The long term impact of health on economic growth in Pakistan.
Subject: Economic growth (Economic aspects)
Authors: Akram, Naeem
Padda, Ihtsham ul Haq
Khan, Mohammad
Pub Date: 12/22/2008
Publication: Name: Pakistan Development Review Publisher: Pakistan Institute of Development Economics Audience: Academic Format: Magazine/Journal Subject: Business, international; Social sciences Copyright: COPYRIGHT 2008 Reproduced with permission of the Publications Division, Pakistan Institute of Development Economies, Islamabad, Pakistan. ISSN: 0030-9729
Issue: Date: Winter, 2008 Source Volume: 47 Source Issue: 4
Topic: Computer Subject: Company growth
Accession Number: 228122161
Full Text: The paper investigates the impacts of different health indicators on economic growth in Pakistan. Cointegration and Error Correction techniques have been applied on the time series data of Pakistan for the period of 1972-2006. We find that per capita GDP is positively influenced by health indicators in the long run and health indicators cause per capita GDP. However, in short run the health indicators fail to put significant impact on per capita GDP. It reveals that health indicators have a long run impact on economic growth. It also suggests that impact of health is only a long run phenomenon and in the short run there is no significant relationship exists between health variables and economic growth. The major policy implication of the study is that if we desire high levels of per capita income, we can achieve it by increasing and improving stock of health human capital, especially when current stocks are at lower end. Moreover, study also points out a rather diminutive role of public health expenditure in determining the per capita GDP.

JEL classification: I12, O15, C22

Keywords: Health Human Capital; Economic Growth; Cointegration, Error Correction

1. INTRODUCTION

Human capital plays pivotal role for sustainable economic Growth. As different growth theories suggest the role of human capital as a significant for growth process. The concept of human capital in economic literature defined broadly by including education, health, training, migration, and other investments that enhance an individual's productivity. However, the growth economists that have incorporated human capital in the growth studies, paid greater attention on analysing the impact of education on economic growth, while ignoring the role of health human capital. It is only in very recent times that studies have started looking at health and tried to estimate the relationship between health status and economic growth.

There exists a two-way relationship between improved health and economic growth. Health and other forms of human and physical capital increases the per capita GDP by increasing productivity of existing resources coupled with resource accumulation and technical change. Furthermore, some part of this increased income is spent on investment in human capital, which results in further per capita growth. According to Fogel (1994), approximately one third of GDP of Britain between 1790 and 1980 is the outcome of improvements in health especially improvement in nutrition, public health, and medical care facilities and these improved health facilities should be considered as labour enhancing technical change.

On the other hand, Economic development results in improved nutrition, better sanitation, innovations in medical technologies; all this increases the life expectancy, reduces the infant mortality rate. World Development report 2007 depicts the situation by concluding that Average life expectancy at birth worldwide rose from 51 years to 65 in less than 40 years. Similarly Average life expectancy in developing countries was only 40 years in 1950 but had increased to 63 years by 1990 [World Bank (1993)]. Preston (1976) has analysed various determinants of life expectancy and emphasised that economic development is the most important factor.

The purpose of this paper is to analyse the long-term relationship between health and per capita GDP, by using Cointegration and Granger Causality. Long-term analysis of health and economic growth would be helpful in determining the possible magnitudes of fully accumulated effects of health on economic growth. Hypothesis that 'health affects economic growth' is a long run phenomenon would be tested.

The organisation of the paper is as follows: Section 2 reviews some of the previous studies conducted on the subject of the relationship between economic growth and health status. Section 3 describes the status of human capital situation in Pakistan. In Section 4 Econometric Model and data used in the study is discussed which make the Analytical Framework of the paper. Section 5 discuses the results and main findings of the analysis and in Section 6 the conclusion emerges from the study are highlighted.

2. LITERATURE REVIEW

As mentioned in introduction that numerous studies have been conducted on the relationship between human capital development and economic growth. The main conclusion of these studies is that there exists a positive relationship between human capital and economic growth. (1) It is only last decade that there is a flurry of studies exploring the relationship between health and economic growth.

By using the adult survival rate as an indicator of health status, Bhargava, et al. (2001) finds positive relationship between adult survival rate and economic growth. Results remains similar when adult survival rate is replaced by life expectancy. However, fertility rate have a negative relationship with economic growth. Because life expectancy is highly influenced by the child mortality. Growth in workforce is mostly lower than population growth. Resultantly high fertility rate reduce the economic growth by putting extra burden on scare resources.

