Chapter 4

Chapter 4: Insurance Benefits

4.1 Overview
• This chapter introduces the International Actuarial Notations and Formulas for the Expected
Present Value (EPV) for different life insurances. Second moment and variance are also
covered.

• For each insurance type, we will look at the timing of when the Death Benefit (DB) is paid:

1. Payable at the moment of death (continuous);

2. Payable at the end of year of death (discrete);
3. Payable at the end of mth period of death (also discrete).

• We will also cover:

4. Links between discrete and continuous insurances, for example, using Uniform Distri-
bution of Deaths (UDD) within each year of age;
5. Varying insurances — in theory, any type of insurances in this chapter can have a varying
benefit;
6. Recursive formulas;
7. Consideration of selected rates.

• The lecture ordering of topics differs from the textbook, but all topics will be covered.

AS 2427 Long Term Actuarial Math I - Winter 2025 46

 Chapter 4

4.2 Introduction
- Insurance contracts:

Insurance type Description (DB = Death Benefit)

A lump sum DB is paid on the death of the insured
1. Whole life
regardless of when death occurs.
A lump sum DB is paid on the death of the insured only
2. n-year term life
if the insured dies within n years.
3. Pure endowment A survival benefit is paid upon survival to n year
Combination of n-year term life insurance and a pure
endowment payment.
4. n-year endowment
A death benefit is paid if death occurs within n years
or a survival benefit is paid upon survival to n years
No coverage in the deferred period. DB is paid only
5. u-year Deferred life if the insured dies after u years. In theory, any policy
type can be deferred.
DB varies by time. It can be any of the aforementioned
insurance types. Applications include:
6. Varying life (a) Step rate insurance coverage;
(b) Insurance return of premium rider;
(c) Indexed life insurance, reversionary bonuses.

- Benefit amount:

(1) level benefit model: the benefit amount remains constant

(2) varying/non-level benefit model

- Payment timings:

1. benefit paid immediately on death (continuous case);

2. benefit paid at the end of year of death (discrete - annual case);

3. benefit paid at the end of mth year of death (discrete - 1/mthly case).

AS 2427 Long Term Actuarial Math I - Winter 2025 47

 Chapter 4

4.3 Assumptions
• We will mostly assume for Chapters 4 and 5 a constant and fixed interest rate, unless explicitly
stated otherwise.

• Many Chapter 4 textbook examples and exercises use the textbook Standard Ultimate Survival
Model introduced in lecture notes Section 3.4 last example (Or textbook Section 3.9):

– See Appendix D for selected life values.

• Many Theory of Interest concepts from AS2553 used and built upon

• Review of interest rate theory functions

i = annual effective rate of interest

1
v= = the PV of $1 in 1 year
1+i
d =1 − v = discount rate
i(m) =m[(1 + i)1/m − 1] = nominal interest rate compound m times per year
d(m) =m[1 − (1 − d)1/m ] = nominal discount rate compound m times per year
δ = ln(1 + i) = constant force of interest

as m → ∞, i(m) → δ and d(m) → δ.

!m
δ i(m)
e =1+i= 1+ ,
m
1
e−δ = v = = 1 − d,
1+i
v t = e−δt ,

are used extensively throughout the next few chapters.

AS 2427 Long Term Actuarial Math I - Winter 2025 48

 Chapter 4

4.4 Valuation of insurance benefits - Part 1: Continuous case

We will first work with continuous cases and then move to the discrete cases. Assume the benefit
amount is level in this section.

Overview for the continuous case

(a) Assumptions for the continuous case
• Death benefit is payable at the moment of death (m.o.d.).
• The Present Value (PV) of the benefit payment depends on the time of death. Since the
lifetime is a random variable Tx , so is the PV.

(b) Present Value Random Variable (PVRV)

Z is the random variable for the PV of a benefit payment of $1. Notice that Z is a function of the
future lifetime Tx :
PVRV = Z = A function that depends on Tx .
We can use the pdf of Tx to determine the expected value (i.e., first moment) and other moments
of Z (useful for variance calculation). Recall that the pdf of Tx is:
fx (t) = t px · µx+t , t ≥ 0.

(c) Expected Present Value (EPV)

Since Z is a random variable, we are interested in its expected value: E[Z]. This is also known as
Actuarial Value or Actuarial Present Value (APV). Here, we use EPV:
X Z
EPV = E[Z] = or amount × discount × probability.

*note that use integration for the continuous case, and summation for discrete cases.

4.4.1 Whole life insurance (DB payable at m.o.d.)

4.4.2 Term life insurance (DB payable at m.o.d.)
4.4.3 Pure endowment (DB not applicable)
4.4.4 Endowment life insurance (DB payable at m.o.d.)
4.4.5 Deferred life insurance (DB payable at m.o.d.)

AS 2427 Long Term Actuarial Math I - Winter 2025 49

 Chapter 4

4.4.1 Whole life insurance (DB payable at m.o.d.)

Death benefit of $1 is payable at the m.o.d.

(a) The PV of the benefit

PVRV = Z = v Tx = e−δTx , where v = (1 + i)−1 = e−δ .

(b) Expected Present Value

A special Actuarial Notation is used to denote the EPV of whole life insurance per $1 payable at
m.o.d. of (x):
def
h i
Āx = E[Z] = E e−δTx .

- Calculation of Āx
For the EPV of Z, we have
h i Z ∞
Āx = E[Z] = E e−δTx = e−δt fx (t) dt.
0

Therefore,
Z ∞
Āx = e−δt t px µx+t dt. (4.1)
0

- Time diagram argument for Āx

Age x x+t x + t + dt

Time 0 t t + dt
(x) dies
Event (x) survives for t years
“instantly”
Probability t px ≈ µx+t dt
↓ ↓
Present value $1 paid
e−δt

[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 50

 Chapter 4

(c) Calculation of Var[Z]

For the variance of Z, we have
h i  2
Var[Z] = E Z 2 − E[Z]
 2   2
=E e−δTx − Āx
Z ∞  2
= e−2δt fx (t) dt − Āx
Z0∞  2
= e−2δt t px µx+t dt − Āx .
0

- the second moment of Z Z ∞

2
Āx = e−(2δ) t t px µx+t dt, (4.2)
0

where 2 Āx is Āx calculated at 2δ. Therefore,

h i  2
Var[Z] = Var e−δTx = 2 Āx − Āx .

