Evaluating Life Insurance Demand in Iran
Insurance is one of the key tools in modern life which has a great role in reducing risks and providing financial and mental security and finally it can broaden for countries economic development. The main objective of this article is estimate the demanding life insurance and analysis effective factors on it which was investigated in the period 1358–1388. According to the theoretical principals, the model of demanding life insurance in Iran is a factor of real per capita income, savings rate, expects inflation, percent literacy and the rate of interest. In order to providing the model of economic evaluation method, vector auto-regressive (var) has been used. Based on the results, the variables real per capita income, interest rates, savings rates have positive and significant relationship with the demanding life insurance. The variable of expected inflation rate is negative and significant relationship with life insurance demand and there isn’t a significant between the variable of percent literacy and the life insurance demand.
Content
Int. j. econ. manag. soc. sci., Vol(3), No (10), October, 2014. pp. 574-581
TI Journals
International Journal of Economy, Management and Social Sciences
www.tijournals.com
ISSN:
2306-7276
Copyright © 2014. All rights reserved for TI Journals.
Evaluating Life Insurance Demand in Iran
Samaneh Derakhshideh *
Ph.D. Student in Economic, Science and Research branch, Islamic Azad University, Kerman, Iran.
Sayyed Abdolmajid Jalaee
Associate Professor in Economic, Shahid Bahonar University of Kerman-Iran.
*Corresponding author: [email protected]
Keywords
Abstract
life insurance
VAR model
savings rates
interest rates
Per capita income
Insurance is one of the key tools in modern life which has a great role in reducing risks and providing
financial and mental security and finally it can broaden for countries economic development. The main
objective of this article is estimate the demanding life insurance and analysis effective factors on it which
was investigated in the period 1358–1388. According to the theoretical principals, the model of demanding
life insurance in Iran is a factor of real per capita income, savings rate, expects inflation, percent literacy and
the rate of interest. In order to providing the model of economic evaluation method, vector auto-regressive
(var) has been used. Based on the results, the variables real per capita income, interest rates, savings rates
have positive and significant relationship with the demanding life insurance. The variable of expected
inflation rate is negative and significant relationship with life insurance demand and there isn't a significant
between the variable of percent literacy and the life insurance demand.
(This essay has been excerpted from parts of
Samaneh Derakhshideh's PHD thesis.)
1.
Introduction
Generally speaking and regarding to the close relationship between developing of life insurance and the rate of wale fare and economic power of
the countries, there are lots of different kinds of living insurances as a tool for providing financial in our country and they didn't even improved,
such that it can be claimed that growing and developing didn't improved compared with European countries. Considering the great role of life
insurance in savings and investment and also its role in economic growth, it's easy understand the importance of this field in the economy of
different countries.
In recent years, life insurance had a growing share of development in the insurances of country and it has been increased from 6.6% in 2005 to
8.3 percent in 2010, but when it's been compared with developed countries, it shows a weak and most important the growth of insurance industry
in our country. The coefficient of insurance the important factors in developing and growing of this has increased from %1.27 in 2005 to
%1.52010. In spite of that we still feel the need of growing this field of insurance in the country. According to the recent changes in insurance
industry in our country, and in order to removing the absolutes and increasing insurers capacity, it seems that a good condition have been
provided for expanding insurance industry and especially life insurance and investing in this case. Insurance is one of the key tools in today's
world which has a great role in reducing risk and providing financial and mental security of people .It is also makes needed beds for economic
development for the country. Present study has evaluated the affective factors on the demand for life insurance in Iran during 1979 to 2009 by
vector auto-regressive model (VAR).Effective factors on the demand for life insurance in Iran are functions of real per capita income, savings
rates, inflation rates and rate of literacy. In this article at first, we consider the literature of the study in summery then we evaluated and conclude
the article.
2.
A review of the research literature
2.1 Background of the research
During recent decades contemporaneously with the expanding of life insurance, lots of studies have been done to investigate the nature of
demanding and supplying life insurance, affective factors in demanding and other cases and related techniques to this field of insurance. The first
academic and theoretical study which is one person behaviors who is not taking risk and he is trying to manage the risk of death. This is named
Yari article which is very famous such that theoretical works after him all affected by his method and his literature. In the study of Pazhuyan and
Poorpartoy (2002) using static data's gathered of1966to 2001yearsthepattern insurance demand has been estimated and the amount of that has
been developed till the end of 2003. The aim of this study is to evaluate the effects of earnings, expectation inflation, guaranteeing responsibility
and the education on demanding different kinds of life insurance in Iran.
Data's related to the variable of guaranteeing responsibility is gained by dividing the unemployed population to employed population in
considered year and for those years which in there are no static data's we have used: Rn = Rp (1 + r)n. In order to gain the amount of life
insurance demand, like many other studies, received money for insuring have been used, and to removing the effects of inflation, the price index
of consumption goods and services in urban areas CPI, the year 1990have been used as the base year. Headed households income is determined
by per capita national income. The effect of inflation on per capita national income is also removed by the CPI. Probability of death of headed
households is gained by dividing the number of deaths by the total population. The expected inflation is calculated by Pe = 0.7 Pt-1 +0.3 Pt-2 where
Pt-1 is inflation in previous term and Pt-2inflation in the last two term ago.
The final model of life insurance demand is as follows:
575
Evaluating Life Insurance Demand in Iran
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
= 78.49 + 0.41
+ 0.23
+ 1.85
+ 1.31
+ 7.78
− 0.17
−
(5.15) (3.2) (−2.46) (7.05) (5.54) (4) (−2.49)
2 = 0.927 . = 1.98 = 62.1
57
Where REV is the received premium per capita real of insurance life, NI is per capita real income (the fixed prices in 1999), POP is the
population of country, INF is the expected inflation, LITE is the percent literacy, DEP is time dependent and DU57 is dummy variable that
impact of the 1357 revolution on the insurance condition shows. Symbol L which is in front of all variables is natural logarithm. Presented
model is the demand pattern. In this model; despite of the equality between coefficients sign and theoretical expectations, the probability of
headed household's death wasn't effective. This can be the reason of religious and cultural beliefs in Iran. In this research, the demand function is
estimated using OLS. The results show a long-term stable equilibrium relationship between the variables in life insurance demand model.
Coefficient of LNI income shows income elasticity for life insurance demand which is estimated equal 41.0. Coefficient of the expected inflation
mentions that life insurance demand elasticity rather inflation is low. LDEP coefficient for guaranteeing responsibility is 1.58 which show a low
elasticity of demand compared to all studies have been done in other countries. LITE coefficient shows the percent literacy impact on the
demand for life insurance and this coefficient is higher than all coefficients. This can be resulting of increase their risk with higher education.
Another example study had been done conducted by Ebrahim Kardgar1995 about evaluating effective factors in the life insurance demand from
statistic information and national accountings and static calendars of 1986 to 1995. Considered variables for coefficient of life insurance demand
are income breadwinner, probability of breadwinner's death, headed household's responsibility and education and expectational inflation.
Received premium per capita of life insurance is considered as an index for demand these insurances and national income per capita for the
breadwinner's income. Probability of breadwinner's death is gained by dividing who are dead by total population and the population of dead
people. Related information to guaranteeing responsibility is gained by dividing the population under 20 ages by the population over 20 to 64
ages. For expectational inflation here we have used regression of logarithmic pattern of consumption goods and services price index in urban
areas (CPI).
