4.

Life Insurance
4.1 Survival Distribution And Life Tables

Introduction

• X, Age-at-death

• T (x), time-until-death

• Life Table
– Engineers use life tables to study the reliability of complex mechanical and
electronic systems.
– Biostatistician use life tables to compare the effectiveness of alternative
treatments of serious disease.
– Demographers use life tables as tools in population projections.
– In this text, life tables are used to build models for insurance systems
designed to assist individuals facing uncertainty about the time of
their death.

1
 Probability for the Age-at-Death
1 The Survival Function

s(x) = 1 − FX (x) = 1 − Pr(X ≤ x) = Pr(X > x), x ≥ 0.

s(x) has traditionally been used as a starting point for further development
in Actuarial science and demography. In statistics and probability, the d.f.
usually plays this role.

Pr(x < X ≤ z) = FX (z) − FX (x) = s(x) − s(z).

2
Time-until-Death for a Person Age x

The conditional probability that a newborn will die between the ages x and
z, given survival to age x, is

FX (z) − FX (x) s(x) − s(z)

Pr(x < X ≤ z|X > x) = =
1 − FX (x) s(x)

• The symbol (x) is used to denote a life-age-x

• The future lifetime of (x), X − x is denote by T (x).

The symbols of Actuarial science

• tqx = Pr[T (x) ≤ T ], t ≥ 0, t px = 1 − tqx = Pr[T (x) > t], t ≥ 0.

• xp0 = s(x), x ≥ 0.

• qx = Pr[(x) will die within 1 year], px = Pr[(x) will attain age x + 1] 3

• (x) will die between ages x + t and x + t + u.

t|u qx = Pr[t < T (x) ≤ t + u] = t+uqx − tqx = tpx − t+upx.

If u = 1, the prefix is delete in t|u qx and denotes as t|qx.

•
x+t p0 s(x + t)
t px = =
p
x 0 s(x)
s(x + t)
t qx = 1 −
s(x)
•

s(x + t) − s(x + t + u)
t|u qx =
s(x)
  
s(x + t) s(x + t) − s(x + t + u)
=
s(x) s(x + t)
= tpx · uqx+t
4
Curtate-future-life time K(x) A discrete random variable associated with
the future lifetime is the number of future years completed by (x) prior to death.

Pr[K(x) = k] = Pr[k ≤ K(x) < k + 1] = Pr[k < T (x) ≤ k + 1]

= k px − k+1px = k pxqx+k
= k| qx

k
X
FK(x)(y) = h| qx = k+1qx,
h=0
y ≥ 0 and k is the greatest integer in y

5
 Force of Mortality
FX (x + ∆x) − FX (x) fX (x)∆x
Pr(x < X ≤ x + ∆x|X > x) = =
1 − FX (x) 1 − FX (x)

Definition of force of mortality

fX (x) −s0(x)
µ(x)=
ˆ =
1 − FX (x) s(x)

In reliability theory, µ(x) is called failure rate or hazard rate or, hazard
rate function.

6
 Force of mortality can be used to specify the distribution of X.
Z n
n px = exp[− µ(x + s)ds].
0

Z n
n p0 = exp[− µ(s)ds].
0
Z x
FX (x) = 1 − s(x) = 1 − exp[− u(s)ds]
0
Z x
0
FX (x) = fX (s) = exp[− µ(s)ds]µ(x) = xp0µ(x).
0

FT (x)(t) and fT (x)(t) denote, respectively, the d.f. and p.d.f. of T (x), the
future lifetime of (x). Then we have
 
d d s(x + t)
fT (x)(t) = q
t x = 1 −
dt dt s(x)
 0 
s(x + t) s (x + t)
= −
s(x) s(x + t)
= t px µ(x + t), t ≥ 0 7
8
Example Ā refer to the complement of the event A within the sample space
and Pr(Ā) 6= 0, the following expression and identity in probability theory:

P r(A ∪ B) = Pr(A) + Pr(Ā) Pr(B|Ā).