(d) General death benefit amount $S

For a general fixed amount $S for the death benefit, we have

PVRV = Z = S × v Tx ,
EPV = S × Āx ,
  2 
2 2
Var[Z] = S × Āx − Āx .

(e) The cdf of Z

P(Z ≤ y) [TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 51

 Chapter 4

4.4.2 Term life insurance (DB payable at m.o.d.)

Death benefit of $1 is payable at the m.o.d., only if death occurs within the first n years since policy
issue.

(a) The PV of the benefit


e−δTx if Tx ≤ n
PVRV = Z = v Tx 1{Tx ≤n} = .
0 if Tx > n

Note that 1{A} is the indicate function which equals to 1 if event A is true, and 0 otherwise.

(b) Expected Present Value

A special Actuarial Notation is used to denote the EPV:

1 def
Āx:n = E[Z] = EPV for an n-year term insurance per $1 payable at m.o.d. of (x)

1
- Calculation of Āx:n
For the EPV of Z, we have
1
Āx:n = E[Z]
Z n Z ∞
−δt
= e fx (t) dt + 0 × fx (t) dt
n
Z0n
= e−δt fx (t) dt.
0

Therefore,
Z n
1
Āx:n = e−δt t px µx+t dt. (4.3)
0

1
Note that as n → ∞, Āx:n → Āx .

(c) Calculation of Var[Z]

For the variance of Z, we have
h i  2
Var[Z] = E Z 2 − E[Z]
"Z Z ∞ #
n  2
= (v t )2 f x (t) dt +
2 1
0 × fx (t) dt − Āx:n
0 n
Z n  2
= v 2t t px µx+t dt − Āx:n
1

Z0n  2
= e−2δt t px µx+t dt − Āx:n
1 .
0

AS 2427 Long Term Actuarial Math I - Winter 2025 52

 Chapter 4

where the second moment of Z is

h i Z n
EZ 2
= 2
Āx:n
1 = e−2δt t px µx+t dt, (4.4)
0

1
and it is Āx:n calculated at 2δ. Therefore,
 2
Var[Z] = 2 Āx:n
1 1
− Āx:n ,

(d) General death benefit amount $S

For a general fixed amount $S for the death benefit, we have

S × e−δTx if Tx ≤ n
PVRV = Z =  ,
0 if Tx > n
1 ,
EPV = S × Āx:n
  2 
2 2 1
Var[Z] = S × Āx:n
1 − Āx:n .

AS 2427 Long Term Actuarial Math I - Winter 2025 53

 Chapter 4

4.4.3 Pure endowment (DB not applicable)

• Pure endowment provides for a single payment at the end of n year if the insured is still alive
after n years from policy issue.

• Pure endowments are the fundamental building block for life annuities (Chapter 5) and they
can also be used to link different insurance types.

Consider an n-year pure endowment of $1 issued on (x).

(a) The PV of the benefit


0 if Tx ≤ n
PVRV = Z = .
v n if Tx > n

(b) Expected Present Value: Calculation of n Ex

A special Actuarial Notation is used to denote the EPV:

def
n Ex = E[Z] = EPV for an n-year pure endowment of $1.

For the EPV of Z, we have

n Ex = E[Z]
Z n Z ∞
= 0 × fx (t) dt + v n fx (t) dt
0 n
Z ∞
n
=0+v fx (t) dt.
n

Therefore,
n Ex = v n n px . (4.5)
Note: There is also another notation
Ax: n1 = n Ex
For pure endowment, n Ex or v n · n px is notation typically used.

(c) Calculation of Var[Z]

For the variance of Z, we have
h i  2
Var[Z] = E Z 2 − E[Z]

where the second moment of Z is

h i
E Z 2 = n2 Ex = v 2n n px (4.6)

Therefore,
Var[Z] = v 2n n px n qx .

AS 2427 Long Term Actuarial Math I - Winter 2025 54

 Chapter 4

4.4.4 Endowment life insurance (DB payable at m.o.d.)

Benefit of $1 is payable at the m.o.d. if death occurs within the first n years since policy issue (death
benefit), OR, upon survival of the insured after n years (pure endowment).
Endowment = Term life + Pure endowment

(a) The PV of the benefit


v Tx if Tx ≤ n
PVRV = Z =  n or simply PVRV = Z = v min{Tx ,n} .
v if Tx > n

(b) Expected Present Value

A special Actuarial Notation is used to denote the EPV:
def
Āx:n = E[Z] = EPV for an n-year endowment insurance per $1 payable at m.o.d. of (x)

- Calculation of Āx:n
For the EPV of Z, we have
Z n Z ∞
Āx:n = E[Z] = v t fx (t) dt + v n fx (t) dt.
0 n

Therefore,
1
Āx:n = Āx:n + n Ex . (4.7)

(c) Calculation of Var[Z]

For the variance of Z, we have
h i  2
Var[Z] = E Z 2 − E[Z] ,
where the second moment of Z
E[Z 2 ] = 2 Āx:n = 2 Āx:n
1 + n2 Ex (4.8)

note that 2 Āx:n is Āx:n calculated at 2δ.