Mayer (2001) also uses the probability of adult survival by gender and age group as a measure of health status. By using Granger-type, causality test study concludes that health status causes economic growth in Latin America generally, and specifically in Brazil and Mexico. Improvements in adult health are associated with 0.8-1.5 percent increase in annual income. Moreover, the growth impact is higher for improvements in health of female compared with health of male.

Bloom, et al. (2004) by using 2SLS technique finds that life expectancy and schooling have a positive and significant effect on GDP. Improvements in health increase the output not only through labour productivity, but also through the Capital accumulation. Study also finds that improvement of one year in a population's life expectancy resulted into an increase of 4 percent in output.

By using the average height adult survival rate and life expectancy as an indicator of health status Weil (2005) finds that health is an important determinant of income variations in different countries. Approximately 17-20 percent of the cross-country variation in income can be explained by cross-country differences in status of health.

Arora (2001) uses the life expectancy at birth, at ages; five, ten, fifteen, twenty, and structure of adulthood as health indicators for 10 industrial countries. Study concludes that improvement in health status has increased the pace of long-term economic growth by 30-40 percent. It also concludes that high rate of disease prevalence and deaths are among the main reasons for poor long-term growth in developing countries.

Lorentzen, et al. (2005) analysed the impacts of adult mortality rate on economic growth. Study finds that high mortality rate reduce the economic growth by curtailing the time horizon. Resultantly people take actions that yield short-term benefits at the long-term cost. Study also concludes that fertility, investment in physical and human Capital, are the channels b adult mortality rate affects economic growth.

Measuring health status by health status by infant mortality rate, life expectancy rate and crude health rate and per capita GNI as indicator of economic growth, Malik (2005) finds that if OLS is used then there is no significant relationship between health status and economic growth. However, when 2SLS is used then study finds highly significant effect of health indicators on economic growth.

Scheffler (2004) argues that health may not be treated as output (life expectancy, adult survival rate etc.), but it needs to be treated as input (health expenditure). Study finds that elasticity of health care spending with respect to GDP is greater than one. This means that if GDP increases by 10 percent then healthcare spending goes up by more than 10 percent. Consequently, developed countries spend more on health as compared to developing countries.

Tallinn (2006) uses adult mortality rate, fertility rate and life expectancy to analyse the economic costs of ill health along with economic benefits from improving it for Estonia. Study finds that fertility rate and adult mortality rate have a significant and negative impact in both OLS and Fixed effect model specification. Moreover By using survey data Study also concludes that ill health has a statistically robust and negative impact on labour supply and productivity at the individual level.

Zon (2001) concludes that good health is a necessary condition for people to be able to provide labour services. Study finds that an increase in the demand for health services caused by an ageing population will negatively affect the economic growth.

Gyimah-Brempong (2004) finds that investment (health expenditure) and stock (child mortality rate) of health human capital have a positive and significant relationship with growth of per capita income. However, the relationship is quadratic. Study concludes that investment in health in LDCs will boost the economic growth in the short run and increases the level of income in the long run because investment in health become a part of Stock of human capital.

While analysing the contribution of health by measuring it by the survival rate of males between age 15 and age 60 in economic growth, Jamison (2003) finds that better health accounted for about 11 percent of growth. Study concludes that investment in physical capital, education and health plays critical role in boosting the economic growth.

Using different household survey indicators of adult nutrition and health, Schultz (2005) examines the impact of health on total factor productivity. Study finds that better health human capital have a significant and positive impact on wages and workers productivity. Study finds the developing countries often lack the resources for investment in health; on the other hand poor health status slows down the economic growth. Developing countries seems to be in a vicious cycle resulting in persistent underdevelopment.

Fogel (1994) concludes that approximately one third of income growth in Britain during 1790-1980 may credited to improvements in health facilities and better nutrition. Study also concludes that public health and medical care must be recognised as labour-enhancing technological change.

Taking into account initial poverty, economic policy, tropical location, and life expectancy Gallup and Sachs (2001) find that per capita GDP of the countries having intensive prevalence of malaria grew 1.3 percent less compared with other countries. Study also concludes that a 10 percent reduction in malaria incidence would result in 0.3 percentage increase in the growth rate of per capita GDP.

Sachs and Warner (1997) by using life expectancy as indicator of health finds a quadratic relationship between health human capital and the rate of economic growth. Study concludes that health human capital increases economic growth at a decreasing rate.

