Method for constructing expected credit loss model of shipping lease industry

Abstract

The invention belongs to the technical field of data processing, and provides a method for constructing an expected credit loss model of a shipping and renting industry, which can realize risk layering of the shipping and renting industry by dividing credit value reduction grades. The default probability and the default loss rate can be adjusted according to the actual situation of the shipping lease industry, and loss calculation by using a fixed proportion is avoided; the expected credit loss model is constructed through the dynamic default probability, the dynamic default loss rate and the dynamic default risk exposure, more refined and comprehensive loss calculation can be realized through the three dimensions, and the limitation of a single index is avoided; different expected credit loss models are constructed according to the credit value reduction level, so that the situation that a relatively large error exists between the calculated loss and the actual situation due to the fact that the loss is calculated by adopting a fixed proportion in the expected credit loss calculation is effectively avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing, and particularly relates to a method for constructing an expected credit loss model for the shipping leasing industry. Background Art

[0002] Expected Credit Loss (ECL) refers to the loss expected to occur over a future period based on the credit risk in the asset portfolio. Currently, with the further development of regulatory measures, the new financial instrument standards require financial institutions to use ECL in financial reports to reflect the risk of potential decline in the value of their assets, so as to ensure the risk accuracy and reliability of financial statements. However, there is a lack of a unified and scientific calculation model for calculating expected credit losses in the shipping leasing industry, which leads to a large subjective judgment space. In addition, the construction and application of the ECL model require efficient cross-departmental collaboration and the support of complex calculation tools. However, some enterprises still adopt simplified methods (such as fixed ratio provisioning) for the sake of reducing operating costs. This approach ignores write-off differences and lacks necessary risk stratification, making it difficult to meet the high standards of credit risk management in the shipping leasing industry.

[0003] Although some prior arts related to expected credit losses have been disclosed, such as the patent application No. 202510172533.6 which proposes a prediction method and system for expected credit losses at a future point in time, this prediction method cannot solve the problems mentioned above in calculating expected credit losses in the shipping leasing industry. Summary of the Invention

[0004] In order to solve the problem that in the existing credit loss model for the shipping leasing industry in the prior art, a fixed ratio is used to calculate the losses in the shipping leasing industry, and there is a lack of a unified expected credit loss model, which cannot meet the credit risk management requirements of the shipping leasing industry, a method for constructing an expected credit loss model for the shipping leasing industry is proposed, and the above problems are solved.

[0005] The solution of the present invention is as follows: A method for constructing an expected credit loss model for the shipping leasing industry, the specific steps including: S1: Assess the credit impairment level and calculate the cumulative default probability: Rate all shipping lease contracts of a tenant within a period according to the credit impairment status to obtain the credit impairment levels of each contract; Count the number of contracts of the tenant that transfer from a specified credit impairment level to other credit impairment levels within the current period, and calculate the ratio of the number of contracts transferred to each other credit impairment level in the current period to the number of contracts corresponding to the specified credit impairment level to obtain the cumulative default probability of the specified credit impairment level; S2: Determine the dynamic default probability: Calculate the marginal default probability from the cumulative default probability; Screen out the macroeconomic indicators whose relevant values with the marginal default probability meet the preset threshold within the current period; Substitute the macroeconomic indicators into the preset multiple regression equation to calculate the adjustment factor value at the marginal default probability in the current period; Subtract the adjustment factor value in the current period from that in the previous period to obtain the first adjustment value; Adjust the current marginal default probability through the first adjustment value to obtain the dynamic default probability; S3: Determine the dynamic loss given default: Determine the type of guarantee for the shipping lease object; Select the average loan recovery rate corresponding to the type of guarantee of the shipping lease object to calculate the loss given default; Screen out the macroeconomic indicators whose relevant values with the loss given default meet the preset threshold within the current period; Substitute the macroeconomic indicators into the preset multiple regression equation to calculate the adjustment factor value at the loss given default in the current period; Subtract the adjustment factor value in the current period from that in the previous period to obtain the second adjustment value; Adjust the loss given default through the second adjustment value to obtain the dynamic loss given default; S4: Determine the dynamic exposure at default: Select the shipping lease contract data according to the business type of the shipping lease to calculate the dynamic exposure at default; S5: Construct the expected credit loss model: Construct the expected credit loss model corresponding to each credit impairment level in the current period through the dynamic default probability, the dynamic loss given default, and the dynamic exposure at default.

[0006] Preferably, the credit impairment levels include: Level 1: When the credit risk value calculated from the shipping lease contract data does not exceed the set threshold after the credit risk exists since the initial recognition, the credit risk level of the shipping lease contract is rated as Level 1; Level 2: When the credit risk value calculated from the shipping lease contract data exceeds the set threshold after the credit risk exists since the initial recognition, but no credit impairment has occurred on the reporting date, the credit risk level of the shipping lease contract is rated as Level 2; Level 3: When the credit risk value calculated from the shipping lease contract data exceeds the set threshold after the credit risk exists since the initial recognition, and credit impairment has occurred on the reporting date, the credit risk level of the shipping lease contract is rated as Level 3.

[0007] Preferably, the calculation method of the cumulative default probability in S1 is: A1: Calculate the transfer ratio by dividing the number of contracts transferred to other credit impairment levels in the current period by the number of contracts corresponding to the specified credit impairment level; The calculation formula of the transfer ratio is: Transfer ratio of specified credit impairment level to a certain credit impairment level = Number of contracts transferred from the specified credit impairment level to the certain credit impairment level in the current period / Number of contracts corresponding to the specified credit impairment level in the current period * 100%; A2: Designate the transfer ratio calculated by A1 as the cumulative default probability of the specified credit impairment level in the 1st period; Calculate the cumulative default probability of the specified credit impairment level in the tth period based on the transfer ratio calculated by A1 and the cumulative default probability in the (t - 1)th period.

