Chapter 2 - Survival Models

Section 2.2 - Future Lifetime Random Variable and

the Survival Function
Let

Tx = ( Future lifelength beyond age x of an individual who has

survived to age x [measured in years and partial years])

The total lifelength of this individual will be x + Tx , i.e. this is the age
at which the individual dies [including partial years].

The additional years of life Tx beyond x is unknown and therefore is

viewed as a continuous random variable. The distribution of this
random variable is described by

2-1
or by

where, of course,
Z t
Fx (t) = fx (s)ds.
0

Either of the functions fx (t) or Fx (t) are used to describe the future
lifetime distribution beyond age x. Clearly, Fx (t) = P[Tx ≤ t] is the
probability that someone who has survived to age x will not survive
beyond age x + t. Therefore,

is the probability that someone age x does survive t additional

years. All of the properties of the future lifetime distribution are in the
survival function Sx (t).

2-2
Properties of a Survival Function Sx (t)
Property 1:

Sx (0) = 1.

Everyone who survived to age x is alive at the beginning of the time

period beyond x.
Property 2:

No one lives infinitly long beyond x.

Property 3: If t1 < t2 then

Sx (t1 ) ≥ Sx (t2 ).

The function Sx (t) is non-increasing.

2-3
Let T0 denote the total lifelength from birth of an arbitrary individual.
The density of its distribution is f0 (t).
Note that

Fx (t) = P[Tx ≤ t] = P[x < T0 ≤ x + t | T0 > x]

P[x < T0 ≤ x + t]
= = (2.1)
p[T0 > x]

2-4
Taking a derivative with respect to t produces

So the fx (·) density is proportional to the f0 (·) density at the

corresponding time point.

From expression (2.1) we also see that

S0 (x) − S0 (x + t) S0 (x + t)
Fx (t) = =1−
S0 (x) S0 (x)

Therefore

which is the fraction alive at x who continue to be alive at x + t.

2-5
Rewriting this expression produces

S0 (x + t) = S0 (x)Sx (t)

which shows that the probability of surviving x + t years is the

probability of surviving x years times the conditional probability of
surviving t additional years given survival to time x.

More generally, the same reasoning produces

which shows that the probability of surviving t + u years beyond x is

the probability of surviving t years beyond x times the conditional
probability of surviving u additional years given survival to time x + t.

2-6
2-7
Assumptions for a Survival Function Sx (t) that are useful when
finding expected values
Assumption 1:

The survival function Sx (t) is a smooth nonincreasing function of t.

Assumption 2:

lim tSx (t) = 0.

t→∞

The right-hand tail of the survival function goes to zero sufficiently

fast as t goes to infinity.
Assumption 3:

lim t 2 Sx (t) = 0.
t→∞

The right-hand tail of the survival function goes to zero even faster
as t goes to infinity.
2-8
Example 2-1: Let ω denote some upper age limit (e.g. 120) and
(  2  
12 t t
1 − for 0 < t < ω
f0 (t) = ω ω ω
0 elsewhere

Find F0 (t), Sx (t) for general ω and S40 (10) when ω = 120.

2-9
Section 2.3 - Force of Mortality
Concept - At any age, what is the rate of death among persons who
have survived to that age?
Large positive number −→ hazardous age
Small positive number −→ less hazardous age
Define the Force of Mortality at age x to be

P[x < T0 < x + dx | T0 > x]

µx = lim
dx&0 dx
F0 (x+dx)−F0 (x)
limdx&0 dx
= or
S0 (x)

2-10
Force of Mortality is a function of the age x of the individual. It is
also called the hazard function or the failure rate function. Note that
F00 (x)|t=x
µx =
S0 (x)
 
d
1 − S (t)

dt 0 
t=x
= or
S0 (x)

This shows that the survival function characterizes the force of

mortality. Note also that
d h i
µx = − ln(S0 (x)) so
dx
Z t
µx dx = − ln(S0 (t)) + ln(S0 (0)).
0
2-11
It follows that

Therefore the force of mortality function characterizes the survival

function.
Note also that
R x+t
S0 (x + t) e− 0 µr dr
Sx (t) = = Rx
S0 (x) e− 0 µr dr

R x+t Rt
= e− x µr dr
= e− 0 µx+r dr

In the same manner we see

2-12
 −S00 (x + t)
µx+t =
S0 (x + t)

 
S0 (x+t+∆)−S0 (x+t)
− lim∆&0 ∆
=
S0 (x + t)

 
S0 (x)Sx (t+∆)−S0 (x)Sx (t)
− lim∆&0 ∆
=
S0 (x + t)

−S0 (x) Sx (t + ∆) − Sx (t)

= lim or
S0 (x + t) ∆&0 ∆

2-13
 1
Example 2-2: Given µx = 100−x for 0 < x < 100,
find S50 (10) = P[T50 > 10].

2-14
Example 2-3: Given µx = 2x for 0 < x,
find f0 (t), F0 (t), S0 (t) and fx (t).

2-15
Gompertz Law of Mortality (1825):

where 0 < B < 1 and C > 1.