3. SCENARIO OF HUMAN CAPITAL IN PAKISTAN

Human Capital shows a dismal picture in Pakistan. On the human poverty index, Pakistan ranked 77th among 108 countries and 136th among 177 countries on the human development index. (2) It is the outcome of extremely low expenditure on health over the last 60 years. Health expenditure in Pakistan remains at low band of 0.5-0.8 percent of GNP during 1970-2007. In FY 2006-07 health expenditure was only 0.6 percent of GNP, which was very low comparing with other developing countries. Not only the health expenditures are low but also delivery of available healthcare facilities is inefficient. Moreover, primary healthcare and rural health services were ignored and the priority was given to hospitals, medical colleges and curative services in the urban areas. In Pakistan, infant mortality rate was high at 77 per thousand live births; life expectancy was low at 65 years in 2006. Comparing the indicators in 2000, 85 per thousand live births and life expectancy of 62 years, there is improvement in health indicators but pace is rather slow. Trend in the health indicators over the years, summarised in Table 1.

Education sector also shows the same situation. Public expenditure on education was on average 0.8 percent of GNP in 1980s, 2.3 percent of GNP in 1990's, lowest in FY 2004-05 of only 1 percent of GNP and 2.4 percent in FY 2006-07, that is much lower than other low income countries of the region. Moreover as in the case of health expenditure, most of spending on education goes largely to the recurring expenditure. Historically, priority was given to the higher education, whereas primary education was ignored. As a result, literacy rate was just 55 percent and gross primary enrolment rate was 87 percent in 2006. Pakistan's health and education indicators represent a depressing picture when it is compared with the countries with same level of development such as India, Bangladesh, China and Sri Lanka. There is a dire need to increase the expenditure on health and education.

The relationship between health expenditure and economic growth can be analysed with the help of following Graph 1:

[GRAPHIC 1 OMITTED]

The graph reveals that we cannot find out a pattern in the relationship between health expenditure and economic growth and it is very difficult to draw a conclusion that health expenditure positively affects per capita GDP.

[GRAPHIC 2 OMITTED]

The Graph 2 shows that there exist a negative relationship between fertility rate and economic growth. Similarly Graph 3 depicts a positive relationship between life expectancy and economic growth. These are according to the theory.

[GRAPHIC 3 OMITTED]

This discussion reveals that health input variables (health expenditure) do not have a relationship with economic growth in Pakistan. However, health output variables (life expectancy, fertility rate etc.) are significantly affecting economic growth in Pakistan.

4. DATA AND METHODOLOGY

In order to determine relationship between health and economic growth different health variables can be used. There are two categories of health indicators, health input indicators and health output indicators. Health input indicators comprises of expenditure on health services, availability and quality of health facilities etc. While health output indicators includes life expectancy, Infant mortality rate and Adult survival rate, fertility rate etc.

Depending upon availability of time series data; life expectancy and Infant mortality are used as health indicators. As data for these variables are not available in a consistent time series, so data was interpolated by using DigDB 7.1.3.3 an excel Add inn. The major output variable used is health expenditure as percentage of GDP. The independent variable of the model is per capita GDP and is used as a proxy for economic growth. There are certain other explanatory variable. A Brief description of all the variables used in the study is presented in Table 2. The data of all the variables is used ranging from 1972 to 2006.

4.1. Theoretical Model

Numerous models have been developed to incorporate impact of human capital in economic growth. Romer (1990), and Barro (1991) have emphasised that human capital is the most important factor in determining the economic growth.

As the focus of study is to analyse the effects of health human capital on economic growth so the human capital is separated into two parts health human capital (H) and other forms of human capital i.e. education human capital (E). Per capita income (Y) is assumed as a function of the stocks of physical capital (K), health human capital (H), education human capital (E) and a vector of other variables (Z) that include technology and other environmental variables.

Y = f (K, H, E, Z)

Where Y is per capita GDP, H is health human capital, E is Education human capital and Z represents all other explanatory variables. H in time t is the sum of the stock of health human capital in the previous period and accumulation to the stock in the current period. It is assumed that accumulation in the health human capital stock ([DELTA]H) depends on the amount of resources devoted to health care and the efficiency by which this expenditure is converted into health stock. It is further assumed that quantity of resources devoted to health investment is a product of the proportion of income devoted to health care ([Y.sub.h]) and the level of income. The stock of health human capital evolves in the following way

[H.sub.t] = [H.sub.t-1+] [DELTA][H.sub.t], and [DELTA]H = [lambda][Y.sub.h] Y,

where [lambda] is the productivity parameter of health expenditure and all other variables. The ability to transform health expenditure into health stock is assumed to be dependent on the stock of health human capital. The health technology equation can be written as : [lambda] = [lambda](H). Substituting [lambda] into the [DELTA]H equation and that in turn into the production function, the income growth equation become.

