Multicollinearity between different variables of human development index for better inclusive growth policy
’The transition of Indian economy from underdeveloped to developing economy traced various social economic problems which have not been conquered by an increased rate of economic growth. Apart from the inequality of income, various other benefits were not extended to vulnerable sections of the society like adequate education, improved health facility and better employment. Various policies have been implemented by the government to bridge the wide gap between the two groups of society and initiated a move towards human development from economic growth. The relationships of life expectancy, education, poverty, and safe water index have been evaluated with HDI. The relationship between HDI and its determinant variables has been investigated with the linear multiple regression model. The education index and safe drinking water have been found as the significant variables in the determination of HDI. These two variables also have the multicollinearity with other variables.
Key Words: Human Development Index, Inclusive Growth, Multicollinearity, Safe Drinking Water. Abbreviations: GDP: Gross Domestic Product, HDI: Human Development Index, SHG: Self Help Group, UNDP: United Nation Development Program.
The development of nation has been the prime objective of every economy. Over the last six and half decades the Indian economy has passed through various growth−phases, starting from backward economy with low growth rate to developing and sustainable economy. Now, the Indian economy has entered into a different orbit endorsed as high rate of expansion with combined objective of inclusive growth. The growth experiences of the Indian economy have been remarkable following positive trend from the second five year plan onwards. By overcoming the bane of British Government rule, the economy had achieved average 4.1 percent gross domestic product (GDP) growth rate in the foundational years that is 1951−1965. With the problem of food crisis and defensive actions the average growth rate of the economy declined a little in the period 1965−1975 at 3 percent. After this period, the Indian economy attained decadal average growth rate ranging from 5.6 percent to 8.2 percent.
A growth process must yield and ensure equal beneficial opportunities for all . But the assumption of economic growth practice, with an increase in industrial growth rate thereby increasing per capita income and ultimately uplifting standard of living of the masses at large, had been a failure which emphasized the issues of human capabilities. Human capabilities refer to ’ well−being’ of the general masses ( Kindleberger, 1965). It is a socio−economic issue which is not articulated in economic growth, for example, income disparity, illiteracy, unemployment, poverty, poor health facility, scanty living standard and low life expectancy ( Gasper and Staveren, 2010). Sen (1980)
identifies basic capabilities to achieve well− being for the general masses. The capabilities to be well−nourished and well−sheltered, to escape premature mortality, to be educated and in good health, and to be able to participate in social interaction without shame are some examples of basic capabilities. Various economists identified the capability approach in their own way. Alkire ( 2002) focused on ’capability to meet basic needs’, Robeyns (2001) defined fundamental capabilities and divided into basic and non−basic capabilities . To overcome these problems, government interfered to improve the living standard and socio−economic behavior of vulnerable sections of the society by implementing various policies. The government has made interference, firstly, by providing direct benefits for example subsidies, public distribution system (fair price shops), free education, free vaccination, and various policies on employment opportunities. Secondly, by providing assistance to the deprived but entrepreneurial section of the society for example Kisan Credit Card scheme for peasants, loans to small scale/ cottage industries or, self help groups (SHGs) through micro finance and banking institutions etc. The self reliance and enhancing the job opportunities are the major attributes of inclusive growth strategy as adopted by the government.
Human capital is implicated in the process of growth not merely as a cause but also as an effect of economic growth and contributes to economic development. The reciprocal relation between economic growth and the growth of human capital is likely to be an important key to sustained economic growth ( Mincer, 1995) .
To achieve economic development, it has to ensure that all segments of the society become part of this growth process; otherwise these socio−economic problems could derail the economy from growth (development) track. The term inclusive growth is often used interchangeably with a suite of other terms like ’broad−based growth’, ’shared−growth’, ’pro−poor growth’, and ’ equitable growth’ etc. The term inclusive growth is finding its way increasingly in the lexicon of government leaders, economists, planners, and academicians not just in India but even in pan−Asia
2. RATIONAL OF THE STUDY
Indicators of Human Development Index (HDI) such as literacy, education, maternal and infant mortality rates, show steady improvement, but the progress is slow and Indian economy is still behind several other Asian countries.
HDI is the aggregate result of the different variables of social development viz., poverty index, education index, life expectancy index, employment opportunities, water resources and GDP index etc. It means the performance of HDI dependents on the performance of individual index of poverty, education, life expectancy, employment opportunities, improved water resources etc. In the present study, the relation of dependency has been investigated between HDI and individual index performance of various variables.
3. METHODOLOGY OF RESEARCH
In the present study, the influence of five predictors viz., life expectancy, education, poverty, safe water availability and GDP index have been considered to analyze the dependency of HDI. The individual factor performance also influences the other factor, for example improvement in the poverty index shows that there is improvement in the economic condition of the masses and increase in the accessing capability of health and sanitation facilities ultimately increases the life expectancy of the people. Similar correlation could be found in cases of other variables also. The significant correlation between the predictor variables creates the problem of multicollinearity. To evaluate the correlation and regression of predictor and dependent variables, multiple− regression analysis has been used to analyze the dependency of HDI on these predictor variables. The statistical equation of the linear model of multiple−regression for four predictors (life expectancy index, education index, poverty index and safe water availability index) and one dependent variable (HDI) is as follows:
Y= β0+ β1X1+ β2 X2 + β3 X3 + β4 X4 + β5 X5 + ε
Where Y = Value of dependent variable (HDI)
= a constant, the value of Y when all Xi values are zero. = the slope of the regression surface ( It represents the regression coefficient associated with each Xi) X1 = life expectancy index. X2 = education index. X3 = poverty index. X4 = safe drinking water index. X5 = GDP index
= an error term, normally distributed about a mean of 0.