Here the force of mortality is increasing exponentially. It follows that
the survival function is:

Rt
Sx (t) = e− 0 µx+r dr

x
Rt
C r dr
= e−BC 0
h it
Cr
−BC x ln(C)
=e 0
 
x
BC
− ln(C) C t −1
=e

2-16
Makeham Law of Mortality (1860):

where A > 0, 0 < B < 1 and C > 1.

The coefficient B is part of what determines the rate of ascent of the
force of mortality. It is also part of the value of the force of mortality
when x = 0. The addition of the coefficient A allows an adjustment
to the force of mortality at x = 0 that is not part of its rate of ascent.
The survival function is now:
Rt Rt
dr −BC x C r dr
Sx (t) = e−A 0 0

 
BC x
−tA− ln(C) C t −1
=e

2-17
Section 2.4 - Actuarial Notation
Having survived to age x, the probability of surviving t additional
years is:

Having survived to age x, the probability of NOT surviving t

additional years is:

Having survived to age x, the probability of surviving u additional

years and then dying within t years after x + u, is:

u t qx = Sx (u) − Sx (u + t) = P[u < Tx < u + t]


This is referred to as a deferred mortality (here deferred u years).

2-18
It follows that

u t qx = u px − u+t px = u px (t qx+u ).


Also,

Note that

S00 (x) − d (x p0 )
µx = − = dx .
S0 (x) x p0

In the same manner,

2-19
But since
d d
t px = Sx (t) = −fx (t),
dt dt
we also get

Using the material from section 2.2, we see that

Likewise, we have

2-20
Section 2.5 - Properties of Tx
The future lifetime at age x, Tx , is a continuous random variable. We
are interested in the properties of this random variable. In particular,
its mean is called the complete expectation of life and is equal to

Z ∞
= t (t px )µx+t dt
0
Z ∞
d

=− tt px dt
0 dt
∞ Z ∞
= −t (t px ) + t px dt,

0 0
producing the computation formula

2-21
In a similar manner, the computation formula for the second moment
of Tx is

It follows that the variance of Tx is computed via

◦
Var [Tx ] = E[Tx2 ] − (ex )2 .

and, of course the standard deviation of Tx is

p
StD[Tx ] = Var [Tx ].

2-22
The percentiles of the distribution of Tx are of interest.
In particular, the Median, m(x), (the 50th percentile) is the value
which satisfies

Another concept is:

◦
ex:n| ≡ Average number of years lived within the next n years

It can be computed with

Z n
◦
ex:n| = t fx (t)dt + nP[Tx > n].
0

2-23
The Central Death Rate
Rt
0 µx+s s px ds
t mx = Rt
0 s px ds

is a weighted average of the Force of Mortality values over the

interval from x to x + t.

2-24
Example 2-4: Continuing example 2-1, find
◦
ex , StDev(Tx ), and m(x).

2-25
Section 2.5.5 - Some Important Mortality Models

Uniform Distribution or DeMoivre’s Law

1

 ω if 0 < t < ω
f0 (t) =
0 elsewhere


t
t q0 = F0 (t) = for 0 < t < ω
ω
ω−t
t p0 = for 0 < t < ω.
ω

With this model, if we assume the person has already lived to age x,
then

2-26
The force of mortality under DeMoivre’s Law is

Note that it is an increasing function of the age x. That is, life is more
hazardous as we get older under this model.

Note also that Tx is also uniform ( 0, ω − x ) and thus, for example,

◦ ω−x ω−x
ex = E[Tx ] = , m(x) = and
2 2
(ω − x)2
Var [Tx ] = .
12

2-27
An important property of the DeMoivre Law (Uniform Distribution) is
its reproducibility. If a future lifelength is uniform, then the future
lifelength beyond any future age is also uniform. That is, if Tx is
uniform (0, ω − x), then Tx+y is uniform (0, ω − x − y ). So future life
length distributions stay within the class of uniform distributions, it
merely changes the parameter of the distribution (the length of the
interval in this case).

2-28
Exponential Distribution

Z t
1 −s
t q0 = F0 (t) = e θ ds
0 θ
s t

= −e− θ 
0
t
= 1 − e− θ for 0 < t

t
t p0 = e− θ = S0 (t) for 0 < t

2-29
Now suppose this exponential function describes survival from birth
and that the person has already lived to age x > 0. The density of
future life length beyond x is

This clearly shows that the future life length beyond x has exactly
the same distribution as the original life length from birth.
The exponential distribution has an even stronger reproducibility
property than the uniform distribution had. Under the exponential
distribution for future life length, the life length distribution beyond
any point in the future is exactly the same exponential distribution
that is applicable beyond today (same distribution AND the same
parameter value).

2-30
For the exponential distribution:
Z ∞ Z ∞
◦ t
ex = E[Tx ] = t px dt = e− θ dt
0 0

Z ∞ Z ∞ t
2
E[Tx ] = 2 t(t px )dt = 2 te− θ dt
0 0
Z ∞
h
− θt ∞
 2 1 −t i
= 2 − θte 0
+θ e θ dt
0 θ
2
= 2θ .
Therefore

Var [Tx ] = 2θ2 − θ2 = θ2 and

2-31
For the exponential, the force of mortality is

d 1 t 1
Sx (t)t=0 = e− θ t=0 = .