[0008] Preferably, the period includes monthly, quarterly, annual, and multi - year periods.

[0009] Preferably, the preset multiple regression equation is constructed by adjustment factors and macro - economic indicators; The expression form of the multiple regression equation is: (1) Where, represents the adjustment factor in the tth period; is the ith macro - economic indicator; is the coefficient of the ith macro - economic indicator; is the normal random disturbance term; is the constant term; LN represents the logit transformation.

[0010] Preferably, for , and in the multiple regression equation, the determination method of the parameter values is: Construct a relationship equation between the adjustment factor and the marginal default probability or default loss rate, expressed as: (2) Where, C represents the marginal default probability or default loss rate; Substitute formula (2) into formula (1) to get: (3) Substitute the values of historical macro - economic indicators and historical marginal default probability values into formula (3) to obtain the parameter values of , and in the multiple regression equation for the marginal default probability in the current period; Then substitute the values of historical macro - economic indicators and historical default loss rate values into formula (3) to obtain the parameter values of , and in the multiple regression equation for the default loss rate in the current period.

[0011] Preferably, the specific method for constructing the dynamic default loss rate parameter in step S2 is as follows: S21: Calculate the marginal default probability: Calculate the marginal default probability through the cumulative default probability; S22: Calculate the adjustment value: Screen the macroeconomic indicators whose relevant values with the marginal default probability meet the preset threshold within the current period; Substitute the macroeconomic indicators into the preset multiple regression equation to calculate the adjustment factor value under the marginal default probability; Subtract the adjustment factor value of the current period from that of the previous period to obtain the first adjustment value; S23: Calculate the forward-looking marginal default probability: Adjust the marginal default probability using the first adjustment value to obtain the adjusted marginal default probability; The adjusted marginal default probability is the forward-looking marginal default probability; S24: Calculate the forward-looking default probability: Calculate the forward-looking default probability through the forward-looking marginal default probability; S25: Determine different shipping lease scenario models; Determine the macroeconomic indicators whose relevant values with the marginal default probability meet the preset threshold within the current period under different shipping lease scenario models, and repeat the calculation process of steps S22 - S24 to calculate the forward-looking default probabilities under different shipping lease scenario models; S26: Construct the dynamic default probability: Set the weights of the forward-looking default probabilities corresponding to each shipping lease scenario model; Calculate the dynamic default probability through weighting the forward-looking default probabilities corresponding to different shipping lease scenarios and the set weights.

[0012] Preferably, the marginal default probability is expressed as: (4) is the cumulative default probability in the t-th period; is the marginal default probability in the t-th period; is the cumulative default probability in the (t - 1)-th period.

[0013] Preferably, the expression form of the forward-looking marginal default probability in S23 is: (5) (6) Wherein, is the forward-looking marginal default probability in the t-th period, is the marginal default probability in the t-th period; is the adjustment value in the t-th period, is an intermediate parameter.

[0014] Preferably, the expression form of the forward-looking default probability in S24 is: (7) Among them, is the forward-looking default probability in the t-th period, represents the forward-looking marginal default probability in the t-th period. When t = 1, .

[0015] Preferably, the shipping lease scenario model includes: a benchmark, an optimistic, and a pessimistic scenario; according to the weights of the shipping lease scenario model set by the shipping lease company in the current period, calculate the dynamic default probability in the current period, which is expressed as: (8) Among them, each scenario represents the forward-looking default probability in the t-th period under different shipping lease scenario models, represents the dynamic default probability in the t-th period.

[0016] Preferably, the method for determining the dynamic default loss rate in S3 is as follows: C1: Classify the shipping lease objects into secured and unsecured categories to obtain the classification results; C2: Statistically calculate the average loan recovery rate of the classification results in C1 in the shipping lease measurement basic data; calculate the default loss rate according to the average loan recovery rate corresponding to the classification results, which is expressed as: (9) is the default loss rate in the t-th period; C3: Screen the macroeconomic indicators whose relevant values with the default loss rate in the current period meet the preset threshold; substitute the macroeconomic indicators that meet the preset threshold into the preset multiple regression equation to calculate the adjustment factor value under the default loss rate; subtract the adjustment factor value in the current period from that in the previous period to obtain the second adjustment value; C4: Use the second adjustment value to adjust the default loss rate to obtain the adjusted default loss rate; the adjusted default loss rate is the forward-looking default loss rate; the forward-looking default loss rate is expressed as: (10) (11) is an intermediate parameter; is the forward-looking default loss rate in the t-th period, is the adjustment value in the t-th period; C5: Determine different shipping lease scenario models; determine the macroeconomic indicators whose relevant values with the default loss rate in the current period meet the preset threshold under different shipping lease scenario models, and repeat the calculation process of steps C3 - C4 to calculate the forward-looking default loss rates under different shipping lease scenario models; C6: Set the weights of the forward-looking default loss rates corresponding to different shipping lease scenario models; calculate the dynamic default probability parameter through weighting the forward-looking default loss rates corresponding to the different shipping lease scenario models and the set weights, expressed as: (12) where each scenario represents the forward-looking default loss rate under different shipping lease scenario models in the t-th period, represents the dynamic default loss rate in the t-th period.

