Professional Documents
Culture Documents
Life Table 41
TABLE 3.1a
0 10,000
97,408
2 97,259
3 97,160
4 97,082
109
110 0
The basic advantage of the traditional form of the life table is its suscepti-
bility to interpretation. If we view o = 100,000 as a hypothetical cohort
group of new-borm lives, then each value of represents the survivors of
that group to age x, according to the model. This is a convenient, determin
istic, interpretation of the model. Of course, since e = lo S(r), and S(x) is
probability, then e is the ekpected number of survivors to agex out ofan
ctiginal grouý of to nèw-borns This connection between S(x) and , is also
giveninChapter I of Jordan [41], and in Chapter 3 of Bowers, et al. [12].
The distinction between Str) and i s sufficiently subtle to justify a
second look at them. We first recognize that mathematically the two func-
tions are identical. They embody the same information. The basic difference
between them lies in their interpretation. Sx) is a probability function, and
therefore, has a probabilistic interpretation. , is interpreted as a number of
persons living at age x (out of an original cohort of o), and thus has both a
probabilistic and a deterministic interpretation.
Although the basic representation of the tabular survival model is in
terms of the values of E, it is customary for the table to also show the value
of several other functions derived from . We define
d= - +1 (3.2)
or, more generally
(3.3)
Since , represents the size of the cohort at age x, and +n.is the number of
them still surviving at age xtn, then clearlyd; gives thè iiimber who die
betweenages i and x+n. (This portrayal of number dying might be the
reason for the frequent historical reference to these models as mortality
tables.) Furthermore.
42 The Life Table
(3.4)
or, more generally,
(3.5)
nP:=1-ng: = = (3.6)
P. = (3.7)
Recall that the conditional probabilities nP, and ,gx were defined in
Section 2.4 in terms of S(x). The consistency of those definitions with the
ones presented in this section is easily seen since ly is simply Eo S(x). We
redefined p and nqx in terms of , here simply to
complete our description
of the life table.
EXAMPLE 3.2 From Table 3.la, find (a) the number who die between
ages 2 and 4; (b) the probability that a life age l will survive to age 4.
SOLUTION
(a) This is given by 2dh = e2 - la = 17.
(3.8)
which we call the force of mortality at age x. Now since ly = lo S(x), then
we see that (3.8) is the same as
A) = S)= (3.9)
Thus the hazard rate and the force of mortality are identical.
Ifwe multiply both sides of Equation (2.11) by lo, replace with x,
and substitute 4, for A(y), we obtain
'A
e, = lo S(r) = lg exp
(3.10)
44 The Life Table
in the life table context, S(x) = sPo = exp- Jo, dy can be interpreted as
decremental factor which reduces the initial cohort of size o to size l, at
age x
xts (3.8a)
a form in which the force of mortality will be
frequently expressed.
EXAMPLE 3.3| Show that the force of mortality is the limiting value of
the probability of death over an interval. divided by the interval length (in
years), as the interval length approaches zero.
lim A
by Equation (3.8).
3.3.2 The Probability Density Function ofX
Sx) =
H-(l/lo) =
Potr, X20. (3.11)
Since, from (3.8), = -
Po Poux (3.12)
Other Functions Derived from 45
With the PDF in hand, we can now find E[X], which we recall is de-
noted by êo. (Throughout this and the following section, all expectations are
assumied to exist.) We have
& ELX) =
Pod = ,dx. (3.14)
If we define
To (3.15)
then we have
o =EX] = (3.16)
T 4, dy. (3.17)
Yo 1,d. (320)
obtaining
2 Yo (.21)
We have already discussed the conditional probabilities nP, and ,4. in terms
of both Sx) and E,.
Another conditional probability of some interest is denoted by
nimf
lt represents the probability that a person known to be alive at age x will die
between ages xt+n and x+nt+m. In terms of the formal probability notation
of Chapter 2, nlmgs = Pr [(xt+n) < X < (r+n+m)| X> x}. This can also be
expressed as the probability that a person age x will survive n vears, but then
die within the next m years. This way of stating the probability suggests that
we can write
Img t .tn
x+n (3.24)
EXAMPLE 3.5| Show that aln4 nd 4 mpr, and give an interpretation
of this result.
SO1X>x) = (3.25)
Letting S=y-X, so y =
X+s, we have
Ss1X> r) =
sPsks (3.26)
where the random variable S denotes the length of future lifetime of a person
alive at age x. The conditional PDF given by (3.26) is a very useful function
for developing other results.
If both numerator and denominator of (3.8a) are divided by e, we
obtain
sPx 3.27)
which shows that
-sPx4- (3.28)
ds (3.30)
from (3.17).
By steps parallel to those leading from (3.18) to (3.21), we find
ES] = (3.31)
where
T,dy (3.32)
TJ , dy =
10,000 (+1) dy =
5000(x+1)
and
Y, =
7,dy
dy=5000 (y+1)dy =
5000(r+1)T