Life insurance demand function is as follows:
1 = −9.69 + 0.635
+ 2.28
_2.29
+ 0.523
−
(−7.47) (3.35) (2.5) (−2.48) (4.26)
2 = 0.86 . = 1.98 = 37.26
PIN1R Where PIN1R is received premium insurance real per capita of life insurance which is gained by dividing received premium insurance
real per capita of life insurance by consumption goods and services price index in urban areas (CPI). INRN is net income per capita in constant
prices in 1999, RB is the literacy rate, PH is the expected inflation and DUM is a dummy variable that it shows life insurance act and the events
of government employee in 1997.For evaluating gained information similar to previous studies, is used t-student statistics for testing the
significance of individual coefficients and F statistics is used for testing the significance of all coefficients. Individual coefficients are significant
at the 5% level. According to the LINRN coefficient, income elasticity of life insurance demand is 0.635.Thus, the income elasticity of life
insurance demand is less than one and it can be concluded that the sensitivity of life insurance demand in contrast to family income can be
increased by using advertising and various selling methods. LRB coefficient shows that elasticity of percent literacy rate for life insurance
demand is larger than unity. The reason for this is the increase in risk aversion and the result is an increasing percent literacy .Because risk
averse individuals shall have demand more for life insurance. LPH coefficient shows that expected inflation elasticity of the demand for life
insurance is negative. Therefore, increasing expectational inflation led to reducing the demand for life insurance. In this study, the influence of
guaranteeing responsibility and probability of headed household's death variables is rejected at the 5% level and remove these two variables have
not a significant impact on the other variables. Its effect relays on issues of faith and culture of people in society.
2.2 Theoreticalprincipals of insurance demand
Since, insurance usual moves in the way returning lost peace which is caused by non-confidence, so in order to argue about the performance of
insurance function in theoretical point of view, we need a framework that is accepted the principle of non-confidence. Model individual demand
for insurance is based on maximizing the expected utility. There is no problem in safe mode, because the level of expected utility is equal to
assumed wealth. I.e. If the W * is person definitive wealth, we have: E [U (W *)] = U (W *). Thus, we can be choosing between the confidence
and non-confidence results. In non-confidence condition two modes can happen. Because of the governing non-confidence condition, the
meaning of utility in the way that there is in complete confidence mode would be insufficient. Since it is impossible to mention all possible
conditions. Thus, in the analysis of consumer behavior in non-confidence condition we can define the "expected utility function", with primary
wealth (W) which is with probability (P), facing with financial risk Rial (L) and L <W. Expected utility function from equation (1) is obtained.
=
(
− ) + (1 − )
(1)
In this case, the person has two choices: 1) buying insurance contract with an insurance premium of d and when a happening occurred receives
the compensations. 2) Doesn't buy the insurance contract and in the event of damage, person pays for all compensations. In order to simplify the
discussion, we assume that insurance contract completely covers all damages and there is only this contract or in other words insurance demand
is an "all or nothing" demand. If person chose the first mode (buying contrast), his expected utility function should be defined as fallow:
=
( − ) + (1 − ) ( − ) = ( − )
(2)
Where U1 is the utility of the insured person, and d insurance premium is paid. Maximum insurance that a person is ready to pay that is d *
which is obtained from equation (3).
Samaneh Derakhshideh *, Sayyed Abdolmajid Jalaee
576
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
( )( ( −
∗) =
( − ) + (1 − )
(3)
According this equation, result utility of complete insurance or the utility for removing non-confidence in wealth is equal to the weighted
average of the utility from total wealth and the utility from net wealth after one financial loss. Therefore, we can concluded that the expected
utility of the insured person is less than uninsured individual utility that nothing bad happened, and also it is more than uninsured individual
utility that something bad happened. So we have:
( ) < ( − ∗) < ( − ) In this relation, L is the maximum premium that a
utility maximizing expected for completely covers insurance is willing to pay which increases as the increase of risk probability and the volume
of risk or Or in other words, the demand for insurance increases with increased risk probability and large financial losses (Maqary, Fahimi, 155,
1993).
2.3 Theoretical principals of life insurance demand
Breadwinner demand for life insurance depends on the number of family members. Lewis has considered this relation by expanding theorical
construction of Yari life insurance and also by considering other members of family's preferences. In this mode, the life insurance is demanded
by members who are under dependent person that bread winner is facing to income non-confidence lifetime. His demand for life insurance based
on the lifetime of breadwinner of the family, is based on the life cycle model in which income is non-confidence (Frank D Lewis 1989).Most of
new theorical studies on the life insurance demand have chosen Yari studies as a starting point. In the meaning of life cycle model with nonconfidence lifetime, Yari shows that a person increases his expected utility by purchasing life insurance and receiving annual. Lewis method is
distinctive because he considers the life insurance demand from heir's point of view. In order word, life insurance is demanded for maximizing
expectational utility of heirs. In Yari's model structure one consumer buys life insurance to increase the expected utility of own lifetime.
E[U(T)] = ∫ α(t)g[c(t)]dt − β(t)Ψ[s(t)]
(4)
It is assumed, T is consumer's lifetime which is a random variable. Ψ(s (t)) is instantaneous utility of inherits, g[c (t)] is instantaneous utility of
consumption, and α(t) and β(t) are the discount factor (adjusted). When consumers get married or have children then β(t) increases significantly.
So these events can more explain the changes in ownership (purchase) of life insurance. According to the equation the change in ownership life
insurance is more depended on out moving of consumer's utility function. Lewis with development of the Yari model has gained movement in
consumer function by children and wives preferences, at least in part, as an endogenous event. Therefore, further analysis is part, in order to we
aren't looking why increase the number of people dependent effect on the utility function .Lewis also assumes that the utility function of each
member of family is separable. This assumption allows us to discussion from the insured person's point of view to wife and children who are
personal heirs. Therefore Lewis explains the more to demand by children for life insurance. A supervisor provides income as exogenous for their
children and this income is allocated so that the expected utility of his children is maximized. Children also their utility, taking into account the
limitations of exogenous income transfer from father to max. Maximizing utility by children can involve buying insurance for family supervisor.
Because children are faced with non-confidence income that comes from lifetime non-confidence, this kind of performance about life insurance
is specific seem. In other words, children may rarely consider these kinds of insurances for themselves. But this is suitable because parents
paying instead of their children can be considered as paying for life insurance and this can be considered as other expenses such as clothes and
other necessities for their children, and because utility children depends on the kind of costs, so we can be analyzed life insurance demand from
function of children. Children buy life insurance because they are not sure about how long their father is alive and so non-confidence income is
faced them. They stay at home till the age of a. until that time they receive determined converting payment each year but when their father dies,
they don't receive any other such kind payment except for their share of inheritance. Before the age of a, children are not allowed to receive a
loan against their probable future income. Although they are allowed to save money, the pattern of converting payments from father are
considered by the way that in fact children can't have any savings till they are at home. The d-age children, the expected utility with respect to its
costs in connection with life insurance premiums, di, to maximize (Frank D Lewis, 543, 1989).
If the father stays alive, children consume at the amount of ti-di in which ti is the earned income from converting payment. If the father dies, the
children earn equal ti + bi-di in which ti is the price of living insurance paper, and bi is the share of receiving inheritance. This question can be
written as follows:
maxEUi = (1 − pi) [ui(ti − di) + EUi + 1] + pi[ui( i + bi − di)]
(5)
In this relationship: EUk is the expected utility from age of k to a, pk is the probable of father's death at age of k for children, uk(0) is future
utility at age of k["uk (0) <0, 'uk> 0] and UK(0 ) is the utility from the age of K to a with the assume of one optimistic consuming pattern. The
relationship between the price of insurance paper and the insurance premium is
=
(insurance premium is a percentage of the price of
insurance paper). L is the factor of overhead. By emphasis on above assumes, under these optimal conditions occur:
U´(ti − di) =
(
)
, U´(∗ i + bi −∗ di) (6)
In the above equation, factors star show optimizing value. In order to simplifying the analysis, we use the following relations which is almost
true when mortality rate for father is low.
( −
∗) =
(
−
∗)
(7)
Where TK and DK show the present value of transfer payments and life insurance premiums from age of k to a, and this is when father is alive.