Example 4.1 The difference between term insurance and endowment insurance:
1
Āx:n v.s. Āx:n .
[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 55

 Chapter 4

Examples for Sections 4.4.1 - 4.4.4

Example 4.2 Assume Z is the present value random variable for a whole life insurance of $b
payable at the moment of death of (x). Given that
(a) δ = 0.04
(b) µx (t) = 0.02
(c) E(Z) = V ar(Z)
Calculate b. [TBC]

Example 4.3 Assume Z is the present value random variable for a 15-year pure endowment of $10
on (x). Given that
(a) v = 0.9
(b) the force of mortality is constant over the 15-year period
(c) V ar(Z) = 0.65E(Z)
Calculate qx . [TBC]

Example 4.4 (Exercise) Assume Z is the present value random variable for a n-year term life
insurance of $5000 payable at the moment of death of (x). Given that
(a) δ = 0.05
(b) µx (t) = 0.007
(c) E(Z) = 266.7826
Calculate V ar(Z).
[Answer: n=10; V ar(Z) = 50002 × 0.04013]

AS 2427 Long Term Actuarial Math I - Winter 2025 56

 Chapter 4

4.4.5 u-year Deferred life insurance (DB payable at m.o.d.)

In the following, we will see the deferred life insurance. In theory, any type of insurance can be
deferred.
u-year Deferred life insurance: Insurance does not begin to offer a benefit coverage until the end
of the deferred period.

1) u-year Deferred whole life insurance (DB payable at m.o.d.)

Death benefit of $1 is payable at the m.o.d., only if (x) dies at least u years after policy issue.

(a) PVRV

0 if Tx ≤ u
PVRV = Z = −δTx
.
v T x =e if Tx > u

(b) A special Actuarial Notation is used to denote the EPV:

def
u| Āx = E[Z] = EPV for a u-year deferred whole life per $1 payable at m.o.d. of (x)

For the EPV of Z, we have

Z u Z ∞
u| Āx = E[Z] = 0 × fx (t) dt + v t fx (t) dt.
0 u

Therefore,
Z ∞
u| Āx = e−δt t px µx+t dt. (4.9)
u

(c) Two alternative ways to calculate the EPV:

u| Āx = u Ex Āx+u (4.10)

and
1 .
= Āx − Āx:u
u| Āx (4.11)
[TBC]

(d) Second moment

Z ∞
2
u| Āx = e−2δt t px µx+t dt. (4.12)
u

note that u|2 Āx is u| Āx calculated at 2δ.

AS 2427 Long Term Actuarial Math I - Winter 2025 57

 Chapter 4

2) u-year Deferred term life insurance (DB payable at m.o.d.)

Death benefit of $1 is payable at the m.o.d., only if (x) dies between u and u + n years after policy
issue.
(a)

0 if Tx ≤ u or Tx > u + n
PVRV = Z = .
v Tx if u < Tx ≤ u + n

(b) A special Actuarial Notation is used to denote the EPV:

1 def
u| Āx:n = E[Z] = EPV for a u-year deferred, n-year term life per $1 payable at m.o.d. of (x)

For the EPV of Z, we have

1 = E[Z]
u| Āx:n
Z u Z u+n Z ∞
t
= 0 × fx (t) dt + v fx (t) dt + 0 × fx (t) dt.
0 u u+n

Therefore,
Z u+n
1
u| Āx:n = e−δt t px µx+t dt.
u

(c) Two alternative ways to calculate the EPV:

1 1
u| Āx:n = u Ex Āx+u:n (4.13)

and
1 = Āx:u+n 1 .
− Āx:u
u| Āx:n
1 (4.14)
[TBC]

3) u-year Deferred n-year pure endowment (Survival benefit)

It is the same as a u + n-year pure endowment.

AS 2427 Long Term Actuarial Math I - Winter 2025 58

 Chapter 4

4) u-year Deferred endowment life insurance (DB payable at m.o.d.)

(a) PVRV of a u-year deferred, n-year endowment life insurance which provides for $1 payable at
the m.o.d. or upon survival. [TBC]

The special Actuarial Notation for the EPV is

u| Āx:n = EPVof a u-year deferred n-year endowment life insurance per $1 (DB payable at m.o.d.).

Show that
1
= u| Āx:n + u+n Ex
u| Āx:n

[TBC]

Examples of deferred life insurance (DB payable at m.o.d.)

Example 4.5 The whole life insurance and term life insurance can be expressed as a sum/combination
of a series of one-year deferred term life insurance policies.
Prove the following two identities:

n−1
X ∞
X
1
Āx:n = Āx =
k| Āx:1 , and k| Āx:1 .
1 1
k=0 k=0

what if there is the limiting age ω? (Suppose that both x and ω are integers.)
ω−x−1
X
Āx = k| Āx:1 .
1
k=0

[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 59

 Chapter 4

Example 4.6 You are given the following assumptions:

• S0 (x) = (100 − x)/100.

• δ = 6%.
• A whole life insurance policy is issue on (50).
• The death benefit is $50,000, payable at the moment of death.

Answer the following questions:

(i) What is the limiting age ω?

(ii) What is the Present Value Random Variable (PVRV) Z for this policy.
(iii) Find the Expect Present Value (EPV) E[Z].
(iv) Find Var[Z]
(v) Now, assume that δ = 6%qonly for the first 20 years since policy is issued, and δ = 4%
thereafter. Find E[Z] and V ar(Z) under this new interest rate.

[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 60

 Chapter 4

Solution of Example 4.6.

AS 2427 Long Term Actuarial Math I - Winter 2025 61

 Chapter 4

4.5 Valuation of insurance benefits - Part 2: Annual case

Assume the benefit amount is level in this section.

Overview of Discrete Life insurance - Annual case

(a) Assumption and the curtate future lifetime
• Death benefit is payable at the end of year of death.
• Recall that the curtate future lifetime random variable Kx = ⌊Tx ⌋ is the integer complete
years that (x) lives.
• The Present Value (PV) of the benefit payment depends on the complete years of future life
of (x), so Kx + 1 is the time to the end of year of death.
- For example, if (x) dies at time 3.4 years, then Tx = 3.4 and Kx = 3. The death benefit of
an annual life insurance will be paid at the end of the year of death, i.e., the payment is made
at time t = 3 + 1 = 4.
• Since the curtate lifetime Kx is a random variable, we see that Kx + 1 is also a random
variable, and therefore so is the PV.