For the purpose of present study ten years data have been collected on life expectancy index, education index, poverty index and safe water index starting from the year 1998 to 2007. The data have been collected from the United Nation Development Program (UNDP) reports on Human Development. The data pertaining to Indian economy have been considered for the above said period. The following vectors set the relationship of regression equation as discussed above for all four variables with their coefficients.
The above stated vectors are representing the each equation of multiple regression model for ten years index values. The values for ß0 will be equal to 1, used to obtain the regression constant. The values for remaining ßi pertaining to four predictor variables and contain scores (coefficients) for these subjects. The log values of HDI and predictor variables have been considered to generalize the actual values.
4. MULTIPLE REGRESSION BETWEEN HDI AND PREDICTOR VARIABLES: EMPIRICAL FINDINGS AND ANALYSIS
The results on multiple−regression have been presented in the following tables. Table 1 shows the correlation between the HDI and predictor variables. Table 2 reveals the predictor variables those which have been either entered or removed for the analysis. Table 1 states the model summary with the help of residual sum square. Table 2 demonstrates the coefficients and collinearity statistics, and Table 3 shows the collinearity diagnostic.
Table 1 states the model summary. The value of R−Square for education index is 0.845 and for safe water availability index is 0.941. Similarly, R−Square of variables LI, PI, and GI is 0.128, 0.112, and 0.903, respectively. The value of Adjusted R−square shows the percent of variation. It shows that 81.9 percent variation has been explained by the education index and 91.7 percent variation has been explained by the both education index and safe water availability index. All the variables are statistically significant. The value of R−square is close to 1 which states the robustness of the model.
Table 2 reveals the standardized coefficients of beta values. It also shows the collinearity statistics as tolerance and variance inflated factor (VIF) . The coefficient of education index and safe drinking water index are significant at 0.01 and 0.05 level. Tolerance shows the percent of the variance in a given predictor that cannot be explained by other predictors. The small value of tolerance shows that variance of a given predictor is explained by other predictors, and large value shows that variance is explained by the variable itself. The values of coefficients are significant. The education index coefficient is significant at the 0.01 level and safe water availability index coefficient is significant at the 0.05 level. The tolerance values are close to ’1’ it means very less percentage of variance of education and safe water availability index is shown by other predictors. This is the reason of selecting these two predictors in the model.
The coefficient of all the variables is not significant which shows that these variables play a less important role in the determination of HDI as compared to education index and safe water availability. The collinearity statistics tolerance shows that around 30 to 85 percent of the
information is explained by the education index and safe water availability index as tolerance score ranges from 0.142 to 0.708. The least value of tolerance of GDP index leads to inflated value of variance of poverty index and shows the problematic situation of collinearity.
Table 3 reveals the diagnostics for the multiple correlated variables. It shows that eigenvalue of the variables education index and availability of safe water is around one and condition index is below 5 which states that these two variables are least influenced by each other. From the above discussion it can be concluded that education index is the most important predictor of HDI. It shows that an increase in education among the general masses increases the awareness of health, career, better employment opportunities, and expectation of good living conditions that result into the improvement in poverty situation. The increasing awareness about health among the general masses, the accessibility of safe drinking water is also a predictor of improved human development index. The results also indicate that the safe drinking water ensures the good health of the general masses. As an implication, government should take effective measures to ensure successful education system and availability of safe drinking water.
5. FINDINGS AND CONCLUSION
It has been observed that the education index also represents other indexes which mean that HDI could be improved with the improvement in the education index. Awareness about health, increased production, arising issues of externalities and economic growth are some of the benefits of educating masses of population ( Nichloas, 1987). Broadly human capital is related with knowledge and skills embodied in humans that are acquired through schooling, training and experience are useful in the production of goods, services and further knowledge. Plausibly education is key factor to human capital and also supplemented by health conditions (Shekhar, 2006).
Barro and Martin (1995) estimated a log linear relationship between years of education and annual wage income. They suggest that an additional year of schooling increases wages at the individual level by around 6.5 per cent in European Union . The positive role of the state in education can overcome problems associated with human well−being, both, economic and social. The education should not just be left
either to the whims of market mechanism rather State should interfere to increase human capabilities (Deitz and Cypher, 2009). However, there are number of policy measures which already have been undertaken by the government to ensure better education index, for example, Sarva Shiksha Abhiyan, free elementary education, and scholarships for female child etc. Besides increasing the number of primary school enrolment, government should formulate the policies which could sustain the primary school enrolment percentage and be helpful for increasing the percentage of secondary and higher education (Tilak, 2007).
