Subir Sen & Dr.

S Madheswaran

DRAFT not to be QUOTED

DETERMINANTS OF LIFE INSURANCE DEMAND

In Selected 12 Asian Economies
Subir Sen1 and S Madheswaran2
Institute for Social and Economic Change
Nagrabhavi (PO)
Bangalore 560072
Email: sens@isec.ac.in
Abstract
The growth of the services sector in the Asian economies, led to substantial changes in the
financial sector. The Asian Financial Crisis, affecting the ASEAN economies in particular
resorted to more regulatory measures to enhance delivery of products with minimal risks
and failures. The countries surrounding the ASEAN economies also went through a phase
of economic-restructuring and the most notable event being the impact of Chinas
accession to WTO. The insurance industry, in most of the Asian economies were publicly
owned and operated. Government monopoly kept this segment of the financial market
isolated from domestic private or foreign participation. Barring few exceptions, the
insurance market on an average remained underdeveloped in terms of insurance density
and penetration. Although the regions insurance industry was not so badly affected by the
financial crisis, but reforms started after 1990s. Regulatory changes allowing private and
foreign entry has been luring global heavyweight insurance companies to enter these
economies. As more and more suppliers enter these markets, the important issue is to
examine the factors that enhance demand for insurance products. This study based on 4
SAARC countries, 2 countries from Greater China Region and 6 ASEAN Countries, deals
with this particular issue. After reviewing existing theoretical as well as empirical
literature, we find that determinants can be grouped as economic, demographic, legal,
socio-political, etc. variables. We restrict our selves in identifying only variables falling
under the first two groups. Above all, considering the luxury good characteristics of life
insurance, we do not consider variable like population growth, income, etc. which
1

Doctoral Fellow, Centre for Economic Studies and Policy, ISEC, Nagarbhavi, Bangalore 72; e-mail:
sens@isec.ac.in
2
Associate Professor, Centre for Economic Studies and Policy, ISEC, Nagarbhavi, Bangalore 72; e-mail:
madhes@isec.ac.in

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

intuitively raises demand. Annual data on selected economic and demographic variables
for the period 1994-2004 was collected from World Banks Global Development
Indicators-06 and figures relating to insurance business were collected from various issues
of Sigma from Swiss Re. Although country specific factors explaining insurance
consumption behavior cannot be fully assumed away, we analyze selected determinants of
life insurance in these 12 countries for a period of 11 years. Our balanced panel
estimation results shows that gross domestic savings, financial sector development, and
inflation are significant economic factors along with demographic variables like old
dependency ratio, urbanization, education and life expectancy. Balanced panel estimation
also suggests that in presence of country specific factors, some of these variables
contradict earlier findings and the relation between some of these variables with insurance
density and penetration didnt came in line with our expectation. We have tried to explain
these differences in relation to density and penetration, and important policy implications
are drawn from our conclusions.

Introduction

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

The growth of insurance companies as a part of the growing domestic financial sector in
both the developed and the developing economies, has led to increase in studies looking at
the economic significance of insurance sector. The strengths and weaknesses of insurance
sector do differ substantially when we compare the developed and the developing
economies. But, the fact that the growth determining factors, which might have once
boosted the developed markets, might motivate insurance consumption in the developing
economies, may lead to contradictory results. For, example, in a panel where we pool both
the developed and developing economies across time for analytical purposes, results may
be in favour of the developed markets. Putting it in more simpler form, the objective is to
see whether the significant different determinants of demand for insurance and life in
particular for studies relating to the developed insurance markets holds true for the selected
12 developing Asian markets or not. The next section is the review of literature where we
have considered the theoretical and empirical contributions towards analysis of insurance
demand separately. Following that is our researchable issues, questions and our expected
results. The methodology briefly puts forth an overview of the 12 economies selected, data
sources and the research design. Results are presented next and we conclude the major
findings with limitations of the study in the last section.
Review of Literature
(a) Theoretical Studies
Studies on life insurance consumption dates back to Heubner (1942) who postulated that
human life value has certain qualitative aspects that gives rise to its economic value. But
his idea was normative in nature as it suggested how much of life insurance to be
purchased and not what will be purchased. There were no guidelines regarding the kind
of life policies to be selected depending upon the consumers capacity and the amount of
risk to be carried in the product.

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

Economic value judgments are made on both the normative as well as positive issues.
Later studies3 on insurance gradually incorporated these issues via assimilating
developments in the field of risk and uncertainty following works by von Neumann and
Morgenstern (1947), Arrow (1953), Debreu (1953) and others. The economics on
insurance demand became more focused on evaluating the amount of risk to be shared
between the insured and the insurer rather than evaluation of life or property values. This
emerged because it was risk associated with individual life or property that called for an
economic valuation of the cost of providing insurance.
Life insurance is essentially a form of saving, competing with other forms of saving in the
market. The theory of life insurance demand thus developed through the life-cycle
model(s) of saving. Let is a persons income rate and his consumption plan are represented
by a continuous function of time y t and c t respectively. Thus, net saving (positive or
negative) at time t is given by4
t

s t e t e t y s c s ds --- (1)
0

where, is the rate of interest.

Even though all individual wants to consume as much possible, but this does not happen in
reality and the above expression of saving is constrained by number of possibilities. For
example, the person has no debt [ s t 0 ], solvent at time of his death at time T [
s T 0 ] and leaves a bequest of amount B [ s T B ].
Uncertainty associated with time of death is represented by a random variable with density
f t t , where t is the probability that a person at age x shall still be alive after
a time t. Taking expectation of (1) and discontinuing, we get

Yaari (1963), Mossin (1968), Hakansson (1969), Fisher (1973), Borch (1977), Pissarides (1980), Campbell
(1980), Karni and Zilcha (1985, 1986), Lewis (1989), Bernheim (1991) to mention few.
4
Borch(1990): Chapter 4

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

y s c s e

e t s t t dt

ds t dt

--- (2)

s e s y s c s ds --- (3)
0

If the individual takes pure endowment insurance, single premium being

t

Ex t e

E x s ds ,
0

Thus, net saving equals E x s ds y t c t dt .