Rewrite this identity in actuarial notation for the events A = [T (x) ≤ t] and
B = [t < T (x) ≤ 1], 0 < t < 1.

Solution: Pr(A ∪ B) becomes Pr[T (x) ≤ 1] = qx0 Pr(A) is tqx, and Pr[B|Ā]
is 1−tqx+t, hence
qx = tqx + tpx1−tqx+t.

9
 Life Tables
Ration of Life Table Functions to the Survival Function

s(x + t) s(x + 1)
t qx = 1 − ⇒ qx = 1 −
s(x) s(x)

A group of l0 = 100, 000 newborns. Let L(x) denote the group’s number of
survivors to age x. We index these lives by j = 1, 2, . . . , l0 and observe that

l0
X
L(x) = Ij
i=0

where Ij is an indicator for the survival of life j. Hence

l0
X
lx = E[L(x)] = = l0s(x).
j=1
10
 n Dx denotes the number of deaths between ages x and x + n from among
the initial l0 lives.

n dx = E[nDx] = l0[s(x) − s(x + n)] = lx − lx+n

From above equations, we also have

1 dlx 1 ds(x)
− =− = µ(x) and − dlx = lxµ(x)dx.
lx dx s(x) dx

Since
lxµ(x) = l0xp0µ(x) = l0fX (x)
We note further that Z x
lx = l0 exp[− µ(y)dy],
0
Z x+n
lx+n = lx exp[− µ(y)dy],
x
Z x+n
lx − lx+n = ly µ(y)dy
11
x
Life Table Example

Life Table for the Total Population: United States, 1979-81

• The 1979-81 U.S. Life Table was not construct by observing 100,000 newborns
until the last survival died.

• Instead, it was base on estimates of probabilities of death, given survival to

various ages, derived from the experience of entire U.S. population in the
years around 1980 census.

• In using the random survivorship group concept with this table, we must
make the assumption that the probabilities derived from the table will be
appropriate for the lifetimes of those who belong to the survivorship group.

• Most studies of human mortality for insurance purpose use the representation
s(x) = lx/l0. Since 100, 000s(x) is displayed only for integer value of x,
there is a need to interpolate in evaluating s(x) for noninteger values.
12
Example On the basis of The 1979-81 U.S. Life Table, evaluate the probability
that (20) will

a. Live to 100

b. Die before 70

c. Die in the tenth decade of Life.

Soltion:

s(100) l100 1,150

a. s(20) = l20 = 97,741 = 0.0118.

s(70)−s(20) 68,248
b. s(20) = 1 − ll70
20
= 1 − 97,741 = 0.3017

s(90)−s(100) l90 −l100 14,154−1,150

c. s(20) = l20 = 97,741 = 0.1330.

13
14
15
16
• The function lxµ(x) is proportional to the p.d.f of the age-at-death of a
newborn. Since lxµ(x) is the expected density of death at age x, under the
random survivorship group idea, the graph of lxµ(x) is called the curve of
deaths.

• There is a local minimum of lxµ(x) at about age 10. The mode of the
distribution of deaths — the age at which the maximum of the curve of
death occurs — is around age 80.

• The function lx is proportional to the survival function s(x). It can also be

interpreted as the expected number living at age x out of an initial group of
size l0.

• Local extreme points of lxµ(x) correspond to points inflection of lx since

d2
 
d d d
lxµ(x) = − lx = − 2 lx .
dx dx dx dx

17
 2. Life Insurance
Introduction In this section we develop models for life insurances designed to
reduce the financial impact of the random event of untimely death.

Due to the long-term nature of the insurance, the amount of investment

earnings, up to the time of payment, provides a significant element of uncertainty.

• The unknown rate of earning over.

• The unknown length of, the investment period.

In this chapter a deterministic model is used for the unknown investment

earnings. Our model will be built in terms of function T , the insured’s future-
lifetime random variable.

