Sudiono, Angela Cindy (2026) Implementasi Expected Credit Loss (ECL) Berbasis Faktor Makroekonomi untuk Pengukuran Non-Performance Risk Aset Reasuransi dalam Kerangka IFRS 17 dan IFRS 9 pada PT XYZ. Other thesis, Institut Tekonologi Sepuluh Nopember.
![[thumbnail of 5006221086-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5006221086-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy |
Abstract
Implementasi IFRS 17 menyaratkan pengakuan non-performance risk (NPR) dalam pengukuran kontrak reasuransi yang dimiliki, yang secara konseptual sejalan dengan pendekatan Expected Credit Loss (ECL) pada IFRS 9. Pada PT XYZ, pengukuran NPR atas aset reasuransi belum dilakukan meskipun IFRS 17 sudah berjalan beberapa tahun. Penelitian ini bertujuan mengestimasi ECL aset reasuransi PT XYZ dengan mengintegrasikan Probability of Default dan faktor makroekonomi. Data yang digunakan berupa piutang klaim reasuransi bulanan periode 2021–2025 dari 46 perusahaan reasuransi. Estimasi Probability of Default (PD) dilakukan menggunakan pendekatan rantai Markov melalui matriks transisi bucket aging untuk memperoleh PD Through-the-Cycle (TtC), yang selanjutnya ditransformasikan menjadi PD Point-in-Time (PiT) dengan mengintegrasikan proyeksi variabel makroekonomi melalui regresi linier berbasis Ordinary Least Squares (OLS). Proyeksi variabel makroekonomi dilakukan menggunakan model Long Short-Term Memory (LSTM) untuk menghasilkan estimasi yang bersifat forward-looking. Exposure at Default (EAD) diproyeksikan menggunakan model LSTM berdasarkan data historis EAD masing-masing perusahaan reasuransi, sedangkan Loss Given Default (LGD) diestimasi sebagai rata-rata historis tingkat pemulihan pada masing-masing perusahaan reasuransi. Nilai ECL dihitung sebagai hasil perkalian PD, LGD, dan EAD serta diagregasi pada tingkat portofolio. Hasil penelitian menunjukkan bahwa pengaruh variabel makroekonomi terhadap PD berbeda pada setiap perusahaan reasuransi. Model LSTM menghasilkan proyeksi makroekonomi dengan tingkat akurasi yang baik, sementara probabilitas transisi menunjukkan bahwa sebagian besar perusahaan reasuransi memiliki tingkat risiko kredit yang tinggi yang tercermin dari nilai PD dan LGD yang besar. Berdasarkan estimasi PD, LGD, dan EAD, diperoleh nilai ECL yang merepresentasikan NPR pada aset reasuransi PT XYZ. Pendekatan ECL dapat digunakan untuk mengukur NPR secara forward-looking dan mendukung penerapan IFRS 17.
=======================================================================================================================================
The implementation of IFRS 17 requires the recognition of non-performance risk (NPR) in the measurement of reinsurance contracts held, which is conceptually aligned with the Expected Credit Loss (ECL) approach under IFRS 9. At PT XYZ, NPR measurement for reinsurance assets has not yet been implemented despite IFRS 17 having been in effect for several years. This study aims to estimate the ECL of PT XYZ’s reinsurance assets by integrating credit Probability of Default and macroeconomic factors. The data used consist of monthly reinsurance claim receivables from 46 reinsurance companies during the period 2021–2025. Probability of Default (PD) is estimated using a Markov Chain approach through an aging bucket transition matrix to obtain Through-the-Cycle (TtC) PD, which is subsequently transformed into Point-in-Time (PiT) PD by incorporating projected macroeconomic variables through Ordinary Least Squares (OLS) regression. Macroeconomic variables are projected using a Long Short-Term Memory (LSTM) model to generate forward-looking estimates. Exposure at Default (EAD) is projected using an LSTM model based on historical EAD data for each reinsurance company, while Loss Given Default (LGD) is estimated as the average historical recovery rate of each reinsurance company. ECL is calculated as the product of PD, LGD, and EAD and then aggregated at the portfolio level. The results indicate that the impact of macroeconomic variables on PD varies across reinsurance companies. The LSTM model projected macroeconomic projections with satisfactory accuracy, while the probability transition shows that most reinsurance companies exhibit high credit risk, as reflected in their relatively high PD and LGD values. Based on the estimated PD, LGD, and EAD, the resulting ECL represents the NPR of PT XYZ’s reinsurance assets. The ECL approach can therefore be used to measure NPR in a forward-looking assesment and support the implementation of IFRS 17.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Expected Credit Loss, IFRS 17, Non-performance risk, LSTM, Markov Chain |
| Subjects: | H Social Sciences > HG Finance > HG3751 Credit--Management. H Social Sciences > HG Finance > HG8051 Insurance H Social Sciences > HG Finance > HG8054.5 Risk (Insurance) |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis |
| Depositing User: | Angela Cindy Sudiono |
| Date Deposited: | 18 Jul 2026 15:45 |
| Last Modified: | 18 Jul 2026 15:45 |
| URI: | http://repository.its.ac.id/id/eprint/135283 |
Actions (login required)
 |
View Item |