By replacing and given a utility function with constant elasticity, we have:
577
Evaluating Life Insurance Demand in Iran
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
∗ i + bi −∗ dj =
(
)
1. α (Ti −∗ Di)
(8)
Where a (negative) is elasticity of marginal utility of consumption or Arrow Pratt relative risk aversion is. Finally, the replacement and limit of
their children to life insurance assets non-negative, we have:
(1 − Lpi)f ∗ = max [
Where
=(
−
(
c∗ − b
)
(9)
) the present value in consuming flows from age k to a when father is alive. This equation has a fairly simple
interpretation. Assume that the children don't receive inheritance (i.e. bi = 0). So, if probability of pi death is low then we have:
∗
=( )
.
c ∗ .In
this case, assuming that father will be alive to till the age of a, then the value of the life insurance paper in above case simply is a part of present
value of children consuming. This is related the reverse with the overhead factor (L) and directly with the child's degree of risk aversion (a). The
case of spouse is like children. It is assumed that spouse is alive till the age of T in which she should leave the share of inheritance B. The first
order condition in the age i for maximizing expected utility is just like equation (10).
(
(
)
)
(
−
∗
∗
−
)=
((
)
)
+
∗
−
∗
(
−
)
(10)
Where vk is future utility for widow at the age of k, VK is the utility from ages K to T with assume of an optimal program, Y is spouse's income,
R is the discount rate, KKis the present spouse's value of capital stock at the time of husband's death in k age. By following the same procedure
which taken into account for life insurance demand for children, spouse's demand is as follows:
(1 − lpi)f ∗ = max [
(
)
c ∗ − ki + (
)
]
(11)
Where C* is value of wife's consuming from ages K to T, if the husband is alive till the age of T. Total issued insurance based on husband's
lifetime is equal to the total buying which are done by wife and each one of children with the same rate of not taking risk, and also regarding this
fact that non-negative limiting factor in life insurance assets are related to all the member or none of them, then we can mix above equation to
gain the total life insurance assets.
(1 − lpi)f ∗ = max [
(
)
W − TC
(12)
In this equation, F is the face value of life insurance related to his father's lifetime. TC is present value of each child consuming from current
period to the age of K, assuming that they stay alive. W is family wealth without accountings wife's share of inheritance (Frank D.lewis.1989).
Above equation is mental demand which explains the clear accounting that many families do during their shopping. Therefore, the results of
discussion taken from above equation are as these that the demand for insurance life with the probable of breadwinner's death, present value of
family consuming and families not taking risk have positively and negatively associated with household wealth and overhead costs.
3.
Estimation model
In this model, insurance demand in Iran is a function of national income, amount of saving, the inflation rate, rate of interest and the rate of
literacy, study period was for 1979 to 2009. Sample volume in the study was the total number of insurance papers in Iran between years 1979 to
2009, regarding this fact that life insurance demand with accounting received insurance expenses in national income of life insurance is
evaluated.
LREV = β1NI + β2 RS + β3 INF + β4 LITE + β5 RB + u t
(13)
Where LREV is real per capita premium life insurance, NI is real per capita income, RS is rates of saving, INF is expectational inflation, LITE is
percent literacy and RB is rate of interest.
Study assumptions based on economical theories include:
1.
By increasing in real per capita income, life insurance demand increases. (
> 0)
2.
By increasing the rate of inflation, demand for life insurance reduces. (
3.
There is a positive relationship between literacy rates and life insurance demand.(
4.
There is a positive relationship between rates of saving and life insurance demand.(
5.
There is a positive relationship between interest rates and life insurance demand.(
< 0)
> 0)
> 0)
> 0)
Because of the lack of the Stats of the interest rates in market or interest rate payable by some limited partnership companies has used from
opportunity cost of money, with this definition that opportunity cost of money is an annual maximum benefit of Short-term which are paid by
commercial banks. This information is gathered from central bank of Iran. Independent variables that are used in this study, are used in many
Samaneh Derakhshideh *, Sayyed Abdolmajid Jalaee
578
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
Iranian and foreign studies. The variable income has a positive relationship in lots of studies such Mehrara and Rajabian (2005), Ebrahim
Kardgar (1996), Firooze Azizi (2004), Jalali Lavasani (2004). Relationship between savings rate with life insurance demandhas been almost
positive. The reason is that income has had positively related to savings in economic theory. Inflation in Pazhuyan and Mirpartoy studies (2000),
Ebrahim Kardgar (1996), Jalali Lavasani (2004), had a negative effect on the insurance demand and it had a positive in Firooze Azizi's studies
(2004). The variable of education also had a positive relationship in studies, like studies Pazhuyan and Mirpartoy (2000), Mehrara and Rajabian
(2005), Ebrahim Kardgar (1996).
To investigate the model at first, we evaluate the stationary test by two tests Dickey Fuller (DF) and Augmented Dickey Fuller (ADF). In the
appendix Table 1, unit root test results Augmented Dickey - Fuller shown.
result
non-stationary
non-stationary
stationary
stationary
non-stationary
non-stationary
Significance
level
0.05
0.05
0.05
0.05
0.05
0.05
Table 1. Unit root test results Augmented Dickey- Fuller
The test statistic
MacKinnon's Critical Value
definition
with trend
without trend
with trend
without trend
0.41
2.07
-3.56
-2.96
real per capita premium life insurance
0.01
-1.04
-3.56
-2.96
real per capita income
-3.19
-3.41
-3.61
-2.97
savings rate
-3.92
-3.98
-3.57
-2.96
expects inflation
-0.47
-2.24
-3.57
-2.96
percent literacy
-1.02
-0.84
-3.57
-2.96
rate of interest
variable
LREV
LNI
RS
INF
LITE
RB
It has to be said that the expected inflation rate from the real inflation rate is calculated as follows:
= 0.7
+ 0.3
It can be seen
that except for variables of the savings rate and inflation rates of economy, other variables are non-stationary in level. Now, the variables are not
stationary, we are taking first-order differencing to determine whether these variables by once difference will be stationary or not. First-order
difference variables are defined as follows:
=
−
=
−
IT =
−
REV = REV − REV
RS =
RS − RS
In appendix table 2, unit root test results Augmented Dickey – Fuller for first-order differencing shown.
result
stationary
stationary
stationary
stationary
Significance
level
0.05
0.05
0.05
0.05
Table 2. Unit root test results Augmented Dickey – Fuller
The test statistic
MacKinnon's Critical Value
definition
with trend
without trend
with trend
without trend
-5.15
-4.21
-3.57
-2.96
real per capita premium life insurance
-3.58
-3.23
-3.57
-2.96
real per capita income
-5.61
-4.46
-3.57
-2.96
percent literacy
-4.87
-4.85
-3.57
-2.96
rate of interest
variable
LREV
LNI
LITE
RB
In Table 3 indicate that the variables remaining the by once difference were stationary. Therefore, briefly order cointegration variables as
summarized in Table 3.
Table 3. Order of integration
Definition
Variable
I(1)
real per capita premium life insurance
LREV
I(1)
I(0)
I(0)
I(1)
I(1)
real per capita income
savings rate
expects inflation
percent literacy
rate of interest
NI
RS
INF
LITE
RB
Order of integration
In order to Volume observations less than 100, to determine the degree from Schwarz-Bayesian, Hannan-Quinn and Akaike criteria which
consider the least lag and prevent reduce the degrees of freedom, are used. Appendix Table 4 shows the cases.
Table 4. Order determination of var
HQ
SC
AIC
FPE
LR
LogL
Lag
21.64
11.54
11.47*
21.83
12.90*
14
21.55
10.92
10.32*
92.31
0.002
0.001*
NA
288.32*
49.36
-306.5
-116.47
-71.74
0
1
2
579
Evaluating Life Insurance Demand in Iran
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
According to the output software, we use Schwartz-Bayesian criterion and lags of one, and to determine the number of vector integration have
used from the trace test and the maximum eigenvalue. Trace test and the maximum eigenvalue test results are shown in Appendix table 5 and 6.
According to tables 5 and 6, it can be seen that there is one optimal vector.