18
 Insurances Payable at the Moment of Death

• bt, a benefit function

• vt, a discount function

• zt = btvt, the present-value function.

• T = T (x), The elapsed time from policy issue to the death of the issued is
the insured’s future-lifetime random variable.

• Z = bT vT , the present value, at policy issue, of the benefit payment which

is a random variable.

The first step in our analysis of a life insurance will be to define bt and vt. The
next step is to determine some characteristics of the probability distribution of
Z that are consequences of an assumed distribution for T , and we work through
these steps for several conventional insurances.
19
 Level Benefit Insurance
n-year term life insurance A payment only if the insured dies within the n-year
term of an insurance commencing at issue.

If a unit is payable at the moment of the death (x), then

vT , T ≤ n
 
1, t ≤ n t
bt = vt = v , t ≥ 0, Z=
0, t > n, 0, T > n,

Where the force of interest is assumed to be constant.

Actuarial present value The expectation of the present-value random variable

Z in insurance. ( The expected loss was called the pure premium. This word is
commonly used in property-liability insurance.)

20
Expected value of Z
Z ∞ Z n
Ā1x:n| = E[Z] = E[zT ] = ztfT (t)dt = v ttpxµx(t)dt.
0 0

Rule of Moments
Z n Z n
E[Z j ] = (v t)j tpxµx(t)dt = e−(δj)ttpxµx(t)dt
0 0

Hence
E[Z j ]@δt = E[Z]@jδt
and
V ar(Z) = 2Ā1x:n| − (Ā1x:n|)2
where 2Ā1x:n| is the actuarial present value for an n-year term insurance for a
unit amount calculated at force of interest 2δ.

21
Whole life insurance provides for a payment following the death of the insured
at any time in the future. If the payment is to be a unit amount at the moment
of death (x), then

bt = 1, vt = v t, Z = v T , t ≥ 0, T ≥ 0.

The actuarial present value is

Z ∞
Āx = E[Z] = v ttpxµx(t)dt.
0

For a life selected at x and now age x + h, the expression would be

Z ∞
Ā[x]+h = v ttp[x]+hµx(h + t)dt.
0

Whole life insurance is the limiting case of n-year term insurance as n → ∞.

22
Example 1. The p.d.f of the future lifetime T , for (x) is assumed to be

1/80, 0 ≤ t ≤ 80
fT (t) =
0, elsewhere.

At a force of interest δ, calculate Z, the present value random variable for a

whole life insurance of unit amount issued to (x):

a. The actuarial present value

b. The variance

c. The 90th percentile , ξZ0.9.

23
24
Example 2. For the assumption in Example 1, determine

a. Z’s d.f.

b. Z’s p.d.f

Example 3. Assume that each 100 independent lives

• Is age x

• Is subject to a constant force of mortality, µ = 0.04, and

• Is insured for a death benefit amount of 10 units, payable at the moment of

death.

The benefit payments are to be withdrawn from an investment fund earning

δ = 0.06. Calculate the minimum amount at t = 0 so that the probability is
approximately 0.95 the sufficient funds will be on hand to withdraw the benefit
payment at the death of each individual.
25
Solution:
bt = 10, vt = v t, Z = 10v T , t ≥ 0, T ≥ 0
100
X
S= Zj
1
where Zj is the present value at t = 0 for the payment to be made at the death
of the life numbered j.
Z ∞
10µ 0.04
E[Z] = 10Āx = 10 e−δte−µtµdt = = 10 = 4,
0 µ+δ 0.1

0.04
E[Z 2] = 102 2Āx = 100 = 25.
0.04 + 2(0.06)
and Var(Z) = 9. Hence
E[S] = 100(4) = 400, Var(S) = 100(9) = 900.

Analytically, the required minimum amount is a number, h, such that

" #
S − E[S] h − 400
Pr(S ≤ h) = Pr p ≤ = 0.95
Var(S) 30 26
 h−400
Using normal approximation, we obtain 30 = 1.645 ⇒ h = 449.35.

