«AN ECONOMETRIC ANALYSIS OF THE DETERMINANTS OF EDUCATIONAL EXPENDITURES IN THE URBAN AREAS OF DISTRICT PESHAWAR Zahoor Khan1,Zulfiqar ...»
AN ECONOMETRIC ANALYSIS OF THE DETERMINANTS OF
EDUCATIONAL EXPENDITURES IN THE URBAN AREAS OF
Zahoor Khan1,Zulfiqar Ali2,Asmatullah3
The present study has been undertaken in the year 2009 to find out the key determinants
of educational expenditures. For this purpose three sample areas namely, Hayatabad, Tehkal and Cantonment area were selected. A sample of 150 respondents was selected on the basis of systematic random sampling technique. Fifty (50) respondents were selected from each sampled area. A structured questionnaire was then used to collect first hand information from the respondents about their expenditures on education. For analysis of the data Ordinary Least Square (OLS) method (ANCOVA model) and binary logit maximum likely hood models were used to find out the determinants of expenditures on education. The findings of the study show that 80% of the respondents do high expenditure on education due to there high level of income and high expected rate of returns from education. Further more regression analysis revealed that if income per household increases by one unit (Rs. 1000) on average Educational expenditure will increase by 0.456 units ( Rs. 456) per household. The binary logit model revels that households with positive demonstration effect are likely to increase their educational expenditure by more than 23.77 % than the household with zero demonstration effect.
The Mc Fadden R-square =0.56 which a measure of goodness of fit, used for qualitative response models. The probability value of the likely hood ratio (LR) Statistic is 0.00012 which shows that all the parameters of the models are significantly different from zero Based upan the findings of the study it is recommended that job opportunities should be provided to unemployed educated people, for increasing the level of income and high expected rate of return from education, so as to increase education level in this area.
1. Lecturer in Economics, Department of Economics University of Peshawar.
Zahoor_660@hotmail.com phone: +92 91 9216733.
2. Lecturer in Economics in Government Postgraduate College Nowshera.
INTRODUCTIONIt is generally believed that education is one of the basic rights of every human being, irrespective of sex, age, creed, religion, etc. Education is key to economic development.
It refers to primary, secondary and College/University (vocational, technical and nontechnical) level education. In the present economic situation, provision of education has become very expensive. Huge expenditures have been made on fees, hostel charges, transportation, uniforms, tuitions, books, teachers training and provision of infrastructure.
These expenditures increase with personal income, demographic factors, occupation type, area of residenceand family background (Jandhyala, 2002).
In developing countries, household level of income and expected return from education are significant determinants of any increase in expenditure on education (Zaidi, 2006).
Dostie & Rajshri (2006) studied the determinants of school enrollment in India. They found that parental education, wealth, village caste composition and aggregate deprivation influence individual enrollment decisions. Aslam (2006) studied effects of parental education on children educational expenditure. He found that parental education has a robust and positive effect on child learning, a result that is often attributed to more educated parents making greater investments in their children's human capital. It is found that parents that are more educated make greater educational investments in both goods and time and that these relationships are generally robust to a rich set of controls. He suggested that making greater investments in both goods and time stems both from higher expected returns to education for children and from different preferences for education among more educated parents. Moreover, he found that the marginal effect of mother's education on educational investments is greater than father education. Kirchsteiger and Sebald (2006) studied the behavior of educated in making investment in their children education. They found that parents with higher levels of education generally also attach a higher importance to the education of their children. This implies an intergenerational chain transmitting the attitude towards the formation of human capital from one generation to the next. They also found that privately provided create inefficiency in economic system. Therefore, they suggested a permanent public subsidy for education to overcome this inefficiency in education provision. Deon (2004) studied the household decision of investment in human capital. He found that the decision is based on benefit and cost of investment in human capital. Benefits are associated with higher productivity.
However, the return differs for boys and girls. Cost of education are direct like user fees, transport costs, textbook fees, drug costs and indirect (children often contribute to household income by working inside or outside the home). Since richer families could cope better with these costs, they made huge investment in human capital. Investments in the human capital of children were also sensitive the allocation of power within households: Families in which the bargaining power of women is stronger tend to invest more in health and education. Daughters’ education might be less valuable to parents if sons look after them in old age, so parents might be less willing to send girls to school.
Nasir and Nazli (1995) examined the role of education, technical training, school quality and literacy and innumeracy skills on the earnings of wage earners and salaried persons in Pakistan by using the data of PIHS (1995-96). They found that education has a positive effect on earning of labour. In addition, the effect of literacy and innumeracy skills was observed to be large and significant. The returns were 15 percent higher for those who had all three skills as compared to those who did not possess any of these skills. The impact of technical training and private schools was found to be positive and significant.
The present study is different from all the rest of studies conducted previously, because it focuses on the analysis of the factors, which are responsible for the expenditures on education and bringing change in the behavior of the household towards education. The research will provide sound information for the educationists and policy makers.
MATERIAL AND METOHDSThe present study has been conducted in year 2009 to find out key determinates of educational expenditure in the urban areas of district Peshawar. The universe of the study is the city of Peshawar. Three areas of Peshawar namely Cantonment Area, Tehkal and Hayatabad are selected. A structured questionnaire is used for collecting primary data on different variables such as income of households’ number of children who attend, educational institutions, expected rate of return from education and easy availability of educational facilities. Cross sectional data set has been used about the mentioned variables. Regression technique (ANCOVA model) has been used to quantify the impact and role of the above-mentioned variables in the determination of educational expenditures. There are other variables which may affect the educational expenditures but here the scope of the study is confined to the mentioned variables only. The ANCOVA
model is specified as:
EdExp = Educational expenditures per household.
