Understanding Life Annuities Explained
Understanding Life Annuities Explained
• There are many different life annuity types and applications, including
– Retail product
– Life insurance premium & benefit determinations (Chapter 6)
∗ Most policy premiums are payable annually or monthly
– Pension plans benefit (and funding of benefits) at retirement
∗ Covered later in textbook (and in upper-year courses)
– Disability insurance
• We will review various types of life annuities, their IAN (International Actuarial Nota-
tion), how to value them, relationship to insurances, and relationship between different
annuity types and approximations used
• Majority of life annuities are payable more frequently than annually, so often need approxi-
mations for fractional periods
• Similar to life insurance, annuity values can be generated recursively and select mortality
rates can be used
• Life annuities are also often sold with a guarantee period and this can be readily handled for
any type of life annuity
Let’s review the formulas for annuity-certain (including annuity-due and annuity-immediate):
1
For example, 1-year annuity-due (certain) payable m
thly:
1
For example, 1-year annuity-immediate (certain) payable m
thly:
• Formulas
1 − vn i
ä n = 1 + v + v 2 + v 3 + · · · + v n−1 = where d = ,
d 1+i
1 − vn
an = v + v2 + v3 + v4 + · · · + vn = = ä n − 1 + v n ,
i
(m) 1 1 1 1 1 1 − vn
ä n = + v 1/m + v 2/m + v 3/m + · · · + v n−1/m = (m) ,
m m m m m d
(m) 1 1/m 1 1 1 1 1 − vn
an = v + v 2/m + v 3/m + v 4/m + · · · + v n = (m) ,
m m m m m i
Z n
1 − vn
ā n = v t dt = .
0 δ
• Others:
¯ n =
Z n
ā n − n v n
(Iā) t v t dt = ,
0 δ
as well as (Ia) n , (Iä) n , (Da) n , (Dä) n .
(12)
Example 5.1 For an annuity notation 1500 ä 5 , how many payments are there, and how much is
each payment?
Life annuities notation (IAN) builds upon some of notation above. Also, life annuities with a
guarantee period will be valued in part with a term certain.
Throughout this Chapter, we use Y to denote the Present Value Random Variable of a life annuity.
0 1 2 3 ··· Kx Kx + 1 Time
$1 $1 $1 $1 $1
Note that
# of life annuity payments = Kx + 1.
The PVRV for a $1 whole life annuity-due is given by
Kx 1 − v Kx +1
PVRV = Y = 1 + v + · · · + v = ä Kx +1 = . (5.1)
d
def
h i 1 − v Kx +1
äx = EPV = E[Y ] = E ä Kx +1 = E[ ].
d
Ax = 1 − däx , or 1 = däx + Ax .
Therefore,
∞
X ∞
X
äx = ä k+1 k| qx = ä k+1 k px qx+k .
k=0 k=0
Therefore,
∞
v k k px .
X
äx =
k=0
k=0
Therefore,
∞
Var(Y ) = E(Y 2 ) − (EY )2 = (ä k+1 )2 k| qx − (äx )2 .
X
k=0
cov(Yi , Yj ) ̸= 0.
[TBC]
or equivalently
1 − v min(Kx +1,n)
PVRV = Y = ä min(Kx +1,n) = . (5.6)
d
Therefore,
n−1
X
äx:n = ä k+1 k| qx + ä n n px .
k=0
Therefore,
n−1
v k k px .
X
äx:n =
k=0
Example 5.2 For a whole life annuity-due issued to (50) that pays $10,000 annually. Determine
the standard deviation of the PVRV for this policy given i = 6% and the corresponding calculated
values:
2
ä50 = 13.26683, A50 = 0.2490475, A50 = 0.0947561.
[TBC]
x
Example 5.3 You are given S0 (x) = 1 − 80 for 0 ≤ x ≤ 80 and i = 5%. Y is the PVRV for a
3-year temporary life annuity-due of $1 per year.
Find (i) E(Y ) and (ii) Var(Y ) when x = 30.
[TBC]
Calculate the change in the EPV of this annuity-due if px+1 increased by 0.03.
[TBC]
def
u| äx = EPV = E[Y ],
Alternative methods:
u| äx = äx − äx:u . (5.11)
Based on the two formulas above, what is the relationship between term life annuity and whole life
annuity? [TBC]
Example 5.5 A whole life annuity-due is issued to (50) that provides for annual payments. The first
10 payments are $1,000 and every payment thereafter is $2,000.
You are given life annuity-due and pure endowment factors (based on i = 6%)
[TBC]
def
äx:n = EPV = E[Y ],
we have
äx:n = ä n + n| äx , (5.14)
where the n-year deferred whole life annuity-due is discussed in the previous subsection.
*Note: any type of life annuity can have a guarantee period (e.g., see Example 5.8).
More Examples:
Example 5.6 Calculate the EPV of a three-year term life annuity-due of $100 on (75). You are
given i = 5% and Z x
µt dt = 0.01x1.2 , x > 0.