[??] = [??] ([DELTA]H + [DELTA]K + [DELTA]E + [H.sub.t-1] + Z)

The per capita output equation that is estimated and the empirical model developed can be written in the following form.

Per capita [GDP.sub.t] = [alpha] + [beta] Age Dependency + [gamma] Health Expenditure + [omega] Openness + v Population per bed + [lambda] Life Expectancy + [GAMMA] Investment + [rho] Mortality Rate + [OMEGA] Secondary Enrolment + [[xi].sub.t]

5. EMPIRICAL FINDINGS

In order to find long run relationship between variables, cointegration technique is used, however before examining the long-term relationship between the variables, the first Step is to determine whether time series is univariate or not.

5.1. Unit Root Test

Unit root test is used to check whether or not data is stationary. A process is said to be stationary if its probability distribution remains unchanged as time proceeds and we can say that data generation process does not changed. To test the unit root most widely used test is Augmented Dickey Fuller (ADF) test. The general form of ADF test can be written at level and first difference form as follows.

[DELTA][x.sub.1] = [alpha][x.sub.t-1] + [k.summation over (i=1)] [[beta].sub.i] [DELTA][x.sub.t-1] + [??] + [[gamma].sub.t] + [[epsilon].sub.t]

[DELTA][DELTA][x.sub.1] = [alpha][DELTA][x.sub.t-1] + [k.summation over (i=1)] [[beta].sub.i] [DELTA][DELTA][x.sub.t-1] + [??] + [[gamma].sub.t] + [[epsilon].sub.t]

In the Table 3 Null Hypothesis of unit root against alternative of stationarity is tested. Results reveals that all the variables are non-stationary at level so the null hypothesis of unit root at level cannot be rejected. However, at first difference null hypothesis of unit root is rejected for all the variables and all the variables are I (1).

As variables are I(1) therefore most appropriate technique for the analysis is contegration.

5.2. Cointegration

With the aim of determining long run, relationship between variables cointegration technique is used. To test cointegration among the variables, there exist two main techniques; Engle and Granger (1987) approach and Johansen (1988) approach. In order to test cointegration among variables, the procedure developed by Johansen (1988) is used. This technique depends on direct investigation of cointegration in the vector autoregressive (VAR) representation. It yields maximum likelihood estimators of the unconstrained cointegration vectors and it allows one to explicitly test for number of cointegration vectors so that the weaknesses of Engle-Granger (1987) two-step procedure are overcome. Engle and Granger (1987) technique is a two-step methodology and stability deviations from the relationship is examined by using the coefficients estimated after fitting static regression. However, the test suffers from a number of shortcomings. The basic assumption of the technique is that the cointegrating vector is unique, bounding to a model that is a linear combination of independent cointegrating vectors. However, if cointegrating vector is not unique it fails to address the situation. Moreover, it examines only the dominant cointegrating vector between series.

If there is a VAR of order p

[y.sub.t] = [[alpha].sub.1] [y.sub.t-1] + [[alpha].sub.2] [y.sub.t-2] ... [[alpha].sub.p] [y.sub.t-p] + [beta][x.sub.t] + [[epsilon].sub.t]

Where [y.sub.t] is a k-vector of non-stationary I(1) variables, is a [x.sub.t] is a d-vector of deterministic variables, and [[epsilon].sub.t] is a vector of innovations. We may rewrite this VAR as,

[DELTA]yt = [universal] [y.sub.t-1] + [p-1.summation over (i=1)] [v.sub.i] [DELTA][y.sub.t-i] + [beta][X.sub.t] + [[epsilon].sub.t]

Where

U = [p.summation over (i=1)] [A.sub.i] - I

[V.sub.i] = - [p.summation over (j=i+1)] [A.sub.j]

Granger's representation theorem asserts that if the coefficient matrix U has reduced rank r, k then there exists kxr matrices [alpha] and [beta] each with rank r such that U = [alpha][beta]' and [beta]'[y.sub.t] is I(0). r is the number of cointegrating relations (the cointegrating rank) and each column of [beta] is the cointegrating vector. The elements of are [alpha] known as the adjustment parameters. Johansen's method is to estimate the matrix from an unrestricted VAR and to test whether we can reject the restrictions implied by the reduced rank of U.