0
0

This shows that saving through life insurance takes place at a higher rate of interest than
conventional saving. In the determination of optimal insurance consumption, the
conventional utilitarian theory is adopted which reflects individuals preferences over
different consumption patterns. Let us consider the utility function of the form
v c

e
0

u c t dt --- (4)

where, is the impatience to consume.

Assuming the person had no debt at time T i.e., s T 0 , the problem is to maximize its
t
utility (4). The solution gives us, v c t ke
. Within this framework, various forms

of life insurance are introduced; the probability associated with death or number of
maximum life years is considered; and then the expected utility of consumption is
maximized subject to any one of the three restrictions on the net savings as described
above.
The role of insurance in the above model has been predominantly to smoothen out
consumption over time, make bequests, and repay debts or to insure a constant income
stream after retirement. The ongoing discussion also reveals that individuals current
income and future anticipated consumption expenditure plays a crucial role in determining
the amount of insurance purchased (we are, for a while ignoring the form in which
insurance is purchased). The importance of rate of interest () or the impatience factor ()

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

is also worth considering. Preferences over different consumption pattern vary from person
to person and there are qualitative factors which affects such preferences.
Using the expected utility framework in a continuous time model, Yaari (1965) studied the
problem of uncertain lifetime and life insurance. Including the risk of dying in the life
cycle model he showed conceptually that an individual increases expected lifetime utility
by purchasing fair5 life insurance and fair annuities. Simple models of insurance demand
were proposed by Pratt (1964), Mossin (1969), Smith (1968) and others; considering a risk
averse decision maker with an initial wealth W. The results indicate that demand for life
insurance varies inversely with the wealth of the individuals. Hakansson (1969) used a
discrete-time model of demand for financial assets and life insurance purchase in particular
to examine bequest motive in considerable detail. Pissarides 6 (1980) further extending
Yaaris work proved that life insurance was theoretically capable of absorbing all
fluctuations in lifetime income. Lewis (1989) found out that the number of dependents as
an influence on the demand for life insurance.
To sum up, there are certain macroeconomic variables like income, rate of interest, and
accumulated savings in wealth form; along with a set of demographic or social variables
having potential impact on an individuals decision to opt for or not to demand life
insurance. Life insurance consumption increases with the breadwinners probability of
death, the present level of familys consumption and the degree of risk aversion.
(b) Empirical Studies
With this brief theoretical understanding we explore some of the pioneering empirical
exercises on determinants of life insurance. Most of these studies had focused on both the
demand side factors and the supply side factors.
Headen and Lee (1974) studied the effects of shortrun financial market behaviour and
consumer expectations on purchase of ordinary life insurance and developed structural
5

Fair actuarially means the premium or price of insurance that exactly equals expected value of payments
in case of claims or loss by the insurer, without charges for expenses or profits.
6
Refer Skipper (2003) page 20.

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

determinants of life insurance demand. They considered three different sets of variables:
first, variables stimulating demand as a result of insurer efforts (e.g. industry advertising
expenditure, size of the sales force, new products and policies, etc.); second, variables
affecting household saving decision (e.g. disposable, permanent and transitory income,
expenditure expectation, number of births, marriages, etc.) and lastly, variables
determining ability to pay and size of potential markets (e.g. net savings by households,
financial assets, and consumer expectation regarding future economic condition). They
concluded that life insurance demand is inelastic and positively affected by change in
consumer sentiments; interest rates playing a role in the shortrun as well as in the longrun.
Using an international dataset (12 countries over a period of 12 years) to examine the
relationship between property liability insurance premiums and income, Beenstock et al.
(1988) found out that marginal propensity to insure i.e., increase in insurance spending
when income rises by 1$, differs from country to country and premiums vary directly with
real rates of interest. Assuming a two period simple model; they considered the case when
wealth W is reduced by G following a loss and no insurance purchased in the first period.
If there had been some insurance purchased than wealth in the first period equals (Wpremium paid) and assuming loss, end period wealth is (G sum insured). Thus, again the
decision of consumer and his/her initial wealth status too are significant factors when
shortrun or longrun consumption of insurance is considered.
The study by Truett et al. (1990) discussed the growth pattern of life insurance
consumption in Mexico and United States in a comparative framework, during the period
1964 to 1984. They assumed that at an abstract level demand depends upon the price of
insurance, income level of individual, availability of substitute and other individual and
environment specific characteristics. Further, they experimented with demographic
variables like age of individual insured(s) and population within the age group 25 to 64 and
also considered education level to have some bearing on insurance consumption decision.
They concluded the existence of higher income inelasticity of demand for life insurance in
Mexico with low income levels. Age, education and income were significant factors
affecting demand for life insurance in both countries.

