Analisis prediksi pengaruh tingkat wanprestasi 90 hari (TWP90) di Indonesia menggunakan panel regression dan dynamic panel regression = Analysis and predicting impact of 90-day default rate in Indonesia using panel regression and dynamic panel regression

Rangkang, Divania Pruesten Shallom (2024) Analisis prediksi pengaruh tingkat wanprestasi 90 hari (TWP90) di Indonesia menggunakan panel regression dan dynamic panel regression = Analysis and predicting impact of 90-day default rate in Indonesia using panel regression and dynamic panel regression. Bachelor thesis, Universitas Pelita Harapan.

Preview	Text (Title) Title.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (191kB) | Preview
Preview	Text (Abstract) Abstract.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (570kB) | Preview
Preview	Text (ToC) ToC.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (574kB) | Preview
Preview	Text (Chapter1) Chapter1.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (671kB) | Preview
	Text (Chapter2) Chapter2.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (662kB)
	Text (Chapter3) Chapter3.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (628kB)
	Text (Chapter4) Chapter4.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (773kB)
	Text (Chapter5) Chapter5.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (559kB)
Preview	Text (Bibliography) Bibliography.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (568kB) | Preview
	Text (Appendices) Appendices.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB)

Abstract

Penelitian ini menganalisis dan memprediksi tingkat wanprestasi 90 hari berdasarkan aktivitas pinjaman di Jawa dan luar Jawa menggunakan metode panel regression dan dynamic panel regression. Data bulanan selama lima tahun (2018-2023) diolah dengan modifikasi, transformasi, dan imputasi menggunakan multiple imputation by chained equations. Sebanyak 11 model panel regression dan empat model dynamic panel regression dievaluasi melalui uji seleksi model, uji asumsi, analisis residual, MAPE, dan R-Square. Hasil penelitian menunjukkan rekening lender tidak berpengaruh signifikan, tetapi aktivitas pinjaman lainnya memiliki hubungan signifikan. Model terbaik dari panel regression adalah model fixed effect dengan efek tahun, dengan MAPE 0,0012% dan R-Square 0,3973. Model terbaik dalam dynamic panel regression adalah estimator Arellano-Bond, yang mengontrol endogenitas dengan MAPE 11,18% dan R-Square 0,5651. Meskipun dynamic panel regression mampu menangani pelanggaran asumsi, panel regression lebih sesuai dengan data ini. Model terbaik menunjukkan tahun berpengaruh signifikan terhadap tingkat wanprestasi 90 hari. Daerah Jawa memiliki hubungan positif yang signifikan dengan tingkat wanprestasi, sementara daerah luar Jawa menunjukkan korelasi negatif. Ini menyimpulkan bahwa peminjam di luar Jawa lebih sering menunaikan kewajiban, sehingga mengurangi tingkat wanprestasi, sementara peminjam di Jawa lebih banyak yang gagal memenuhi kewajibannya. / This research analyzes and predicts 90-day default rates based on loan activity in Java and outside Java using the panel method regression and dynamic panel regression. Monthly data for five years (2018-2023) processed with modification, transformation and imputation using multiple imputation by chained equations. A total of 11 panel regression models and four dynamic panel regression models were evaluated through model selection tests, assumption test, residual analysis, MAPE, and R-Square. The research results show lender accounts do not have a significant effect, but other lending activities do have a significant relationship. The best model of panel regression is the model fixed effect with year effects, with MAPE 0.0012% and R-Square 0.3973. The best model in dynamic panel regression is the Arellano-Bond estimator, which controls endogeneity with a MAPE of 11.18% and R-Square of 0.5651. Although dynamic panel regression is able to handle violations of assumptions, panel regression fits these data better. The best model indicates the year significant effect on the 90 day default rate. Java area has a significant positive relationship with the level of default, while areas outside Java show a negative correlation. This concludes that borrowers outside Java fulfill their obligations more often, thereby reducing default rate, while more borrowers in Java fail fulfill its obligations.

Actions (login required)

View Item