[0017] Preferably, determine the accounting subject type according to the business type of the shipping lease; select the shipping lease contract data according to the accounting subject type to calculate the dynamic default risk exposure; when the business type is a finance lease business, specify the accounting subject as the receivable finance lease payment; when the business type is an off-balance-sheet loan commitment business, specify the accounting subject as the off-balance-sheet loan commitment.

[0018] Preferably, the expected credit loss models corresponding to each credit impairment level in S5 are: S61: When the credit impairment level is level one, the corresponding expected credit loss model is; (13) where ECL is the expected credit loss; EIR is the actual interest rate per period; is the dynamic default probability in the first period; is the dynamic default loss rate in the first period; is the dynamic default risk exposure in the first period; S62: When the credit impairment level is level two, the corresponding expected credit loss model is; (14) where: ECL is the expected credit loss; is the dynamic default loss rate in the t-th period; is the dynamic default probability in the t-th period; is the dynamic default risk exposure in the t-th period; S63: When the credit impairment level is level three, the corresponding expected credit loss model is; (15) where ECL is the expected credit loss, PD = 100%, is the dynamic default loss rate in the t-th period, is the dynamic default risk exposure in the t-th period.

[0019] Beneficial effects: The invention proposes a method for constructing an expected credit loss model for the shipping leasing industry. By classifying the credit impairment levels of the shipping lease contracts of tenants, risk stratification management of each shipping lease contract can be achieved; macroeconomic indicators with high correlation values related to the marginal default probability or default loss rate are selected to calculate the adjustment factor values, avoiding the reduction of the calculation results of the expected credit loss by irrelevant or weakly correlated indicators; when calculating the adjustment values under the marginal default probability or default loss rate each time, relevant macroeconomic indicators are newly selected according to the calculation time, which can effectively avoid the problem of data lag and ensure that the calculation results of the constructed expected credit loss model are closer to the actual situation of the shipping leasing company, so as to be able to more accurately evaluate potential losses and provide a more reliable basis for the subsequent decision-making of the shipping leasing company.

[0020] By classifying the shipping lease objects into secured and unsecured categories, only the default loss rates of the secured and unsecured types need to be calculated, avoiding the need for each department to calculate the default loss rates for different types of shipping lease objects separately, which can effectively improve the calculation efficiency; according to the business type of each shipping lease contract, shipping lease contract data is selected to calculate the dynamic default risk exposure, so that shipping lease contracts of the same business type can select the same data and the same calculation formula for calculation, which can avoid each department calculating the dynamic default exposure risk for each contract based on experience, wasting manpower and material resources.

[0021] An expected credit loss model is constructed through the dynamic default probability, dynamic default loss rate, and dynamic default risk exposure, and a unified expected credit loss model is constructed, avoiding subjective judgment space in the calculation of the expected credit loss; at the same time, it can achieve more refined and comprehensive calculation of the expected credit loss through these three dimensions of the dynamic default probability, dynamic default loss rate, and dynamic default risk exposure, avoiding the limitations of a single indicator; different expected credit loss models are constructed according to the credit impairment levels, effectively avoiding the situation where in the calculation of the expected credit loss, using a fixed ratio to calculate the loss ignores the influence of the credit impairment level on the calculation results of the expected credit loss, resulting in a large error between the calculated loss and the actual situation.

[0022] Furthermore, by clarifying the specific method for calculating the adjustment factor value, the cumbersome process of multiple departments investing a large amount of manpower and material resources to determine the manual adjustment value can be avoided; by calculating the cumulative default probability through the transfer ratio and then calculating the marginal default probability through the cumulative default probability, the impact of the expected credit rating on the marginal default probability can be effectively reflected, and the sensitivity loss problem in risk stratification using a fixed ratio can be effectively avoided, thus significantly improving the accuracy of the calculation results; by calculating the coefficient value in the multiple regression equation through the marginal default probability or the loss given default, the obtained adjustment value will show a stronger correlation with the marginal default probability or the loss given default, and the marginal default probability or the loss given default adjusted by the adjustment value is also more in line with the actual situation; at the same time, by performing weighted summation of the default probability and the loss given default in the shipping lease scenario mode, the default probability and the loss given default in different scenarios can be fully considered, avoiding the one-sidedness of single-scenario analysis and being closer to the actual loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flowchart of a method for constructing an expected credit loss model for the shipping lease industry. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] The present invention will be further described below with reference to the drawings and embodiments.

[0025] Embodiment 1: As Figure 1 shown, a method for constructing an expected credit loss model for the shipping lease industry specifically includes the following steps: S1: Assess the credit impairment level and calculate the cumulative default probability: Rate all shipping lease contracts of a tenant within a period according to the credit impairment status to obtain the credit impairment level of each contract; count the number of contracts of the tenant that transfer from a specified credit impairment level to other credit impairment levels within the current period, and calculate the ratio of the number of contracts transferred to each other credit impairment level in the current period to the number of contracts corresponding to the specified credit impairment level to obtain the cumulative default probability of the specified credit impairment level; S2: Determine the dynamic default probability: Calculate the marginal default probability through the cumulative default probability; screen out the macroeconomic indicators whose relevant values with the marginal default probability meet the preset threshold within the current period; substitute the macroeconomic indicators into the preset multiple regression equation to calculate the adjustment factor value under the marginal default probability in the current period; subtract the adjustment factor value of the current period from that of the previous period to obtain the first adjustment value; adjust the current marginal default probability through the first adjustment value to obtain the dynamic default probability; S3: Determine the dynamic default loss rate: Determine the type of guarantee for the shipping lease object; Select the average loan recovery rate corresponding to the type of guarantee for the shipping lease object to calculate the default loss rate; Screen the macroeconomic indicators whose relevant values for the default loss rate meet the preset threshold within the current period; Substitute the macroeconomic indicators into the preset multiple regression equation to calculate the adjustment factor value at the default loss rate for the current period; Subtract the adjustment factor value for the current period from that for the previous period to obtain the second adjustment value; Adjust the default loss rate through the second adjustment value to obtain the dynamic default loss rate; S4: Determine the dynamic default risk exposure: Select the shipping lease contract data according to the business type of the shipping lease to calculate the dynamic default risk exposure; S5: Construct the expected credit loss model: Construct the expected credit loss model corresponding to each credit impairment level for the current period through the dynamic default probability, the dynamic default loss rate, and the dynamic default risk exposure. Embodiment

