Baltica Insurance Company LTD., Ballerup, Denmark: by Henrik Ramlau-Hansen

DISTRIBUTION OF SURPLUS IN LIFE INSURANCE
BY HENRIK RAMLAU-HANSEN
Baltica Insurance Company Ltd., Ballerup, Denmark
ABSTRACT
This paper discusses distribution of surplus in life insurance within a general

Markov chain framework. A conservative interest rate and a conservative set
of transition intensities are used for reserving purposes whereas more realistic
assumptions are used for the purpose of distributing surplus. The paper
examines various actuarial aspects of distributing surplus through either cash
bonuses, terminal bonuses or increased benefits. The results are illustrated by
some examples.
KEYWORDS
Distribution of surplus; bonus; with profits annuity policy; with profits

disability policy.
1. INTRODUCTION
The traditional life policy is a participating policy with margins of safety built
into the valuation elements to allow for protection for adverse deviations.
Surplus or profit can, therefore, in most cases be expected to emerge over the
life of a portfolio of business. A large proportion of the surplus is usually
distributed to the policyholders as bonuses or dividends. This distribution of
surplus may be carried out in various ways. One method provides cash
payments or reduction of premiums as the surplus arises, or the accumulated
value of the cash bonuses may be paid when the policy becomes a claim or
expires. By this method, a separate savings account is attached to the policy
and the surplus is credited to the account as it emerges. Another way of
distributing surplus is through terminal bonuses paid only when the policy
expires. By this method, only survivors get a share of the accumulated surplus.
The third method, and perhaps the most widely used, is one in which the profit
is distributed to the policyholders by means of increasing the insurance
benefits. This method provides a gradual increase in the benefits granted under
the policy.
It is believed that these three different ways of distributing surplus cover
many of the methods used in practice. We shall in this paper discuss various
actuarial aspects of the mentioned distribution methods. The idea is that the
ASTIN BULLETIN, Vol. 21, No. 1
Downloaded from https://www.cambridge.org/core. IP address: 92.253.57.250, on 17 Nov 2020 at 15:26:22, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.2143/AST.21.1.2005401
58 HENRIK RAMLAU-HANSEN
surplus should be distributed to those policyholders who contributed to the

profit. Moreover, the distribution should be equitable, and the actuarial
present value of the surplus generated by a policy should equal the actuarial
present value of the bonuses paid to that same policy.
The results are discussed within a general Markov chain framework where
an insurance policy is modelled as a time-inhomogeneous Markov chain, see
e.g. HOEM (1969, 1988). The paper is motivated by BERGER (1939), SVER-
DRUP (1969) and SIMONSEN (1970), who discussed some aspects of accumula-
tion and distribution of surplus. Moreover, RAMLAU-HANSEN (1988) analysed
the emergence of surplus using a general Markov chain and counting process
framework.
2. THE MARKOV CHAIN MODEL
We shall in the following consider life insurance policies which can be modelled
by time-inhomogeneous Markov chains with finite state spaces. Hence, let S(.)
denote the right-continuous sample path function of a time-inhomogeneous
Markov chain with finite state space /, and assume that the process starts in a
state 1 e / at time 0. The transition probabilities are denoted by
PUs, t) = P(S(t) =j\S(s) = i), i, jel, s < t, and the forces of transition
fiij(-) are defined by
4 ( 0 = l i m P» (/, / + h)/h, i, j el, i+j.

h—*0 +
The intensities are assumed to be integrable on compact intervals.