Observations:

• The 49.35 difference between this initial fund of 449.35 and the expectation
of the present value of all payments, 400, is the risk loading.

• This example used the individual risk model and a normal approximation to
the probability distribution S.

• A graph of the amount in the fund during the first 2 years for a payout
pattern when one death occurs at each time 1/8, 7/8, 9/8,13/8 and 15/8,
and two deaths occur at time 10/8 is shown in Figure 1. Between the
benefit payments, represented by the discontinuities, are exponential arcs
representing the growth of the fund at δ = 0.06.

• There are infinitely many payout patterns, each with its own graph. Both the
number of claims and times of those claims affect the fund. For example, had
the seven claims all occurred within the first instant, instead of the payout
pattern of the Figure, the fund would have dropped immediately to 379.35
27
and grown to 427.72 by the end of the second year.
28
 Endowment Insurance
An n-year pure endowment provides for a payment at the end of the n years
if and only if the insured survives at least n years from the time of policy issue.
 
0, t ≤ n 0, T ≤n
bt = vt = v n, t ≥ 0, Z=
1, t > n, v n, T > n,

A1x:n| = E[Z] = v nE[Y ] = v nnpx

and
Var(Z) = v 2nVar(Y ) = v 2nnpx nqx = 2A1x:n| − (A1x:n|)2.
where Z = v nY , Y is the indicator of the event of survival to age x + n. Y
has the value 1 if the insured survives to age x+n and has the value 0 otherwise.

29
N-year endowment insurance provides for an amount to be payable either
following the death of the insured or upon the survival of the insured to the end
of the n-year term, whichever occurs first. If the insurance is for a unit amount
and the death benefit is payable at the moment of death, then

t
vT , T ≤ n
 
v , t≤n
bt = 1, t ≥ 0, vt = Z=
v n, t > n, v n, T > n,

The actuarial present value is denoted by Āx:n| and

Var(Z) = 2Āx:n| − (Āx:n|)2.

and
Āx:n| = Ā1x:n| + A1x:n|
Var(Z) = 2Ā1x:n| − (Ā1x:n|)2 + 2A1x:n| − (A1x:n|)2 − 2Ā1x:n|A1x:n|.

30
 Deferred Insurance
M-year deferred insurance provides for a benefit following the death of the
insured only if the insured dies at least m years following policy issue.

vT , T > m
 
1, t > m t
bt = vt = v , t ≥ 0, Z=
0, t ≤ m, 0, T ≤ n,

The actuarial present value is

Z ∞
m| Āx = v ttpxµx(t)dt.
m

31
Example 4. Consider a 5-year deferred whole life insurance payable at the
moment of the death of (x). The individual is subject to a constant force of
mortality µ = 0.04. For the distribution of the present value of the benefit
payment, at δ = 0.10:
• Calculate the expectation
• Calculate the variance
• Display the distribution function
• Calculate the median ξZ0.5.
Observations:
1. The largest value of Z with nonzero probability density in this example is
e−0.1(5) = 0.6065, corresponding to T = 5.
2. The distribution of Z in this example is highly skewed to the right. While
its total mass is in the interval [0, 0.6065] and its mean is 0.1419, its median
is only 0.0573. This skewness in the direction of large positive values is
characteristic of many claim distributions in all field of insurance.

32
33
34
 3. Life Annuities
Introduction

Life Annuity: a series of payments made continuously or at equal interval

(such as months, quarters, years) while a given life survives.

• It may be temporary, that is, limited to a given term of years. Or it may be

payable for the whole of life.

• The payment intervals may commence immediately or, alternatively, the

annuity may be deferred.

• Payment may be due at the beginning of the payment intervals(annuities-

due) or at the end of such interval (annuities-immediate).

35
Life annuities play a major role in life insurance operations

• Life insurances are usually purchased by a life annuity of premiums rather

than by a single premium.