Safe drinking water availability index is the second most important predictor variable for HDI. It is an important variable because it has not been presented by other predictor variables. In simple words, it has no significant correlation with other variables. Tadaro ( 1998) has estimated that more than one billion people world− wide have no access to clean water and an additional one billion live in areas with chronic water shortages. And many of the poor collect drinking water from rivers and canals that are polluted with human excreta and chemical which is contributing to the spread of diseases.
The increased life expectancy might not ensure the availability of safe drinking water rather safe drinking water availability may ensure the better life expectancy.
The water quality monitoring is now being considered as an important part of the government program. Since the year 2000, water quality monitoring has been accorded a high priority and institutional mechanisms have been developed at national, state, district, block and panchayat levels. The government has also outlined requisite mechanisms to monitor the quality of drinking water and devise effective Information, Education and Communication (IEC) interventions to disseminate information and educate people on health and hygiene. The Government of India launched the National Rural Drinking Water Quality Monitoring and Surveillance Program in February 2006. This envisages institutionalization of community participation for monitoring and surveillance of drinking water sources at the grassroot levels by Gram Panchayats and Village Water and Sanitation Committees, followed by checking the positively tested samples at the district and state level laboratories. One major problem when it comes to addressing the problems related to water is that the provisions for water are distributed across various ministries and institutions. With several institutions involved in water supply, inter−sectoral coordination becomes critical for the success of the program.
The trends of government budget on health and family welfare has a significant role in the improvement of various predictor variables of HDI. As an implication to the government and policy maker, it is suggestive that the education percentage should be increased by implementing more feasible policies. The initiative taken for availability of safe drinking water should be continued. The employment opportunities generated under the social services are not only solving the problem of unemployment rather these are also helpful to increase the standard of living because the major aims of these employment opportunities are developing hygiene conditions around the living areas and self−development of the vulnerable sections of the society.
1. Planning Commission of India ( 2008), Eleventh Five Year Plan, 2007−2012, Vol.I.
2. Kindleberger, C. P. (1965), Economic Development. (2nd Edition), ch.1.
3. Gasper, Des and Staveren, Irene Van, Development as Freedom and as what else? Edited in, Aggarwal, Bina., Humphries, Jane and Robeyns, Ingrid (2010) Capabilities, Freedom and Equality: Amartya Sen’s Work
from a Gender Perspective, Oxford University Press, New Delhi. 4. Sen, Amartya (1980), Equality of what? In The Tanner Lecture on Human Values, edited by S. McMurrin. Cambridge: Cambridge University Press. 5. Alkire, Sabina (2002), Valuing freedom: Sen’s capability approach and poverty reduction. Oxford University Press, Oxford. 6. Robeyns, Ingrid (2001), Understanding Sen’s capability approach. Available from http:// www.ingridrobeyns.nl. 7. Mincer, Jacob (1995), Economic Development, Growth of Human Capital and the Dynamices of the Wage Structure, Journal of Economic Growth, Vol. 1, pp. 29− 48. 8. Eleventh Five Year Plan (2007−12), Inclusive Growth, Planning Commission, Government of India, Vol. 1, 2008. 9. Stevens, James P. ( 2009), Applied Multivariate Statistics for the Social Sciences, Taylor & Francis, USA, pp. 63− 93 10. Multiple regression is used as a descriptive tool in three types of situations. First, it is often used to develop a self−weighting estimates equation by which to predict values for a criterion variables from the predictor variables. Second is used for controlling for confounding variables to better evaluate the contribution of other variables. The third type referred to describe an entire structure of linkages that have been advanced from casual theory. 11. The variance inflation factor (VFI) for a predictor indicates whether there is a strong linear association between it and all the remaining predictors. Myers (1990) offered the following suggestion: "Though no rule of thumb on numberical values is foolproof, it is generally believed that if any VIF value exceeds 10, there is reason for at least some concern; then one should consider an alternative to least squares estimation to combat the problem". 12. Barr, Nicholas (1987), The Economics of the Welfare State, Weidenfeld and Nicolson, (Londan), pp. 288−293. 13. Kumar, C. Shekhar (2006), Human Capital and Growth Empirics, The Journal of Developing Areas, Vol. 40, No. 1, pp. 153− 179. 14. Barro and Sala−I−Martin (1995), Technological Diffusion, Convergences and Growth, NBER Working Paper, 5151. 15. Dietz, James L. and Cypher, James M. ( 2009), The Process of Economic Development, Routledge (USA), pp. 391− 415. 16. Tilak, Jandhyala B. G. ( 2007), Post− elementary Education, Poverty and Development in India, International Journal of Educational Development, Vol.27, issue 4, pp. 435−445. 17. Tadaro, Michael P. (1998), Economic Development, Addison Wesley Longman (USA), p.365. 18. Chandraiah, C. Ramesh (2001), Drinking Water as a Fundamental Right, Economic and Political Weekly, Vol. 36, No.8, pp. 619− 621. 19. Rober J. Tata and Ronald R. Schnltz (1988), World Variation in Human Welfare: A New Index of Development Status, Annals of the Association of America Geographers, Vol. 78, No.4, pp.580−593.