نوع مقاله : علمی - پژوهشی
نویسندگان
1 دانشجوی دکتری مهندسی مالی، گروه مدیریت، دانشگاه اصفهان، اصفهان، ایران.
2 استادیار گروه مدیریت، دانشگاه اصفهان، اصفهان، ایران
3 دانشیار، گروه ریاضی، دانشگاه تحصیلات تکمیلی علوم پایه زنجان، ایران.
چکیده
کلیدواژه‌ها
عنوان مقاله [English]
نویسندگان [English]
Purpose: Credit risk, particularly the risk of borrower default on obligations, is a significant concern for banks. Two primary variables used in modeling this risk and determining the expected loss from lending are the "Probability of Default" (PD) and the "Loss Given Default" (LGD). Given the importance of LGD in credit risk management and credit rating applications, this study utilizes a structural model approach to estimate LGD for both secured and unsecured debts, emphasizing the roles of debt structure and collateral liquidity. The proposed methodology is applied to selected firms in the Iranian capital market.
Method: To estimate "Expected LGD" and "Downturn LGD" for secured debts (with both liquid and less liquid collateral) and unsecured debts, a structural model and Monte Carlo simulation approach were employed. In this model, the market value of a firm’s total assets follows a Jump-Diffusion process, while each collateral, as part of the firm’s tangible assets, is modeled with a separate but dependent stochastic process relative to the firm's total assets. A liquidity penalty is applied for less liquid collateral. Additionally, different values were assigned to the ratio of tangible assets to total assets under default and non-default conditions. For eleven selected firms, the parameters of the stochastic processes for assets and collaterals were estimated using data up to the end of the 2017 fiscal year (1396 in the Iranian calendar). The LGD for these firms’ debts at the end of 2018 (1397) was then estimated using the structural model and Monte Carlo simulation. The simulation generated 500,000 paths for the firm's total assets and collaterals. For paths involving default, the LGD values were first estimated at the debt level and subsequently aggregated at the firm level. A sensitivity analysis was also conducted to assess the impact of alternative debt structures on LGD.
Findings: For firms with a single secured debt backed by liquid collateral, where recovery is solely from the collateral, the expected LGD ranged from 14% to 14.5%. However, if the collateral type was changed to a less liquid one, the expected LGD increased to a range of 30% to 33%. For the total LGD (where debt recovery includes both collateral and other remaining assets), the average expected values under these two scenarios were 9% and 12%, respectively. For firms with two secured debts, the expected LGD values, when recovery is solely from collateral, were similar to those of firms with a single secured debt using the same type of collateral. However, the average total LGD for debts backed by liquid collateral was 10%, compared to 15.8% for less liquid collateral. At the firm level, as the proportion of secured debts backed by liquid collateral relative to total debts decreased, the expected LGD tended to increase on average. In downturn scenarios, LGD for secured debt with liquid collateral was less influenced by the debt structure and more affected by the dynamics of assets and collateral, as well as their values under critical conditions.
Conclusion: The results highlight the significant role of collateral liquidity in reducing the expected LGD for secured debts. Additionally, the debt structure, particularly the priority of secured debt payments, impacts the LGD for unsecured debts, the total LGD for secured debts, and the overall LGD at the firm level. These findings emphasize the importance of considering collateral liquidity and debt structure in credit risk modeling to enhance financial decision-making and risk management.
کلیدواژه‌ها [English]
)