• The amount payable at the time of claim may be converted through a

settlement option into some form of life annuity for the beneficiary.

• Some type of life insurance carry this concept even further and instead of
featuring a lump sum payable on death, provide stated forms of income
benefits. Thus for example, there may be monthly income payable to a
surviving spouse or to a retired insured.

Annuities are even more central in pension systems

In fact, a retirement plan can be regards as a system for purchasing deferred
life annuities (payable during retirement) by some form of temporary annuity of
contributions during active service.

36
Life annuities also have a role in disability and worker’s compensation
insurances

• In the case of disability insurance, termination of the annuity benefit by

reason of recovery of the disabled insured may need to be considered.

• For surviving spouse benefits under workers’ compensation, remarriage may

terminate the annuity.

— In this chapter, we express the present value of benefits to be received by the

annuitant as a function of T , the annuitant’s future lifetime random variable.

— As in the preceding chapter on life insurance, unless otherwise stated we

assume a constant effective annual rate of interest i

37
 Continous Life Annuities

• Annuities payable continuously at the rate of 1 per year, this is as a practical

matter closely approximates annuities payable on a monthly basis.

• Whole life annuity provides for payments until death.

The present value of payment is Y = āT | for all T ≥ 0 where T is the future
lifetime of (x).

FY (y) = Pr(Y ≤ y) = Pr(āT | ≤ y) = Pr(1 − v T ≤ δy)

 
− log(1 − δy)
= P r(v T ≥ 1 − δy) = Pr T ≤
δ
 
− log(1 − δy) 1
= FT for 0 < y < .
δ δ
 
d − log(1 − δy) fT ([− log(1 − δy)]/δ) 1
fY (y) = FT = for 0 < y <
dy δ 1 − δy δ
38
 The actuarial present value for a continous whole life annuity
Z ∞ Z ∞ Z ∞
āx = E[Y ] = āt|tpxµ(x + t)dt = v ttpxdt = t Ex dt.
0 0 0

The current payment technique for determining an actuarial present

value for an annuity gives
Z ∞
APV = v t Pr[payments are being made at time t]×[Payment rate time t]dt.
0

Backward recursion formula

Z 1 Z ∞ Z ∞
āx = v ttpxdt + v ttpxdt = āx:1| + vpx v sspx+1ds = āx:1| + vpxāx+1
0 1 0

39
 A relationship, familiar from compound interest theory

1 = δāt| + v t ⇒1 = δāT | + v T ⇒1 = δāx + Āx.

Variance of āT |

T
Var(v T ) 2Āx − (Āx)2
 
1−v
Var(āT |) = Var = 2
= 2
.
δ δ δ

Example 1. Under the assumptions of a constant force of mortality, µ, and of

a constant force of interest, δ, evaluate

a. āx = E[āT |]

b. Var(āT |)

c. The probability that āT | will exceed āx.

40
N-year temporary life annuity of 1 per year, payable continuously while (x)
survives during the next n year, is

āT |, 0 ≤ T < n
Y =
ān|, T ≥ n

The actuarial present value of an n-year temporary life annuity

Z n Z n
āx:n| = E[Y ] = āt|tpxµ(x + t)dt + ān| npx = v ttpxdt.
0 0

(
1−v T 1 − Āx:n|
 
āT | = δ , 0≤T <n 1−Z
Y = 1−v n
=⇒E[Y ] = āx:n| = E =
ān| = δ , T ≥n δ δ
2
VarZ Āx:n| − (Āx:n|)2 2 2 2
Var(Y ) = 2
= 2
= (āx:n| − āx:n| ) − (āx:n| ) .
δ δ δ

41
N- year deferred whole life annuity

0 = āT | − āT |, 0≤T <n


Y =
v nāT −n| = āT | − ān|, T ≥ n

Z ∞
n| āx = E[Y ] = v nāt−n|tpxµ(x + t)dt
n
Z ∞
= v n n px ās|spx+nµ(x + n + s)ds = nExāx+n.
0

