Pemodelan Produk Domestik Bruto (PDB) dengan Metode Seemingly Unrelated Regression (SUR) di Indonesia, Malaysia, Singapura, Filipina, dan Vietnam
Abstract
Abstract. Seemingly Unrelated Regression (SUR) is a system consisting of two or more linear regression equations that can be used to address violations of assumptions in panel data regression, namely heteroskedasticity in the variance of residuals and the presence of correlation between residuals of different individuals, also known as contemporaneous correlation. The independent variables used are Foreign Direct Investment (FDI), inflation, and the labor force, while the dependent variable is the Gross Domestic Product (GDP) in Indonesia, Malaysia, Singapore, Philippines, and Vietnam from 2000 to 2022. In the assumption check for the best panel data regression model, the Random Effect Model (REM) with different intercepts and slopes for each country, heteroskedasticity and contemporaneous correlation were found. Therefore, an alternative method, SUR, was used. Based on the SUR model, it can be concluded that GDP in Indonesia and the Philippines is influenced by FDI and the labor force. GDP in Malaysia is only influenced by the labor force. Additionally, GDP in Singapore and Vietnam is influenced by all independent variables, namely FDI, inflation, and the labor force. The McElroy's coefficient of determination (R2) for the SUR model is 97.68%, and the coefficient of determination for REM is 96,49%. This means that, because the R2 value of the SUR model is closer to 100%, it can be said that the SUR model is better to apply than the REM.
Abstrak. Seemingly Unrelated Regression (SUR) adalah sistem yang terdiri dari dua atau lebih persamaan regresi linear yang bisa digunakan untuk mengatasi pelanggaran asumsi pada regresi data panel yaitu heteroskedastisitas pada ragam residual dan terdapat korelasi antar residual pada individu yang berbeda atau yang disebut korelasi kesebayaan (contemporaneous correlation). Variabel independen yang digunakan yaitu Foreign Direct Investement (FDI), inflasi, dan angkatan kerja, serta variabel dependen yaitu Produk Domestik Bruto (PDB) di Negara Indonesia, Malaysia, Singapura, Filipina, dan Vietnam tahun 2000-2022. Dalam pemeriksaan asumsi pada model regresi data panel terbaik, yaitu Random Effect Model (REM) dengan intercept dan slope berbeda untuk setiap negara, ditemukan adanya heteroskedastisitas dan korelasi kesebayaan. Oleh karena itu, digunakan metode alternatif lain yaitu SUR. Berdasarkan model SUR, dapat disimpulkan bahwa PDB di Indonesia dan Filipina dipengaruhi oleh FDI dan angkatan kerja. PDB di Malaysia hanya dipengaruhi oleh angkatan kerja. Serta PDB di Singapura dan Vietnam dipengaruhi oleh semua variabel independen yaitu FDI, inflasi, dan angkatan kerja. Diperoleh juga koefisien determinasi (R2) McElroy’s untuk model SUR sebesar 97,68%, dan koefisien determinasi untuk REM sebesar 96,49%. Artinya karena nilai R2 dari model SUR lebih mendekati 100%, maka dapat dikatakan bahwa model SUR lebih baik untuk diterapkan daripada REM.
References
Gujarati, D.N. (2003). Basic Econometrics. (Fourth Edition). New York: McGraw-HiII.
Zellner, Arnold. (2006). Seemingly Unrelated Regression [Internet], H.G.B. Alexander Research Foundation, University of Chicago. Available from: https://www.researchgate.net/publication/226411188_Seemingly_Unrelated_Regressions
Baltagi, B.H. (2005). Econometric Analysis of Panel Data. (Third Edition). New York: John Wiley & Sons Inc.
Savitri, C., S.P. Faddila, Irmawartini, H.R. Iswari, C. Anam, S. Syah, S.R. Mulyani, P.R. Sihombing, E.R. Kismawadi, A. Pujianto, A. Mulyati, Y. Astuti, W.C. Adinugroho, R. Imanuddin, Kristia, A. Nuraini, and M.T. Siregar. (2021). Statistik Multivariat dalam Riset. Bandung: Widina Bhakti Persada Bandung.
Jannah, W., I.D. Sulvianti, Indahwati, P. Silvianti, & A. Kurnia. (2020). Analysis of Credit Bank Distribution with Seemingly Unrelated Regression Method on Panel Data. Journal of Physics: Conference Series, 1863. DOI: 10.1088/1742-6596/1863/1/012057.
Dufour, J.M. & L. Khalaf. (2002). Exact Tests for Contemporaneous Correlation of Disturbances in Seemingly Unrelated Regressions. Journal of Econometrics, 106(1). https://doi.org/10.1016/S0304-4076(01)00093-8
Halunga, A.G., C.D. Orme, & T. Yamagata. (2017). A Heteroskedasticity Robust Breusch–Pagan Test for Contemporaneous Correlation in Dynamic Panel Data Models. Journal of Econometrics, 198(2). https://doi.org/10.1016/j.jeconom.2016.12.005
Greene, W.H. (2002). Econometric Analysis. (Fifth Edition). New York: Macmillan Publishing Company.
Srihardianti, M., M. Mustafid, & A. Prahutama. (2016). Metode Regresi Data Panel untuk Peramalan Konsumsi Energi di Indonesia. Jurnal Gaussian, 5(3). https://doi.org/10.14710/j.gauss.5.3.475-485
McElroy, M.B. (1977). Goodness of fit for seemingly unrelated regressions. Journal of Econometrics, 6. https://doi.org/10.1016/0304-4076(77)90008-2
Andini, F.N., & Wachidah, L. (2023). Penerapan Regresi Spasial Panel Random Effect pada Kasus Kemiskinan di Provinsi Jawa Tengah Tahun 2011-2020. Jurnal Riset Statistika, 3(1). https://doi.org/10.29313/jrs.v3i1.1885
Agustin Nuriani Sirodj D, Made Sumertajaya I, Kurnia A. Analisis Clustering Time Series untuk Pengelompokan Provinsi di Indonesia Berdasarkan Indeks Pembangunan Manusia Jenis Kelamin Perempuan [Internet]. Vol. 23. 2023. Available from: https://www.bps.go.id/indicator/40/462/1/indeks-pembangunan-manusia-ipm-menurut-
Yulianti NA, Cahyawati D, Susanti E, Jurusan ), Fakultas M. Penggunaan Metode Double Exponential Smoothing Tipe Holt pada Peramalan Kasus Covid-19 di Provinsi Sumatera Selatan [Internet]. Vol. 23. Available from: http://corona.sumselprov.go.id/.
Esra R, Nohe DA, Fathurahman M, Program ), Fakultas SS. Pemilihan Model Terbaik pada Generalized Poisson Regression Menggunakan Akaike Information Criterion. Vol. 23. 2023.