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

Starting with a brief review of Lewiss theoretical study and an assumption that inhabitants
of a country are homogeneous relative to those of other countries, the study by Browne et
al. (1993) expanded the discussion on life insurance demand by adding newer variables
namely, average life expectancy and enrollment ratio of third level education. The study
based on 45 countries for two separate time periods (1980 and 1987) concluded that
income and social security expenditures are significant determinants of insurance demand,
however, inflation has a negative correlation. Dependency ratio, education and life
expectancy were not significant but incorporation of religion 7, a dummy variable, indicates
that Muslim countries have negative affinity towards life insurance.
Based on a cross-sectional analysis of 45 developing countries, Outreville (1994) analysed
the demand for life insurance for the period 1986. In his study he considered variables such
as agricultural status of the country represented by the percentage of agricultural labour
force; health status of the country in terms of amenities like percentage of population with
access to safe drinking water; percentage of labour force with higher education and the
level of financial development as factors explaining insurance demand other than the
variables we have discussed above. Two dummy variables were used to reflect the extent
of competition in the domestic market and foreign participation in the countries
considered. The analysis shows that personal disposable income and level of financial
development significantly relates to insurance development. Since the political philosophy
regarding market openness varies from country to country, market structures dummy
appeared to be significant.
Taking into account the expansion of the service sector during the early nineties and
growth of insurance services in particular, Browne et al. (2000), tried to explain the
differences in property liability insurance consumption across countries. They considered
individuals income and wealth, degree of risk aversion, loss probability and price of
insurance as variables affecting property-liability insurance demand, similar to those used
for life insurance demand analysis. The analysis was focused on the OECD countries and
7

Religion can provide weights into individuals and life insurance consumption is less in predominantly
Islamic countries studies by Douglas et al. (1982), Henderson et al. (1987), Zelizer (1979) and Warsaw
(1986): cited in Browne et al (1993) page 621.

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

concluded that in general, insurance purchase is influenced by various economic and

demographic conditions. Another study based on nine OECD countries examined the short
run and long run relationship exhibited between economic growth and growth in the
insurance industry. This study by Ward et al. (2000) is a co-integration analysis using
annual data for real GDP and total real premiums for the period 1961 to 1996. Results give
an indication that country specific factors 8 influence the causal relationship between
economic growth and insurance market development.
Allowing income elasticity to vary as GDP grows for an economy, Enz (2000) proposed
the S-curve relation between per-capita income and insurance penetration. Using this one
factor model one can generate longrun forecast for life insurance demand. Observing the
outlier countries or countries distant from the S-curve plot, it is possible to identify
structural factors like insurance environment, taxation structures, etc. resulting in such
deviations.
There are two detailed studies on the determinants of life insurance demand, one taking
into consideration only the Asian countries and the other based on 68 economies. The
former study by Ward et al. (2003) and the later by Beck et al. (2003) evolves around the
issue of finding the cause behind variations in life insurance consumption across countries.
After almost three decades of empirical work in this direction, it is still hard to explain the
anomalous behaviour of Asian countries with higher savings rate, large and growing
population, relatively low provision for pensions or other security and a sound capital
market but comparatively low per-capita consumption of insurance. Except Japan, most of
the Asian countries have low density and penetration figures.
The two main services provided by life insurance: income replacement for premature death
and long-term savings instruments, are the starting point for Beck et al. (2003). They
considered three demographic variables (young dependency ratio, old dependency ratio
and life expectancy), higher levels of education and greater urbanization as independent
factors in explaining insurance demand. Economic variables like Gini index and Human
8

Country specific factors referred here to are attitudes towards risk and risk management, regulatory factors,
legal environment, other modes or availability of financial intermediation, etc.

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

development index are new additions along with institutional variables reflecting political
stability, access to legal benefits and an index of institutional development were used. The
analysis considering the time period 1961 to 2000 shows that countries with developed
banking system, high income and lower inflation have higher life insurance consumption.
The association of insurance demand with demographic is not statistically strong however
older the population, higher tends to be insurance consumption. The luxury good nature of
insurance did not reflect through its association with income distribution.
In contrast to Beck et al. (2003) results, the study by Ward et al (2003) is indicative of the
fact that improved civil rights and political stability leads to an increase in the consumption
of life insurance in the Asian region as well in the OECD region. Following Laporta et al
(1997, 1998, and 2000) works relating to supportive aspect of legal environment for
finance, they too considered the same in determining insurance demand. Analyzing the
data from 1987 through 1998 for OECD and Asian countries, they observed that income
elasticity between developed economies and emerging economies are consistent with Scure insurance growth findings by Enz (2000).
Issues and Questions
The major issue to start with would be to re-examine the significant variables that can be
best fit as determinants of life insurance demand. Recently, there are number of studies on
single economies: Hwang and Gao (2003) study were on the Chinese economy; Lim and
Haberman (2004) focuses on Malayasis; and Hwang and Greenford (2005) on Mainland
China, Hong Kong and Taiwan. Lenten and Rulli (2006) explored the time series properties
of the demand for life insurance in Australia using a novel statistical procedure that allows
unobservable components to be extracted.
Zietz (2003) has reviewed the efforts of researchers to explain consumer behaviour
concerning the purchase of life insurance for almost 50 years. The review of earlier studies
concludes that bulk of the empirical studies undertaken finds a positive association
between increase in savings behaviour, financial services industry and demand for life

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

insurance. Taking this forward, our first issue is to see whether or not per capita gross
domestic savings and financial depth influences life insurance consumption. GDP and Percapita GDP are often highly correlated with the proxy variables measuring insurance
demand- density and penetration. We therefore ignore these two variables and assume that
as income grows, it will add to insurance demand via rise in the savings component i.e.,
GDS.
The demographic factors dependency ratios, adult literate population, life expectancy and
adult literate population are considered and in line with earlier studies we expect these to
be significant in explaining life insurance demand. It is expected that Young dependency
will be negatively related and the rest three demographic variables are expected to have a
positive relation. We have assumed first that consumer price index (CPI) will be the best
proxy for inflation although we have considered log difference of CPI as an alternative
measure for inflation. Lastly we have the real interest rate, which is in our case the deposit
interest rate minus inflation. In Table 1, we briefly present the identified potential
determinants of life insurance consumption with the expected signs.
Table 1: Determinants, Research Questions and Expected Relationship
Determinants
GDS per-capita

Questions
Is level of saving a significant factor in explaining life insurance

Expected Sign
+

consumption?
Financial Depth

Does the level of financial sector development has any spill-over

Urbanization

effect on enhancing insurance consumption?