µx = −
dt θ θ

Moreover, a constant force of mortality characterizes an exponential

distribution. Let µ∗ denote a constant force of mortality. Then

This is, of course, the survival function of an exponential distribution

with
1
µ∗ = .
θ
While a constant force of mortality throughout life is unrealistic, MLC
exam questions frequently assume different constant forces of
mortality over various segments of a lifetime.

2-32
Weibull Distribution
This family of distributions has two parameters: a scale parameter
θ > 0 and a shape parameter τ > 0. Its survival function takes the
form:

This produces a density function of the form

  τ −1
τ
t
 τθ θt
 e− θ for 0 < t
f0 (t) =

for t ≤ 0

0

and a distribution function

τ
t
−
t q0 = F0 (t) = 1 − e θ

2-33
The Weibull force of mortality function is:

When τ > 1, this is an increasing function of x (proper for mortality)

though it does spread mortality over the whole positive part of the
real line.
When τ < 1, this is an decreasing function of x (generally improper
for mortality).
When τ = 1, the Weibull is the exponential distribution and is only
appropriate for relatively short periods of time.

2-34
Using the Weibull distribution to describe mortality from birth, the
future lifelength beyond age x satisfies

τ
t+x
−
S0 (x + t) e θ
Sx (t) = =
τ
S0 (x) x
e− θ

which is not a survival function of a Weibull distribution (it lacks

reproducibility).
Also note that
◦
τ + 1
ex = θΓ and
τ

n  τ + 2  h  τ + 1 i2 o
Var [Tx ] = θ2 Γ − Γ
τ τ

2-35
Generalized DeMoivre (Beta)

Here ω, the maximum age, is essentially a scale parameter and α is

a shape parameter.

2-36
When α = 1, this is DeMoivre’s Law, ie it is a uniform (0, ω)
distribution. We also note that for the generalized DeMoivre
distribution
 ω − t α
t q0 = F0 (t) = 1 − for 0 < t < ω and
ω

Suppose The generalized DeMoivre applies form birth, but the

individual has survived to age x > 0. The density of the future
lifelength beyond x is:
  α−1
α ω−x−t
f0 (x + t)

 ω−x ω−x if 0 < t < ω − x
fx (t) = =
x p0 

0 elsewhere

2-37
We see that this conditional distribution is also a member of the
generalized DeMoivre family with scale parameter ω − x and the
same shape parameter α. So this family has a reproducibility
property.
The force of mortality function for the generalized DeMoivre is

This is a decreasing function of x for all α > 0. Like the DeMoivre

Law this generalized family is best applied to relatively short periods
of time.
Also note that
◦ ω−x
ex = and
α+1

(ω − x)2 α
Var [Tx ] = .
(α + 1)2 (α + 2)
2-38
Example 2-5: You are given that there is a constant force of mortality
◦
µ∗ and that e30 = 41. Find µ∗ .

2-39
  α
Example 2-6: You are given S0 (t) = 1 − ωt for 0 < t < ω and
◦ ◦
α > 0. Derive ex and then find µx ex .

2-40
Section 2.6 - Curtate Future Lifetime

When describing a number of features of a policy, e.g. the number of

future annual premium payments, it is useful to model the integer
which represents the whole number of future years lived by a person
who is currently age x. This is the discrete random variable

where btc is the largest integer that is less than or equal to t.

We note that

P[Kx = k ] = P[individual survives k years but not k + 1 years]

= P[k ≤ Tx < k + 1]
= k px − k +1 px = k px − k px px+k
= k px (1 − px+k ) = k px qx+k .

2-41
The expected value of Kx is denoted by ex and can be computed via
∞
X
ex = E[Kx ] = k P[Kx = k ]
k =0

= 1(1 px − 2 px ) + 2(2 px − 3 px ) + 3(3 px − 4 px ) + · · ·

= 1 px + 2 px + 3 px + · · ·
∞
X
= k px
k =1

2-42
Likewise the second moment is:
∞
X
E[Kx 2 ] = k 2 P[Kx = k ]
k =0
∞
X
= k 2 (k px − k +1 px )
k =0
= 12 (1 px − 2 px ) + 22 (2 px − 3 px ) + 32 (3 px − 4 px ) + · · ·

∞
X
= (2k − 1) k px
k =1
∞
X ∞
X
=2 k k px − k px
k =1 k =1
X∞
=2 k k px − ex .
k =1
2-43
Therefore,

Because

Tx ≥ Kx > Tx − 1,

0 ≤ Tx − Kx < 1.
As an approximation, it is sometimes assumed that in a short period
of time (eg one year) deaths occur uniformly. Thus it is assumed
1
(Tx − Kx ) ∼ uniform(0, 1) and therefore E[Tx − Kx ] = .
2
Based on this assumption
◦
ex = E[Tx ] = E[Kx + (Tx − Kx )]

. 1
= ex + .
2
2-44
Example 2-7: Suppose T0 ∼ DeMoivre with ω = 100. Find the
curtate mean e20 .

2-45