[0026] I. Data preparation and processing The data for calculating the expected credit loss model of the shipping lease industry includes: shipping lease contract data and external data. Among them, the shipping lease contract data is sourced from the information exported from the business system, and some fields need to be obtained through manual supplementary recording; The main external data publishing agencies are the National Bureau of Statistics and various commercial banks.

[0027] (1) Shipping lease contract data Data date, contract number, lease start date, lease expiration date, lessee name, industry, contract interest rate, currency, remaining margin, remaining principal, unloaned amount, five - level classification, manual adjustment stage division, etc.

[0028] (2) External data, i.e., the measurement basis data for shipping leases External data includes historical data of macro - indicators required for impairment measurement, non - performing loan ratios of the shipping lease industry published by state - owned large commercial banks and joint - stock commercial banks, as well as multi - year default probabilities and loan recovery rates, etc.

[0029] II. Calculate the expected credit loss for the contracts with the accounting subject of finance lease receivables: For the financial leasing business of a certain shipping leasing company, the customer types are all corporate customers; The specific steps for building the ECL model for the financial leasing business are as follows: (I) Classification criteria for credit impairment levels Level 1: After the initial recognition of credit risk, if the overdue days of the principal or interest are less than 30 days, the credit risk level of the shipping lease contract is rated as level 1; Level 2: After the initial confirmation of the existence of credit risk, the overdue days of principal or interest are greater than 30 days (inclusive) and less than 90 days, but there is no credit impairment on the reporting date. The credit risk rating of the shipping lease contract is rated as Level 2; Level 3: After the initial confirmation of the existence of credit risk, the overdue days of principal or interest are greater than 90 days (inclusive), and there is credit impairment on the reporting date. The credit risk rating of the shipping lease contract is rated as Level 3.

[0030] (2) Model establishment and parameter determination 1. Determination of dynamic default probability (1) Calculation of transition ratio Using the shipping lease contract data of the financial leasing business of the shipping leasing company as the modeling sample data, calculate the transition ratio. Statistics are carried out in units of the number of contracts of the same tenant. The transition ratio is calculated with an observation period window of "interval of 1 year" and an annual statistical frequency. The calculation method is as follows: Transition ratio from Level 2 to Level 3 = Number of contracts that enter the 3rd stage at the end of the current statistical period among the total number of Level 2 contracts at the beginning of the current statistical period / Total number of Level 2 contracts at the beginning of the current statistical period * 100%; The transition ratios of moving from Level 3 to Level 1, from Level 3 to Level 2, and from Level 3 to Level 3 are 0%, 0%, and 100% respectively.

[0031] (2) Calculation of adjustment value Use the Wilson model to construct a multiple regression equation. The construction process is as follows: ① Obtain the non-performing loan rate data of the shipping leasing industry announced by large state-owned commercial banks and joint-stock commercial banks listed on the stock market from the first half of 2016 to the first half of 2024, and divide them according to the cycle.

[0032] ② Perform a logit transformation on the above data to obtain an adjustment factor To complete the data preparation for building the multiple regression equation.

[0033] (2) Among them: C represents the marginal default probability or default loss rate.

[0034] ③ Select macroeconomic indicators that have a long-term stable equilibrium relationship with the adjustment factor to establish a multiple regression equation. The macroeconomic indicators are selected as those whose relevant values related to the marginal default probability or default loss rate meet the preset threshold; the multiple regression equation is expressed as: (1) Among them, represents the adjustment factor in the t-th period; is the i-th macroeconomic indicator; is the coefficient of the i-th macroeconomic indicator; m represents the number of macroeconomic indicators that meet the preset threshold; is the normal random disturbance term; is the constant term; LN represents the logit transformation; the frequencies of the independent variable and the dependent variable need to be consistent.

[0035] Substitute formula (2) into formula (1) to obtain: (3) Specifically, substitute the historical values of the macroeconomic indicators that meet the preset threshold corresponding to the marginal default probability and the values of the historical marginal default probability into formula (3) to obtain the parameter values of in the multiple regression equation at the marginal default probability of the current period 、 and ; substitute the historical values of the macroeconomic indicators that meet the preset threshold corresponding to the loss given default and the values of the historical loss given default into formula (3) to obtain the parameter values of in the multiple regression equation at the loss given default of the current period 、 and .

[0036] ④ Use the multiple regression model established in ③ to predict the values of the adjustment factors in future periods; Specifically, in all the formulas in Embodiment 1, the t-th period is obtained by the following method: The remaining duration from the current period to the end of the shipping lease contract is split according to the duration corresponding to the current period, and it can be split into n parts, so as to determine that there are n periods from the current period to the end of the shipping lease contract; t is any one of these n periods; when t = 1, it corresponds to the current period, indicating the 1st period.

[0037] ⑤ Calculate the adjustment value .