What relation exists between the rate of urbanization and insurance

consumption?
The impact of rising population in the 0-14 years age group on

insurance consumption
The impact of rising population above 65 years of age on insurance

consumption
Whether or not level of education, as an indicator of awareness,

affect insurance consumption?

The effect of life expectancy on insurance consumption

Do price fluctuations affect insurance consumption?

What is the effect of changes in interest rate on insurance

Young
Dependency Ratio
Old Dependency
Ratio
Adult Literate
Population
Life Expectancy at
Birth
Inflation
Real Interest Rate

consumption?

Methodology

10

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

The 12 Asian Economies: The study focuses on 12 selected economies starting with 4
from the 7 SAARC9 countries (India, Bangladesh, Pakistan and Sri Lanka); 2 from the
Greater China region (China and Hong Kong) and 6 from the 10 nations ASEAN 10
(Indonesia, Malaysia, Philippines, Singapore, Thailand, and Vietnam). These 12 nations
apart from being the most promising emerging and growing economies, most of them are
also struggling to overcome poverty. Except Hong Kong and Singapore, most of the
countries have less than USD 5000 GDP per capita. To be little more precise, 6 economies
in our sample have GDP per capita less than USD 1000.
Table 2: Life Insurance Premiums and Penetration Figures for the 12 Asian
Economies
1994
1999
#
Premiums*
Density
Premiums*
Density#
Bangladesh
0.011
0.0001
103.71
0.81
China
2027.41
1.71
10752.17
8.58
Hong Kong
2536.44
422.88
4973.26
752.73
India
3652.70
3.99
6306.50
6.31
Indonesia
625.08
3.30
780.79
3.83
Malaysia
1222.39
60.81
1892.14
83.32
Pakistan
211.00
1.77
186.49
1.38
Philippines
375.08
5.61
538.99
7.20
Singapore
1476.57
431.62
2837.70
718.04
Sri Lanka
0.011
0.0006
73.80
4.05
Thailand
1235.20
21.20
1550.48
25.74
Vietnam
0.011
0.0002
35.62
0.46
Source: Sigma (various issues), Swiss Re.
*in Millions of USD at constant 2000 prices
# Life Premiums as percentage of total mid-year population
1
Figures are extrapolated

2004
Premiums*
Density#
231.48
1.66
30584.31
23.60
14643.54
2122.25
14716.36
13.63
1496.19
6.88
3798.01
152.53
230.14
1.51
795.78
9.75
6437.32
1532.70
115.46
5.95
2936.49
46.10
442.46
5.38

The overall spending on insurance and related products in these economies is very low
when compared to the OECD economies. With population growth rate above or near 1
percent and a comparatively low level of insurance consumption measured by per capita
premium, indicates the existing economies of scope and the potential for growth. Although,
some of these economies were griped by the Asian Financial Crisis, the issue is to create
stronger infrastructure for financial services and thereby boost the underdeveloped
insurance market.
9

South Asia Association for Regional Cooperation; the rest 3 countries are Maldives, Nepal and Butann
The Association of Southeast Asian Nations and the other members are Brunei Darussalam, Cambodia,
Laos and Myanmar.
10

11

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

Table 3: Correlations
Per Capita GDP
GDS
Life Premium Penetration
Density
1.000
-0.080
0.205
0.811**
0.879**
Per Capita GDP
1.000
0.817**
0.050
-0.057
GDS
1.000
0.453**
0.305
Life Premium
1.000
0.910**
Penetration
1.000
Density
**Correlation is significant at the 0.01 level (2-tailed).

These economies display immense diversity in ethnicity, culture and religion. The legal
system too varies a lot with most of them following the European legal system due to their
colonial influence. Following the General Agreement on Trade and Services, most of these
economies agreed to open up there financial sector in general and insurance in particular,
to either domestic private companies or foreign companies or to both. Table 5 shows that
most of these countries were dominated by large number of insurers but there insolvencies
called for nationalization. It was in the beginning of 90s that regulatory changes were
initiated and implemented which once again opened the insurance industry but with more
cushions in terms of regulation and supervision.
The 11 years period considered for this study from 1994 to 2004, witnessed important
regulatory changes. During this period, in most of the economies expansion of life segment
of the insurance market particular gained momentum. This helped to rejuvenate insurance
density. The only exception from this increasing trend was Pakistan, where actually density
has dropped.

12

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

Data Source: The panel for 12 selected Asian countries studied for 11 years starting 1994
to 2004 is constructed using annual aggregate data from different secondary source. The
Insurance premium figures are collected from various issues of Sigma, a publication from
Swiss Re. The Economic as well as demographic variables used are collected from the
International Financial Statistics 2006 and the World Development Indicators 2006. The
explanatory variables in the model are the economic and demographic variables (refer
Appendix: 1)
Research Design: We estimate two different panel data regression models. The models are
different since we are trying to represent life insurance demand by two different dependent
variables: insurance penetration and insurance density. To address the problems of
estimation and inference associated with panel data, we estimate both the fixed effects
model (FEM) and the random effect model (REM). The decision regarding the best model
between FEM and REM is arrived at using the Hausman test.
The specification of the fixed-effects model as well as the random effect model is as
follows:
log( Penetration)it i 1 log(GDSPC )it 2 log( FIND)it
3URBit 4 log(YD)it 5 log(OD)it 6 log( ADLP )it 7 log(TLEX )it
8 RIRit 9 (1/ CPI )it it
And the second regression equation estimated is,
log( Density )it i 1 log(GDSPC )it 2 log( FIND)it
3URBit 4 log(YD)it 5 log(OD)it 6 log( ADLP )it 7 log(TLEX )it
8 RIRit 9 (1/ CPI )it it
Here, subscript i denotes the country, the subscript t represents time, i are cross-sectional
intercept terms,s are the slope parameters and it are random error term. The model
argues that variations in the independent variables in the 12 selected economies may
contribute to insurance penetration and insurance density.
Results and Interpretation