Consider an w-year insurance policy characterized by the following condi-
tions :
1. While the policy stays in state /, premiums are paid continuously to the
company at the rate 7t,(.), i.e. 7tj(t)dt is paid during [t, t + dt). Annuity
benefits received by the insured while in state i are denoted by &,(•)•
2. If the policy moves from state / to state j at time t, a lump sum benefit
Bjj(t) is paid to the insured immediately after time t.
3. When the policy expires at time n, the insured receives an amount 2?,(«) if
the policy is in state i at the maturity date.
The quantities nt(t), bt{t), By{t), and Bt(n) are all assumed to be non-
random. It should also be noted that we have restricted ourselves to continuous
payment of premiums and annuities, benefits tied to transitions between
different states and to maturity benefits. However, single premiums and other
types of non-random payments can be incorporated easily. Note also that we
have introduced different notation for premiums paid and annuity benefits
received because the two types of payments are affected differently by surplus
distribution. Moreover, we shall refer to the "standard" benefits (b^t), Btj{t),
Bj(n), i,jel, i ± j) as one unit of benefits, because one of the distribution
methods operates by increasing all benefits proportionally. Finally, expenses
are not included explicitly but can be regarded as separate benefits.
DISTRIBUTION OF SURPLUS IN LIFE INSURANCE 59
It is assumed that the company is making its valuations on the basis of a

constant force of interest 5 and a set of transition intensities //,}•(•)• The basis (3,
fiy, i,je I, i =£ j) is often called the valuation basis of the first order, and we
shall assume that the company is required to use this set of (conservative)
assumptions in determining reserves and premiums. However, we shall assume
that the actual force of interest is 3° (d° > d) and that the actual behaviour of
the Markov chain is governed by the intensities //"•(•). The elements (d°, /ifj,
i,je I, i 41 j) are often called the second order basis, and we shall assume that
surplus is distributed according to this set of (realistic) assumptions.
Given that the policy is in state / at time t, let Vt{t) denote the prospective
premium reserve corresponding to the valuation basis of the first order.
Moreover, let SPt{t) be the single premium or the actuarial present value of
one unit of future benefits, provided that the policy is in state i at time t. We
shall also assume that the equivalence principle is followed, i.e. Vx (0) = 0. The
reserve Vt(t) is given by
j
f
J,
f P9(t,u)[bj{u)-nj(u)]du
it
where the Py(s, t)'s are the transition probabilities corresponding to the
intensities //,;,•(•)• A similar expression holds for SPt(t); just substitute 0 for
7ij(u) in (2.1). It is well known, see e.g. HOEM (1969), that Vt(t) satisfies
Thiele's differential equation
(2.2) - V,(t) = 8 VM + n^-bM - X Hij(t)Rij(t),

dt j+i
where Ryit) = Vj(t) + By(t)- Vt(t) denotes the amount at risk associated with
a transition from state i to state j at time t. Similarly, SPt{t) satisfies
(2.3) —SPt{t) = dSPM-biit)

dt j
3. ACCUMULATION OF SURPLUS
Assume in this section that no bonuses are paid and that the company just pays
the promised benefits bf(t), By(t), and 2?,-(n) in return for the premiums nt{t).
The average surplus or profit realized over the term of the policy may then be
derived in the following way. Assume that the policy is in state / at time t and
that the amount Vt(t) has been reserved. Then during [t,t + dt) the actual
interest earned is 3°dt Vt{t), the premiums and the annuity benefits are nt{t) dt
and bj(t)dt, respectively, and the expected net loss due to transitions out of
state i is Z $j(t) dt Ryit). However, the reserve needed at time t+dt,
assuming the policy is still in state /, is Vt{t + dt), and hence the net profit
becomes
yt(t)dt = (l+d°dt) K,
This leads to
4(W0
j+ dt
and using (2.2) we get
(3.1) y,(r) = (S°-S) V,{t) + Z (MaW-rijit)) Ryit)
= AS V,(t) +
introducing AS — S° — S and Afiy(t) — fiy(O~/4(0- Thus, assuming that the

policy is in state i at time t, surplus accumulates at the rate yt(t), which,
according to (3.1), is the sum of the excess interest earnings and the profit or
loss associated with transitions out of state /. The actuarial present value at
time 0 of the total surplus accumulated over [0, t] during stays in the state / is
given by
(3.2) r,-(0= f e~s'sFDli(0ts)yl(s)ds,