Prob.**
Table 5. Effect test for determining the number of convergence vector
Critical
Max-Eigen
Hypothesized
Eigenvalue
Value (0.05)
Statistic
No. of CE(s)
0.000
0.000
0.08
0.41
0.40
0.66
95.75
69.81
47.85
29.79
15.49
3.84
137.02
89.87
45.40
20.14
8.54
0.19
0.80
0.78
0.58
0.32
0.25
0.0006
None
At most 1*
At most 2
At most 3
At most 4
At most 5
Table 6. The maximum eigenvalue test to determine the number of convergence vector
Critical
Max-Eigen
Hypothesized
Eigenvalue
Value (0.05)
Statistic
No. of CE(s)
Prob.**
0.0006
0.0005
0.096
0.588
0.343
0.661
40.07
33.87
27.58
21.13
14.26
3.84
47.14
44.46
25.26
11.59
8.35
0.19
0.80
0.78
0.58
0.32
0.25
0.0006
None
At most 1*
At most 2
At most 3
At most 4
At most 5
At first, we consider all variables in the model, and then we remove those variables that are not significant and estimate the model. Thus, the
convergence of the vector is determined by table 7: Significant is at 95% confidence level.
Table 7. Model estimation
LITE
RB
INF
RS
LNIS
LREV
0.004
0.954143
0.05
2.363343
-0.01
-5.8069
0.05
6.131461
1.47
3.418091
coefficient
T test statistic
R2=0.96
Finally, the model is estimated as follows:
LREV= 1.47 NI + 0.05 RS -0.01 INF + 0.004 LITE + 0.05 RB +ut
t
(3.418091) (6.131461) (-5.8069) (0.954143) (2.363343)
As we can see from the results of the estimated model, all variables except percent literacy are significant, thus by remove this variable again
estimate the model which its results are presented in table 8 with a 95% confidence level.
Table 8. Model estimation
INF
RB
RS
LNI
LREV
-0.02
-5.6
0.06
6,85
0.06
5.05
1.86
5.31
coefficient
T test statistic
LREV= 1.86NI +0.06RS -0.02 INF +0.06 RB +ut
t (5.31) (5.05)
(-5.6)
(6,85)
The research results show that the per capita income variable has a positive and significant effect on real per capita life insurance premium. It
means that by increasing in per capita income will increase premiums received. In considering this fact that these variables are presented as
logarithmic, its coefficient is represents elasticity of the dependent variable relative to this variable. So, it should be said that by increasing even
one percent in per capita income, life insurance demand increases by 1.86 percent and knowing this that this, this coefficient is bigger than one
,so we can say that life insurance shows very sensitive to this variable. Interest rate and the savings rate coefficients are statistically significant
and have positive effect on insurance premiums. The significant point here is that the coefficient of the two variables is equal. So these two
variables have equal impact on dependent variable. Also, regarding the positive and significant coefficient of the literacy rate which is equal
0.021, it can be stated that increase literacy will also increase premiums received. Rate of inflation has negative and significant impact on
insurance premiums. As shown by the results among independent variables, real per capita income has the greatest effect on insurance
premium.At next step, to evaluate the variables effects on real per capita premium of life insurance use variance analyzing method. To do so,
using the output of the software is seen that during 10 periods, reaction the variations dependent variable comparing variations independent
variable is some percent: the result of variance analysis is come at fallowing table:
Samaneh Derakhshideh *, Sayyed Abdolmajid Jalaee
580
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
Table 9. Variance analysis
As result show, effect of income, expected inflation rate and interest rate variables increases gradually, but the effect of the saving rate from
fourth period and the percentage of literacy from sixth period are decreasing. At the next stage, we evaluated the impulse response function
graphs. Since the interpretation of the single coefficients in estimating model of VAR is often difficult, in practice, usual the response function
(IRF) is estimated. In the VAR, IRF determines response dependent variable to shocks on the error terms. Therefore, evaluate the impulse
response functions of the per capita premium relative to each of the independent variables. Impulse response functions are shown in appendix
9.According to results taken from impulse response functions see that the impulse response functions of the per capita premium compared with
per capita income during 2 periods was fixed, and after than increases . Also sudden increase in the rate of saving variable causes the decrease in
dependent variable in 3 periods, and its negative effects increases gradually. And after 3rd period see that response was so small and it finishes at
next period. At first, the effect of the expected inflation rate increases with a regular tone, such that has the process increasing and decreasing
and after eighth periods, it starts to decrease to a mild tone. Sudden increase in the literacy rate at first causes reduce in insurance premium, but
after fifth period the amount of that reduce and finally in the period of 10 years, it vanished completely. About the rate of interest, as see the
effect of one increases dependent variable in 2 periods which is minor in primary periods, but in other periods causes reducing this variable. And
during 10 years, it doesn't vanish.
Chart 1. Impulse response function of real life
insurance premium per capita in relation to savings rate.
Chart 3. Impulse response function of real life
insurance premium per capita in relation to expects inflation
Chart 2. Impulse response function of real life
insurance premium per capita in relation to real per capita income.
Chart 4. Impulse response function of real life
insurance premium per capita in relation to percent literacy
Chart 5. Impulse response function of real life
insurance premium per capita in relation to rate of interest
581
Evaluating Life Insurance Demand in Iran
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
4. Conclusions
In this study, was estimated the appropriate model for life insurance .According to the model results, per capita income variable has positive and
significant effect on insurance premiums received and with increasing per capita income and improving purchasing power, and finally allocate a
greater share of income to the life insurance premium also increases According to negative coefficient of expected inflation during the period
under study, the expected inflation has a negative effect on premium. The expectation that inflation in future will be more motivated people for
life insurance decreases. Because people will think that the price increase in future will be causing decrease the purchasing power and the
becoming worth less of their money. Therefore less willing for savings will have and also the demand for insurance will reduce. Also, the
savings rate coefficient has statistically significant and positive effect on insurance premiums, i.e. by increasing savings rate increases premium
.Because of interest rates has Positive relationship to the premium received Can say when interest rates increase and this interest rate with life
insurance rates are different; people will prefer to participate on other investments. In this study was shown that by increasing literacy will also
increase premiums received. Thus, estimates showed that the effective variables on the life insurance demand what are, that this issue can be
effective on determine of appropriate strategy for life insurance.
References
Abrishami, H., & Mehrara, M. (2002). Applied econometrics: modern approaches. Tehran: Tehran University.
Alpha C. C., 1992, Elements of Dynamic optimization, McGraw-hill, Inc.USA.
Anderson, D. R. and J. R. Nevin, 1975, Determinants of Young MarriedLife insurance Purchasing Behavior: an empirical investigation, TheJournal of Risk and
Insurance, vol. 42,no 3, pp.375-387.
Babbel, D.F. and E.Ohtsuka, 1989, Aspects of Optimal Multi-period LifeInsurance, The Journal of Risk and Insurance, vol.56, pp.460-481.
Babbel, D.F. , 1985, The Price Elasticity of Demand for Whole LifeInsurance, The Journal of Finance, vol. 40, no.1, pp. 225-239.
Gharebaqian, M. (1997). The role of life and social welfare insurances in economic development. Sanaat-e-Bimeh Quarterly, 47.
Jalali Lavasani, E. (2005). A survey on the effect of macroeconomic variables on the demand for individual insurance. Sanaat-e-Bimeh Quarterly, 87.
Jafari Samimi, A., & Kardgar, A. (2006). Does insurance development support economic growth?. Pazhuheshhay-e-Eghtesadi Quarterly, 2.
Mahdavi, Q. (2009). A survey on the quantitative and qualitative factors affecting the demand for life insurance and ways to expand its influence in the insurance
industry (Research proposal). Bimeh Research Institute.
Mehrara, M., & Rajabian, M. A. (2006). Demand for life insurance in Iran and the oil-exporting countries. Journal of Tehran University Tahghighat-e-Eghtesadi,
74.
Pasban, F., & Azizi, F. (1997). Relationship between life insurance and the country’s economic growth. Sanaat-e-bimeh Quarterly, 43.
Pazhuyan, J., & Purpartoy, M. T. (2003). Estimate and forecast of demand for life insurance. Sanaat-e-bimeh Quarterly, 69.