13

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

As already discussed, we have dropped GDP per-capita to reduce the high level of multicollinearity in the fixed effects model. For similar reason we have taken the inverse of CPI.
The discussion in this section is based on the final results of the models and the summary
results are reproduced in Table 4.
Table 4: Empirical Model Estimation
Dependent Variable
Independent Variables
Constant
GDS per-capita
Financial Depth
Urbanization
Young Dependency Ratio
Old Dependency Ratio
Adult Literate Population
Total Life Expectancy at Birth
Inflation
Real Interest Rate

Model 1
Model 2
Insurance Penetration
Fixed Effect
Random Effect
-161.993*
-59.125*
(39.091)
(27.491)
2.997*
0.868
(1.0540)
(0.855)
4.268*
4.296*
(1.352)
(1.200)
-0.088*
-0.021
(0.0432)
(0.016)
-6.697
-1.617
(5.968)
(3.901)
-8.448*
-8.869*
(3.639)
(3.388)
-11.103*
0.752
(3.212)
(0.495)
119.749*
20.271
(0.249)
(13.940)
-8.500*
-10.485*
(4.620)
(3.911)
0.119
-0.151
(0.249)
(0.284)
126
126
13.19
101.57
0.6467
0.5288
0.0829
0.5514
0.0628
0.4627
101.89

Observations
F Test Statistics / 2
R2
Within
Between
Overall
Hausman Test Statistics
*Statistically significant
Figures in parentheses are standard errors

Model 3
Model 4
Insurance Density
Fixed Effect
Random Effect
-161.854*
-66.253*
(39.378)
(28.595)
3.190*
1.374
(1.062)
(0.875)
4.392*
4.495*
(1.362)
(1.209)
-0.087*
-0.020
(0.043)
(0.017)
-6.804
-1.746*
(6.012)
(4.042)
-8.152*
-8.816
(3.666)
(3.447)
-10.775*
0.624
(3.236)
(0.523)
118.448*
23.716
(22.374)
(14.556)
-8.414*
-10.149*
(4.654)
(3.935)
0.109
-0.133
(0.251)
(0.279)
126
126
13.15
134.96
0.6598
0.5639
0.3265
0.7086
0.2588
0.6362
91.03

Our results based on the balanced panel, allows us to exploit both cross-country and timeseries variation in the data. The fixed effect model (Model 1) shows that variation in
insurance penetration is explained by GDS per-capita, financial depth, urbanization, old
dependency ratio, adult literacy, life expectancy and measure of inflation. But, in the
random effect model (Model 2), financial depth, old dependency ratio and inflation turns
out to be significant. From the fixed effects regression, the F statistics for testing the joint
significance of the country effect is 13.19 {F [11,105]}. This is greater than the critical

14

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

value and the evidence is strongly in favour of country effect in the data. Also, the
Hausman test for the fixed and random effects regressions gives the test statistics 101.89
with 9 degrees of freedom. The critical value of chi-square with 9 degrees of freedom is
16.92 at 95 percent level of significance. Thus, we can reject the fact that the individual
effects are uncorrelated with other regressors in the model. This further suggests that of the
two alternatives we have considered, the fixed effects model (Model 1) is a better choice.
Similarly, when we use insurance density as the dependent variable to see the impact of
variations of the independent variables, fixed effects model (Model 3) turns out to be a
better choice. Model 3 also suggest that there is a significant relation between GDS percapita, financial depth, urbanization, old dependency ratio, adult literacy, life expectancy
and inflation.
The regression results from Models 1 and 3 shows that GDS per-capita has a positive
relationship with insurance density and penetration. This suggests that, as savings activity
increases, it will push up per capita insurance expenditure and there by enhance insurance
reach. The significant positive relation between density and penetration with financial
depth suggest that with strength in the other segments of the financial sector, insurance
sector too will benefit. Although urbanization is significant, its sign is contradictory to
earlier studies and our expectation that the level of urbanization in an economy raises
insurance consumption. In our case, it suggests that urbanization will decrease insurance
consumption. This may be as a result of rural-urban migration and a rising share of poor
population in the urban centres in most of the selected economies.
Young dependency ratio has the expected sign but its insignificant in all the 4-models. Old
dependency ratio is significant but again its sign does not corroborate with earlier studies.
It suggests that the growth of the old dependency ratio will decrease insurance density as
well as insurance penetration. Interesting, the education variable in the model, adult literate
population is significant but not in line with expectation. This might probably give rise to
the issue whether higher education or a certain minimum, guarantees awareness of
insurance benefits. The last of the demographic variable, the life expectation of the total

15

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

population at birth is highly significant and suggests that as living conditions and longevity
of life improves, demand for insurance products would go up.
As expected, inverse of inflation is inversely related with density and penetration and is
significant but real interest rate is not. The relation between inflation with the demand
proxies, do not corroborate with earlier studies and hence we conclude that current interest
rate or price situation does not affect insurance consumption decisions.
Conclusions
Overall, our cross-country analysis confirm that if we exogenously consider income to be a
crucial factor in explaining insurance consumption, economic variables of importance
would be gross domestic savings, level of financial sector development and inflation. As
specialized financial institutions turns to financial conglomerates, one important policy
implication can be strength and weaknesses of banking and other non-banking institutions
will have a positive or negative spillover effect on the insurance industry. As more and
more banks line up for insurance service provision, the entry of these institutions will also
push up demand. Our analysis suggests that as the savings increases, insurance too would
increase but as such insurance is not purely savings and hence its purchase can be to
smoothen the income or wealth status. If savings plus life risk insurance products are sold
then it might boost insurance consumption. Although, real interest rate was not significant
in our study, nevertheless, some variant of it may also play an important role in explaining
individuals choice between insurance and other savings instruments. Our main results
based on panel of 12 economies over 11 years supports the fact that demographic variables
like life expectancy, young and old dependency, adult literacy rate and rate of urbanization
significantly determines insurance demand. However, the study can be extended
considering more variables and dummies to look for the country and time specific factors
affecting demand. The results are of importance to the policy makers as most of the
selected economies have undergone changes, particularly in terms of regulatory reforms,
recently. Identification of variables, which affect insurance consumption decisions, may
be of prime interest and more detailed country specific variables with dummy for legal and