Jo
and the present value of the total surplus accumulated over [0, t] is
(3.3)
It should also be noted that
(3.4) r(0 = Z f e-s>'P°u(P,s)[7ii(s)-bl(s)]ds

i Jo
Z Sf e-s>'Poll(0,s)fil(s)Ba(s)ds
' j+i i0
and
(3.5) r W> = I f
k + i Jo
-e"'P°u(P,s)[7li(S)-bl(s)]ds
Jo
Jo
s
- e~ °' ^i,(o,0^(0,
see e.g. RAMLAU-HANSEN (1988) formulas (4.1) and (4.10). Hence, F(t) may be
interpreted as the actuarial present value of the difference between the
premiums received and the benefits and reserves that have to be provided. The
gain /"",•(/) may be interpreted similarly.
For a broader discussion of surplus accumulation and in particular various
stochastic aspects, see RAMLAU-HANSEN (1988). However, note that in RAM-
LAU-HANSEN (1988) F(t) and Ft{t) are random variables and not actuarial
values.
4. DISTRIBUTION OF SURPLUS
4.1. Cash bonuses

It was shown in the previous section that the surplus accumulates at the rate
yt{t) in state i at time t. Hence, the surplus may be distributed by simply paying
the policy holder an annuity }>,(?) while the policy is in state i. These dividend
payments may then supplement annuity benefits or partly offset premiums
paid under the terms of the policy. The present value at time 0 of the total
bonuses paid during [0, /] is
(4.1) f
Jo
where Yj(s) = 1 if S(s) = i and 0 otherwise. Note that the amount C{t) is
random, but EC(t) = F{t). In practice, companies that pay cash bonuses do
not pay the continuous annuities yt(t), but they may distribute the surplus
through annual instalments or by other means, cf. Section 5.1.
The amount C(t) may also be interpreted as the present value of the amount
in a savings account attached to the insurance policy. During stays in state i,
the account is then credited continuously at the rate y,-(/). Some companies do
follow this procedure by deferring the payment of the cash bonus until the
policy becomes a claim or expires. If the policy becomes a claim or expires at,
say time /, then the amount exp(<50/) C(t) is paid in addition to the policy
benefits. If two or more lump sum payments are possible under the policy, the
surplus may be distributed through a series of payments.
It should be noted that the distribution of surplus through periodic payments
allows all policyholders to share in the profit.
4.2. Terminal bonuses

In this subsection we discuss a distribution method according to which the
surplus is distributed to the policyholders only when the policies expire. No
additional benefits are paid during the term of the policy, except at the
maturity date. Hence, terminal bonuses may be used to enhance the maturity
value of the policy.
It was shown in Section 3 that the actuarial present value of the total surplus
accumulated during stays in a state i is /",-(«) given by (3.2). Hence, if this profit
is to be distributed as a payment to those policyholders who are in state i at
time n, each should receive
(4.2) Ti(n) = ri(n)/[e-d°"Pou(O,n)].
One might also limit the payment of bonuses to those survivors who are in the
initial state at time n. Depending on the design of the policy, this practice may
favour those policyholders who have not made any claims under the policy. In
this situation, each of the survivors in state 1 should receive
(4.3) T(n) = r(n)/[e~sa"Pou(O,n)].
at time n.
However, it should be noted that by applying terminal bonuses only
survivors are rewarded, and those who have died do not get a share of the
profit, although they may actually have contributed to it. Hence, the method
resembles in a way a tontine scheme, and this may explain why terminal
bonuses are only used in connection with policies with a strong savings
element.
4.3. Increased benefits

In this section we assume that the surplus is used to increase the policy benefits.
This is one of the most common ways of distributing surplus in practice. We
shall assume that all benefits are increased proportionally so that the original
relationship between the benefits is preserved. Hence, the surplus is used as a
single premium to purchase additional units of benefits, cf. Section 2.
At issue, the net premium reserve is V{ (0) = 0 and the policy provides the
benefits bt(s), Btj{s), for s > 0, and #,(«). Let us now assume that the policy is in
state i at time t and that the policy entered this state at some time t,. Moreover,
assume that past surplus has been used to buy D(t) units of additional benefits
so that they are now promised to be bf(s) = bj(s) (1 +D(t)), Bfk{s) =
Bjk{s) (1 +D(t)), for s > t, and Bf(n) = Bj(n) (1 +D(t)). The rate of increase
of benefits at time u is denoted by d(u), i.e. D(t) — fI d(u)du.