There are four different steps involved while testing cointegration, in the first step order of stationarity is determined and variable must be stationary at same level. We have already found that variables are stationary at first difference i.e. series of the model are I (1). Therefore, the cointegration can be determined between the variables. Second step involves choosing the optimal lag length. To determine the lag length VAR model is used. According to AIC criteria, we determine the lag length of one for the model. Next step deals with determining the number of cointegrating vectors. In the study, both trace statistic and eigenvalue statistic are used. The results of both of the statistics are summarised in Table 4 and Table 5.

Results of trace static suggest that there exist seven cointegrating vectors while the results of maximum Eigenvalue value suggest the six cointegrating vectors.

Empirical evidence presented in Table 6 reveals that in the long run age dependency is negatively and significantly affects per capita GDP, as more people become idle due to age or other factors then these people would definitely have negative impacts on economic growth. The public health expenditure is also having positive but insignificant impact on per capita GDP. These results show that public health expenditure a major health input variable does not have a relationship with per capita GDP. This result confirms the poor allocation and utilisation of public health expenditure. It also depicts that public health expenditures are so mere that they fail to put a significant impact on economic Growth. Nevertheless, other health status indicators like life expectancy, mortality rate and population per bed all are having significant impacts on economic growth. It means that improvement in health status is the result of private sector spending, whereas public health expenditure are very little and are utilised in such a way that do not affect economic growth significantly.

These results confirm the vital and significant contribution of private sector in improving the health conditions. As the public sector fails to contribute in provision of health facilities it is the private sector that contributes considerably in improving the health status. Openness to trade is having positive and significant impact on economic growth. The population per bed is negatively effecting the economic growth. When population per bed increases, it means that less health facilities are available to the people which ultimately affect economic growth in the long run. Secondary education remains highly significant implying that more educated workers are more likely to adapt with prevailing technologies and to contribute to economic growth. Contradicting with theory gross capital formation has failed to have a significant impact on economic growth in the long run, however relationship is positive.

Results reveals that in the long run indicators of human capital i.e. health and education both have significant impact on economic growth. Therefore, we can say that for sustainable economic growth, policies should aim to improve the standards of health and education. As the public health expenditure does not have significant impact on economic growth, the policies regarding health should be directed in such a way that they should give more incentives to private sector for investing in health facilities.

5.3. Error Correction Model

If there a long run relationship between different variables exists then an error correction process is also taking place. Error correction model indicates the speed of adjustment towards the long run equilibrium after a short run shock. In order to check error correction following equation is estimated.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

The results show that estimated lagged error correction term is negative and significant suggesting error correction is happening in the model. The coefficient of feedback coefficient (Error Correction term) is -0.68, suggesting approximately 68 percent of disequilibrium in previous year is corrected in the current year. Other estimated coefficients shows that in the short run only age dependency and secondary education have significant impact on per capita GDP. No health indicators have the significant impact on economic growth. It reveals that impact of health is only a long run phenomenon and in the short run there exist no significant relationship between health variables and economic growth.

6. CONCLUSIONS AND POLICY IMPLICATIONS

The main objective of this paper is to analyse the short run and long run dynamic of health human capital on economic growth. To attain that objective Cointegration coupled with Error Correction techniques has been used. The results shows that age dependency, openness, population per bed, secondary school enrolment, life expectancy, mortality rate are affecting per capita GDP but health expenditure have no relationship with per capita GDP.

The result confirms that health variable plays a very significant role in determining the long run economic growth. As all the health indicators have a significant impact on the long run economic growth. However, results obtained from Error Correction model reveal that health indicator does not have significant impact on economic growth in the short run. It suggests that impact of health is only a long run phenomenon and in the short run there is no significant relationship between health variables and economic growth.

The policy implications of the study is that country like Pakistan that desire a high levels of per capita income, they can achieve it by increasing and improving the stock of health human capital, especially if current stocks are at lower end. Moreover, study also points out a rather diminutive role of public health expenditure in determining per capita GDP.

From a research perspective, results implies that health human capital must be included in the growth equations as it is also a very important part of human capital. Moreover there is dire need of study, which analyse the dynamics of health demand in Pakistan, as such study is lacking for many years. Similarly, there is also need for a comparative study on the role of private and public health care facilities in improving the health human capital.

REFERENCES

Arora, S. (2001) Health, Human Productivity, and Long-term Economic Growth. The Journal of Economic History 699-749.