(b) Present Value Random Variable (PVRV)

Z is the random variable for the PV of a benefit payment of $1. Notice that Z is a function of the
future lifetime Kx + 1:
PVRV = Z = A function that depends on Kx + 1.
We can use the probability mass function (pmf) of Kx to determine the expected value and other
moments of Z. Recall that the pmf of Kx is:
P(Kx = k) = P[ (x) dies between times k and k + 1 ] = k| qx , k = 0, 1, 2, . . .

(c) Expected Present Value (EPV)

Since Z is a random variable, we are interested in its expected value: E[Z]. This is also known as
Actuarial Value or Actuarial Present Value (APV). Here, we use EPV:
X Z
EPV = E[Z] = or amount × discount × probability.

*note that use integration for the continuous case, and summation for discrete cases.
4.5.1 Whole life insurance (DB payable at the end of year of death)
4.5.2 Term life insurance (DB payable at the end of year of death)
4.5.3 Endowment life insurance (DB payable at the end of year of death)
4.5.4 Deferred life insurance (DB payable at the end of year of death)

AS 2427 Long Term Actuarial Math I - Winter 2025 62

 Chapter 4

4.5.1 Whole life insurance (DB payable at the end of year of death)
Death benefit of $1 is payable at the end of year of death.

(a) The PV of the benefit

PVRV = Z = v Kx +1 .

(b) Expected Present Value

A special Actuarial Notation is used to denote the EPV of whole life insurance per $1 payable at
the end of year of death of (x):
def
h i
Ax = E[Z] = E v Kx +1 .

- Calculation of Ax
For the EPV of Z, we have
h i ∞
v k+1 P(Kx = k).
X
Ax = E[Z] = E v Kx +1 =
k=0

Therefore,
∞
v k+1 k| qx
X
Ax = (4.15)
k=0

AS 2427 Long Term Actuarial Math I - Winter 2025 63

 Chapter 4

(c) Calculation of Var[Z]

For the second moment of Z, we have
h i  2  ∞
2 2
v 2(k+1) k| qx .
X
Ax = E Z =E v Kx +1 =
k=0

where 2 Ax is Ax calculated at v 2 (or equivalently, replace δ with 2δ). Therefore,

h i  2
Var[Z] = Var v Kx +1 = 2 Ax − Ax .

(d) General death benefit amount $S

For a general fixed amount $S for the death benefit, we have

PVRV = Z = S × v Kx +1 ,
EPV = S × Ax ,
  2 
Var[Z] = S 2 × 2
Ax − Ax .

(e) Recursive formulas for Ax and 2 Ax

Ax = v qx + v px Ax+1 . (4.16)
Proof.
∞ ∞
k+1
v k+1 k px qx+k
X X
Ax = v k| qx =
k=0 k=0
= v 1 0 px qx + v 2 1 px qx+1 + v 3 2 px qx+2 + v 4 3 px qx+3 + · · ·
 
= v qx + v px v qx+1 + v 2 px+1 qx+2 + v 3 2 px+1 qx+3 + · · ·
 
= v qx + v px v 0| qx+1 + v 2 1| qx+1 + v 3 2| qx+1 + · · ·
= v qx + v px Ax+1 .

Interpretation: We separate the EPV of the whole life insurance into the value of the benefit due in
the first year, followed by the value at age x + 1 of all subsequent benefits multiplied by px to allow
for the probability of surviving to age x + 1 and by v to discount the value back from age x + 1 to
age x. [TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 64

 Chapter 4

Example 4.7 (Exercise) Prove the following recursive formula for 2 Ax :

2
Ax = v 2 qx + v 2 px · 2 Ax+1 .

Example 4.8 You are given the following mortality rates:

q88 = 0.15981, q89 = 0.17375, q90 = 0.18877,

and the following factors:

2
A90 = 0.79346, A90 = 0.64496, i = 6%.

Find the EPV and the variance of a whole life insurance issued on (88) that provides for a death
benefit of $20,000 payable at the end of year of death.

[TBC]

(f) How to calculate Ax and 2 Ax from the life table

Answer: Using backwards recursion.

Suppose that the limiting age is ω. This means that a person aged ω − 1 would certainly die within
the next year, i.e., qω−1 = 1. This also means that the DB is certainly paid at the end of the first
year. Therefore,
Aω = 0,
or equivalently
Aω−1 = v.
Then we use the recursive formular

Ax = v qx + v px Ax+1 .

to calculate Ax backwards from Aω−1 back to Ax0 , where x0 is the minimum age in the table.

AS 2427 Long Term Actuarial Math I - Winter 2025 65

 Chapter 4

Example 4.9 (Spreadsheet Exercise) Create the life table of Standard Select & Ultimate Survival
Models (see Example 3.13) on your own. In your life table, you need to include the functions px ,
p[x] , p[x]+1 , l[x] , l[x]+1 , lx+2 , qx , Ax and 2 Ax for x = 18, 19, . . . , 130. Assume the radix l20 = 100, 000
and the interest rate i = 5%.

• The solution spreadsheet “AS2427 Makeham SSSM Model” (except for 2 Ax ) is on OWL.

• To calculate the column Ax , in practice, we choose a very large age, say, 130 as the “limiting
age”, and then we set A130 = 0.
From there, we can calculate all Ax values backwards; that is A129 , A127 , A126 , . . . For
example A129 = v q129 + v p129 A130 .

• How do you calculate the value for 2 Ax ?

[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 66

 Chapter 4

4.5.2 Term life insurance (DB payable at end of year of death)

Death benefit of $1 is payable at the end of year of death, only if death occurs within the first n years
since the policy is issued.

(a) The PV of the benefit


v Kx +1 if Kx ≤ n − 1
PVRV = Z = .
0 if Kx ≥ n

(b) Expected Present Value

A special Actuarial Notation is used to denote the EPV of an n-year term insurance per $1 payable
at the end of year of death of (x):
1 def
Ax:n = E[Z]

1
- Calculation of Ax:n
For the EPV of Z, we have

n−1 n−1
v k+1 P(Kx = k) = v k+1 k| qx .
X X
1
Ax:n = (4.17)
k=0 k=0

(c) Calculation of Var[Z]

For the variance of Z, we have
h i  2
Var[Z] = E Z 2 − E[Z]
n−1  2
v 2(k+1) k| qx − Ax:n
X
= 1 .
k=0

Therefore,
 2
Var[Z] = 2 Ax:n
1 1
− Ax:n ,

where the second moment 2 Ax:n

1 1
is Ax:n calculated at v 2 .