Rajabian, M. A. (2005). Demand for life insurance in Iran and the oil-exporting countries (Master’s thesis). Tehran Economics Faculty.
Tajdar, R. (1996). A survey on the reasons for the lack of growth in the life insurance (Master’s thesis). Tehran University Faculty of Management.
Yaari, M., 1965,Uncertain Lifetime, Life Insurance and the Th eory of the Consumer, Review of Economic Studies, no 32
TI Journals
International Journal of Economy, Management and Social Sciences
www.tijournals.com
ISSN:
2306-7276
Copyright © 2014. All rights reserved for TI Journals.
Evaluating Life Insurance Demand in Iran
Samaneh Derakhshideh *
Ph.D. Student in Economic, Science and Research branch, Islamic Azad University, Kerman, Iran.
Sayyed Abdolmajid Jalaee
Associate Professor in Economic, Shahid Bahonar University of Kerman-Iran.
*Corresponding author: [email protected]
Keywords
Abstract
life insurance
VAR model
savings rates
interest rates
Per capita income
Insurance is one of the key tools in modern life which has a great role in reducing risks and providing
financial and mental security and finally it can broaden for countries economic development. The main
objective of this article is estimate the demanding life insurance and analysis effective factors on it which
was investigated in the period 1358–1388. According to the theoretical principals, the model of demanding
life insurance in Iran is a factor of real per capita income, savings rate, expects inflation, percent literacy and
the rate of interest. In order to providing the model of economic evaluation method, vector auto-regressive
(var) has been used. Based on the results, the variables real per capita income, interest rates, savings rates
have positive and significant relationship with the demanding life insurance. The variable of expected
inflation rate is negative and significant relationship with life insurance demand and there isn't a significant
between the variable of percent literacy and the life insurance demand.
(This essay has been excerpted from parts of
Samaneh Derakhshideh's PHD thesis.)
1.
Introduction
Generally speaking and regarding to the close relationship between developing of life insurance and the rate of wale fare and economic power of
the countries, there are lots of different kinds of living insurances as a tool for providing financial in our country and they didn't even improved,
such that it can be claimed that growing and developing didn't improved compared with European countries. Considering the great role of life
insurance in savings and investment and also its role in economic growth, it's easy understand the importance of this field in the economy of
different countries.
In recent years, life insurance had a growing share of development in the insurances of country and it has been increased from 6.6% in 2005 to
8.3 percent in 2010, but when it's been compared with developed countries, it shows a weak and most important the growth of insurance industry
in our country. The coefficient of insurance the important factors in developing and growing of this has increased from %1.27 in 2005 to
%1.52010. In spite of that we still feel the need of growing this field of insurance in the country. According to the recent changes in insurance
industry in our country, and in order to removing the absolutes and increasing insurers capacity, it seems that a good condition have been
provided for expanding insurance industry and especially life insurance and investing in this case. Insurance is one of the key tools in today's
world which has a great role in reducing risk and providing financial and mental security of people .It is also makes needed beds for economic
development for the country. Present study has evaluated the affective factors on the demand for life insurance in Iran during 1979 to 2009 by
vector auto-regressive model (VAR).Effective factors on the demand for life insurance in Iran are functions of real per capita income, savings
rates, inflation rates and rate of literacy. In this article at first, we consider the literature of the study in summery then we evaluated and conclude
the article.
2.
A review of the research literature
2.1 Background of the research
During recent decades contemporaneously with the expanding of life insurance, lots of studies have been done to investigate the nature of
demanding and supplying life insurance, affective factors in demanding and other cases and related techniques to this field of insurance. The first
academic and theoretical study which is one person behaviors who is not taking risk and he is trying to manage the risk of death. This is named
Yari article which is very famous such that theoretical works after him all affected by his method and his literature. In the study of Pazhuyan and
Poorpartoy (2002) using static data's gathered of1966to 2001yearsthepattern insurance demand has been estimated and the amount of that has
been developed till the end of 2003. The aim of this study is to evaluate the effects of earnings, expectation inflation, guaranteeing responsibility
and the education on demanding different kinds of life insurance in Iran.
Data's related to the variable of guaranteeing responsibility is gained by dividing the unemployed population to employed population in
considered year and for those years which in there are no static data's we have used: Rn = Rp (1 + r)n. In order to gain the amount of life
insurance demand, like many other studies, received money for insuring have been used, and to removing the effects of inflation, the price index
of consumption goods and services in urban areas CPI, the year 1990have been used as the base year. Headed households income is determined
by per capita national income. The effect of inflation on per capita national income is also removed by the CPI. Probability of death of headed
households is gained by dividing the number of deaths by the total population. The expected inflation is calculated by Pe = 0.7 Pt-1 +0.3 Pt-2 where
Pt-1 is inflation in previous term and Pt-2inflation in the last two term ago.
The final model of life insurance demand is as follows:
575
Evaluating Life Insurance Demand in Iran
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
= 78.49 + 0.41
+ 0.23
+ 1.85
+ 1.31
+ 7.78
− 0.17
−
(5.15) (3.2) (−2.46) (7.05) (5.54) (4) (−2.49)
2 = 0.927 . = 1.98 = 62.1
57
Where REV is the received premium per capita real of insurance life, NI is per capita real income (the fixed prices in 1999), POP is the
population of country, INF is the expected inflation, LITE is the percent literacy, DEP is time dependent and DU57 is dummy variable that
impact of the 1357 revolution on the insurance condition shows. Symbol L which is in front of all variables is natural logarithm. Presented
model is the demand pattern. In this model; despite of the equality between coefficients sign and theoretical expectations, the probability of
headed household's death wasn't effective. This can be the reason of religious and cultural beliefs in Iran. In this research, the demand function is
estimated using OLS. The results show a long-term stable equilibrium relationship between the variables in life insurance demand model.
Coefficient of LNI income shows income elasticity for life insurance demand which is estimated equal 41.0. Coefficient of the expected inflation
mentions that life insurance demand elasticity rather inflation is low. LDEP coefficient for guaranteeing responsibility is 1.58 which show a low
elasticity of demand compared to all studies have been done in other countries. LITE coefficient shows the percent literacy impact on the
demand for life insurance and this coefficient is higher than all coefficients. This can be resulting of increase their risk with higher education.
Another example study had been done conducted by Ebrahim Kardgar1995 about evaluating effective factors in the life insurance demand from
statistic information and national accountings and static calendars of 1986 to 1995. Considered variables for coefficient of life insurance demand
are income breadwinner, probability of breadwinner's death, headed household's responsibility and education and expectational inflation.
Received premium per capita of life insurance is considered as an index for demand these insurances and national income per capita for the
breadwinner's income. Probability of breadwinner's death is gained by dividing who are dead by total population and the population of dead
people. Related information to guaranteeing responsibility is gained by dividing the population under 20 ages by the population over 20 to 64
ages. For expectational inflation here we have used regression of logarithmic pattern of consumption goods and services price index in urban
areas (CPI).
Life insurance demand function is as follows:
1 = −9.69 + 0.635
+ 2.28
_2.29
+ 0.523
−
(−7.47) (3.35) (2.5) (−2.48) (4.26)
2 = 0.86 . = 1.98 = 37.26
PIN1R Where PIN1R is received premium insurance real per capita of life insurance which is gained by dividing received premium insurance
real per capita of life insurance by consumption goods and services price index in urban areas (CPI). INRN is net income per capita in constant
prices in 1999, RB is the literacy rate, PH is the expected inflation and DUM is a dummy variable that it shows life insurance act and the events
of government employee in 1997.For evaluating gained information similar to previous studies, is used t-student statistics for testing the
significance of individual coefficients and F statistics is used for testing the significance of all coefficients. Individual coefficients are significant
at the 5% level. According to the LINRN coefficient, income elasticity of life insurance demand is 0.635.Thus, the income elasticity of life
insurance demand is less than one and it can be concluded that the sensitivity of life insurance demand in contrast to family income can be
increased by using advertising and various selling methods. LRB coefficient shows that elasticity of percent literacy rate for life insurance
demand is larger than unity. The reason for this is the increase in risk aversion and the result is an increasing percent literacy .Because risk
averse individuals shall have demand more for life insurance. LPH coefficient shows that expected inflation elasticity of the demand for life
insurance is negative. Therefore, increasing expectational inflation led to reducing the demand for life insurance. In this study, the influence of
guaranteeing responsibility and probability of headed household's death variables is rejected at the 5% level and remove these two variables have
not a significant impact on the other variables. Its effect relays on issues of faith and culture of people in society.