d(u)du. It
It should
Jo
be noted that D{.) is actually a stochastic process since it is a function of the
sample path of the Markov chain. At time 0, D{t) is unknown because the
future course of the policy is unknown.
Taking the increased benefits into account, the policy reserve is now
(4.4) Vl*{t)=Vl{t) + D(t)SPi{t),
where both Vt(t) and SPt{t) are calculated using the first order valuation basis,
cf. (2.2)-(2.3). Hence, using arguments similar to the ones in Section 3, the
average surplus that emerges at time t is given by the rate
yf(t) = AS V?{t)
= AS Vt{t) + X
+ D(t)[AdSP,(t)
{
using (4.4). Thus,

(4.5) yT(t) =
if we introduce «:,-(/) = A5 SPt{t) + £ Anb (t) [SPj (t) + Bt]{t) - SP,(t)].
The surplus yf{t) is used to buy d{t) units of additional benefits at a cost of
SPj(t) per unit. Thus, we must have that
or
(4.6) D'(t) = d(t) = q,(t)+ D(t)r,(t),
where q,(t) = y,-(0/5'P,-(0 and rt(t) = K^/SP^t). Equation (4.6) is a linear
differential equation with solution
(4.7) D(t) = f q,(s)exp f r,., ( M ) Jw ds
»•/(*) * ,
which yields, in a closed form, an expression for the total increase of the
benefits due to the emerging surplus.
It should be noted that (4.7) holds only during the stay in state i. If the
policy at some later time t}- moves to state j then a similar formula holds with tj
and j substituted for t,- and i, respectively. Thus, the rate of increase of benefits
depends on the current state of the policy, but the policyholder should not
expect any sudden changes in the benefits because D(-) is a continuous
function.
It should also be noted that in this section additional benefits are granted as
the surplus is earned. In order to make this a prudent distribution method, it
requires that at any time the future safety margins are sufficient to safeguard
the company against any adverse experience. Moreover, since companies
normally cannot reduce bonuses once they have been declared, it also requires
surplus always to be positive, i.e. yf (t) has to be positive. If this is not the case,
distribution of surplus will have to be deferred, and the method above will have
to be modified.
If the original policy is a single premium policy, then Vt{t) = SPj(t),
K
i(0 = ?i(t)> a Q d 9,(0 = rt{t). In this case, it follows from (4.7) that
(4.8) 1 + D(t) = (1 + D(?,)) exp ( f r,(u)du\, t>tt.
Finally, we shall see that Vj*(t) satisfies a second order differential equation
although it was defined as a first order premium reserve, cf. (4.4). The reason is
that the benefits are adjusted continuously. According to (4.4),
- V,*(t) = — VM + D'WSPiW + Dit)- SP,(t),

dt dt dt
and using (2.2)-(2.3) and (4.6) we get after some simple arithmetic the
equation
— K,*(/) = <5° ^ ( 0 + 1 / ( 0 - 6 ? ( 0 -
dt j
5. EXAMPLES
To illustrate some of the results, we shall consider two examples: A single-

premium annuity policy and a disability policy. The first example focuses on
ways of distributing interest surplus, whereas the other example is a discussion
of surplus distribution in a three-state model. We have not included an example
of a typical endowment policy, because we feel that the two other examples are
more interesting.
5.1. An annuity policy