Z ∞
Var(Y ) = v 2n(āt−n|)2tpxµ(x + t)dt − (n|āx)2
n
2 2n
= v npx(āx+n − 2āx+n) − (n|āx)2
δ

42
N-year certain and life annuity This is a whole life annuity with a guarantee
of payments for the first n years.

ān|, 0 ≤ T < n
Y =
āT |, T ≥ n

Z n Z ∞
āx:n| = E[Y ] = ān|tpxµ(x + t)dt +āt|tpxµ(x + t)dt
0 n
Z ∞ Z ∞
= nqxān| + āt|tpxµ(x + t)dt = ān| + v ttpxdt.
n n

Y is the sum of a constant ān| and the random variable for the n-year deferred
annuity. Hence

āx:n| = ān| + n|āx = ān| + nExāx+n = ān| + (āx − āx:n|)

Further more, since Var(Y − ān|) = Var(Y ), the variance for the n-year certain
and life annuity is the same as that of the n-year deferred annuity given above.
43
 d
An Expression of dx āx

Z ∞   Z ∞
d ∂
ā = vt t px dt = v ttpx[µ(x) − µ(x + t)]dt
dx 0 ∂x 0
= µ(x)āx − Āx = µ(x)āx − (1 − δāx)
= [µ(x) + δ]āx − 1.

Example 2. Assuming that probabilities come from an aggregate table, obtain

formulas for
∂ ∂
a. āx:n| b. n| āx
∂x ∂n
Solution:

∂
a. ∂x āx:n| = [µ(x) + δ]āx:n| − (1 − nEx).

∂
b. ∂n n| āx = −v nnpx

44
45
46
 Discrete Life Annuities
Whole life annuity-due: This annuity pays a unit amount at the beginning of
each year that the annuitant (x) survives. The present value random variable is
denoted by
Y = äK+1|
where the random variable K is the curtate-future lifetime of (x).
The actuarial present value of the annuity:

∞
X
äx = E[Y ] = E[äK+1|] = äk+1|k pxqx+k
k=0
∞
X X∞
= 1+ v k+1k+1px = v k k px .
k=0 k=0

The last term of the equation above is the current payment form of the actuarial
present value for a whole life-annuity-due where the k px term is the probability
of a payment of size 1 being made at time k.
47
 We can also obtian in succession
K+1
 
1−v 1 − Ax
äx = E =
d d

and
äx = ä∞| − ä∞|Ax, 1 = däx + Ax.
and the variance formula is
K+1
Var(v K+1) 2Ax − (Ax)2
 
1−v
Var(äK+1|) = Var = = .
d d2 d2

48
N-year temporary life annuity-due of 1 per year is

äK+1|, 0 ≤ K < n
Y =
än|, K≥n

Its actuarial present value is

n−1
X n−1
X
äx:n| = E[Y ] = äk+1|k pxqx+k + än|npx = v k k px .
k=0 k=0

Since Y = (1 − Z)/d, where

v K+1, 0 ≤ K < n

Z=
v n, K≥n

is the present -value random variable for a unit of endowment insurance, payable
at the end of the year or at maturity, we have

1 − E[Z] 1 − Ax:n|
äx:n| = = =⇒1 = däx:n| + Ax:n|. 49
d d
To calculate the variance, we can use

Var(Z) 2Ax:n| − (Ax:n|)2

Var(Y ) = = .
d2 d2

N-year deferred whole life annuity-due of 1 payable at the beginning of

each year while (x) survives from age x + n onward, the present-value random
variable is 
0, 0≤K<n
Y =
n |äK+1−n| , K ≥ n