2.2 Theoreticalprincipals of insurance demand
Since, insurance usual moves in the way returning lost peace which is caused by non-confidence, so in order to argue about the performance of
insurance function in theoretical point of view, we need a framework that is accepted the principle of non-confidence. Model individual demand
for insurance is based on maximizing the expected utility. There is no problem in safe mode, because the level of expected utility is equal to
assumed wealth. I.e. If the W * is person definitive wealth, we have: E [U (W *)] = U (W *). Thus, we can be choosing between the confidence
and non-confidence results. In non-confidence condition two modes can happen. Because of the governing non-confidence condition, the
meaning of utility in the way that there is in complete confidence mode would be insufficient. Since it is impossible to mention all possible
conditions. Thus, in the analysis of consumer behavior in non-confidence condition we can define the "expected utility function", with primary
wealth (W) which is with probability (P), facing with financial risk Rial (L) and L <W. Expected utility function from equation (1) is obtained.
=
(
− ) + (1 − )
(1)
In this case, the person has two choices: 1) buying insurance contract with an insurance premium of d and when a happening occurred receives
the compensations. 2) Doesn't buy the insurance contract and in the event of damage, person pays for all compensations. In order to simplify the
discussion, we assume that insurance contract completely covers all damages and there is only this contract or in other words insurance demand
is an "all or nothing" demand. If person chose the first mode (buying contrast), his expected utility function should be defined as fallow:
=
( − ) + (1 − ) ( − ) = ( − )
(2)
Where U1 is the utility of the insured person, and d insurance premium is paid. Maximum insurance that a person is ready to pay that is d *
which is obtained from equation (3).
Samaneh Derakhshideh *, Sayyed Abdolmajid Jalaee
576
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
( )( ( −
∗) =
( − ) + (1 − )
(3)
According this equation, result utility of complete insurance or the utility for removing non-confidence in wealth is equal to the weighted
average of the utility from total wealth and the utility from net wealth after one financial loss. Therefore, we can concluded that the expected
utility of the insured person is less than uninsured individual utility that nothing bad happened, and also it is more than uninsured individual
utility that something bad happened. So we have:
( ) < ( − ∗) < ( − ) In this relation, L is the maximum premium that a
utility maximizing expected for completely covers insurance is willing to pay which increases as the increase of risk probability and the volume
of risk or Or in other words, the demand for insurance increases with increased risk probability and large financial losses (Maqary, Fahimi, 155,
1993).
2.3 Theoretical principals of life insurance demand
Breadwinner demand for life insurance depends on the number of family members. Lewis has considered this relation by expanding theorical
construction of Yari life insurance and also by considering other members of family's preferences. In this mode, the life insurance is demanded
by members who are under dependent person that bread winner is facing to income non-confidence lifetime. His demand for life insurance based
on the lifetime of breadwinner of the family, is based on the life cycle model in which income is non-confidence (Frank D Lewis 1989).Most of
new theorical studies on the life insurance demand have chosen Yari studies as a starting point. In the meaning of life cycle model with nonconfidence lifetime, Yari shows that a person increases his expected utility by purchasing life insurance and receiving annual. Lewis method is
distinctive because he considers the life insurance demand from heir's point of view. In order word, life insurance is demanded for maximizing
expectational utility of heirs. In Yari's model structure one consumer buys life insurance to increase the expected utility of own lifetime.
E[U(T)] = ∫ α(t)g[c(t)]dt − β(t)Ψ[s(t)]
(4)
It is assumed, T is consumer's lifetime which is a random variable. Ψ(s (t)) is instantaneous utility of inherits, g[c (t)] is instantaneous utility of
consumption, and α(t) and β(t) are the discount factor (adjusted). When consumers get married or have children then β(t) increases significantly.
So these events can more explain the changes in ownership (purchase) of life insurance. According to the equation the change in ownership life
insurance is more depended on out moving of consumer's utility function. Lewis with development of the Yari model has gained movement in
consumer function by children and wives preferences, at least in part, as an endogenous event. Therefore, further analysis is part, in order to we
aren't looking why increase the number of people dependent effect on the utility function .Lewis also assumes that the utility function of each
member of family is separable. This assumption allows us to discussion from the insured person's point of view to wife and children who are
personal heirs. Therefore Lewis explains the more to demand by children for life insurance. A supervisor provides income as exogenous for their
children and this income is allocated so that the expected utility of his children is maximized. Children also their utility, taking into account the
limitations of exogenous income transfer from father to max. Maximizing utility by children can involve buying insurance for family supervisor.
Because children are faced with non-confidence income that comes from lifetime non-confidence, this kind of performance about life insurance
is specific seem. In other words, children may rarely consider these kinds of insurances for themselves. But this is suitable because parents
paying instead of their children can be considered as paying for life insurance and this can be considered as other expenses such as clothes and
other necessities for their children, and because utility children depends on the kind of costs, so we can be analyzed life insurance demand from
function of children. Children buy life insurance because they are not sure about how long their father is alive and so non-confidence income is
faced them. They stay at home till the age of a. until that time they receive determined converting payment each year but when their father dies,
they don't receive any other such kind payment except for their share of inheritance. Before the age of a, children are not allowed to receive a
loan against their probable future income. Although they are allowed to save money, the pattern of converting payments from father are
considered by the way that in fact children can't have any savings till they are at home. The d-age children, the expected utility with respect to its
costs in connection with life insurance premiums, di, to maximize (Frank D Lewis, 543, 1989).
If the father stays alive, children consume at the amount of ti-di in which ti is the earned income from converting payment. If the father dies, the
children earn equal ti + bi-di in which ti is the price of living insurance paper, and bi is the share of receiving inheritance. This question can be
written as follows:
maxEUi = (1 − pi) [ui(ti − di) + EUi + 1] + pi[ui( i + bi − di)]
(5)
In this relationship: EUk is the expected utility from age of k to a, pk is the probable of father's death at age of k for children, uk(0) is future
utility at age of k["uk (0) <0, 'uk> 0] and UK(0 ) is the utility from the age of K to a with the assume of one optimistic consuming pattern. The
relationship between the price of insurance paper and the insurance premium is
=
(insurance premium is a percentage of the price of
insurance paper). L is the factor of overhead. By emphasis on above assumes, under these optimal conditions occur:
U´(ti − di) =
(
)
, U´(∗ i + bi −∗ di) (6)
In the above equation, factors star show optimizing value. In order to simplifying the analysis, we use the following relations which is almost
true when mortality rate for father is low.
( −
∗) =
(
−
∗)
(7)
Where TK and DK show the present value of transfer payments and life insurance premiums from age of k to a, and this is when father is alive.
By replacing and given a utility function with constant elasticity, we have:
577
Evaluating Life Insurance Demand in Iran
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
∗ i + bi −∗ dj =
(
)
1. α (Ti −∗ Di)
(8)
Where a (negative) is elasticity of marginal utility of consumption or Arrow Pratt relative risk aversion is. Finally, the replacement and limit of
their children to life insurance assets non-negative, we have:
(1 − Lpi)f ∗ = max [
Where
=(
−
(
c∗ − b
)
(9)
) the present value in consuming flows from age k to a when father is alive. This equation has a fairly simple
interpretation. Assume that the children don't receive inheritance (i.e. bi = 0). So, if probability of pi death is low then we have:
∗
=( )
.
c ∗ .In
this case, assuming that father will be alive to till the age of a, then the value of the life insurance paper in above case simply is a part of present
value of children consuming. This is related the reverse with the overhead factor (L) and directly with the child's degree of risk aversion (a). The
case of spouse is like children. It is assumed that spouse is alive till the age of T in which she should leave the share of inheritance B. The first
order condition in the age i for maximizing expected utility is just like equation (10).