Let us consider a single-premium annuity policy where a benefit b is paid
continuously throughout the life of an individual (x). The first order premium
reserve is
= bdx+I, t>0,
using standard actuarial notation. We assume that the actual force of interest is
a constant 3° > 3 and that the interest earnings are the only source of surplus,
i.e. //>(.) = fi(.).
Then, according to (3.1), surplus is accumulated at the rate y(t) = A3 V(t),
and we may, therefore, pay the insured the adjusted benefit
(5.1) bl(t) = b + A3V(t).
Alternatively, (4.8) shows that the surplus may also be distributed by means of
the increased benefits
(5.2) b2[t) = exp[(3°-3)t]b.
It is interesting to note that (5.1) is typically a decreasing function of time/age,
whereas (5.2) is increasing exponentially. Thus, the two formulas represent two
completely different ways of distributing the same surplus.
In practice, however, it is not possible to adjust the benefits continuously as
it is assumed in (5.1) and (5.2). In Denmark, for instance, pensions are adjusted
only annually. Therefore, there is a need for more practical versions of (5.1)
and (5.2). If, for example, the total surplus accumulated during year t,
t = 0,1,..., has to be distributed through a level benefit b3(t) payable
continuously during year t, then b3 (t) has to be determined by
(5.3) V(t) = b3(t)d°x+,:T\ + v°px+lV(t+l), t= 0,l,...,

where the superscript " 0 " indicates that the values are based on S°. Hence,
&3(0) is the level benefit that is paid continuously during year 0, 63(1) is paid
during year 1 etc. It follows from (5.3) that the series of benefits b3(0),
b3(l),... serves the same purpose as the function bl(').
Similarly, the function b2(-) may be replaced by level annual benefits in the
following way. Assume that the benefit is a level amount b4(t) during year /.
Then b4(t+ 1) is determined by the equation
hit) [ e-sa(s-!)s_tPx+lAdax+sds+v»px+tb<{t)ax+t+x
= v°px+tb4(t+l)dx+l+i.
Thus, we see that the surplus accumulated over the year is used to grant an
increase of the benefit from b4(t) to b4(t+ 1).
Table 1 gives examples for an annuity of 10,000 issued to a male aged 60.
The valuation rate of interest is 4.5%, 3 = log (1.045), whereas the actual
interest rate is assumed to be 8%, i.e. 3° = log (1.08). Moreover, the mortality
is fi(t) = 0.0005+10 0 0 3 8 ( x + o ~ 4 1 2 which is the standard assumption used by
Danish life companies.
TABLE 1
COMPARISON OF VARIOUS WAYS OF DISTRIBUTING SURPLUS FOR AN ANNUITY OF
10,000 ISSUED TO A MALE AGED 6 0
Age
x+t MO b2(t) 63(O 64(0
60 13,885 10,000 13,835 10,000

61 13,784 10,335 13,734 10,350
62 13,682 10,681 13,632 10,713
63 13,580 11,039 13,529 11,089
64 13,477 11,409 13,426 11,479
65 13,373 11,791 13,322 11,884
70 12,853 13,902 12,803 14,141
75 12,345 16,391 12,297 16,861
80 11,869 19,326 11,825 20,161
The table highlights the difference between the payment schemes b3(t) and
b4(t). The calculations show that b3(t) is larger than b4(t) during the first
8 years after which bA{t) exceeds b3(t). The distribution method that leads to
b4(t) is widely used in Denmark, primarily because it provides some protection
against inflation. However, one might also argue that in years with low
inflation, many retirees are presumably prepared to forfeit inflation protection
in return for higher benefits while they are healthy and the quality of life is
higher. Thus, b3(t) should perhaps be recommended more widely than it has
been until now.
5.2. A disability policy

We shall in this section consider an n-year disability policy issued on an able
male aged x. The policy may be described by the three-state Markov model
depicted in Figure 1. It is assumed that the policy provides a continuous
o°(t) Disabled
U°(t)
FIGURE 1. The disability model.
annuity of 1 as long as the insured is disabled. Premiums are waived during

disability, and it is assumed that the premium payments cease after m = n — 5
years in order to avoid negative reserves close to maturity.
Danish companies assume in their valuations that the transition intensities
are given by
H(t) = v (f) = 0.0005+10 0038( * +() ~ 412
and
<T(0 = 0.0004+
The rate of interest is still assumed to be 4.5%, i.e. S = log (1.045). We shall
study surplus distribution under the somewhat more realistic assumptions that
the actual behaviour of the policy is governed by
H°(t) = 9lft(t),
°
and
where 6_ = (6l,92, 63) is given below. Moreover, the actual rate of interest is
also in this example 8%, i.e. 3° = log (1.08).
The premium n and the first order reserves are given by
where dx^ = f v'spax'ds, a^ = f vsspaxads, dux.^ =dx-\ = f vs sPxds,

