Penerapan Accelerated Failure Time (AFT) Shared Frailty Model dalam Analisis Ketahanan Sektor Perbankan Indonesia terhadap Risiko Financial Distress

Safinatunnajah, Auly Zuhdia (2026) Penerapan Accelerated Failure Time (AFT) Shared Frailty Model dalam Analisis Ketahanan Sektor Perbankan Indonesia terhadap Risiko Financial Distress. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006221062-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Ketahanan sektor perbankan terhadap risiko financial distress merupakan faktor penting bagi stabilitas makroekonomi nasional. Penelitian ini bertujuan untuk memodelkan waktu ketahanan (survival time) 59 Bank Umum Konvensional di Indonesia selama periode Kuartal I 2018 hingga Kuartal III 2025 menggunakan pendekatan Analisis Survival, khususnya Accelerated Failure Time (AFT). Penelitian ini membandingkan performa model AFT Lognormal Standar dengan perluasan Shared Frailty (distribusi Gamma dan Inverse Gaussian) untuk menginvestigasi keberadaan heterogenitas tak teramati (unobserved heterogeneity) pada kelompok kepemilikan bank (BUMN, BPD, Swasta Nasional, dan KCBA). Berdasarkan algoritma seleksi variabel, Capital Adequacy Ratio (CAR) dan ukuran aset (Size) ditetapkan sebagai prediktor utama. Hasil estimasi parameter menunjukkan bahwa peningkatan rasio CAR dan ukuran Size secara konsisten memperpanjang waktu ketahanan perbankan secara signifikan. Pengujian Likelihood Ratio Test (LRT) mengonfirmasi bahwa heterogenitas laten antar-kelompok kepemilikan terdeteksi secara statistik (p-value = 0,037). Namun, parameter varians frailty yang diestimasi bernilai sangat kecil (θ≈ 0,044 - 0,046; korelasi intra-klaster 2,15%), yang mengindikasikan bahwa kontribusi kelompok kepemilikan sangat marjinal dibandingkan kekuatan fundamental individu bank. Berlandaskan prinsip kesederhanaan (parsimony), AFT Lognormal Standar ditetapkan sebagai model utama. Evaluasi model sebagai instrumen Early Warning System (EWS) menunjukkan kemampuan diskriminasi yang tinggi (C-Index = 0,8730) dan tingkat kalibrasi prediksi yang akurat (Rasio Observed/Expected = 0,991). Proyeksi kuantil waktu mengidentifikasi bahwa kelompok BPD dan Bank Swasta memiliki tingkat kerentanan tertinggi, sementara BUMN dan KCBA lebih resilien karena didukung oleh skala ekonomi (economies of scale).
=====================================================================================================================================
The resilience of the banking sector against the risk of financial distress is a critical factor for national macroeconomic stability. This study aims to model the survival time of 59 Conventional Commercial Banks in Indonesia during the period from the First Quarter of 2018 to the Third Quarter of 2025 using Survival Analysis, specifically the Accelerated Failure Time (AFT) approach. This research compares the performance of the Standard Lognormal AFT model with the Shared Frailty extension (Gamma and Inverse Gaussian distributions) to investigate the presence of unobserved heterogeneity across bank ownership groups (State-Owned Banks, Regional Development Banks, National Private Banks, and Foreign Bank Branches). Based on the variable selection algorithm, the Capital Adequacy Ratio (CAR) and asset size (Size) were determined as the main predictors. Parameter estimation results show that increases in the CAR and Size consistently and significantly extend the survival time of the banks. The Likelihood Ratio Test (LRT) confirmed that latent heterogeneity among ownership groups was statistically detected (p-value = 0.037). However, the estimated frailty variance parameter is very small (θ≈0.044 - 0.046; intra-cluster correlation of 2.15%), indicating that the contribution of ownership clusters is marginal compared to the fundamental strength of individual banks. Based on the principle of parsimony, the Standard Lognormal AFT was designated as the primary model. Evaluation of the model as an Early Warning System (EWS) instrument demonstrates a high discrimination ability (C-Index = 0.8730) and an accurate prediction calibration level (Observed/Expected Ratio = 0.991). Time quantile projections identify that Regional Development Banks (BPD) and National Private Banks have the highest vulnerability levels, while State-Owned Banks (BUMN) and Foreign Bank Branches (KCBA) are more resilient as they are supported by economies of scale.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Accelerated Failure Time, Analisis Survival, Financial Distress, Perbankan, Shared Frailty, Accelerated Failure Time, Financial Distress, Banking, Shared Frailty, Survival Analysis.
Subjects:	Q Science > QA Mathematics > QA273.6 Weibull distribution. Logistic distribution. Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA401 Mathematical models.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Auly Zuhdia Safinatunnajah
Date Deposited:	20 Jul 2026 02:10
Last Modified:	20 Jul 2026 02:10
URI:	http://repository.its.ac.id/id/eprint/135293

Actions (login required)

View Item