approximation of how the life settlement and associated pool of
policies is expected to perform – based upon a curve of the specified mortality
table (e.g. 2008 VBT) and applicable adjustments (e.g. for impaired health,
etc.).
o
Because LE is the point where
approximately half (50%) of a group will be alive, when simulating one policy –
there will be a significant difference between the best and worst iterations
(i.e. a wide standard deviation).
Accordingly, a Stochastic model (i.e. Monte Carlo analysis) is best used
on a group of policies; used on one policy the mean answer produced simply
approximates the answer by using the Deterministic (A.K.A. probabilistic)
approach. The Stochastic methodology
is the most complex since it requires actuarial knowledge, modeling knowledge
and computer time to run the repeated simulations. It does, however, help to better describe the risks and
rewards of a given portfolio.
·
Why does using both the
Deterministic and Stochastic methods and using the distribution of outcomes lead
to a better pool?
·
Since a policy purchase
should be based on more than simply an IRR (and all IRRs are not equal) using a
combination of these approaches (Deterministic and Stochastic) enables a
portfolio of policies to be produced which have desirable characteristics –
such as expected duration, costs to carry and return profile. While being less visually simple, the probabilistic
approach provides a better IRR approximation of a policy and the Monte Carlo
approach shows how the entire pool should be expected to perform while
providing details on tail risk, expected premium costs and duration, for
example.