Jo Jo Jo
/ r / v \
and where spx" = exp - I n{u) + a{u)du , spx — exp - I /n(u)du ,
\ Jo \ Jo /
and spx' = spx —sPx"• The corresponding amounts at risk are Rai(t) =
Vi{t)- Va(t), Rad{t) = - Va(t), and Rid(t) = - V^t). Here a denotes the state
able, i the state disabled (invalid), and d the state dead.
According to (3.1), surplus accumulates at the rates
ya{t) = A3 Va(
= (AS-Afi(t)-Aa(t)) Va(t) + Aa(t) Vt(t),
and
= {A5-Av{t))Vi{t)
during stays in the states able and disabled, respectively. Here Ad - S° — d,

An(t) = /t(t)-M°(t), Ao(t) = a(t)-a°(t), and. Jv(r) = v(/)-v°(r). Hence,
the present values at time 0 of the total accumulated surpluses are
(5.3) ra(n) = f"

Jo^ oCXP( 3 S) PX y {S)dS
' "
ai
(5.4) r,(n)= I exP(-S°S)sPx yi(s)ds,
and
(5.5) r{ri) = ra(n
cf. (3.2). Here, sp°xaa and sp°xai are second order values of spxa and spaj,
respectively. The corresponding possible terminal bonuses Ta(n), r,(«), and
T(n) are given by (4.2) and (4.3).
We have in Table 2 shown examples of (5.3)-(5.5) for policies with
x + n = 65 and x + m = 60. Moreover, it is assumed in these examples that
9\ = 0.7, 02 — 0.8, and #3 = 1 which are close to what currently is used by
many Danish companies. The figures illustrate clearly the size of the surplus
inherent in the policies. Take as an example the policy issued at age 30. Here
the actuarial present value of the total surplus is 0.144 compared with the total
value of the premium payments n aaxa^\ which equals 0.423. The surplus might
be distributed through trie terminal dividends given in Table 2. However, it is
hard to argue that only paying 2.13 and 5.12 to the lives that are able and
disabled at age 65 is an equitable way of distributing the profit. It is also
difficult to justify that large amounts should be paid to the disabled lives who
have already collected benefits under the terms of the policy.
Table 3 shows for the example x = 30 the possible benefits if the surplus is
used to continuously increase the benefits. We have shown the rates of surplus
accumulation y*(t) and yf{t), cf. (4.5), together with 1 +Da(t) and 1 +£>,(/),
respectively. Here 1 + Da(t) is the basic disability annuity that becomes payable
if disability occurs at time t. This quantity and y*(t) have been calculated
assuming that the policy has remained in the state able during [0, t). Similarly,
TABLE 2
EXAMPLES OF PRESENT VALUES OF ACCUMULATED SURPLUSES AND POSSIBLE TERMINAL BONUSES FOR
VARIOUS DISABILITY POLICIES WITH 0 = (0.7, 0.8, 1 )
Issue
age
1000 % ra(n) /» Ta(n) 7»
20 19.0 0.086 0.037 0.123 3.97 9.29 5.65

30 26.8 0.101 0.043 0.144 2.13 5.12 3.03
40 40.8 0.110 0.049 0.159 1.05 2.77 1.51
50 65.5 0.103 0.040 0.143 0.43 1.13 0.60
TABLE 3
RATES OF SURPLUS ACCUMULATION AND SIZE OF INCREASED BENEFITS.
AGE AT ISSUE X = 30 AND 6 = (0.7,0.8, 1)
Age
yiif) yf(t) l + Da(t) .•«o
x+t
30 0.002 0.560 1.00 1.00