[0038] (16) (3) The specific steps for a shipping leasing company to determine the dynamic default probability are as follows: 1) Determine the macroeconomic indicator pool of a shipping leasing company The macroeconomic indicator pool of a certain company is shown in Table 1.

[0039]

[0040] The above data can be uniformly obtained through the WIND database.

[0041] In the above table, data needs to be obtained quarterly. If the indicator is announced monthly / daily, the data at the end of March, June, September, and December is taken.

[0042] 2) Forecast of macroeconomic indicators A shipping leasing company chooses to use the internal model forecasting method combined with expert judgment to predict macroeconomic indicators. A shipping leasing company uses the ARIMA model to predict the value of macroeconomic indicators for the next year.

[0043] 3) Calculate the cumulative probability of default Specifically, the migration ratio calculated in (1) is designated as the cumulative default probability of the first period. The cumulative default probability of the first period contract with a credit impairment rating of level 1 is the migration ratio from level 1 to level 3. The cumulative default probability of the first period contract with a credit impairment rating of level 2 is the migration ratio from level 2 to level 3. The cumulative default probability of the first period contract with a credit impairment rating of level 3 is 100%.

[0044]

[0045] The cumulative default probability of other periods is calculated through Table 2; the cumulative default probability matrix is constructed through the cumulative default probability, as shown in Table 3.

[0046] From Table 2, the calculation rules for the cumulative default probability matrix in Table 3 are as follows: ①B51= B48, B52= C48, B53= D48; ②C51= SUMPRODUCT(B51:B53,$B$46:$B$48); ③C52= SUMPRODUCT(B51:B53,$C$46:$C$48); ④C53= 100%; ⑤The formulas in columns D to F are similar.

[0047] 4) Calculate the marginal probability of default The marginal default probability is calculated by the cumulative default probability. The expression of marginal default probability is: (1) is the marginal default probability in period t; is the cumulative default probability in period t; is the cumulative default probability in period t-1, .

[0048] 5) Construct multiple regression equation Select from the macroeconomic indicator pool the macroeconomic indicators whose values related to the marginal probability of default of the tenant contracts corresponding to each credit impairment level meet the preset threshold, and then substitute the macroeconomic indicators into formula (1).

[0049] Specifically, for contracts of tenants in the same credit impairment grade, the same macroeconomic indicators are selected to calculate the adjustment factor value corresponding to the marginal default probability in the current period of the contract.

[0050] The adjustment factor is calculated through the multiple regression equation after substituting the macroeconomic indicators ; the macroeconomic indicator coefficients, normal random disturbance terms, and constant terms in the multiple regression equation are determined through the aforementioned external data and historical marginal default probabilities, or by using statistical software to build a model. The obtained multiple regression equation is shown in Table 4:

[0051] In Table 4, Lag3 and Lag12 represent a lag of 3 months and 12 months respectively. At the same time, a certain shipping leasing company updates the macroeconomic indicators, macroeconomic indicator coefficients, constant terms, and normal random disturbance terms in the multiple regression equation at least once every six months.

[0052] The adjustment value corresponding to the marginal default probability is calculated through the multiple regression equation determined in Table 4.

[0053] 6) Adjust the marginal default probability through the adjustment value.

[0054] Adjust the marginal default probability in the t-th period through the adjustment value to obtain the adjusted marginal default probability in the t-th period; the adjusted marginal default probability is the forward-looking marginal default probability It is expressed as: (5) Where (6) 7) Calculate the forward-looking default probability in the t-th period through the forward-looking marginal default probability (7) Where: .

[0055] For each year over 20 years , take the value of the 20th year .

[0056] 8) Scenario patterns and weights A. Scenario patterns A certain shipping leasing company sets multiple shipping leasing scenario patterns according to regulatory requirements, based on its own business conditions and the macro political and economic situation, and clarifies the weights of different scenarios.

[0057] The current economic scenario settings include three shipping lease scenario models, namely the baseline, optimistic, and pessimistic scenarios. According to the economic definitions of the macro data used in each group of equations, one standard deviation is added or subtracted from the macroeconomic indicators in the finally determined multiple regression equation as the predicted value of the macroeconomic indicators in the optimistic or pessimistic scenarios.

[0058] B. Scenario Weighting Based on the weights of the optimistic and pessimistic scenarios determined by the approval of a certain shipping lease company, the dynamic default probability is calculated. The expression of the dynamic default probability is: (8) Among them, each scenario represents the forward-looking default probability in different shipping lease scenario models in the t-th period, and represents the dynamic default probability in the t-th period.

[0059] 2. Determination of Dynamic Default Loss Rate (1) Determination of Default Loss Rate, i.e., LGD The default loss rate LGD is calculated using the average loan recovery rates of secured and unsecured banks from 1983 to 2023 in the statistical external data; (9)

[0060] The data in Table 5 should be the latest reported data when calculating the default loss rate.

[0061] The shipping lease items are classified into secured and unsecured. Among them, ship assets are secured, and other shipping lease items are unsecured; the secured default loss rate is used for ship assets; the unsecured default loss rate is used for other lease items.

[0062] The default loss rates corresponding to different types of shipping lease items of a certain shipping lease company are shown in Table 6:

[0063] (2) Determination of Adjustment Value Based on Wilson Model Among the macroeconomic indicator pool, select the macroeconomic indicators whose relevant values corresponding to the default loss rates of the tenant contracts in each category meet the preset thresholds. Then substitute the macroeconomic indicators into formula (1) to obtain the multiple linear regression equation of the default loss rate and the relevant forward-looking information indicators, calculate the adjustment value, and then adjust the default loss rate through the adjustment value.

