46.2. Credit risk

[Financial notes are presented in PLN thousand]

Credit risk is one of the basic risks associated with activities of the Group. The percentage share of credits and loans in the Group’s statement of financial position makes the maintenance of this risk at safe level essential to the Group’s performance. The process of credit risk management is centralized and managed mainly by Risk Management Division units, situated at the Bank Head Office or in local units.

Risk management process covers all credit functions – credit analysis, making credit decisions, monitoring and loan administration, as well as restructuring and collection.

These functions are conducted in compliance with the Bank’s credit policy, adopted by the Bank’s Management Board and the Bank’s Supervisory Board for a given reporting year. The effectiveness and efficiency of credit functions are achieved using diverse credit methods and methodologies, supported by advanced IT tools, integrated into the Bank’s general IT system. The Bank’s procedures facilitate credit risk mitigation, in particular those related to transaction risk evaluation, to establishing collateral, setting authorization limits for granting loans and limiting of exposure to some areas of business activity in line with current client’s segmentation scheme in the Bank.

Credit granting authorizations, restrictions on crediting the specific business activities as well as internal and external prudential standards include not only credits, loans and guarantees, but also derivatives transactions and debt securities.

The Bank’s lending activity is limited by the restrictions of the external regulation as well as internal prudential standards in order to increase safety. These restrictions refer in particular to credit exposure concentration, credit quality ratios and exposure limits for particular foreign countries, foreign banks and domestic financial institutions.

The Bank established the following portfolio limits in the Bank’s credit policy:

• exposure limits for sectors of economy,
• limits on the concentration of the largest exposures to entities / groups of related entities,
• limits for main business lines and currency receivables,
• product limits (mortgage loans to private individuals, exposures to business entities secured by mortgage, inculidng financing commercial real estate).

The internal limits system operating in the Bank also includes a number of detailed limits supporting key limits set out in the credit policy.

Moreover, the Bank limits higher risk credit transactions, marked by excess risk by restricting the decision-making powers in such cases to higher-level decision-making bodies.

The management of the Bank’s credit portfolio quality is further supported by regular reviews and continuous monitoring of timely loan repayments and the financial condition of the borrowers.

Armed conflict in Ukraine

In connection with Russia’s armed attack on Ukraine, which has been ongoing since 2022, the Group identifies the following threats in the area of credit risk:

• credit loss risk for exposures to entities from Russia, Belarus and Ukraine, with the Group’s exposure in this regardmostly covered by KUKE policies,
• the risk that the conflict will translate into deterioration of the economic and credit conditions for the rest of the portfolio (through the raw material price growth channel, disruption of economic relations, deterioration of consumer sentiment, etc.).

As at 31 December 2022, the Group’s balance sheet net exposure to countries involved in the conflict amounted to PLN 225 million (which represents 0.14% of the Group’s total exposure).

The tables below present the Group’s exposures to countries involved in the armed conflict in Ukraine as at 31 December 2022 and 31 December 2021.

For credit risk management purposes, the Group uses the internal rating models depending on the client’s segment and/or exposure type.

The rating process is a significant element of credit risk assessment in relation to clients and transactions, and constitutes a preliminary stage of the credit decision-making process of granting a new credit or changing the terms and conditions of an existing credit and of the credit portfolio quality monitoring process.

In the credit risk measurement the following three parameters are used: PD, LGD and EAD. PD is the probability of a client’s failure to meet its obligations and hence the violation of contract terms and conditions by the borrower within one year horizon, such default may be subject-matter or product-related. LGD indicates the estimated value of the loss to be incurred for any credit transaction from the date of occurrence of such default. EAD reflects the estimated value of credit exposure as at such date.

The risk parameters based on the rating models are designed for calculation of the expected losses resulted from credit risk.

The value of expected loss is one of the significant assessment criteria taken into consideration by the decision-making bodies in the course of the crediting process. In particular, this value is compared to the requested margin level.