40 0.012 0.654 1.14 1.39
50 0.029 0.655 1.51 1.93
60 0.045 0.381 2.68 2.69
61 0.044 0.324 2.97 2.78
62 0.041 0.258 3.38 2.87
63 0.036 0.183 4.02 2.97
64 0.027 0.098 5.38 3.07
64.5 0.019 0.050 7.17 3.12
65 0 0 oo 3.17
1 + D,(0 is the annuity payable at time t and yf (?) measures the rate of surplus
accumulation, provided that the insured became disabled just after time 0. It is
interesting to note that (4.7) leads to
f' f' \ / f'
- 1
= ^(•«)exp ri(u)du\ds = exp\
Jo Js I \ Jo
with qi(s) = yi(s)/SPi(s) = Ad-Av(t), SP,(t) = V,(t), and r,(u) = ?,-(«).
Hence, Dt{t) is in general easy to compute, and in the example in Table 3
Av(t) = 0, so 1+A(O = exp(ASt), cf. (5.2).
It is interesting to note that l+Da(t) and 1+ /),(?) increase at different
rates. In particular, the sharp increase in 1 +Da(t) close to maturity should be
noted. Actually, it is easily seen that l+Da(t)-> oo as t -> n. It may be
explained by the fact that close to maturity, the surplus is of the size O (h),
h = n-t, whereas the price of providing additional benefits is
^S'+c^Tl ~ O{h2). In practice, these excessive benefits should, of course, be
avoided, and it may be achieved by shifting to a system with cash or deferred
bonuses when the policy approaches maturity.
In Table 3, 1 + Dt(t) yields the annuity at time / if the disability occurred at
time 0. However, if disability occurs at some later time, say tt, then it follows
from (4.7) that the benefit at time t > tt is given by
f ri(u)du\
= {\+Da(?,)) (1 + A(0)/0 + A(
Thus, if for example disability occurs at age 40, then the initial annuity is 1.14,
which after 10 years of disability will have risen to (1.14) (1.93)/1.39 = 1.58. It
illustrates that the benefits while disabled depend on the duration of the
disability.
70 HENRIK. RAMLAU-HANSEN
TABLE 4
EXAMPLES OF DISABILITY ANNUITIES 1 + / ) , ( / ) IN THE SITUATIONS WHERE 9} = 1,2 AND 5.
AGE AT ISSUE X = 30 AND (0,, 82) = (0.7,0.8)
£f «3 = 1 03 = 2 ff
3 =5
30 1.00 1.00 1.00

40 1.39 1.42 1.52
50 1.93 2.07 2.53
60 2.69 3.18 5.27
65 3.17 4.11 9.01
TABLE 5
PRESENT VALUES OF ACCUMULATED SURPLUSES FOR DIFFERENT VALUES OF #.
A G E AT ISSUE X = 30
= (ex,e2,ei) ra(n) r,(«)

(0.7, 0.8, 1) 0.101 0.043 0.144
(0.7, 1, 1) 0.051 0.054 0.104
(0.7, 1, 2) 0.051 0.062 0.113
(0.7, 1, 5) 0.051 0.085 0.136
(0.7, 1, 10) 0.051 0.112 0.163
We have also shown in Table 4 the kind of disability annuities that can be
offered if it is further taken into account that disabled lives often have a much
higher mortality than able lives. We have shown examples of 1 +/),(?) in the
situations where 93 = 1,2, and 5. Otherwise, the assumptions are the same as
in Table 3. It is clear that substantial mortality gains on the disabled lives
might be used to increase the disability benefits further.
However, in some cases mortality gains on disabled lives would rather be
used to offset unsatisfactory disability experience among able lives. In this way
all get a share of the " favourable" mortality among disabled lives. To give an
impression of to what extent an unfavourable value of 62 can be offset by a
favourable value of #3, we have shown in Table 5 some examples where
#2 = 0.8 and 1, and where 03 = 1,2,5, and 10. Hence, taking
0. = (#i > 02. 03) = (0-7, 0.8, 1) as our basis, it is seen that even 03 = 5 is not
sufficient to eliminate the overall effect of 62 — 1, whereas 03 = 10 more than
compensates for the effect of 62 = 1 -
REFERENCES
BERGER, A. (1939) Mathematik der Lebensversicherung, Springer, Wien.
HOEM, J. M. (1969) Markov chain models in life insurance. Blatter der Deutschen Gesellschaft fiir
Versicherungsmathematik IX, 97—107.
HOEM, J.M. (1988) The versatility of the Markov chain as a tool in the mathematics of life
insurance. Transactions of the 23rd international congress of actuaries, vol. R, 171-202.
RAMLAU-HANSEN, H. (1988) The emergence of profit in life insurance. Insurance: Mathematics and
Economics 7, 225-236.
SIMONSEN, W. (1970) Forsikringsmatematik, hefte III. Kobenhavns Universitets Fond til tilveje-
bringelse af lsremidler.
SVERDRUP, E. (1969) Noen forsikringsmatematiske emner. Statistical memoirs No. 1, Institute of
Mathematics, University of Oslo.
HENRIK RAMLAU-HANSEN
Baltica Insurance Company Ltd., Klausdalsbrovej 601, DK-2750 Ballerup,
Denmark.