[0064] Specifically, if the guarantee categories of the shipping lease assets corresponding to different contracts among the tenants are the same, the adjustment factor values for the current period calculated using consistent macroeconomic indicators are adopted for the default loss rates of the contracts with the same guarantee type of these shipping lease assets.

[0065] The specific adjustment process is the same as that of the marginal default probability. The final multiple regression equation for the default loss rate is shown in Table 7 as follows:

[0066] Note: In Table 7, Lag12 represents a lag of 12 months, and Lag6 represents a lag of 6 months. (3) Scenario patterns and weights A. Scenario patterns According to regulatory requirements, a shipping leasing company sets multiple shipping leasing scenario patterns based on its own business conditions and the macro political and economic situation, and clarifies the weights of different scenarios.

[0067] Currently, three shipping leasing scenario patterns are set for the economic scenario, namely the baseline, optimistic, and pessimistic scenarios. According to the economic definitions of the macro data used in each group of equations, one standard deviation is added to or subtracted from the macroeconomic indicators in the finally determined multiple regression equation as the predicted value of the macroeconomic indicators under the optimistic or pessimistic scenario.

[0068] B. Scenario weighting Based on the optimistic and pessimistic scenario weights approved by a shipping leasing company, the dynamic default loss rate is calculated. The formula is as follows: (12) Among them, each scenario represents the forward-looking default loss rate under different shipping leasing scenario patterns, and represents the dynamic default loss rate in the t-th period.

[0069] 3. Determination of dynamic default risk exposure EAD is the risk exposure at the time of default of a financial asset, which is the present value of all the cash flows receivable under the contract discounted to the time of default at the actual interest rate. The calculation formula is shown in Table 8.

[0070]

[0071] Among them, the principal of the receivable financial lease payment is the amount of the receivable specified in the contract; the interest of the receivable financial lease payment is calculated according to the interest rate and interest calculation method agreed in the contract; the handling fee of the receivable financial lease payment is the amount of the fee item clearly agreed in the contract; the unearned finance income refers to the income that has not been realized during the process of the contract; the received guarantee deposit is a certain amount paid by the tenant to fulfill the contract obligation.

[0072] 4. Establishment of the Expected Credit Loss Model Construct the corresponding expected credit loss model according to the credit impairment level; construct the expected credit loss model through the dynamic default probability, dynamic loss given default, and dynamic exposure at default. Each contract selects the expected credit loss model corresponding to its own credit impairment level for calculation.

[0073] The formula for the expected credit loss model is as follows: The duration corresponding to the current period in the following formula is one year, that is, one period represents one year: When the credit impairment level is at level 1, for the corresponding expected credit loss model, it is necessary to calculate the expected credit loss for the next 12 months (13) Among them, ECL is the expected credit loss; EIR is the effective interest rate per period; is the estimated dynamic default probability for the first period in the future at the current time point; is the dynamic loss given default for the first period; is the dynamic exposure at default for the first period; When the credit impairment level is at level 2, for the corresponding expected credit loss model, it is necessary to calculate the expected credit loss for the remaining term of the shipping lease contract from the current time point to the end of the contract as agreed: (14) Among them: ECL is the expected credit loss; is the dynamic loss given default for the t-th period; is the dynamic default probability for the t-th period; is the dynamic exposure at default for the t-th period; When the credit impairment level is at level 3, for the corresponding expected credit loss model, it is necessary to calculate the expected credit loss for the remaining term of the shipping lease contract from the current time point to the end of the contract as agreed: (15) Among them, ECL is the expected credit loss, PD = 100%, is the dynamic loss given default for the t-th period, where the LGD parameter applies the LGD results for the next 1 year in all three stages of the 1st, 2nd, and 3rd stages, and t can be unified as 1; is the dynamic exposure at default for the t-th period.

[0074] III. Calculate the expected credit loss for contracts with off-balance-sheet loan commitments as accounting subjects: (I)Stage division According to the characteristics of the business's credit-granted person, conduct stage division in accordance with the stage division rules for receivables from finance leases.

[0075] (2) Model establishment and parameter determination 1. Determination of dynamic default probability Adopt the same PD parameter as the receivable financial lease payment.

[0076] 2. Determination of dynamic loss given default Adopt the same LGD parameter as the receivable financial lease payment.

[0077] 3. Determination of dynamic exposure at default risk (1) Basic formula When calculating the dynamic exposure at default of off - balance - sheet credit commitments, estimate the utilization of loan commitments (i.e., the credit conversion factor CCF), which is expressed as:

[0078] Among them, the off - balance - sheet credit balance is calculated from the off - balance - sheet credit limit signed in the contract and the off - balance - sheet credit limit already used; the margin balance is obtained by subtracting the used margin from the received margin.

[0079] (2) Credit conversion factor CCF for off - balance - sheet credit commitment items A certain shipping leasing company adopts a credit conversion factor of 40% corresponding to the current irrevocable off - balance - sheet loan commitment business. The CCF parameter needs to be updated and adjusted according to different usage periods.

[0080] 4. Establishment of the expected credit loss model Since there is no clear repayment plan for off - balance - sheet business, when calculating the expected credit loss of off - balance - sheet business, instead of estimating the change in future exposure, it is calculated based on the current balance. The ECL calculation formula for off - balance - sheet business sets the EIR parameter and the actual interest rate value of each period to 0 during the calculation process, and the calculation formula is: (17) According to the impairment stage, the specific model for calculating the off - balance - sheet business ECL is as follows: When the credit impairment level is at level 1, for the corresponding expected credit loss model, it is necessary to calculate the expected credit loss for the next 12 months: (18) Among them, is the dynamic default probability of the first period; is the dynamic loss given default of the first period; the dynamic exposure at default of the first period; When the credit impairment level is at level 2, for the corresponding expected credit loss model, it is necessary to calculate the expected credit loss for the remaining term of the shipping lease contract from the current time point to the end agreed in the contract: (19) Where: ECL is the expected credit loss; is the dynamic default loss rate in the t-th period; is the dynamic default probability in the t-th period; is the dynamic default exposure in the t-th period; Note: For off-balance-sheet credit commitments with a corresponding grade of secondary, the remaining term is 1 year.