The level of minimum margins for given products or client segments is determined based upon risk analysis, taking into consideration the value of risk parameters assessed.

The client and transaction rating, as well as other credit risk parameters hold a significant role in the Credit Risk Management Information System. For each rating model, the credit risk reports provide information on the comparison between the realized parameters and the theoretical values for each rating class.

Credit risk reports are generated on a monthly basis, with their scope varying depending upon the recipient of the report (the higher the management level, the more aggregated the information presented). Credit risk reports are being used in the credit risk management process.

For internal purposes, within the Group the following rating models are used, developed in accordance with provisions of Regulation (EU) no 575/2013 of the European Parliament and of the Council of 26 June 2013 on prudential requirements for credit institutions and investment firms:

• For the retail clients, the Group uses the following models applicable for:
  • micro-enterprises,
  • private individuals, dividing clients into:
    • mortgage loans (secured by mortgage),
    • consumer loans (consumer),
    • credit cards,
    • renewable limits.
• For the corporate clients, the Group uses rating models dividing clients into:
  • • corporate clients (corporations),
    • small and medium enterprises (SME),
    • local government units.
• For the corporate clients, Pekao Bank Hipoteczny S.A. uses the SOP rating model (Point Rating System) under the Internal Ratings Based Approach, which involves the use of supervisory classes in the process of assigning risk weights.
• For specialized lending the Group uses a slotting criteria approach to the Internal Ratings Based Approach, which consists of the use of supervisory classes in the process of assigning risk weights..

In 2021, the Group started the process of adjusting the rating scale for internal rating models in line with the rating scale applicable to external ratings – called Master scale.

The masters scale is presented in the table below:

At the end of 2022, the rating models within the corporate client / enterprise segment and the private individuals within retail clients segment were mapped to the Masterscale.

The following exposure types are not covered by internal rating models:

• retail exposures immaterial in terms of size and perceived risk profile:
  • overdrafts,
  • exposures related to the Building Society (Kasa Mieszkaniowa) unit,
  • other loans.
• corporate clients:
  • exposures to stock exchanges and other financial intermediators,
  • exposures to insurance companies,
  • project financing,
  • purchased receivables,
  • exposures to investment funds,
  • exposures to leasing companies and financial holding companies,
  • other loans immaterial in terms of size and perceived risk profile.
• exposures to regional governments and local authorities which are not treated as exposures to central governments, for which the number of significant counterparties is limited.

The tables below present the quality of the loan portfolio.

The distribution of rated portfolio for retail client segment (excluding impaired loans)

The distribution of rated portfolio for individual client segment (excluding impaired loans)

The distribution of rated portfolio for corporate client segment (excluding impaired loans)

The distribution of rated portfolio for corporate client segment (excluding impaired loans)

The distribution of rated portfolio for local government units segment (excluding impaired loans)

The distribution of the portfolio exposure to specialized lending (excluding impaired loans)

Portfolio of exposures not covered by the rating model (excluding impaired loans), broken down by delays in repayment

Portfolio of impaired exposures, broken down by delays in repayment

Decision-making level connected with transaction approval is directly dependent upon the client’s rating.

Decision-making entitlement limits are associated with the position held, determined in accordance with the Bank’s organizational structure. The limits are determined taking the following matters into consideration:

• the Bank’s total exposure to a client, including the amount of the requested transaction,
• type of a client,
• commitments of persons and entities associated with the

The internal validation of models and risk parameter assessments is focused on the quality assessment of risk models and the accuracy and stability of parameter assessments, applied by the Bank. Validation is carried out at the level of each risk model, although the Bank may apply several models for each class of exposures.

Moreover, the internal audit unit is obligated to review the Bank’s rating systems and their functionality at least once a year. In particular, the internal audit unit reviews the scope of operations of credit division and estimations of risk parameters.