Barro, R. (1991) Economic Growth in a Cross-section of Countries. Quarterly Journal of Economics 106:2, 403-443.

Bhargava, A., D. T. Jamison, L. J. Lau, and C. J. Murray (2001) Modeling the Effects of Health on Economic Growth. Journal of Health Economics 423-440.

Bloom, D. E., D. Canning, and J. Sevilla (2004) The Effect of Health on Economic Growth: A Production Function Approach. World Development 1-13.

Engle, R. F. and C. W. Granger (1987) Co-integration and Error Correction: Representation, Estimation, and Testing. Econometrica 55, 251-276.

Fogel, R. W. (1994) Economic Growth, Population Health and Physiology: The Bearing of Long-term Processes on the Making of Economic Policy. American Economic Review 84, 369-395.

Gallup, J. L. and J. D. Sachs (2001) The Economic Burden of Malaria. The American Journal of Tropical Medicine and Hygiene 85-96.

Gyimah-Brempong, K. and M. Wilson (2004) Health Human Capital and Economic Growth in Sub-Saharan African and OECD Countries. The Quarterly Review of Economics and Finance 296-320.

Jamison, D. T., L. J. Lau, and J. Wang (2003) Health's Contribution to Economic Growth in an Environment of Partially Endogenous Technical Progress. (Disease Control Priorities Project Working Paper No. 10).

Johansen, S. (1988) Statistical Analysis of Cointegration Vactors. Journal of Economic Dynamics and Control 231-254.

Lorentzen, P., J. McMillan, and R. Wacziarg (2005) Death and Development. (NBER Working Paper 11620).

Malik, G. (2006) An Examination of the Relationship between Health and Economic Growth. (ICRIER Working Paper No. 185).

Mankiw, N. G., D. Romer, and D. Weil (1992) A Contribution to the Empirics of Economic Growth. Quarterly Journal of Economics 407-437.

Mayer, D. (200l) The Long Term Impact of Health of Economic Growth. World Development 1025-1031.

Preston, S. H. (1976) Mortality Patterns in National Populations. New York: Academic Press.

Romer, P. M. (1990) Endogenous Technical Change. Journal of Political Economy 71-102.

Sachs, J. D. and A. M. Warner (1997) Fundamental Sources of Long Run Economic Growth. The American Economic Review 183-188.

State Bank of Pakistan (Various Issues) SBP Annual Reports. Karachi: State Bank of Pakistan.

Scheffler, R. M. (2004) Health Expenditure and Economic Growth: An International Perspective. Occasional Papers on Globalisation, University of South Florida.

Schultz, T. P. (2005) Productive Benefits of Health: Evidence from Low-income Countries. (Economic Growth Centre Discussion Paper No. 903).

Tallinn (2006) The Economic Consequences of Ill-health in Estonia. PRAXIS Centre for Policy Studies.

Weil, D. N. (2005) Accounting for the Effect of Health on Economic Growth. (NBER Working Paper No. 11455).

World Bank (1993) World Development Report. Washington, DC.

World Bank (2007) World Development Report. Washington, DC.

World Bank World Development Indicator CD-ROM. Washington, DC.

Zon, A. V. and J. Muysken (2001) Health and Endogenous Growth. Journal of Health Economics, 169-185.

(1) For more details see Barro (1991), Mankiw, et al. (1992), Sachs and Warner (1997), etc.

(2) Human Development Report 2007-08.

Naeem Akram, Ihtsham ul Haq and Mohammad Khan are PhD Students of Federal Urdu University of Arts Science and Technology, Islamabad.

Authors' Note: The authors are highly indebted to Prof. Syed Nawab Haider Naqvi for his encouragement and providing research oriented environment at the FUUAST, Islamabad. The idea of this paper emerged when authors were working as research assistants with Ather H. Akbari who had obtained research grant for a health and education related project from Higher Education Commission of Pakistan. The authors thank Ather H. Akbari for guidance and permission to publish this research. The viewpoints presented in the paper are authors' own and do not represent the viewpoint of the affiliated institutions.
Table 1
Health Indicators

        Life Expectancy at Birth,    Infant Mortality Rate
Years        Total (Years)          (Per 1,000 Live Births)

1960              44                         139
1970              49                         120
1980              55                         110
1985              57                         105
1990              59                         100
1995              61                          93
2000              63                          85
2005              65                          79
2006              65                          78

Source: World Development Indicator.