IPHH = Income per household.
D1 = Dummy variable used for expected rate of return from education.
D2 = Dummy variable used for easy availability of educational facilities.
ei = usual residuals.
The mentioned variables have very strong correlation with educational expenditures and used by many prominent educationists the worth mentioning of them are Jandhyala, B.G Tilak (2002), Jandhyala, B.G Tilak (1994).For advanced empirical investigation the MAXIMUM LIKELIHOOD BINARY LOGIT (ML-Binary Logit) model for individual data has been used. The estimation of the model will help us in calculating odd ratios which will enable us to assign a chance/ probability to each regressor that in turn will help us to investigate a unit increase in a given regressor (holding other regressors constant) will change the estimated Logit.
MAXIMUM LIKELIHOOD BINARY LOGIT (ML-BINARY LOGIT) MODEL
As the conventional R2 is particularly meaningful in binary regressand model therefore finally we will compare the actual and fitted values of the regression model with each other to find out how many correct predictions are there out of total number of observations. For
this end we shall use Count R2:
Count R2 = Number of correct prediction /Total number of observations.
Since the regressand in the Logit model takes a value of 1 or zero, if the predicted probability is greater than 0.5, we classify that as 1, but if it is less than 0.5, we classify that as 01. To test the null hypothesis that all slope coefficients are simultaneously equal to zero the equivalent of F test is likelihood ratio (LR) Statistic for binary logit model has been used.
Basic econometrics 4th edition Tata McGraw-Hill publishing company limited (New Dehli) pp. 618-619.
RESULTS AND DISCUSSIONThe estimates of the regression model suggest that if income per household, expected rate of return (D1) and easy availability of educational facilities (D2) become equal to zero then average educational expenditures per household in urban areas of district Peshawar is Rs. 3458. Average educational expenditures with zero educational facilities and zero income per household but high-expected rate of return will be Rs. 7708. Easy availability of educational facilities has significantly negative effects on the educational expenditures of the households. If easy availability of educational facilities exists as a result average educational expenditure per household decrease by Rs. 535. Regression result revels that t and p-values of the estimates are very large and small respectively, based on these criteria we can conclude that all the explanatory variables are statistically significant. The value of R-square =0.67 which shows that 67% of the total variation in dependent variable (Educational expenditures) are being explained by the explanatory variables (IPHH,D1 and D2). Further more the R-square suggests that the model is a good fit.The value of DW-Test is approximately 2 (2.11) which shows that there is no severe serial or autocorrelation.The probability value of F-statistic is extremely low which shows that the model is over all significant.
CONCLUSION AND RECOMMENDATIONS
Major findings of the study can be summarized as under:
Regression analysis revealed that if income per household increases by one unit (Rs.
1000) on the average educational expenditures will increase by 0.456 units ( Rs. 456) per household. The findings of the study show that 80% of the respondents do high expenditures on education due to there high level of income and high expected rate of returns from education. The study further shows that income per household, in the
sampled areas is Rs: 43460 and average educational expenditure per household is Rs:
9105. The results of econometric model show that easy availability of educational facilities has significantly negative effects on the educational expenditures of the households. If easy availability of educational facilities exists, as a result average educational expenditure per household decrease by Rs. 535.
From the Binary logit model we can deduce some important results. The model revels that all the regressands except D1 have positively related with the estimated logit. The parameters in term of odds can be interpreted meaningfully. The coefficient of D2 is
3.21by taking the anti log (e3.21) = 23.77 shows that households with positive demonstration effect are likely to increase their educational expenditure by more than 23.77 % than the household with zero demonstration effect. The Mc Fadden R-square =0.56 which a measure of goodness of fit, used for qualitative response models. The probability value of the likely hood ratio (LR) Statistic is 0.00012 which shows that all the parameters of the models are significantly different from zero. The value of Count R2 = Number of correct prediction /Total number of observations =123/150 = 0.82.
Based on the findings of the study it is recommended that job opportunities should be provided to unemployed educated people, for increasing the level of income and high expected rate of return so as to increase education level in this area. To tackle the problem of low returns from education, we should bring awareness in general public about the importance of education and existing job opportunities in Khyber Pukhtoon Khwa in particular and in Pakistan in general. The relationship of availability of educational facilities and educational expenditures can remedy if excellent educational institutions are provided in that particular district. In that case students will get admissions in the nearest educational institutions.
Asalm. (2006). Public Education Expenditure and Other Determinants of Private Investment in Pakistan.Available at SSRN. http:// ssm.corn.come/abstract=883864 Deon. F. (2004). Determinants of Health and Education Outcomes: Background Note for World Development Report 2004: Making Services Work for Poor People.
Dostie, Benoit, & Jayaraman, Rajshri. (2006). Determinants of School Enrollmenin Indian Villages. Economic Development and Cultural Change. 54 (2), 405-21 Jandhyala, B. G. Tilak. (1994). Relationship between Education Expenditures, Social and Human Capital. Education in Rural India. NCAER working paper No. 56 Jandhyala, B. G. Tilak. (2002). Determinants of Household Expenditures on School Education in Rural India. NCAER working paper No. 88. (2002).
Khan, S, Mohsin. (2005). Human Capital and Economic Growth. The Pakistan Development Reviews, 44(4), 455-478.