Division of loans and advances to customers for covered and not covered by internal rating models

Division of off-balance sheet exposures to customers (loan commitments and financial guarantee contracts) for covered and not covered by internal rating models

Classification of loans and advances to banks according to Fitch ratings

Classification of exposures to debt securities according to Fitch ratings (*)

Classification of exposures to derivative financial instruments according to Fitch ratings

The Group has recognized impairment allowance in accordance with the IRFS 9. IFRS 9 assumes the calculation of impairment losses based on expected credit losses and taking into account forecasts and expected future economic conditions in the context of credit risk exposure assessment.

Expected credit loss model applies to financial assets classified, in accordance with the IFRS 9, as financial assets at amortized cost or at fair value through other comprehensive income, with the exception of equity instruments (except for equity instruments), as well as off-balance sheet commitments.

Expected credit loss model in accordance with IFRS 9 is based on the allocation of exposure to one of the three stages, depending on credit quality changes compared to the initial recognition of assets in the accounting records. How to calculate the impairment loss depends on the stage.

In addition, financial assets that were classified as POCI at the time of initial recognition are treated as POCI (i.e. purchased or originated credit-impaired) in all subsequent periods until they are derecognised. This rule applies even if, in the meantime, the asset has been healed. In other words, assets once recognized as POCI remain in this status regardless of future changes in estimates of their cash flows.

In the case of instruments with the POCI status, life-time expected credit losses are recognized throughout the lifetime of these instruments.

Calculation of expected credit losses

For the purpose of calculating the credit loss in accordance with IFRS 9, the Group compares cash flows that it should receive pursuant to the agreement with the borrower and flows estimated by the Group that it expects to receive. The difference is discounted using the effective interest rate.

Expected credit losses are determined in the contractual maturity period with the exception of products meeting the criteria of IFRS 9 para. 5.5.20, for which the Group determines the expected losses in the period in which it is exposed to credit risk (ie in the economic maturity).

Methodology for calculating group parameters – PD, RR and EAD.

The lifetime ECL calculation requires the use of long-term risk parameters.

Multi-year PD parameters are an assessment of the probability of a default event in the next annual intervals in the lifetime horizon. The long-term PD curve for a given exposure depends on the current value of the 12M PD parameter (and the appropriate rating class) determined based on the internal PD models of the Group. In the estimation, the Group:

• estimates unbiased PD parameters without taking into account additional margins of conservatism (IFRS 9, paragraph5.5.17 (a)),
• takes into account current and forecasted macroeconomic conditions (IFRS 9, paragraph 5.5.17 (c)).

The calculation of expected recovery rates (RR) is based on the ‘pool’ model, in which, within homogeneous groups, average monthly recoveries are calculated conditionally against the months since default (MSD). Homogeneous groups of accounts were separated on the basis of the following characteristics:

• the type of a borrower,
• product type,
• ranges of the LTV parameter (for mortgages and housing loans) or credit amount (for chosen products).

As part of defined homogeneous groups, average monthly recovery rates are calculated, which consist of repayments and recoveries resulting from both the secured part and the unsecured exposure, weighted by the value of outstanding capital observed at the beginning of a given MSD.

For products for which a repayment schedule is available, the Group sets the exposure value at the moment of default (EAD, Exposure at Default) and principal at the moment of default (PAD, Principal at Default) in the lifetime (ie for future repayments) based on contractual payment schedules and taking into account the following effects:

• the effect of arrears on principal and interest installments related to the expected non-payment of the last installments prior to the occurrence of the default,
• the effect of arrears of payments (principal and interest) on the date of calculation of the provision,
• the effect of settlement of the EIR adjustment over time.

For products for which a repayment schedule is not available, the Group sets the long-term EAD and PAD using the CCF (Credit Conversion Factor) and parameters. CCF parameters vary depending on the portfolio and the time horizon of EAD / PAD estimation.

For exposures for which it is not possible to determine risk parameters based on internal models, the Group adopts an approach based on using parameters from other portfolios with similar characteristics.