Its actuarial present value and variance are

∞
X
E[Y ] = n|äx = nExäx+n = äx − äx:n| = v k k px .
k=n

2 2n
Var(Y ) = v npx(äx+n − 2äx+n) + n|2äx − (n|ä)2.
d

50
N-year certain and life annuity-due

än|, 0≤K<n
Y =
äK+1|, K ≥ n

Its actuarial present value is

∞
X ∞
X
äx:n| = E[Y ] = än| nqx+ äk+1| k px qx+k = än|+ v k k px = än|+äx−äx:n|.
k=n k=n

The procedures used above for annuities-due can be adapted for annuities
immediate where payments are made at the ends of the payment period.
For instance, for a whole life annuity-immediate, the present-value random
variable is Y = aK|. Then

∞ ∞
X X 1 − (1 + i)Ax
ax = E[Y ] = k px qx+k ak| = v k k px = .
i
k=0 k=1

51
52
 4. Benefit Premiums
Introduction

In practice individual life is usually purchased by a life annuity of contract

premiums—the insurance contract specifies the premium to be paid.

Contract premiums provides for

• benefit

• expense of initiating and maintaining the insurance,

• and margins for profit and for offsetting possible unfavorable experience.

Determination of the insurance premium requires the adoption of a premium

principle. Example 1 illustrates the application of three such premium principles.

53
Example 1. An insurer is planning to issue a policy to a life age 0 whose
curtate-future-lifetime, K is governed by the p.f.

k| q0 = 0.2, k = 0, 1, 2, 3, 4.

The policy will pay 1 unit at the end of the year of death in exchange for
the payment of a premium P at the beginning of each year, provided the life
survives. Find the annual premium, P , as determined by:
a. Principle I: P will be the least annual premium such that the insurer has
probability of a positive financial loss of at most 0.25.
b. Principle II: P will be the annual premium such that the insurer, using a
utility of wealth function u(x) = x will be indifferent between accepting and
not accepting the risk.
c. Principle III: P will be the annual premium such that the insurer, using
a utility of wealth function u(x) = −e−0.1x, will be indifferent between
accepting and not accepting the risk.

For all three parts assume the insurer will use an annual effective interest rate
of i = 0.06. 54
55
• Percentile premiums Although the principle is attractive on surface, it is
easy to show that it can lead to quite unsatisfactory premiums.
• Equivalence principle It has many application. If we define the insurer’s
loss, L, as the random variable of the present value of benefit to be paid
by insurer less the annuity of premiums to be paid by the insured. The
requirement of this principle is that
E[L] = 0.
Benefit premiums
E[present value of benefits] = E[present value of benefit premiums]
Single benefit premium When the equivalence principle is used to determine
a single premium at policy issue for a life insurance or a life annuity, the
premium is equal to the actuarial present value of the benefit payments.
• Exponential premiums These premiums are nonproportional in the sense
that the premium for the policy with a benefit level of 10 is more than 10
times the premium for a policy with a benefit level of 1. This is consistent
for a risk averse utility function.
56
 Fully Continuous Premiums

• The equivalence principle.

• Fully continuous level annual benefit premium for a unit whole life insurance
payable immediately on the death of (x).

• For any continuously paid premium, P̄ , consider

l(t) = v t − P̄ āt|,

the present value of the loss to the issuer if death occurs at time t.
– l(0) = 1.
– l(t) → −P̄ /δ as t → ∞.
– if l(t0) = 0, death before t0 results in a positive loss, whereas death after
t0 produces a negative loss.

57
 Consider the loss random variable

L = l(T ) = v T − P̄ āT |.

By the equivalence principle, the premium is denoted by P̄ (Āx) and is such

that
E[L] = 0.
Then
Āx
Āx − P̄ (Āx)āx = 0 =⇒ P̄ (Āx) = .
āx

The variance of L can be used as a measure of the variability of losses on an

individual whole life insurance due to the random nature of time-until-death.
T
 
P̄ (1 − v )
Var(L) = E[L2] = Var(v T − P̄ āT |) = Var v T −
δ
     2
T P̄ P̄ T P̄
= Var v 1+ − = Var(v ) 1 +
δ δ δ
2 2
Āx − (Āx)2

2 2 P̄
= [ Āx − (Āx) ] 1 + = 2
.
δ (δāx) 58
59
Example 2. Calculate P̄ (Āx) and Var(L) with the assumptions that the force
of mortality is a constant µ = 0.04 and the force of interest δ = 0.06.