Table 2
Description of Variables

Sr. No.   Variables                       Data Source

  1       Per Capita GDP                  WDI 2008
            (proxy for economic growth)
  2       Age Dependency                  WDI 2008
  3       Openness (Trade % of GDP)       WDI 2008
  4       Life Expectancy                 WDI 2008
  5       Infant Mortality Rate           WDI 2008
  6       Investment % of GDP             WDI 2008
  7       Secondary Enrolment             SBP Annual Reports
  8       Health Expenditure % of GDP     SBP Annual Reports
  9.      Population per Bed              SBP Annual Reports

Table 3
Results of ADF Test

                                             Level

Name of Variable           Intercept         Trend           None

Per Capita GDP           -1.55 (-2.97)   -0.16 (-3.56)   -1.05 (-1.95)
Age Dependency           -1.55 (-2.97)   -0.16 (-3.57)   -1.59 (-1.95)
Health Expenditure       -2.34 (-2.95)   -2.34 (-3.55)   -0.41 (-1.95)
Investment               -1.98 (-2.95)   -2.03 (-3.55)    0.88 (-1.95)
Life Expectancy          -2.02 (-2.95)   -2.33 (-3.55)    3.83 (-1.95)
Mortality Rate            1.21 (-2.95)   -2.32 (-3.55)   -1.79 (-1.95)
Openness                 -2.68 (-2.95)   -2.71 (-3.55)    0.40 (-1.95)
Population per Bed       -1.01 (-2.95)   -0.79 (-3.55)   -1.38 (-1.95)
Primary Enrolment         2.46 (-2.95)   -0.36 (-3.55)    4.90 (-1.95)
Secondary Enrolment       1.28 (-2.95)   -1.53 (-3.55)    4.28 (-1.95)

                                         1st Difference

Name of Variable           Intercept         Trend           None

Per Capita GDP           -5.66 (-2.95)   -5.59 (-3.56)    -2.8 (-1.95)
Age Dependency           -5.66 (-2.95)   -5.59 (-3.57)   -5.62 (-1.95)
Health Expenditure       -4.18 (-2.95)   -4.57 (-3.57)   -4.28 (-1.95)
Investment               -4.85 (-2.95)   -4.78 (-3.55)   -4.69 (-1.95)
Life Expectancy          -3.73 (-2.95)   -4.27 (-3.56)   -1.34 (-1.95)
Mortality Rate            -3.1 (-2.95)   -3.71 (-3.56)   -1.27 (-1.95)
Openness                 -5.70 (-2.95)   -5.59 (-3.55)   -5.73 (-1.95)
Population per Bed       -5.48 (-2.95)   -5.65 (-3.55)   -5.19 (-1.95)
Primary Enrolment        -2.96 (-2.95)   -3.70 (-3.55)   -1.89 (-1.95)
Secondary Enrolment      -4.34 (-2.95)   -4.68 (-3.55)   -3.08 (-1.95)

Values in parenthesis are MacKinnon critical values for rejection of
hypothesis of a unit root.

Table 4
Unrestricted Cointegration Rank Test (Trace)

Hypothesised                  Trace          0.05
No. of CE(s)   Eigenvalue   Statistic   Critical Value   Prob. **

None *          0.961769    407.4585       197.3709       0.0000
At Most 1 *     0.884889    299.7428       159.5297       0.0000
At Most 2 *     0.861482    228.4013       125.6154       0.0000
At Most 3 *     0.815939    163.1684       95.75366       0.0000
At Most 4 *     0.737361    107.3164       69.81889       0.0000
At Most 5 *     0.580019    63.19614       47.85613       0.0010
At Most 6 *     0.450058    34.56712       29.79707       0.0131
At Most 7       0.334958    14.83502       15.49471       0.0627
At Most 8       0.040786    1.374157       3.841466       0.2411

Trace test indicates 7 cointegrating eqn(s) at the 0.05 level.

* Denotes rejection of the hypothesis at the 0.05 level.

** MacKinnon-Haug Michelis (1999) p-values.

Table 5
Unrestricted Cointegration Rank Test (Maximum Eigenvalue)

Hypothesised                Max-Eigen        0.05
No. of CE(s)   Eigenvalue   Statistic   Critical Value   Prob. **

None *          0.961769    107.7157       58.43354       0.0000
At Most 1 *     0.884889    71.34145       52.36261       0.0002
At Most 2 *     0.861482    65.23291       46.23142       0.0002
At Most 3 *     0.815939    55.85203       40.07757       0.0004
At Most 4 *     0.737361    44.12022       33.87687       0.0022
At Most 5 *     0.580019    28.62902       27.58434       0.0366
At Most 6       0.450058    19.73211       21.13162       0.0775
At Most 7       0.334958    13.46086       14.26460       0.0667
At Most 8       0.040786    1.374157       3.841466       0.2411

Max-eigenvalue test indicates 6 cointegrating eqn(s) at the 0.05
level.