1
(d) Recursive formulas for Ax:n and 2 Ax:n
1

1 1
Ax:n = v qx + v px Ax+1:n−1 . (4.18)

2
Ax:n
1 = v 2 qx + v 2 px 2 Ax+1:n−1
1 .

AS 2427 Long Term Actuarial Math I - Winter 2025 67

 Chapter 4

Example 4.10 You are given

lx = 99,727.3, lx+1 = 99,695.8, i = 5%.

Find Ax:1
1 and 2 Ax:1
1 .

[TBC]

Example 4.11 You are given i = 6% and the following mortality rates:

q50 = 0.00592, q51 = 0.00642, q52 = 0.00697.

You are valuing a 3-year term life insurance policy issued on (50), where the death benefit is $1,000
payable at the end of year of death. Determine

(a) E[Z] or EPV for this policy.

(b) Var[Z].

[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 68

 Chapter 4

Example 4.12 You are given that i = 4% and the following mortality rates:

x 1,000 qx
30 1.5289
31 1.6089
32 1.6965
33 1.7927
34 1.8980
35 2.0136
36 2.1402
37 2.2791
38 2.4313
39 2.5982
40 2.7812

Assume that a 10-year term insurance policy is issued to (30) and the death benefit of $1 is payable
at the end of year of death. Determine the EPV.

[TBC]

Example 4.13 (Relation between whole life and term life insurances) Prove the following relation
between whole life and term life insurances where the DB is payable at the end of year of death.
1
Ax = Ax:n + n Ex Ax+n .

AS 2427 Long Term Actuarial Math I - Winter 2025 69

 Chapter 4

4.5.3 Endowment life insurance (DB payable at end of year of death)

Benefit of $1 is payable at the end of year of death if death occurs within the first n years since
policy issue (death benefit), OR, upon survival of the insured after n years (pure endowment).

(a) The PV of the benefit


v Kx +1 if Kx ≤ n − 1
PVRV = Z = or PVRV = Z = v min{Kx +1,n} .
v n if Kx ≥ n

(b) Expected Present Value

A special Actuarial Notation is used to denote the EPV for an n-year endowment insurance per $1
on (x) with death benefit payable at the end of year of death:

def
Ax:n = E[Z]

- Calculation of Ax:n
For the EPV of Z, we have
1
Ax:n = Ax:n + n Ex . (4.19)

(c) Calculation of Var[Z]

For the variance of Z, we have
 2    2
Var[Z] = 2 Ax:n − Ax:n = 2
Ax:n
1 + n2 Ex − Ax:n
1 + n Ex ,

where 2 Ax:n is Ax:n calculated at 2δ.

Example 4.14 Identify the difference between term insurance and endowment insurance factors:
1
Ax:n v.s. Ax:n .

Example 4.15 Prove the following formula for 2 Ax:n :

2
Ax:n = 2 Ax:n
1 + v n n Ex .

AS 2427 Long Term Actuarial Math I - Winter 2025 70

 Chapter 4

4.5.4 Deferred life insurance (DB payable at end of year of death)

(1) Deferred whole life insurance (DB payable at end of year of death)
Death benefit of $1 is payable at the end of year of death, only if (x) dies at least u years after policy
issue. So 
0 if Kx ≤ u − 1
PVRV = Z =  Kx +1 .
v if Kx ≥ u

A special Actuarial Notation is used to denote the EPV for a u-year deferred whole life per $1
payable at the end of year of death of (x):

def
u| Ax = E[Z]

We have the following two identities:

u| Ax = u Ex Ax+u

and
1 .
= Ax − Ax:u
u| Ax

(2) Deferred term life insurance (DB payable at end of year of death)
Death benefit of $1 is payable at the end of year of death, only if (x) dies between u and u + n years
after policy issue. So

0 if Kx ≤ u − 1 or Kx ≥ u + n
PVRV = Z = .
v Kx +1 if u ≤ Kx ≤ u + n − 1

A special Actuarial Notation is used to denote the EPV for a u-year deferred, n-year term life
insurance per $1 payable at the end of year of death of (x):

1 def
u| Ax:n = E[Z]

We have the following two identities:

1 1
u| Ax:n = u Ex Ax+u:n

and
1 = Ax:u+n 1 .
− Ax:u
u| Ax:n
1

AS 2427 Long Term Actuarial Math I - Winter 2025 71

 Chapter 4

Example 4.16 Find the EPV for a 30-year term insurance policy issued on (50), assuming i = 6%
and the $100,000 death benefit is payable at the end of year of death.
You are also given the following values that were extracted from a life table:

A50 = 0.2490473 A80 = 0.6657528 l50 = 89,509.00 l80 = 39,143.64

[TBC]

Example 4.17 (Exercise) Find the EPV for a 20-year $50,000 endowment insurance policy issued
on (50). Assume that i = 6% and the death benefit is payable at the end of year of death. You are
also given the following values that were extracted from a life table:

A50 = 0.2490473 A70 = 0.5149481 l50 = 89,509.00 l70 = 66,161.54

[Answer: EPV = 50, 000A50:20 = 50, 000 ∗ 0.36083906 = 18, 041.953.]

AS 2427 Long Term Actuarial Math I - Winter 2025 72

 Chapter 4

Example 4.18 An special insurance policy is issued on (65) that has a death benefit of $250,000
if death occurs between ages 70 and 80, and $150,000 if death occurs after age 80. Assume that
i = 6% and the death benefit is payable at the end of year of death.
You are also given the following values that were extracted from a life table:

A70 = 0.5149481 A80 = 0.6657528

l65 = 75,339.63 l70 = 66,161.54 l80 = 39,143.64

Set up an equation that could be used to determine the EPV of this policy. [TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 73

 Chapter 4

4.6 Valuation of insurance benefits - Part 3: 1/mthly case

In this section, we discuss the discrete life insurance in the more frequent cases (monthly, weekly,
etc.) Assume the benefit amount is level in this section.

