Sectoral Employment in Indonesia With Spatial and Seemingly Unrelated Regression (SUR) Modelapproach

Abstract

Employment becomes one of the most important focuses of development in Indonesia. Analysis of employment and its factors could be the consideration in making employment policies. Several studies of employment related to a particular economic sector have been carried out. For a comparison, this paper discussed the model of labor absorption with three economic sectors. The source of data was derived from all the provinces in Indonesia for five years. Spatial model was estimated with Maximum Likelihood Estimation (MLE) for each year of observation. Moran’s I and LM test were used to identify the spatial dependency. SUR model was estimated with Ordinary Least Square (OLS) and General Least Square (GLS). The variables used to estimate labor absorption were the output and real wage. The result indicated that the spatial dependency was significant particularly in the agricultural sector with a spatial error model. Meanwhile, labor absorption was significantly affected by the output and real wage for both OLS estimation and GLS estimation for SUR model. Service sector had the highest R2 value. UR model with GLS estimation was evidenced to be more efficient than OLS estimation, in addition, standard error of parameters using GLS estimation evenly was lower than OLS estimation.