Addressing Corner Solution Effect for Child Mortality Status Measure: An Application of Tobit Model

Abstract

Although various models regarding corner solution data have been suggested, it is known that corner solution data is too complicated to predict the model. A corner solution is a unique solution in the sense that it is characterized by a mass-point at zero and a long right tail. For the present study, the data was taken from MICS collected by Punjab Bureau of Statistics (BOS) during 2011-2012. However, this study is based on the particular data. The child mortality within five-year interval was taken as a response variable. In the data obtained, 65% of the response variable observed to be zero. Tobit model reveals that it is the most accurate and widely used technique in this regard. The implementation of Tobit model shows that the significant predictor of child mortality in a family within five year are women age, postnatal checkups, antenatal care, total children ever born, and wealth index score. Since the majority (i.e. 74.4%) of mother age in the data is between 20 to 35 years. The results indicate that the chances of child survival are relatively higher for the mothers of age group of 20-35 years. Furthermore, Child mortality is highest for the families with low wealth index score. In addition, child mortality percentage tends to decrease with the improvement in postnatal and antenatal care facilities. The specification of Tobit model also used to verify the result of the Tobit model.