Overview of Discrete Life insurance - 1/mthly case

(a) Assumption and the 1/mthly curtate future lifetime
• Death benefit is payable at the end of the 1/mth year of death.

• Define the 1/mthly curtate future lifetime random variable

1
Kx(m) = ⌊m Tx ⌋ = the future lifetime of (x) in years rounded to the lower 1/m of a year
m

where m > 1 is an integer.

• Death benefit is paid at the end of 1/mth year of death; that is,

1
Time of DB payment = Kx(m) + .
m

– m = 2. This is the half-year or semi-annual setup.

– m = 4. This is the seasonal or quarterly setup.
– m = 12. This is the monthly setup:
1
DB is payable at the end of month of death (at Kx(12) + 12
).
– m = 52. This is the weekly setup:
1
DB is payable at the end of week of death (at Kx(52) + 52
).
– m → ∞:
This implies that Kx(m) → Tx and also Kx(m) + m1 → Tx . Therefore, m → ∞ reduces to
the continuous case where DB is payable at the moment of death.

• Example 1: DB payable at end of half-year of death

A life insurance policy is issued on (50). The death benefit is payable at the end of half-year
of death. What does this mean?

Scenario 1: (50) dies at age 52.3, T50 = 2.3

(2)
K50 T50
Complete half-years survived by (50)

0 1 2 3 Time

AS 2427 Long Term Actuarial Math I - Winter 2025 74

 Chapter 4

Scenario 2: (50) dies at age 52.9, T50 = 2.9

(2)
K50 T50
Complete half-years survived by (50)

0 1 2 3 Time

(1) Note that

(2) 1
K50 = ⌊2 T50 ⌋.
2
(2) Death benefit is paid at the end of half-year of death; that is,

(2) 1
Time of DB payment = K50 + .
2
• Example 2: If Tx = 23.675,
2
Kx = 23, Kx(2) = 23.5, Kx(4) = 23.5, Kx(12) = 23 = 23.6667.
3

(b) Present Value Random Variable (PVRV)

Z is the random variable for the PV of a benefit payment of $1. In the 1/mthly case, we have:
1
PVRV = Z = A function that depends on Kx(m) + .
m
Note that Kx(m) is a discrete random variable.
Example 4.19 In the monthly setup, what values can Kx(12) take?
Answer. Kx(12) can take discrete values:
1 2 3 4
0, , , , , ...
12 12 12 12

The probability mass function of Kx(m) is:

" # " #
k k k+1
P Kx(m) = = P (x) dies between times and = | qx ,
k 1 k = 0, 1, 2, . . .
m m m m m

(c) Expected Present Value (EPV)

Since Z is a random variable, we are interested in its expected value: E[Z]. This is also known as
Actuarial Value or Actuarial Present Value (APV). Here, we use EPV:
X Z
EPV = E[Z] = or amount × discount × probability.

*note that use integration for the continuous case, and summation for discrete cases.

AS 2427 Long Term Actuarial Math I - Winter 2025 75

 Chapter 4

4.6.1 Whole life insurance (DB payable at end of 1/m-year period of death)
Death benefit of $1 is payable at the end of 1/m-year period of death.

(a) The PV of the benefit

(m) 1
PVRV = Z = v Kx +m
.

(b) Expected Present Value

def
h (m) 1
i
A(m)
x = EPV = E[Z] = E v Kx +m
. (No bar on the top!)

To calculate the EPV of Z, we have

∞
" #
h (m) 1
i k 1 k
A(m) +m
Kx(m)
X
Kx +m
x = E[Z] = E v = v m P = .
k=0 m

Therefore,
∞
k+1
A(m)
X
x = v m k 1
| qx . (4.20)
m m
k=0

(c) Calculation of Var[Z]

h (m) 1
i  2
Var v Kx +m
= 2 A(m)
x − A(m)
x ,

where 2 A(m)
x is A(m)
x calculated at 2δ (or equivalently, v replaced with v 2 ).

Other insurances (including the deferred insurances) with DB payable at the end of 1/m-year
of death can be derived accordingly. For Example:
- Term life insurance with $1 DB payable at the end of 1/m-year of death:
 (m)
1 1
v Kx + m if Kx(m) ≤ n −
PVRV = Z = m .
0 if Kx(m) ≥ n

nm−1
(m) def X k+1
A1 = E[Z] = v m
| qx .
k 1 (4.21)
x:n m m
k=0

- Endowment life insurance with $1 DB payable at the end of 1/m-year of death:

 (m)
1 1
v Kx + m if Kx(m) ≤ n −
PVRV = Z = m .
v n if Kx(m) ≥ n

(m) (m)
Ax:n = A1 + n Ex . (4.22)
x:n

AS 2427 Long Term Actuarial Math I - Winter 2025 76

 Chapter 4

Example 4.20 (Exercise) Show that the following recursion formula holds:
1 1 (m)
A(m)
x = vm 1 qx + v m 1 px Ax+ 1
m m m

Example 4.21 Consider a whole life insurance of $1 on (x). Given that

δ = 0.04 and µx (t) = 0.02.

(1) Assume the death benefit is payable at the end of month of death. Find the EPV.
[TBC]

(2) What is the EPV if the death benefit is paid at exactly a month after the date of death?
[TBC]

Example 4.22 Compare the following three EPVs:

Āx A(m)
x Ax

[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 77

 Chapter 4

4.7 Relationships between discrete and continuous insurances

• Most often life insurance products are valued using life tables with values at integer ages

• Goal: Approximate Āx or A(m)

x using Ax .

• Can approximate mortality rate for fractional periods (using methods covered in Lecture notes
Section 3.2 Fractional Age Assumptions.)