Baltica Insurance Company LTD., Ballerup, Denmark: by Henrik Ramlau-Hansen

Uploaded by

Copyright:

Available Formats

Baltica Insurance Company LTD., Ballerup, Denmark: by Henrik Ramlau-Hansen

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Baltica Insurance Company LTD., Ballerup, Denmark: by Henrik Ramlau-Hansen

Uploaded by

Copyright:

Available Formats

DISTRIBUTION OF SURPLUS IN LIFE INSURANCE

This paper discusses distribution of surplus in life insurance within a general

Distribution of surplus; bonus; with profits annuity policy; with profits

surplus should be distributed to those policyholders who contributed to the

2. THE MARKOV CHAIN MODEL

4 ( 0 = l i m P» (/, / + h)/h, i, j el, i+j.

The intensities are assumed to be integrable on compact intervals.

It is assumed that the company is making its valuations on the basis of a

(2.2) - V,(t) = 8 VM + n^-bM - X Hij(t)Rij(t),

(2.3) —SPt{t) = dSPM-biit)

state i is Z $j(t) dt Ryit). However, the reserve needed at time t+dt,

(3.1) y,(r) = (S°-S) V,{t) + Z (MaW-rijit)) Ryit)

introducing AS — S° — S and Afiy(t) — fiy(O~/4(0- Thus, assuming that the

(3.2) r,-(0= f e~s'sFDli(0ts)yl(s)ds,

It should also be noted that

(3.4) r(0 = Z f e-s>'P°u(P,s)[7ii(s)-bl(s)]ds

4.1. Cash bonuses

4.2. Terminal bonuses

4.3. Increased benefits

of benefits at time u is denoted by d(u), i.e. D(t) — fI d(u)du.

using (4.4). Thus,

if we introduce «:,-(/) = A5 SPt{t) + £ Anb (t) [SPj (t) + Bt]{t) - SP,(t)].

(4.7) D(t) = f q,(s)exp f r,., ( M ) Jw ds

(4.8) 1 + D(t) = (1 + D(?,)) exp ( f r,(u)du\, t>tt.

- V,*(t) = — VM + D'WSPiW + Dit)- SP,(t),

To illustrate some of the results, we shall consider two examples: A single-

5.1. An annuity policy

(5.3) V(t) = b3(t)d°x+,:T\ + v°px+lV(t+l), t= 0,l,...,

60 13,885 10,000 13,835 10,000

5.2. A disability policy

FIGURE 1. The disability model.

annuity of 1 as long as the insured is disabled. Premiums are waived during

where dx^ = f v'spax'ds, a^ = f vsspaxads, dux.^ =dx-\ = f vs sPxds,

during stays in the states able and disabled, respectively. Here Ad - S° — d,

(5.3) ra(n) = f"

20 19.0 0.086 0.037 0.123 3.97 9.29 5.65

30 0.002 0.560 1.00 1.00

30 1.00 1.00 1.00

= (ex,e2,ei) ra(n) r,(«)

You might also like