“…The last difference is that the prior studies employ the stochastic frontier analysis (Effendi, 2014), the estimated generalized least square (EGLS) (Bato, 2014;Effendi, 2014;Hermawan, 2011;Lembang & Pratomo, 2013;Rizal, 2018), the multiple regression model (DiRienzo & Das, 2020), the pooled least square or the common effect model (Bato, 2014;Effendi, 2014;Hermawan, 2011;Lembang & Pratomo, 2013;Rizal, 2018;Suryanta, 2012;Waristi, 2014), the fixed effect model (FEM) (Aryani et al, 2020;Mahendra & Solikin, 2019;Purnamasari et al, 2020;Ridwannulloh & Sunaryati, 2018;Waristi, 2014) or the random effect model (REM) (Ambarita & Sirait, 2020;Lembang & Pratomo, 2013;Mahendra & Solikin, 2019;Ridwannulloh & Sunaryati, 2018;Rizal, 2018;Waristi, 2014), in which studies that use FEM and REM do not estimate the difference on the constant model and the random effect within countries being studied. Meanwhile, this study applies the mixedeffect model, which enables us to differentiate the constant model and the regression coefficients within the studied countries, both for variables with constant and various values within time.…”