– UDD within each year of age is the assumption used most often
– CF within each year of age can also be used

4.7.1 Using the UDD fractional age assumption

4.7.2 The claims acceleration approach

4.7.1 Using the UDD fractional age assumption

For any Insurance plan with constant benefit in each year and with no pure endowments

i
Ā = A (4.23)
δ

and
i
A(m) = A (4.24)
i(m)
i
Note that δ = ln(1 + i), and i(m)
is given in the Exam LTAM Tables.

Examples: Under UDD,

i i
Āx = Ax , A(m)
x = Ax . (4.25)
δ i(m)
i (m) i
Ā1x:n = A1x:n , A1 = A1x:n .
δ x:n i(m)
i (m) i
u| Āx = u| Ax , u| Ax = u| Ax .
δ i(m)

(m)
What about the approximation for Āx:n or Ax:n ? [TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 78

 Chapter 4

[Proof for (4.25)] (not required)

AS 2427 Long Term Actuarial Math I - Winter 2025 79

 Chapter 4

Example 4.23 Calculate the E(Z) and Var(Z) for a $50,000 whole life insurance policy issued to
(50) where the death benefit is payable at moment of death. Assume UDD over each year of age,
i = 6% and that you are given:
2
A50 = 0.2490475 and A50 = 0.0947561.

[TBC]

Example 4.24 Calculate the EPV for a 30 year $10,000 endowment insurance policy issued to (35)
where the death benefit is paid at the moment of death. Assume UDD over each year of age. You
are given
1
i = 6%, A35:30 = 0.06748178, 30 E35 = 0.1392408.

[TBC]

Example 4.25 You are given i = 6%, p30 = 0.9984711 and p31 = 0.9983911.
1
Determine Ā30:2 , assuming

(a) UDD within each year age

(b) Constant Force of mortality (CF) within each year of age

[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 80

 Chapter 4

4.7.2 The claims acceleration approach

• The claims acceleration approach is also a simple approximation, but less commonly used

• Idea: The only difference between continuous and discrete insurances is the timing of the DB
payment.
For example,

Ax : $1 death benefit is paid at Kx + 1

1 2 3
A(4)
x : $1 death benefit may be paid at Kx + 4 , Kx + 4 , Kx + 4
or Kx + 44 .
- If the death occurs evenly over the year, on average the payment is made at

(Kx + 14 ) + (Kx + 42 ) + (Kx + 34 ) + (Kx + 44 ) 5

= Kx +
4 8
3
which is 8
years earlier than the end of year.
- Therefore, we have an approximation

A(4)
x ≈ (1 + i)
3/8
Ax .

• In general, for an 1/mthly life insurance, the possible timings of payment are
1 2 m
Kx + , Kx + , . . . , Kx + ,
m m m
assuming deaths are uniformly distributed over the year of age, the average timing of DB
payment is
m+1
Kx + ,
2m
which is 1 − m+1
2m
= m−1
2m
years earlier than the end of year.
Therefore, the approximation under this method is
m−1
A(m)
x ≈ (1 + i) 2m Ax . (4.26)

Let m → ∞,
1
Āx ≈ (1 + i) 2 Ax . (4.27)
1
What about Āx:n and Āx:n ?

AS 2427 Long Term Actuarial Math I - Winter 2025 81

 Chapter 4

4.8 Varying Life Insurance (non-level benefit)

• Insurances payable at moment of death, end of year of death or end of 1/mth year of death
can be varying

• can categorize varying insurance, where benefits varies

– in step rate fashion (worked with already - see Example 4.18)

– arithmetically (application - return of premium rider)
– geometrically
– others, e.g., endowment insurance where survival benefit differs from death benefit
would also be considered “varying insurance”

• Note that for all the insurance benefits, the following rules always hold:

PVRV = Z = Benefit amount × discount factor.

X Z
EPV = E[Z] = or amount × discount × probability.

*note that use integration for continuous case, and summation for discrete cases.

Example 4.26 Consider a special insurance that pays $1,000 after 10 years if (x) dies by that time,
and $2,000 after 20 years if (x) dies in the second 10-year period, with no benefit otherwise. Write
out the PV and EPV. [TBC]

Example 4.27 Let Z be the present value random variable of a whole life insurance with the DB
paid at the moment of the death of (x). Given that µx (t) = 0.02, δ = 0.08 and the benefit amount
at time t is bt = e0.03t . Find Var(Z). [TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 82

 Chapter 4

4.8.1 Standard increasing insurance benefits

4.8.2 Standard decreasing insurance benefits
4.8.3 Geometrically increasing insurance benefits

4.8.1 Standard increasing insurance benefits

We use bt to denote the Death Benefit (DB) amount at time t.

• Two types of standard increasing, namely

– Continuously increasing: bt = t (or bTx = Tx ), notation I¯

– Annually increasing: bt = ⌊t⌋ + 1 (or bTx = Kx + 1), notation I
i.e., $1 in Year 1, $2 in Year 2, $3 in Year 3, . . ., $n in Year n, . . .

• For example, the EPV of the continuous life insurance can be calculated as
Z Z
EPV = E[Z] = amount × discount × probability = bt · e−δt · t px µx+t dt

• Whole life insurance with increasing benefit:

     
I¯Ā I Ā (IA)x IA(m)
x x x

(1) Continuously increasing Whole life payable at m.o.d.:

  Z ∞
I¯Ā = E(Z) = E[Tx e −δTx
]= t e−δt t px µx+t dt
x 0

(2) Annually increasing Whole life payable at m.o.d.:

  Z ∞
−δTx
I Ā = E(Z) = E[(⌊Tx ⌋ + 1)e ]= (⌊t⌋ + 1) e−δt t px µx+t dt
x 0
∞
X
= (k + 1)k| Āx:1
1
k=0
X∞
= k| Āx
k=0

(3) Annually increasing Whole life payable at e.o.y. of death:

∞
(IA)x = E(Z) = E[(Kx + 1)v Kx +1 ] = (k + 1)v k+1 k| qx
X

k=0

(4) Annually increasing Whole life payable at the end of 1/mth year of death:
∞

(m)
 (m)
Kx 1
+m
X k k+1
IA = E(Z) = E[(Kx + 1)v ]= (⌊ ⌋ + 1)v m k | 1 qx
x
k=0 m m m