(
(
)
)
(
−
∗
∗
−
)=
((
)
)
+
∗
−
∗
(
−
)
(10)
Where vk is future utility for widow at the age of k, VK is the utility from ages K to T with assume of an optimal program, Y is spouse's income,
R is the discount rate, KKis the present spouse's value of capital stock at the time of husband's death in k age. By following the same procedure
which taken into account for life insurance demand for children, spouse's demand is as follows:
(1 − lpi)f ∗ = max [
(
)
c ∗ − ki + (
)
]
(11)
Where C* is value of wife's consuming from ages K to T, if the husband is alive till the age of T. Total issued insurance based on husband's
lifetime is equal to the total buying which are done by wife and each one of children with the same rate of not taking risk, and also regarding this
fact that non-negative limiting factor in life insurance assets are related to all the member or none of them, then we can mix above equation to
gain the total life insurance assets.
(1 − lpi)f ∗ = max [
(
)
W − TC
(12)
In this equation, F is the face value of life insurance related to his father's lifetime. TC is present value of each child consuming from current
period to the age of K, assuming that they stay alive. W is family wealth without accountings wife's share of inheritance (Frank D.lewis.1989).
Above equation is mental demand which explains the clear accounting that many families do during their shopping. Therefore, the results of
discussion taken from above equation are as these that the demand for insurance life with the probable of breadwinner's death, present value of
family consuming and families not taking risk have positively and negatively associated with household wealth and overhead costs.
3.
Estimation model
In this model, insurance demand in Iran is a function of national income, amount of saving, the inflation rate, rate of interest and the rate of
literacy, study period was for 1979 to 2009. Sample volume in the study was the total number of insurance papers in Iran between years 1979 to
2009, regarding this fact that life insurance demand with accounting received insurance expenses in national income of life insurance is
evaluated.
LREV = β1NI + β2 RS + β3 INF + β4 LITE + β5 RB + u t
(13)
Where LREV is real per capita premium life insurance, NI is real per capita income, RS is rates of saving, INF is expectational inflation, LITE is
percent literacy and RB is rate of interest.
Study assumptions based on economical theories include:
1.
By increasing in real per capita income, life insurance demand increases. (
> 0)
2.
By increasing the rate of inflation, demand for life insurance reduces. (
3.
There is a positive relationship between literacy rates and life insurance demand.(
4.
There is a positive relationship between rates of saving and life insurance demand.(
5.
There is a positive relationship between interest rates and life insurance demand.(
< 0)
> 0)
> 0)
> 0)
Because of the lack of the Stats of the interest rates in market or interest rate payable by some limited partnership companies has used from
opportunity cost of money, with this definition that opportunity cost of money is an annual maximum benefit of Short-term which are paid by
commercial banks. This information is gathered from central bank of Iran. Independent variables that are used in this study, are used in many
Samaneh Derakhshideh *, Sayyed Abdolmajid Jalaee
578
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
Iranian and foreign studies. The variable income has a positive relationship in lots of studies such Mehrara and Rajabian (2005), Ebrahim
Kardgar (1996), Firooze Azizi (2004), Jalali Lavasani (2004). Relationship between savings rate with life insurance demandhas been almost
positive. The reason is that income has had positively related to savings in economic theory. Inflation in Pazhuyan and Mirpartoy studies (2000),
Ebrahim Kardgar (1996), Jalali Lavasani (2004), had a negative effect on the insurance demand and it had a positive in Firooze Azizi's studies
(2004). The variable of education also had a positive relationship in studies, like studies Pazhuyan and Mirpartoy (2000), Mehrara and Rajabian
(2005), Ebrahim Kardgar (1996).
To investigate the model at first, we evaluate the stationary test by two tests Dickey Fuller (DF) and Augmented Dickey Fuller (ADF). In the
appendix Table 1, unit root test results Augmented Dickey - Fuller shown.
result
non-stationary
non-stationary
stationary
stationary
non-stationary
non-stationary
Significance
level
0.05
0.05
0.05
0.05
0.05
0.05
Table 1. Unit root test results Augmented Dickey- Fuller
The test statistic
MacKinnon's Critical Value
definition
with trend
without trend
with trend
without trend
0.41
2.07
-3.56
-2.96
real per capita premium life insurance
0.01
-1.04
-3.56
-2.96
real per capita income
-3.19
-3.41
-3.61
-2.97
savings rate
-3.92
-3.98
-3.57
-2.96
expects inflation
-0.47
-2.24
-3.57
-2.96
percent literacy
-1.02
-0.84
-3.57
-2.96
rate of interest
variable
LREV
LNI
RS
INF
LITE
RB
It has to be said that the expected inflation rate from the real inflation rate is calculated as follows:
= 0.7
+ 0.3
It can be seen
that except for variables of the savings rate and inflation rates of economy, other variables are non-stationary in level. Now, the variables are not
stationary, we are taking first-order differencing to determine whether these variables by once difference will be stationary or not. First-order
difference variables are defined as follows:
=
−
=
−
IT =
−
REV = REV − REV
RS =
RS − RS
In appendix table 2, unit root test results Augmented Dickey – Fuller for first-order differencing shown.
result
stationary
stationary
stationary
stationary
Significance
level
0.05
0.05
0.05
0.05
Table 2. Unit root test results Augmented Dickey – Fuller
The test statistic
MacKinnon's Critical Value
definition
with trend
without trend
with trend
without trend
-5.15
-4.21
-3.57
-2.96
real per capita premium life insurance
-3.58
-3.23
-3.57
-2.96
real per capita income
-5.61
-4.46
-3.57
-2.96
percent literacy
-4.87
-4.85
-3.57
-2.96
rate of interest
variable
LREV
LNI
LITE
RB
In Table 3 indicate that the variables remaining the by once difference were stationary. Therefore, briefly order cointegration variables as
summarized in Table 3.
Table 3. Order of integration
Definition
Variable
I(1)
real per capita premium life insurance
LREV
I(1)
I(0)
I(0)
I(1)
I(1)
real per capita income
savings rate
expects inflation
percent literacy
rate of interest
NI
RS
INF
LITE
RB
Order of integration
In order to Volume observations less than 100, to determine the degree from Schwarz-Bayesian, Hannan-Quinn and Akaike criteria which
consider the least lag and prevent reduce the degrees of freedom, are used. Appendix Table 4 shows the cases.
Table 4. Order determination of var
HQ
SC
AIC
FPE
LR
LogL
Lag
21.64
11.54
11.47*
21.83
12.90*
14
21.55
10.92
10.32*
92.31
0.002
0.001*
NA
288.32*
49.36
-306.5
-116.47
-71.74
0
1
2
579
Evaluating Life Insurance Demand in Iran
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
According to the output software, we use Schwartz-Bayesian criterion and lags of one, and to determine the number of vector integration have
used from the trace test and the maximum eigenvalue. Trace test and the maximum eigenvalue test results are shown in Appendix table 5 and 6.
According to tables 5 and 6, it can be seen that there is one optimal vector.
Prob.**
Table 5. Effect test for determining the number of convergence vector
Critical
Max-Eigen
Hypothesized
Eigenvalue
Value (0.05)
Statistic
No. of CE(s)
0.000
0.000
0.08
0.41
0.40
0.66
95.75
69.81
47.85
29.79
15.49
3.84
137.02
89.87
45.40
20.14
8.54
0.19
0.80
0.78
0.58
0.32
0.25
0.0006
None
At most 1*
At most 2
At most 3
At most 4
At most 5
Table 6. The maximum eigenvalue test to determine the number of convergence vector
Critical
Max-Eigen
Hypothesized
Eigenvalue
Value (0.05)
Statistic
No. of CE(s)
Prob.**
0.0006
0.0005
0.096
0.588
0.343
0.661
40.07
33.87
27.58
21.13
14.26
3.84
47.14
44.46
25.26
11.59
8.35
0.19
0.80
0.78
0.58
0.32
0.25
0.0006
None
At most 1*
At most 2
At most 3
At most 4
At most 5
At first, we consider all variables in the model, and then we remove those variables that are not significant and estimate the model. Thus, the
convergence of the vector is determined by table 7: Significant is at 95% confidence level.
