Bivariate Binary Probit Model for Modeling Human Development
Index ( HDI ) and Poverty Index
(Case Study : in East Java 2012)
Name : Merly Fatriana Bintariningrum
NRP : 1313201907 Supervisor : Dr. Vita Ratnasari, M.Si
ABSTRACT
Indicators that often used by countries in the world to measure the
progress of the country based on the results of development is the Human Development Index ( HDI ) . While the poverty index is considered as a failure to achieve or the shortage forming element of HDI. Indonesia is a developing country which has a very large number of people in poverty. Areas that have high poverty index tend to have a low HDI. East Java has the largest number of people in poverty in Indonesia, and also has low HDI. To minimize poverty and improve HDI need to know the appropriate model of the two variables in order to identify the factors that influence them. Probit model regression is a model to explain the pattern of the relationship of a categorical response variable. Bivariate probit model is a binary probit model involving two response variables, each of which has two categories. In bivariate binary probit model, it is assumed that has relationship between response variable. Estimation of binary bivariate probit model parameters using maximum likelihood estimation (MLE) and Newton Raphson iteration . To get the best model based on the smallest value of AIC ( Akaike Information Criterion ). Based on the best model obtained, the variables that significantly influence the HDI and the poverty index are variable school participation rates and economic growth.
Keywords : AIC, Bivariate Binary Probit Model,Human Development Index