[0081] When the credit impairment grade is tertiary, for the corresponding expected credit loss model, it is necessary to calculate the expected credit loss for the remaining term of the shipping lease contract from the current time point to the end as agreed in the contract: (20) Where: ECL is the expected credit loss; is the dynamic default loss rate in the t-th period, where the LGD parameter applies the LGD results for the next 1 year at all levels of primary, secondary, and tertiary, and t can be uniformly 1; is the dynamic default exposure in the t-th period.

[0082] Finally, it should be noted that the above is only used to illustrate the technical solution of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred arrangement, those of ordinary skill in the art should understand that the technical solution of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for constructing an expected credit loss model for the shipping lease industry, characterized in that: S1: Assess the credit impairment level and calculate the cumulative default probability: Rate all shipping lease contracts of a tenant within a certain period according to the credit impairment status to obtain the credit impairment levels of each contract; Count the number of contracts of the tenant that transfer from a specified credit impairment level to other credit impairment levels within the current period, and calculate the ratio of the number of contracts transferred to each other credit impairment level in the current period to the number of contracts corresponding to the specified credit impairment level to obtain the cumulative default probability of the specified credit impairment level; S2: Determine the dynamic default probability: Calculate the marginal default probability through the cumulative default probability; Screen macroeconomic indicators whose relevant values related to the marginal default probability meet the preset threshold within the current period; Substitute the macroeconomic indicators into the preset multiple regression equation to calculate the adjustment factor value under the marginal default probability in the current period; Subtract the adjustment factor value in the current period from that in the previous period to obtain the first adjustment value; Adjust the current marginal default probability through the first adjustment value to obtain the dynamic default probability; S3: Determine the dynamic loss given default: Determine the type of guarantee for the shipping lease object; Select the average loan recovery rate corresponding to the type of guarantee of the shipping lease object to calculate the loss given default; Screen macroeconomic indicators whose relevant values related to the loss given default meet the preset threshold within the current period; Substitute the macroeconomic indicators into the preset multiple regression equation to calculate the adjustment factor value under the loss given default in the current period; Subtract the adjustment factor value in the current period from that in the previous period to obtain the second adjustment value; Adjust the loss given default through the second adjustment value to obtain the dynamic loss given default; S4: Determine the dynamic exposure at default: Select shipping lease contract data according to the business type of the shipping lease to calculate the dynamic exposure at default; S5: Construct an expected credit loss model: Construct an expected credit loss model corresponding to each credit impairment level in the current period through the dynamic default probability, the dynamic loss given default, and the dynamic exposure at default.

2. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 1, characterized in that, The credit impairment levels include: Level 1: When the credit risk value calculated through the shipping lease contract data does not exceed the set threshold after the credit risk exists since initial recognition, the credit risk level of the shipping lease contract is rated as Level 1; Level 2: When the credit risk value calculated through the shipping lease contract data exceeds the set threshold after the credit risk exists since initial recognition, but no credit impairment has occurred on the reporting date, the credit risk level of the shipping lease contract is rated as Level 2; Level 3: When the credit risk value calculated through the shipping lease contract data exceeds the set threshold after the credit risk exists since initial recognition, and credit impairment has occurred on the reporting date, the credit risk level of the shipping lease contract is rated as Level 3.

3. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 1, characterized in that, The calculation method of the cumulative default probability described in S1 is: A1: Calculate the ratio of the number of contracts transferred to each other credit impairment level in the current period to the number of contracts corresponding to the specified credit impairment level to obtain the transfer ratio; The calculation formula of the transfer ratio is: Transfer ratio of specified credit impairment level to a certain credit impairment level = Number of contracts transferred from the specified credit impairment level to the certain credit impairment level in the current period / Number of contracts corresponding to the specified credit impairment level in the current period * 100%; A2: Designate the transfer ratio calculated in A1 as the cumulative default probability of the first period of the specified credit impairment level; Calculate the cumulative default probability of the t-th period of the specified credit impairment level through the transfer ratio calculated in A1 and the cumulative default probability of the (t - 1)-th period.

4. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 1, characterized in that, The said period includes monthly, quarterly, annual and multi-year periods.

5. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 1, characterized in that, The preset multiple regression equation is constructed by adjustment factors and macroeconomic indicators; The expression form of the multiple regression equation is: （1） Among them, represents the adjustment factor in the t-th period; is the i-th macroeconomic indicator; is the coefficient of the i-th macroeconomic indicator; is the normal random disturbance term; is the constant term; LN represents the logit transformation.

6. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 5, characterized in that, In the multiple regression equation, , and The parameter values are determined as follows: Construct a relationship equation between the adjustment factor and the marginal default probability or default loss rate, expressed as: （2） Wherein, C represents the marginal default probability or default loss rate; Substitute formula (2) into formula (1) to obtain: （3） Substitute the values of historical macroeconomic indicators and historical marginal default probabilities into formula (3) to obtain the parameter values of 、 and in the multiple regression equation for the marginal default probability at the current time period; then substitute the values of historical macroeconomic indicators and historical loss given default into formula (3) to obtain the parameter values of 、 and in the multiple regression equation for the loss given default at the current time period.

7. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 1, characterized in that, The specific method for constructing the dynamic default loss rate parameter in step S2 is: S21: Calculate the marginal default probability: Calculate the marginal default probability through the cumulative default probability; S22: Calculate the adjustment value: Screen the macroeconomic indicators whose relevant values with the marginal default probability meet the preset threshold in the current period; Substitute the macroeconomic indicators into the preset multiple regression equation to calculate the adjustment factor value under the marginal default probability; Subtract the adjustment factor value of the current period from that of the previous period to obtain the first adjustment value; S23: Calculate the forward-looking marginal default probability: Adjust the marginal default probability using the first adjustment value to obtain the adjusted marginal default probability; The adjusted marginal default probability is the forward-looking marginal default probability; S24: Calculate the forward-looking default probability: Calculate the forward-looking default probability through the forward-looking marginal default probability; S25: Determine different shipping lease scenario models; Determine the macroeconomic indicators whose relevant values with the marginal default probability meet the preset threshold in the current period under different shipping lease scenario models, and repeat the calculation process of steps S22 - S24 to calculate the forward-looking default probabilities under different shipping lease scenario models; S26: Construct the dynamic default probability: Set the weights of the forward-looking default probabilities corresponding to each shipping lease scenario model; Calculate the dynamic default probability through weighting the forward-looking default probabilities corresponding to different shipping lease scenarios and the set weights.

8. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 7, characterized in that, The marginal default probability is expressed as: （4） is the cumulative default probability for the t-th period; is the marginal default probability for the t-th period; is the cumulative default probability for the (t - 1)-th period.

9. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 7, characterized in that, The expression form of the forward-looking marginal default probability in S23 is: （5） （6） Among them, is the forward-looking marginal default probability in the t-th period, is the marginal default probability in the t-th period; is the adjustment value in the t-th period, is an intermediate parameter.

10. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 7, characterized in that, The expression form of the forward-looking default probability in S24 is: （7） wherein, is the forward-looking default probability in the t-th period, represents the forward-looking marginal default probability in the t-th period. When t = 1, .

11. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 7, characterized in that, The shipping lease scenario models include: benchmark, optimistic and pessimistic scenarios; According to the weights of the shipping lease scenario models set by the shipping lease company in the current period, calculate the dynamic default probability in the current period, expressed as: （8） Among them, represents the forward-looking default probability in different shipping lease scenario modes in the t-th period, represents the dynamic default probability in the t-th period.

12. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 1, characterized in that, The method for determining the dynamic default loss rate in S3 is: C1: Classify the shipping lease objects into secured and unsecured categories to obtain the classification result; C2: Statistically calculate the average loan recovery rate of the classification result in C1 in the shipping lease measurement basic data; Calculate the default loss rate according to the average loan recovery rate corresponding to the classification result, expressed as: （9） is the loss given default for the t-th period; C3: Screen macroeconomic indicators whose relevant values with the default loss rate in the current period meet the preset threshold; substitute the macroeconomic indicators that meet the preset threshold into the preset multiple regression equation to calculate the adjustment factor value at the default loss rate; Subtract the adjustment factor value in the current period from that in the previous period to obtain the second adjustment value; C4: Use the second adjustment value to adjust the default loss rate to obtain the adjusted default loss rate; the adjusted default loss rate is the forward-looking default loss rate; the forward-looking default loss rate is expressed as: （10） （11） is an intermediate parameter; is the forward-looking default loss rate in the t-th period, is the adjustment value in the t-th period; C5: Determine different shipping lease scenario models; determine the macroeconomic indicators whose relevant values with the default loss rate in the current period meet the preset threshold under different shipping lease scenario models, repeat the calculation process of steps C3 - C4, and calculate the forward-looking default loss rates under different shipping lease scenario models; C6: Set the weights corresponding to the forward-looking default loss rates of different shipping lease scenario models; perform weighting through the forward-looking default loss rates corresponding to different shipping lease scenario models and the set weights to calculate the dynamic default probability parameter, expressed as: （12） Among them, each scenario represents the forward-looking default loss rate under different shipping lease scenario models in the t-th period, represents the dynamic default loss rate in the t-th period.

13. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 1, characterized in that, Determine the accounting subject type according to the business type of the shipping lease; select the shipping lease contract data according to the accounting subject type to calculate the dynamic default risk exposure; when the business type is a finance lease business, the designated accounting subject is the receivable finance lease payment; when the business type is an off-balance-sheet loan commitment business, the designated accounting subject is the off-balance-sheet loan commitment.

14. A method for constructing an expected credit loss model for the shipping leasing industry according to claim 1, characterized in that, The expected credit loss models corresponding to each credit impairment level in S5 are: S61: When the credit impairment level is level one, the corresponding expected credit loss model is; （13） Among them, ECL is the expected credit loss; EIR is the actual interest rate for each period; is the dynamic default probability for the first period; is the dynamic loss given default for the first period; is the dynamic exposure at default for the first period; S62: When the credit impairment rating is level 2, the corresponding expected credit loss model is; (14) Where: ECL is the expected credit loss; is the dynamic default loss rate in the t-th period; is the dynamic default probability in the t-th period; is the dynamic default risk exposure in the t-th period; S63: When the credit impairment level is level three, the corresponding expected credit loss model is; （15） Among them, ECL is the expected credit loss, PD = 100%, is the dynamic default loss rate in the t-th period, is the dynamic default risk exposure in the t-th period.

Images