New Model Averaging on Household Income to Examine Poverty In Malaysia

Siti Aisyah Mohd Padzil, Khuneswari Gopal Pillay, Rohayu Mohd Salleh, Sya Sya Avtar, Azme Khamis

Household income for Malaysian vary due to many factors as Malaysian’s are multiracial and multi religious. Malaysia is divided into two regions which are east and west Malaysia. Due to different geographical area, there’s a contrast on the development between these two regions. Even though the rate of poverty in Malaysia had been reduced yearly, the poverty rate of Malaysian still has not been eradicated. This research aims to examine the consequences of households differences (age, gender, marital status, education, activity and family size), states, region and net income to the cause of poverty among Malaysian. Model Averaging (MA) and New Model Averaging (NMA) will be compared based on 𝐴𝐼𝐶𝐶 and 𝐵𝐼𝐶 model selection criteria. The performance of best model from each method are evaluated using accuracy measures such as residual, RMSE, MSE and MAE. The analysis indicates that computing weights based on 𝐴𝐼𝐶𝑐 produce a slightly better performance model for both MA and NMA approach. Model formed using NMA based on 𝐴𝐼𝐶𝑐 was selected as the best model as it has minimum error. In conclusion, the best model summarizes that state, household differences as well as net income are the causes of poverty among Malaysian. There’s a positive relationship between household sizes with probability of poor which means, risk of poverty will increase as household size increases.

Volume 11 | 12-Special Issue

Pages: 137-144