0
[TBC]
Example 5.7 (Policy with two benefits) For a special 30-year deferred whole life annuity-due of
$200 paid annually on (30), you are given
(i) The single benefit premium is refunded without interest at the end of year of death if death
occurs during the deferral period
(ii) From the mortality table, A30 = 0.10248, A60 = 0.36913, 30 E30 = 0.150044.
(iii) i = 6%.
[TBC]
Example 5.8 A 20-year term life annuity due is issued on (35) that provides for annual payments
of $6,000. The first 5 payments are guaranteed. Calculate the Expected present value (EPV) of this
annuity. Assume i = 5% and that mortality follows the SOA-FAM-L Standard Ultimate Life Table.
[TBC]
Example 5.9 (Exercise) A whole life annuity due that provides for annual payments of $10,000 is
purchased by (60). You are given that i = 5% and that mortality follows the SOA-FAM-L Standard
Ultimate Life Table except in the first two years where the mortality rates are higher denoted by q ∗ .
∗ ∗
You are given that q60 = 0.003 and q61 = 0.0035. Calculate the expected present value (EPV) of
this annuity.
0 1 2 3 ··· Kx Kx + 1 Time
$1 $1 $1 $1
PVRV = Y = v + · · · + v Kx = a Kx .
ax = äx − 1. (5.15)
– Variance
Since Y = a Kx = ä Kx +1 − 1, the variance of whole life annuity-immediate is the same
is the variance of whole life annuity-due:
2
Ax − (Ax )2
Var(Y ) = . (5.16)
d2
– Variance
Using (5.18), the variance of n-year term life annuity-immediate is the same is the
variance of (n + 1)-year term life annuity-due:
2
Ax:n+1 − (Ax:n+1 )2
Var(Y ) = . (5.20)
d2
Example 5.10 Determine the EPV of a life annuity issued to (50) that pays $12,000 at the end of
each year for 10 years. You are given i=6% and the corresponding calculated values:
[TBC]
Find ax:20 .
[TBC]
1 − v Tx
" #
def
h i
āx = EPV = E[Y ] = E ā Tx =E .
δ
1 − Āx
āx = . (5.22)
δ
– Method 2: Aggregate method Recall the density of Tx is f (t) = t px µx+t
h i Z ∞
āx = E ā Tx = ā t t px µx+t dt. (5.23)
0
Z ∞ Z ∞
āx = v t t px dt = e−δt t px dt. (5.24)
0 0
(c) Variance
(b) Expected Present Value (EPV) A special Actuarial Notation is used to denote the EPV:
1 − v min(Tx ,n)
" #
def
h i
āx:n = EPV = E[Y ] = E ā min(Tx ,n) =E .
δ
1 − Āx:n
āx:n = . (5.26)
δ
*Note that when δ = 0, āx:n = e̊x:n the n-year term expectation of life.
(c) Variance
2
Āx:n − ( Āx:n )2
Var(Y ) = . (5.29)
δ2
Example 5.12 Consider a continuous whole life annuity of $1 on (x). Assume the constant force of
mortality µ = 0.06 and the constant force of interest δ = 0.04. Calculate the probability that the
present value of the annuity payments is larger than its EPV.
[TBC]
Example 5.13 Find the Var(Y ) where Y = ā Tx . You are given µx+t = c, δ = 8% and Āx = 0.3443.
[TBC]
1
5.6 Life Annuities Payable m thly
• $1 per year is payable 1
m
thly contingent on survival.
Note that:
1
If annuity (of $1 per year) is payable weekly, each annuity payment is $ 52 ;
1
If annuity (of $1 per year) is payable daily, each annuity payment is $ 365 ;
• Note that when m = 1, it becomes the life annuity with annual payment; when m → ∞, it
becomes the life continuous annuity.
1
5.6.1 Whole life annuity-due payable m
thly:
• Payments of 1
m
made at the beginning of each 1
m
th of a year for as long as (x) lives
1
Kx(m) = ⌊m Tx ⌋ = the future lifetime of (x) in years rounded to the lower 1/m of a year
m
Time 0 1 2 3 ··· 1
m m m Kx(m) Kx(m) + m
1 1 1 1 1
Payments m m m m m
1
The PVRV for a $1 whole life annuity-due payable m
thly is given by
(m) 1
1 1 1 1 (m) (m) 1 − v Kx +m
PVRV = Y = + v m + · · · + v Kx = ä (m) 1 = . (5.30)
m m m Kx + m d(m)
1 − A(m)
x
ä(m)
x = . (5.31)
d(m)
∞
1 k
ä(m)
X
x = vm k px .
k=0 m m
1
a(m)
x = ä(m)
x − . (5.33)
m
1
(Only difference is that m
is not paid at time t = 0.)
1
5.6.2 Term life annuity-due payable m
thly
• Payments of m1 made at the beginning of each 1
m
th of a year until the earlier of n years
(maximum of nm payments) or when (x) dies.
