Page 99 - Annual Report 2023
P. 99

ANNUAL REPORT 2023      30
                               Notes to the Consolidated Financial Statements for the year ended 31 December 2023
                                                                           (amounts in USD, unless otherwise stated)






          1. Probability of Default
          The probability of default has been modelled on the basis of the one factor Merton model. The approach to model the point-in-time
          forward looking and macroeconomic PD is divided into three different steps: - The first step is to calibrate the cumulated TTC PDs;
          - The second step is to include the forward look information and transform the cumulated TTC PDs to PIT PDs; - The final step is to
          describe the conversion of PIT PDs to TTC PDs after a certain horizon.
          Default rates are essential for the calculation of Point-in-Time Probability of Default (PD) and for the estimation of ECL.
          The Bank uses internal default rate data for wholesale and individual segments and  external default rate data for other segments. A
          number of macro-economic variables sourced from the IMF, including historical data starting from 1990 up to today and projections
          for the upcoming 4 years under three scenarios: baseline, best and adverse, were considered in modelling of PIT PD.
          For non-rated accounts, PD are estimated using the nominal weighted average PD of all rated accounts within the same rating system.

          2. Loss Given Default (LGD)
          Loss Given Default is defined as the percentage of the Exposure at Default (EAD) that is ultimately lost in case of default of a counterparty,
          after all possible recoveries through selling of collateral or collection procedures. As regards to the loan to customers’ portfolio, LGD is
          modelled based on collateral value and a number of parameters, such as Collateral Value, Time from Default to Possession, Probability
          of Cure, Probability of Possession, Time from Default to write off, Forced Sale Discount etc. All of these parameters are calculated
          based on internal credit history.


          A defaulted collateralised asset can move through different stages post default. The bank can seize the collateral (“Possession”) in
          order to sell it (“Sale”) and make up for the potential loss due to the default of the counterparty. Possession of collateral can occur
          voluntarily (“handing over the keys”) or via litigation (court proceedings). And sales may be carried out by the institution (after obtaining
          possession) or by means of a customer self-sale. On the date of collateral sale, any shortfall is recognised in the P&L (write-off expense)
          and subsequent recoveries (debt collection either internal or external) may occur. Write-off occurs when there is a shortfall on collateral
          sale. Closed and cures occur when the full outstanding is recovered with the former resulting in the account closing (i.e. no lending). Cure
          refers to both closed and curing accounts. For the unsecured/uncollateralised types of assets the value of the collateral is supposed to
          be 0 and the actual model is still applied taking in account pure debt collection. In the light of this recovery process, the Bank defines
          LGD as the expected severity (loss) given a default.

          For the Treasury and Project and Structured Finance accounts, the Bank applies a fixed LGD of 25% for secured exposures, and
          45% for unsecured exposures, based on the Basel III framework. For the Treasury and Project and Structured Finance portfolios, LGD
          values are assigned on an asset type level.


          3. Exposure at Default (EAD)
          An asset can have a customised, linear or bullet amortisation type. For assets with a customised amortisation type, repayment
          schedules are used to estimate EAD. For assets without any amortisation type, a linear repayment plan is assumed. For off balance
          sheet exposures, it is required that provisions are held against undrawn commitments. BKT’s calculation of the credit conversion factor
          (CCF) values is in line with Basel II requirements under the standardised approach: “Commitments with an original maturity up to one
          year and commitments with an original maturity over one year will receive a CCF of 20% and 50%, respectively.” Early repayment/
          refinance assumptions are also incorporated into the calculation. However the early repayments are considered to be 0 for all assets
          as the Bank’s historical data suggests insignificant material impact.
   94   95   96   97   98   99   100   101   102   103   104