16

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

institutional reforms may reveal the growth trend. Lastly, it will be worth to take a much
bigger sample in terms of countries and periods considered, to understand why some of the
variables behaved so differently than expected.

17

Development of Life Insurance Industry*

Table : 5
Economies

Regulatory Authority

Pre-Regulatory Regime

Banglades
h

The Dept. of Insurance,

headed by the Insurance
Directorate of the Ministry of
Commerce.
China Insurance Regulatory
Commission (CRIC) formed in
1998

After separation from Pakistan in

1972, state-owned monopoly life
insurer Jiban Bima Corp. was
established in 1973
PICCs1 monopoly as insurer ended
in 1985; 5 new public insurer
created; 4 regional insurer existed

Hong Kong

Office of the Commissioner of

Insurance (OCI); 1990

India

Insurance Regulatory and

Development Authority (IRDA);
1999
The Directorate General for
Financial Institutions along with
Ministry of Finance
Ministry of Finance de facto
regulator with administration by
Bank Negara Malaysia (BNM)
Securities and Exchange
Commission of Pakistan
(SECP); 1999
The Insurance Commission
(Komiyon ng Seguro)

Comparatively the most competitive

insurance market with over 204
insurers
The Life Insurance Corp. Act of 1956
merged 256 life insurers to form LIC2

China P.R.

Indonesia
Malaysia
Pakistan
Philippine
s
Singapore
Sri Lanka
Thailand
Vietnam

The Insurance Department of

the Monetary Authority of
Singapore (MAS)
The Insurance Board of Sri
Lanka; 2001
The Department of Insurance
(of the Ministry of Commerce)
independent from 2002
Ministry of Finance

Foreign Ownership

No. of
Insurers

Not Allowed

18
1 Public

Allowed;
2002
Regulations on Admn. of
Foreign-invested Ins. Co.

47

140

100% private participation

allowed since 1999

Allowed;
104
captive
insurers in operating from
25 countries
26% equity share in local
companies

16
1 Public

---

---

Allowed

62

In 1984 Takaful Insurance Act

permitted insurer operations based
on Islamic principles.
In 1972, 34 out of 50 existing
insurers were merged to form State
Life Insurance Corp.2
Private sector was always
emphasized

Always open to Private

Law and Regulatory changes
in 1996.
1990
100% private entry Allowed
since 1992
The Republic Act No. 8179 of
1996 permits 100% foreign
ownership
2000
The market was liberalized to
direct insurers
2001,
100%
private
operation
allowed since 1986
1992
Insurance
business
is
relatively liberal
1999
2000, approved & enacted
new set of laws

Allowed

Allowed

5
1 Public

New foreign insurer is not

allowed to hold a
composite license
In 2000, 49% restriction
on foreign ownership in
local companies was lifted
Up to
90%
foreign
investment
in
local
companies
25% of foreign ownership
in domestic insurers5

34

Until 1960s the market was loosely

regulated
Insurance Corp. Act in 1961
nationalized the life insurance
Industry to form ICS4
Market was blocked for decades till
late 1990s
Baoviet, a state owned life insurer
started operation in 1996 but was
initially created for non-life insurance

Regulatory Change
1984
Allowed
50%
underwriting in 1990

private

1992;
Interim
Mgmt.
Regulation
for
foreign
Insurance Institutions allowed
foreign entry
1997after it became PRCs
special administrative region

Allowed 2 foreign
companies in 1999
followed by 2 in 2000

8
10
1 Public
20
10
3 Public

Source: * Refer to Endnote i

1
Peoples Insurance Company of China; 2 Life Insurance Corporation of India under the control of Govt. of India; 3 The Corporation established under Article 11 of
the Life Insurance (nationalization) Order of 1972 ;4Insurance Corporation of Sri Lanka, renamed the Sri Lanka Insurance Corporation Ltd. in 1993; 5 to be lifted to
49%;

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

Table 6: Selected Economic, Demographic variables and Regulatory Restrictions

Year
2004
Bangladesh
China P.R.

Popn
Growth
Rate#
1.88
0.60

GDP Per
capita@
402.07
1323.14

Religion:
Majority
Popn
Muslim
Confucianism

Not Known
Approval from CIRC

Hong Kong

1.16

27446.32

Confucianism

Self-regulatory

India

1.43

538.31

Hindu

Approval from IRDA

Indonesia

1.35

906.19

Muslim

Not Known

Malaysia

1.86

4289.72

Muslim

Pakistan

2.41

566.03

Muslim

Philippines

1.79

1084.92

Roman
Catholic

Singapore

1.31

24163.91

Buddhist

Sri Lanka

0.86

961.61

Buddhist

1996 Insurance Act &

the
Insurance
Regulation
has
provisions
regarding
pricing
To
follow
SECP
Guidelines
Approval
from
the
commission
of
the
premium rates
Approval from principal
Officers and directors
from Insurance Dept.&
certificate
from
appointed actuary
Approval required

Thailand

0.87

2355.99

Buddhist

Approval required

Vietnam

1.04

501.99

Buddhist

Approval required

Price Regulation

Life Business Regulatory Restrictions

Investment Regulation
Stringent
Strict restrictions, confining the
areas of insurer investment to
bank deposits, Govt. bonds,
financial bonds, etc.
No specific guidelines.
Stringent
Not less than 50% in approved
Securities
Total
invest.
excluding
mortgage loans should be
equal to technical reserves
30% in Bumiputra Shares,
ownership share by ethnic
Malay
Statutory fund for each product
sold
Partial/ varying
Partial/ varying

30% reserve funds in Govt.

securities
Investments to be within
Thailand
Separate for foreign and
domestic companies

# Annual rate of growth in percentage

@ In USD at constant 2000 prices
Source: Refer Endnote I; World Bank (2006) World Development Indicators CD-ROM 2006.

