Pemodelan Middle Income Trap Di Indonesia Dalam Kajian Subnasional Pada Tingkat Provinsi Menggunakan Metode Regresi Probit Panel

Rozi, Fatchur (2026) Pemodelan Middle Income Trap Di Indonesia Dalam Kajian Subnasional Pada Tingkat Provinsi Menggunakan Metode Regresi Probit Panel. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221034-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Dalam perekonomian global, World Bank mengklasifikasikan negara – negara yang ada di dunia berdasarkan Gross National Income (GNI) atau Produk Domestik Bruto (PDB), menjadi empat kelas pendapatan yakni Low Income, Lower-Middle Income, Upper-Middle Income dan High Income. Dari pengklasifikasian tersebut munculah istilah Middle Income Trap. Fenomena Middle Income Trap (MIT) merupakan tantangan pembangunan yang dihadapi ketika pertumbuhan ekonomi terhambat pada level pendapatan menengah dalam periode panjang. Kondisi ini mencerminkan perlambatan daya saing ekonomi nasional maupun daerah, sehingga penting untuk dikaji secara lebih mendalam pada tingkat provinsi. Penelitian ini menganalisis status MIT provinsi di Indonesia periode 2014–2024 dengan menggunakan PDRB per kapita yang dikonversi melalui metode Atlas sebagai dasar penentuan Status MIT. Variabel penjelas yang digunakan meliputi Angka Harapan Hidup (AHH), Pembentukan Modal Tetap Bruto (PMTB), Tingkat Pengangguran Terbuka (TPT), Angka Partisipasi Kasar (APK), dan Rata-rata Lama Sekolah (RLS). Analisis dilakukan melalui statistika deskriptif, pemeriksaan multikolinearitas, pemodelan regresi probit biner data panel, pengujian signifikansi parameter, perhitungan efek marginal, serta evaluasi kinerja klasifikasi. Hasil pemilihan model menunjukkan bahwa pendekatan Random Effect lebih sesuai, dan setelah proses backward elimination diperoleh model terbaik dengan variabel signifikan PMTB, TPT, APK, dan RLS. Secara umum, PMTB dan RLS meningkatkan peluang provinsi tidak terjebak MIT, sedangkan TPT dan APK menurunkannya. Evaluasi model menunjukkan akurasi klasifikasi sebesar 70,07%, sensitivitas 11,71%, dan spesifisitas 92,89%, dengan nilai AUC ROC sebesar 0,7311. Temuan ini menegaskan pentingnya penguatan investasi produktif dan peningkatan kualitas pendidikan, sekaligus pengendalian pengangguran, untuk memperbesar peluang provinsi bertransisi menuju kelompok pendapatan lebih tinggi.
======================================================================================================================================
In the global economy, the World Bank classifies countries based on Gross National Income (GNI) or Gross Domestic Product (GDP) into four income classes: Low Income, Lower-Middle Income, Upper-Middle Income, and High Income. From this classification, the term Middle Income Trap emerged. The Middle Income Trap (MIT) phenomenon is a development challenge faced when economic growth is hampered at the middle income level over a long period. This condition reflects a slowdown in national and regional economic competitiveness, making it important to study it in greater depth at the provincial level. This study analyzes the MIT status of provinces in Indonesia for the period 2014–2024 using GRDP per capita converted using the Atlas method as the basis for determining MIT status. The explanatory variables used include Life Expectancy (LE), Gross Fixed Capital Formation (GFCF), Open Unemployment Rate (OUR), Gross Enrollment Rate (GER), and Average Length of Schooling (ALS). The analysis was conducted using descriptive statistics, multicollinearity testing, binary probit regression modeling of panel data, parameter significance testing, marginal effect calculations, and classification performance evaluation. The model selection results showed that the Random Effect approach was more appropriate, and after the backward elimination process, the best model was obtained with significant variables PMTB, TPT, APK, and RLS. In general, PMTB and RLS increased the probability of provinces not being trapped in MIT, while TPT and APK decreased it. Model evaluation shows a classification accuracy of 70.07%, sensitivity of 11.71%, and specificity of 92.89%, with an AUC ROC value of 0.7311. These findings emphasize the importance of strengthening productive investment and improving the quality of education, as well as controlling unemployment, to increase the chances of provinces transitioning to higher income groups.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Data Panel, Middle Income Trap, Produk Domestik Regional Bruto, Random Effect, Regresi Probit Biner
Subjects:	Q Science Q Science > QA Mathematics Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Fatchur Rozi
Date Deposited:	28 Jan 2026 05:46
Last Modified:	28 Jan 2026 05:46
URI:	http://repository.its.ac.id/id/eprint/130750

Actions (login required)

View Item