Table 7. Model estimation
LITE
RB
INF
RS
LNIS
LREV
0.004
0.954143
0.05
2.363343
-0.01
-5.8069
0.05
6.131461
1.47
3.418091
coefficient
T test statistic
R2=0.96
Finally, the model is estimated as follows:
LREV= 1.47 NI + 0.05 RS -0.01 INF + 0.004 LITE + 0.05 RB +ut
t
(3.418091) (6.131461) (-5.8069) (0.954143) (2.363343)
As we can see from the results of the estimated model, all variables except percent literacy are significant, thus by remove this variable again
estimate the model which its results are presented in table 8 with a 95% confidence level.
Table 8. Model estimation
INF
RB
RS
LNI
LREV
-0.02
-5.6
0.06
6,85
0.06
5.05
1.86
5.31
coefficient
T test statistic
LREV= 1.86NI +0.06RS -0.02 INF +0.06 RB +ut
t (5.31) (5.05)
(-5.6)
(6,85)
The research results show that the per capita income variable has a positive and significant effect on real per capita life insurance premium. It
means that by increasing in per capita income will increase premiums received. In considering this fact that these variables are presented as
logarithmic, its coefficient is represents elasticity of the dependent variable relative to this variable. So, it should be said that by increasing even
one percent in per capita income, life insurance demand increases by 1.86 percent and knowing this that this, this coefficient is bigger than one
,so we can say that life insurance shows very sensitive to this variable. Interest rate and the savings rate coefficients are statistically significant
and have positive effect on insurance premiums. The significant point here is that the coefficient of the two variables is equal. So these two
variables have equal impact on dependent variable. Also, regarding the positive and significant coefficient of the literacy rate which is equal
0.021, it can be stated that increase literacy will also increase premiums received. Rate of inflation has negative and significant impact on
insurance premiums. As shown by the results among independent variables, real per capita income has the greatest effect on insurance
premium.At next step, to evaluate the variables effects on real per capita premium of life insurance use variance analyzing method. To do so,
using the output of the software is seen that during 10 periods, reaction the variations dependent variable comparing variations independent
variable is some percent: the result of variance analysis is come at fallowing table:
Samaneh Derakhshideh *, Sayyed Abdolmajid Jalaee
580
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
Table 9. Variance analysis
As result show, effect of income, expected inflation rate and interest rate variables increases gradually, but the effect of the saving rate from
fourth period and the percentage of literacy from sixth period are decreasing. At the next stage, we evaluated the impulse response function
graphs. Since the interpretation of the single coefficients in estimating model of VAR is often difficult, in practice, usual the response function
(IRF) is estimated. In the VAR, IRF determines response dependent variable to shocks on the error terms. Therefore, evaluate the impulse
response functions of the per capita premium relative to each of the independent variables. Impulse response functions are shown in appendix
9.According to results taken from impulse response functions see that the impulse response functions of the per capita premium compared with
per capita income during 2 periods was fixed, and after than increases . Also sudden increase in the rate of saving variable causes the decrease in
dependent variable in 3 periods, and its negative effects increases gradually. And after 3rd period see that response was so small and it finishes at
next period. At first, the effect of the expected inflation rate increases with a regular tone, such that has the process increasing and decreasing
and after eighth periods, it starts to decrease to a mild tone. Sudden increase in the literacy rate at first causes reduce in insurance premium, but
after fifth period the amount of that reduce and finally in the period of 10 years, it vanished completely. About the rate of interest, as see the
effect of one increases dependent variable in 2 periods which is minor in primary periods, but in other periods causes reducing this variable. And
during 10 years, it doesn't vanish.
Chart 1. Impulse response function of real life
insurance premium per capita in relation to savings rate.
Chart 3. Impulse response function of real life
insurance premium per capita in relation to expects inflation
Chart 2. Impulse response function of real life
insurance premium per capita in relation to real per capita income.
Chart 4. Impulse response function of real life
insurance premium per capita in relation to percent literacy
Chart 5. Impulse response function of real life
insurance premium per capita in relation to rate of interest
581
Evaluating Life Insurance Demand in Iran
International Journal of Economy, Management and Social Sciences Vol(3), No (10), October, 2014.
4. Conclusions
In this study, was estimated the appropriate model for life insurance .According to the model results, per capita income variable has positive and
significant effect on insurance premiums received and with increasing per capita income and improving purchasing power, and finally allocate a
greater share of income to the life insurance premium also increases According to negative coefficient of expected inflation during the period
under study, the expected inflation has a negative effect on premium. The expectation that inflation in future will be more motivated people for
life insurance decreases. Because people will think that the price increase in future will be causing decrease the purchasing power and the
becoming worth less of their money. Therefore less willing for savings will have and also the demand for insurance will reduce. Also, the
savings rate coefficient has statistically significant and positive effect on insurance premiums, i.e. by increasing savings rate increases premium
.Because of interest rates has Positive relationship to the premium received Can say when interest rates increase and this interest rate with life
insurance rates are different; people will prefer to participate on other investments. In this study was shown that by increasing literacy will also
increase premiums received. Thus, estimates showed that the effective variables on the life insurance demand what are, that this issue can be
effective on determine of appropriate strategy for life insurance.
References
Abrishami, H., & Mehrara, M. (2002). Applied econometrics: modern approaches. Tehran: Tehran University.
Alpha C. C., 1992, Elements of Dynamic optimization, McGraw-hill, Inc.USA.
Anderson, D. R. and J. R. Nevin, 1975, Determinants of Young MarriedLife insurance Purchasing Behavior: an empirical investigation, TheJournal of Risk and
Insurance, vol. 42,no 3, pp.375-387.
Babbel, D.F. and E.Ohtsuka, 1989, Aspects of Optimal Multi-period LifeInsurance, The Journal of Risk and Insurance, vol.56, pp.460-481.
Babbel, D.F. , 1985, The Price Elasticity of Demand for Whole LifeInsurance, The Journal of Finance, vol. 40, no.1, pp. 225-239.
Gharebaqian, M. (1997). The role of life and social welfare insurances in economic development. Sanaat-e-Bimeh Quarterly, 47.
Jalali Lavasani, E. (2005). A survey on the effect of macroeconomic variables on the demand for individual insurance. Sanaat-e-Bimeh Quarterly, 87.
Jafari Samimi, A., & Kardgar, A. (2006). Does insurance development support economic growth?. Pazhuheshhay-e-Eghtesadi Quarterly, 2.
Mahdavi, Q. (2009). A survey on the quantitative and qualitative factors affecting the demand for life insurance and ways to expand its influence in the insurance
industry (Research proposal). Bimeh Research Institute.
Mehrara, M., & Rajabian, M. A. (2006). Demand for life insurance in Iran and the oil-exporting countries. Journal of Tehran University Tahghighat-e-Eghtesadi,
74.
Pasban, F., & Azizi, F. (1997). Relationship between life insurance and the country’s economic growth. Sanaat-e-bimeh Quarterly, 43.
Pazhuyan, J., & Purpartoy, M. T. (2003). Estimate and forecast of demand for life insurance. Sanaat-e-bimeh Quarterly, 69.
Rajabian, M. A. (2005). Demand for life insurance in Iran and the oil-exporting countries (Master’s thesis). Tehran Economics Faculty.
Tajdar, R. (1996). A survey on the reasons for the lack of growth in the life insurance (Master’s thesis). Tehran University Faculty of Management.
Yaari, M., 1965,Uncertain Lifetime, Life Insurance and the Th eory of the Consumer, Review of Economic Studies, no 32