19

Solvency Norms

Weakly imposed
The solvency margin is a
function of insurers size of
business measured by
premiums, claims, or both.
4% of mathematical reserves
and 0.3% of capital at risk or
$HK 2million
Excess of the value of insurer
over its liability amount value
1% of Premium Reserves
Actual valuation of liability,
aggregate Insurance
coverage, etc.
Prescribes asset & liability
valuation methods
Aggregate
insurance
Coverage in force, subject to
min. amount
Fund solvency margin &
company solvency margin

Total Assets sufficiently >

Total Liabilities
2% of reserve fund subject to
min. amount
Exist

Appendix 1:
Type of Variable

ECONOMIC

Variable Name

Description

Gross Domestic Product

(GDP)

It is the sum of value added by all resident producers plus any

product taxes (less subsidies) not included in the valuation of output.
Growth is calculated from constant price GDP data in local currency
unit
It is gross domestic product divided by midyear population.

GDP per-capita
Gross Domestic savings
(GDS)
Consumer Price Index
(CPI)
Liquid Liabilities as a % of
GDP (for Financial Depth)
(FIND)
Inflation Rate (INFR)
Real Interest Rate (RIR)
Population

Young Dependency
ratio (YD)
Old Dependency (OD)
DEMOGRAPHIC

Adult Literacy Rate (ADLP)

Life Expectancy at Birth
(TLEX)
Urban Popn (URB)

INSURANCE
BUSINESS

Premiums
Premium Density
Premium Penetration

It reflects changes in the cost to the average consumer of acquiring

a basket of goods and services that may be fixed or may change at
specified intervals, such as yearly.
They include bank deposits of generally less than one year plus
currency. It is the sum of currency and deposits in the central bank,
plus transferable deposits, electronic currency, plus savings in time
and savings deposits, foreign currency transferable deposits,
certificate of deposit and securities repurchase agreements etc.
Log difference of CPI
Deposit interest minus inflation
Total mid-year population
The ratio of dependents-people younger than 15 years of age to the
working age population-those ages 15 to 64
The ratio of dependents-people older than 64 years of age to the
working age population-those ages 15 to 64
The percentage of people ages 15 and older who can, with
understanding, both read and write a short, simple statement about
their everyday life.
The number of years a newborn infant would live if prevailing
patterns of mortality at the time of its birth were to stay the same
throughout its life.
The population of the urban agglomeration, a contiguous inhabited
territory without regard to administrative boundaries.
Total Premium generated (net)
Premiums per-capita
Premium from Business in force as a percentage of Gross Domestic
Product

i Kwon, W. Jean (2001) Toward Free Trade in Services: The ASEAN Insurance Market, IIF
Occasional Paper, No. 3; International Insurance Foundation, Washington, DC.
Kwon, W. Jean (2002) The Insurance Markets of South Asia, IIF Occasional Paper, No. 4;
International Insurance Foundation, Washington, DC.
Kwon, W. Jean (2002) The WTO and Insurance in Greater China: The Peoples Republic, Hong
Kong, Macau, and Taiwan, IIF Occasional Paper, No. 6; International Insurance Foundation,
Washington, DC.

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

References
Arrow, K. J. (1953). Le role des valeurs boursires pour la repartition la meilleure des
risques. Economtrie.Translated as the Role of securities in the Optimal
Allocation of Risk-Bearing, Review of Economic Studies, Vol. 31, 1964, pp 9196.
Arrow, K. J. (1965). Insurance, Risk and Resource Allocation in Foundations of
InsuranceEconomics, G. Dionne and S. E. Harrington (eds.), Kluwer Academic
Publishers.
Babbel, D. F. (1985) The Price Elasticity of Demand for Whole Life Insurance, Journal
of Finance, Vol. 40; pp 225-239.
Beck, T. and I. Webb (2003) Economic, Demographic, and Institutional Determinants of
Life Insurance Consumption Across Countries, World Bank Economic Review,
Vol. 17; pp 51-88.
Beenstock, M., G. Dickinson and S. Khajuria (1986) The Determinants of Life Premiums:
An International Cross-Section Analysis, Insurance: Mathematics and
Economics, Vol. 5; pp 261-270
Beenstock, M., G. Dickinson and S. Khajuria (1988) The Relationship between PropertyLiability Insurance Premiums and Income: An International Analysis, Journal of
Risk and Insurance, Vol. 55; pp 259-272.
Bernheim, B. D. (1991) How Strong are Bequest Motives? Evidence based on Estimates
of the Demand for Life Insurance and Annuities, Journal of Political Economy,
Vol. 99; pp 899-927.
Bernheim, B. D., L. Forni, J. Gokhale and L. J. Kotlikoff (2003) The Mismatch Between
Life Insurance Holdings and Financial Vulnerabilities: Evidence from the Health
and Retirement Study, American Economic Review, Vol. 93; pp 354-365.
Black, K. Jr. and H. D. Skipper Jr. (2003). Life and Health Insurance. Pearson Education,
India.
Browne, M. J. and K. Kim (1993) An International Analysis of Life Insurance Demand,
Journal of Risk and Insurance, Vol. 60; pp 616-634.
Browne, M. J., J-W. Chung and E. W. Frees (2000) International Property-Liability
Insurance Consumption, Journal of Risk and Insurance, Vol. 67; pp 73-90.
Campbell, R. A. (1980) The Demand for Life Insurance: An Application of the Economics
of Uncertainty, Journal of Finance, Vol. 35; pp 1155-1172.
Cargill, T. F. and T. E. Troxel (1979) Modelling Life Insurance Savings: Some
Methodological Issues, Journal of Risk and Insurance Vol. 46; pp 391-410.
Debreu, G. (1953). Une conomie de certain. Mimo, lectricit de France. Cited in
An Introduction to Insurance Economics in Foundations of
InsuranceEconomics, G. Dionne and S. E. Harrington (eds.), Kluwer Academic
Publishers.
Enz, R. (2000) The S-curve Relationship Between Per-Capita Income and Insurance
Penetration, Geneva Papers on Risk and Insurance, Vol. 25; pp 396-406.
Fisher, S. (1973) A Life Cycle Model of Life Insurance Purchases, International
Economic Review, Vol. 14; pp 132-52.
Greene, W. H. (2002) Econometric Analysis, 4th Edition, Pearson Education Asia.