(a) Present Value Random Variable (PVRV)
1
The PVRV for a $1 n-year term life annuity-due payable m
thly is given by
(m) 1
,n)
(m) 1 − v min(Kx +m
PVRV = Y = ä = . (5.34)
min(Kx
(m) 1
+m ,n) d(m)
(m)
(m) 1 − Ax:n
äx:n = . (5.35)
d(m)
– Method 2: Aggregate method
nm−1
(m) X (m) (m)
äx:n = ä k 1
k 1
| qx + ä n n px .
m
+m m m
k=0
nm−1
(m) X 1 k
äx:n = vm k px .
k=0 m m
(m) 2
2 (m)
Ax:n − Ax:n
Var(Y ) = . (5.36)
(d(m) )2
(m) (m) 1 1 n
ax:n = äx:n − + v n px . (5.37)
m m
ax < a(m)
x < āx < ä(m)
x < äx (5.38)
Note: same ordering occurs with term life annuities (see text Table 5.2).
or
u| äx:n = u Ex äx+u:n
• Any n-year term life annuity can be valued as a series of n one year deferred life annuities,
for example
n−1
X
äx:n = u| äx:1 .
u=0
Example 5.14 You are given that 10 p30 = 0.99611, 10 p40 = 0.99233, i = 6%, and also
Example 5.15 (Exercise) You are given the force of interest δ = 0.06 and the force of mortality
0.01, 0<t<5
µx (t) =
0.02, t ≥ 5.
Find āx .
(Hint: separate the whole life annuity into a term life annuity and a deferred whole life annuity.)
[Answer: 13.027]
Example 5.16 (Textbook Example 5.4) A pension plan member is entitled to benefit of $1,000 pear
month, in advance, for life from age 65, with no guarantee. She can opt to take a lower benefit, with
a 10-year guarantee. The revised benefit is calculated to have equal EPV at age 65 to the original
benefit. Calculate the revised monthly benefit using
(12) (12)
ä65 = 13.087, ä75 = 9.854, 10 p65 = 0.876938
Example 5.17 (Exercise) A person (40) wins $10,000 in a lottery. Rather than receiving the money
at once, the winner is offered an equivalent option of receiving an annual payment of $K at the
beginning of each year guaranteed for 10 years and continuing thereafter for the whole life. Assume
payments are made at the beginning of each year.
You are given
1
d = 0.05, A40 = 0.30, A50 = 0.35, A40:10 = 0.09.
Find K.
[Answer: 631.90]
• Life annuities that vary in step-rate fashion can be split and valued using term life annuities
and deferred life annuities, and you should be able to work with these types of problems (e.g.,
Examples 5.5 and 5.20)
∞
(k + 1) v k k px
X
EPV = (Iä)x = (5.39)
k=0
n−1
(k + 1) v k k px
X
EPV = (Iä)x:n = (5.40)
k=0
n−1
X Z k+1
EPV = (Iā)x:n = (k + 1) v t t px dt (5.41)
k=0 k
R1 R2
For example, (Iā)x:2 = 0 v t t px dt + 2 1 v t t px dt.
n−1
(n − k) v k k px
X
EPV = (Dä)x:n = (5.43)
k=0
∞
(1 + j)k v k k px = ä∗x ,
X
EPV = (5.44)
k=0
i−j
where ä∗x:n is the n-year term life annuity-due evaluated at i∗ = 1+j
.
Example 5.18
¯ x . Can you connect it to (I¯Ā)x ?
(a) Write out the PV random variable for (Iā)
¯ x directly.
(b) Given µ = 0.04 and δ = 0.06, calculate (Iā)
[TBC]
Example 5.19 A five year life annuity due is purchased by (50).The first payment is $2,000 and
each successive payment increases by 3%. Assuming i=4.5% and given mortality determine the
EPV for this life annuity.
[TBC]
Example 5.20 Determine an expression for the EPV for a life annuity-due issued to a person age
30 if the benefits are:
$10,000 per year from age 30 to 39;
$25,000 per year from age 40 to 49; and
$40,000 thereafter
[TBC]
äx = 1 + v px äx+1 ,
and
1 1 (m)
ä(m)
x = + vm 1 px äx+ 1 ,
m m m
• If there is an integer limiting age ω so that qω−1 = 1. First set äω−1 = 1. Then use the
backward recursion for x = ω − 2, ω − 3, . . ., we can construct äx in the spreadsheet. (Similar
to Chapter 4)
5.11.2 Approximations
• Approximations for annuities payable more frequently than annually
ä(m)
x = α(m)äx − β(m), (5.46)
where
id i − i(m)
α(m) = , β(m) = .
i(m) d(m) i(m) d(m)
*Note that α(m) and β(m) are given in the distribution table.
• It express ä(m)
x in terms of äx using the first three term from Euler-Maclaurin formula. The
Euler-Maclaurin formula is a numerical integration method (see Appendix B in the textbook).
m − 1 m2 − 1
ä(m)
x ≈ äx − − (δ + µx ), (5.47)
2m 12m2
Example 5.22 Given corresponding life annuity due factors below and assuming i=5%:
(i) Calculate the following assuming UDD (within each year of age)
(12) (12) (12)
(a) ä40 , (b) ä40:30 , (c) 30| ä40 .
(ii) Calculate the above using Woolhouse’s formula with 2 terms (W2). [TBC]
• Woolhouse’s formula with 3 terms (W3) is the best method; gives excellent results and is
more efficient than exact calculations (Less calculations required)
• W2 is worst approximation