AS 2427 Long Term Actuarial Math I - Winter 2025 83

 Chapter 4

• Term
 life insurance
  with increasing benefit
 is defined similarly:
¯ 1
I Ā x:n 1 1
I Ā x:n (IA) x:n IA (m) 1
x:n

(1) Continuously increasing Term life payable at m.o.d.:

  Z n
I¯Ā 1
x:n = t e−δt t px µx+t dt
0

(2) Annually increasing Term life payable at m.o.d.:

  Z n n−1
1
(⌊t⌋ + 1) e−δt t px µx+t dt =
X
I Ā x:n = (k + 1)k| Āx:1
1
0 k=0

(3) Annually increasing Term life payable at e.o.y. of death:

n−1
(IA) 1x:n = (k + 1)v k+1 k| qx
X

k=0

(4) Annually increasing Term life payable at the end of 1/mth year of death:
nm−1
  k k+1
IA(m) 1
X
x:n = (⌊ ⌋ + 1)v m k | 1 qx
k=0 m m m

4.8.2 Standard decreasing insurance benefits

• Term
 life
 insurance
 with decreasing benefits:
 
D̄Ā 1x:n DĀ 1x:n (DA) 1x:n DA(m) 1x:n

• Two types of standard decreasing, namely

– Continuously decreasing: bt = n − t (or bTx = n − Tx ), notation D̄

– Annually decreasing: bt = n − ⌊t⌋ (or bTx = n − Kx ), notation D
i.e., $n in Year 1, $(n-1) in Year 2, $(n-2) in Year 3, . . ., $1 in Year n.

• Note that
   
D̄Ā 1
x:n
1
= n Āx:n − I¯Ā 1
x:n
   
1 1 1
DĀ x:n = (n + 1) Āx:n − I Ā x:n

(DA) 1x:n = (n + 1)Ax:n

1 − (IA) 1x:n
   
(m)
DA(m) 1
x:n = (n + 1)A1 − IA(m) 1
x:n
x:n

AS 2427 Long Term Actuarial Math I - Winter 2025 84

 Chapter 4

Example 4.28 A $100,000 Whole Life Insurance Policy payable at the moment of death is issued to
(x). It provides for the Return of Premiums paid (without interest) if death occurs within 10 years
of policy issue. The policyholder pays an annual premium P at the beginning of each year for life
of the policy.

What is an expression for the EPV of all the policy benefits? [TBC]

Example 4.29 You are given i = 6% and the following annual mortality rates

q55 = 0.0089605, q56 = 0.0097538, q57 = 0.0106230.

1
(a) Calculate 30, 000(IA)55:3 .
1
(b) Determine 30, 000(I Ā)55:3 , assuming UDD of death within each year of age.
1
(c) What does 30, 000(DA)55:3 represent? How would you calculate this value? [TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 85

 Chapter 4

Example 4.30 A whole life insurance policy offers an increasing death benefit payable at the end
of the quarter of year of death. If (x) dies in the first year of the contract, then the benefit is $5,000.
If (x) dies in the second year, the benefit is $10,000. If (x) dies in the third year, the benefit is
$15,000, and so on.

Derive an expression for the EPV of the death benefit. [TBC]

What if the death benefit is $5,000 for the first year, $6,000 for the second year, $7,000 for the third
year and so on?

Example 4.31 (Exercise) Given δ = 5% and

100 − x
S0 (x) = .
100
Determine the EPV for an increasing 30 year term insurance policy issued to (35). Assume the death
benefit is payable at the moment of death and the death benefit at time t is $750t (or bt = 750t).
S0 (x+t) t
[Answer: 2,040.81 ] Note that t px = S0 (x)
=1− 100−x
.

Example 4.32
(a) Derive a recursion formula for (IA)x . (The formula should include (IA)x+1 ).
1 . (The formula should include (IA) 1
(b) Derive a recursion formula for (IA)x:n [TBC]
x+1:n−1 ).

AS 2427 Long Term Actuarial Math I - Winter 2025 86

 Chapter 4

4.8.3 Geometrically increasing benefits (or benefits with inflation)

• Assume benefit increases by j% per year

• We can do calculations at i∗ = i−j

1+j
.

• Example: If DB is $1 in year 1, $(1 + j) in year 2, $(1 + j)2 in year 3, and so on, where DB
is payable at the end of year of death, then the n-year term life insurance has the EPV:
n−1
1
(1 + j)k v k+1 k| qx =
X
EPV = 1
Ax:n i∗ ,
k=0 1+j
(1+i)
where v ∗ = (1 + j)v, or equivalently 1 + i∗ = 1
v∗
= (1+j)
, or

i−j
i∗ = .
1+j

Example 4.33 A 10-year term life insurance issued to (x) provides a death benefit at the end of
year of death. The death benefit is $100,000 if death occurs in the first year, $110,000 if death
occurs in the second year, and so on with the death benefit increasing by 10% each year (over the
previous year’s death benefit).
(a) Define Z.
(b) Derive an expression for the EPV of this policy.

AS 2427 Long Term Actuarial Math I - Winter 2025 87

 Chapter 4

4.9 Functions for select lives

Note: Text exercise 4.16 is a good question that uses select table.

Example 4.34 For a whole life insurance of $1000 on (80), with death benefits payable at the end
of the year of death, you are given

(1) Mortality follows a select and ultimate mortality table with a one-year select period
(2) q[80] = 0.5q80
(3) i = 6%
(4) 1000A80 =679.80 and 1000A81 =689.52

Find 1000A[80] .

[TBC]

Example 4.35 You are given i = 4% and the following select life table

Issue age x q[x] q[x]+1 q[x]+2 qx+3 Attained age x + 3

65 0.08 0.10 0.12 0.14 68
66 0.09 0.11 0.13 0.15 69
1
Calculate 2| A[65]:2 .

[TBC]

AS 2427 Long Term Actuarial Math I - Winter 2025 88