Solution: These assumptions yields āx = 10, Āx = 0.4, and 2Āx = 0.25.
Then we obtain
Āx
P̄ (Āx) = = 0.04,
āx
and
0.25 − 0.16
Var(L) = 2
= 0.25.
(0.6)

60
• Under the constant force of mortality assumption,

µ(µ + δ)−1
P̄ (Āx) = −1
= µ,
(µ + δ)

which does not depend on the force of interest interest or the age at issue.

• For a variety of fully continuous life insurances, the formulas for annual
premiums can be determined by

E[bT vT ]
bT vT − P̄ Y = Z − P̄ Y and E[bT vT − P̄ Y ] = 0 =⇒ P̄ = .
E[Y ]

• For an n-year deferred whole life annuity of 1 per year payable continuously.
In this case bT vT = 0, T ≤ n and bT vT = āT −n|v n, T > n Then

E[bT vT ] = npxE[āT −n|v n|T > n] = v nnpxāx+n = Ax:n|1 āx+n.

61
 Example 3. Express the variance of the Loss L, associated with an n-year
endowment insurance, in terms of actuarial present values.

Example 4. Find the 25th percentile premium for an insured age 55 for the
following plans of insurances:

a. 20-year endowment

b. 20-year term

c. 10-year term.

Assume a fully continuous basis with a force of interest, δ = 0.06 and mortality
following the illustrative life Table.

62
63
 Example 5. This example builds on Example 4.except that T (55) has a De
Movivre distribution with p.d.f

t p55 µ55 (t) = 1/45, 0 < t < 45.

For the three loss variables, display the d.f. of L and determine the parameter
P̄ as the smallest non-negative number such that Pr(L > 0) ≤ 0.25.

64
65
 Fully Discrete Premiums
• Px, the level annual benefit premium for a unit whole life insurance.
• the absence of (Āx) means that the insurance is payable at the end of policy
year of death.
• The loss of this insurance is

L = v K+1 − PxäK+1|, k = 0, 1, 2, . . . .

The equivalence principle requires that E[L] = 0, or

E[v K+1] − PxE[äK+1|] = 0

which yields Ax
Px =
äx
2
and Ax − (Ax)2
Var(L) = .
(däx)2
66
67
 Example 6. If k|qx = c(0.96)k+1, k = 0, 1, 2, . . . where c = 0.04/0.96 and
i = 0.06, calculate Px and Var(L).

Continuing to use the equivalence principle, we can determine formulas for

annual benefit premiums for a variety fully discrete life insurances.

E[bK+1vK+1]
E[bK+1vK+1 − P Y ] = 0 =⇒ P =
E[Y ]

where P is a general symbol for an annual premium paid at the beginning of

each policy year during the premium paying period while insured survives and
Y is a discrete annuity random variable as defined in Section 3.

Example 7. Express the variance of the Loss L, associated with an n-year

endowment insurance, in terms of actuarial present values.

68
 Example 8. Consider a 10,000 fully discrete whole life insurance. Let π
denote an annual premium for this policy and L(π) denote the loss-at issue
random variable for one such policy on the basis of the Illustrative Life Table,
6% interest and issue age 35

a. Determine the premium, πa, such that the distribution of L(πa) has mean 0.
Calculate the variance of L(πa).

b. Approximate the smallest non-negative premium, πb, such that the probability
is less than 0.5 that the loss L(πb) is positive. Find the variance of L(πb).

c. Determine the premium, πc, such that the probability of a positive total loss
on 100 such independent policies is 0.05 by the normal approximation.

69