21

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

Hakansson, N. H. (1969) Optimal Investment and Consumption Strategies Under Risk,

An Uncertain Lifetime, and Insurance, International Economic Review, Vol. 10;
pp 443-466.
Hammond, J. D., B. D. Houston and R. E. Melander (1967) Determinants of Household
Life Insurance Premium Expenditures: An Empirical Investigation, Journal of
Risk and Insurance, Vol. 34; pp 397-408.
Hatekar, N. and A. Singh (2004) Determinants of Life Insurance Consumption in India: A
Cross-State Analysis, BimaquestVol. 4; pp 3-21.
Headen, R. S. and J. F. Lee (1974) Life Insurance Demand and Household Portfolio
Behaviour, Journal of Risk and Insurance, Vol. 41; pp 685-698.
Hwang, T. and B. Greenford (2005) A Cross Section Analysis of the Determinants of Life
Insurance Consumption in Mainland China, Hong Kong, and Taiwan, Risk
Management and Insurance Review, Vol. 8; pp 103-125.
Hwang, T. and S. Gao (2003) The Determinants of Demand for Life Insurance in an
Emerging Economy-The case of China, Managerial Finance, Vol. 29; pp 82-96.
Karni, E. and I. Zilcha (1985) Uncertain Lifetime, Risk Aversion and Life Insurance,
Scandinavian Actuarial Journal, pp 109-123.
Kwon, H-S. and B. L. Jones (2006) The Impact of the Determinants of Mortality on Life
Insurance and Annuities, Insurance: Mathematics and Economics, Vol. 38; pp
271-288.
Kwon, W. Jean (2001) Toward Free Trade in Services: The ASEAN Insurance Market,
IIF Occasional Paper, No. 3; International Insurance Foundation, Washington,
DC.
Kwon, W. Jean (2002) The Insurance Markets of South Asia, IIF Occasional Paper, No.
4; International Insurance Foundation, Washington, DC.
Kwon, W. Jean (2002) The WTO and Insurance in Greater China: The Peoples Republic,
Hong Kong, Macau, and Taiwan, IIF Occasional Paper, No. 6; International
Insurance Foundation, Washington, DC.
Lenten, Lim J. A. and D. N. Rulli (2006) A Time-Series Analysis of the Demand for Life
Insurance Companies in Australia: An Unobserved Components Approach,
Australian Journal of Management, Vol. 31; pp 41-66.
Lewis, F.D. (1989) Dependents and the Demand for Life Insurance, American
Economic Review, Vol. 79; pp 452-466.
Lim, C. C. and S. Haberman (2004) Modeling Life Insurance Demand from a
Macroeconomic Perspective: The Malaysian Case, Research Paper: The 8th
International Congress on Insurance, Mathematics and Economics, Rome.
Mossin, J. (1968) Aspects of rational Insurance Purchasing Journal of Political
Economy, Vol. 79; pp 553-568.
Outreville, J. F. (1990) The Economic Significance of Insurance Markets in Developing
Countries, Journal of Risk and Insurance, Vol. 62; pp 487-498.
Outreville, J. F. (1996) Life Insurance Markets in Developing Countries, Journal of
Risk and Insurance, Vol. 63; pp 263-278.
Pissarides, C. A. (1980) The Wealth-Age Relation with Life Insurance, Economica, Vol.
47; pp 451-457.

22

Subir Sen & Dr. S Madheswaran

DRAFT not to be QUOTED

Truett, D. B. and L. J. Truett (1990) The Demand for Life Insurance in Mexico and the
United States: A Comparative Study, Journal of Risk and InsuranceVol. 57; pp
321-328.
Ward, D. and R. Zurbruegg (2000) Does Insurance Promote Economic Growth? Evidence
From OECD Countries, Journal of Risk and Insurance, Vol. 67; pp 489-506.
Ward, D. and R. Zurbruegg (2002) Law, Politics and Life Insurance Consumption in
Asia, Geneva Papers on Risk and Insurance, Vol. 27; pp 395-412.
Yaari, M. E. (1964) On Consumers Lifetime Allocation Process, International
Economic Review, Vol. 5; pp 304-317.
Yaari, M. E. (1965) Uncertain Lifetime, Life Insurance, and the Theory of the Consumer,
Review of Economic Studies, Vol. 32; pp 137-150.
Zietz, E. N. (2003) An Examination of the Demand for Life Insurance, Risk
Management and Insurance Review, Vol. 6; pp 159-191.

23