* Denotes rejection of the hypothesis at the 0.05 level.

** MacKinnon-Hauo-Michelis (1999) p-values.

Table 6
Normalised Cointegrating Coefficients

Variable              Coefficient   Std. Error   t-Statistic

AGE_DEPENDENCY         -18494.47     5501.205    -3.361895 *
OPENESS                118.7778      46.34173    2.563086 *
HEALTH EXPENDITURE     2209.714      1324.95      1.667769
POPULATION_PER_BED     -12.98682     2.976037    -4.363799 *
SECONDARY_ENROLMENT    0.004666      0.001038    4.493965 *
INVESTMENT             81.81509      77.32077     1.058126
LIFE_EXPECTANCY        526.8660      68.63043    7.676856 *
MORTALITY_RATE         153.2179      74.95079    2.044246 **
R-squared              0.989909
Adjusted R-squared     0.987292
S.E. of Regression     623.8845
Sum Squared Resid      10509261
Log Likelihood         -270.3802

* and ** Indicate significance at the 5 percent level and at 10 percent
level respectively.

Table 7
Error Correction Model Estimation

Variable                     Coefficient   Std. Error   t-Statistic

D(AGE_DEPENDENCY)             112965.4      43997.96     2.567515
D(HELATH_EXPENDITURE)         960.1742      990.0141     0.969859
D(OPENESS)                    49.40765      31.80442     1.553484
D(POPULATION_PER_BED)         -4.265212     2.434431     -1.752036
D(SECONDARY_ENROLMENT)        0.002938      0.001161     2.530879
D(INVESTMENT)                 -24.64672     73.67595     -0.334529
D(LIFE EXPECTANCY)            -41.1806      344.7947     -0.119435
D(MORTALITY_RATE)             -389.4584     445.1464      -0.8749
D(GDP_PER_CAPITA(-1))         0.277605      0.212884     1.304016
D(AGE_DEPENDENCY(-1))         -115873.5     42775.77     -2.708858
D(HELATH_EXPENDITURE(-1))     -387.314      1069.411     -0.362175
D(OPENESS(-1))                -34.21279     36.27845     -0.943061
D(POPULATION_PER_BED(-1))     1.093438      2.658591     0.411285
D(SECONDARY_ENROLMENT(-1))    0.000625      0.001467     0.426117
D(INVESTMENT(-1))             -19.73265     77.5046       -0.2546
D(LIFE_EXPECTANCY(-1))        -182.8239     455.2101     -0.401625
D(MORTALITY_RATE(-1))         115.7301      435.1011     0.265984
ECT(-1)                       -0.684606     0.238475     -2.87077

R-squared                     0.736754        Mean dependent var
Adjusted R-squared            0.438409        S.D. dependent var
S. E. of Regression           356.4104        Durbin-Watson stat
Sum Squared Resid              1905426          Log likelihood

Variable                     Probability

D(AGE_DEPENDENCY)              0.0214
D(HELATH_EXPENDITURE)          0.3475
D(OPENESS)                     0.1411
D(POPULATION_PER_BED)          0.1002
D(SECONDARY_ENROLMENT)         0.0231
D(INVESTMENT)                  0.7426
D(LIFE EXPECTANCY)             0.9065
D(MORTALITY_RATE)              0.3954
D(GDP_PER_CAPITA(-1))          0.2119
D(AGE_DEPENDENCY(-1))          0.0162
D(HELATH_EXPENDITURE(-1))      0.7223
D(OPENESS(-1))                 0.3606
D(POPULATION_PER_BED(-1))      0.6867
D(SECONDARY_ENROLMENT(-1))     0.6761
D(INVESTMENT(-1))              0.8025
D(LIFE_EXPECTANCY(-1))         0.6936
D(MORTALITY_RATE(-1))          0.7939
ECT(-1)                        0.0117

R-squared                     553.7899
Adjusted R-squared            475.5985
S. E. of Regression           2.352009
Sum Squared Resid             -227.7262
Gale Copyright: Copyright 2008 Gale, Cengage Learning. All rights reserved.