The models and parameters used to calculate impairment allowance are periodically validated.

Changes in the methodology of calculation an expected credit losses introduced in 2022

In 2022, the Group introduced changes to the rules for calculating allowances related to the implementation of the updated Recommendation R of the Polish Financial Supervision Authority. The changes included:

• The use of the so-called New Default Definition (‘NDD’) in line with the EBA/GL/2016/07 guidelines,
• Updating the portfolio segmentation to bring it into line with NDD,
• Reconstruction of the default probability (‘PD’) model to better reflect the risk level dependence of the exposure age for,

• Reconstruction of the Transfer Logic (‘TL’) model in order to ensure that the thresholds for classification to Stage 2,
• Other changes related to the above (including recalibration of other models).

The main changes related to the implementation of NDD in the area of calculation an expected credit losses (in terms of the capital adequacy, NDD was implemented at the beginning of 2022) are: adjustment of segmentation (division of the portfolio into retail and non-retail clients with an additional division of retail clients into sub-portfolios by segment or product ) taking into account the relative threshold in the calculation of the days past due, adjustment of the absolute threshold in the calculation of days past due, adjustment of the rules of contagion of default exposures, taking into account the quarantine for qualitative premises and taking into account additional qualitative indications of unlikeliness to pay.

In area of segmentation, the division of portfolios for all relevant models used in the estimation of allowances was adjusted to the segmentation used under NDD, which was not ensured in the previous approach, where the segmentation for each model was independent. The alignment addresses the highest level of segmentation and ensures consistency in the application of NDD and all models used. At the lower level of segmentation, the divisions appropriate for the modeled observation / risk parameter are used.

In area of the PD model, changes are of a different nature for the modeled portfolios (for which the Bank has sufficient historical data and uses them to set risk parameters) and benchmark portfolios (for which it does not have sufficient data and determines risk parameters based on internal benchmarks extrapolated from other portfolios or external data). The main change for the modeled portfolios in relation to the previous approach is the use of the migration matrix, instead of the survival analysis, to estimate the risk of default, which in a more consistent way allows for taking into account time dependencies such as the survival effect (quick entry into default loans with high PD and improvement of quality credit portfolio) or the effect of negative selection (faster repayment of good loans and remaining in the portfolio of loans with an average higher risk in the late years after origination). Additionally, migration matrices allow for effective use of historical data to determine the dependence of PD on credit age and are resistant to potential data disturbances, which is important when using long time series. For benchmark portfolios, the most important change consists in replacing the periodic expert assessment used so far with an algorithmic approach based on the long-term average loss ratio of the analyzed portfolio, or with the external rating of the exposure / client.

As regards the TL model, the approach was completely rebuilt in order to meet the requirements of Recommendation R. The measure of credit risk change was simplified, which is determined as the quotient of the average annual PD value over the exposure life horizon as at the reporting date and the initial recognition date determined in accordance with the previous paragraph. The increase in risk measured by this measure is considered significant if it exceeds the established threshold. This threshold is 2 increased by the calibration parameter ‘a’. Calibration of parameter ‘a’ is performed separately for each homogeneous group of portfolios modeled to correspond to the Group’s risk appetite in the period at the time of origination the transaction. The parameter ‘a’ determined in a given quarter is assigned to all exposures that will be defined as initially recognized in the next quarter of the recognition and parameter is constant during the life of the exposure. The described criterion of classification to Stage 2 allows to minimize the deviations of the exposure valuation from the hypothetical valuation in which the write-offs would be estimated as a change in the lifetime loss expected from the moment of initial recognition. The second quantitative criterion for classification into Stage 2 is the value of one-year PD determined using scoring / rating models above the level of 25%. This criterion results from the fact that the Bank granting loans does not accept the risk higher than approximately 10%. A 25% PD therefore by definition means a significant increase in credit risk. The last quantitative criterion for classification into Stage 2 is the benchmarking results based on NBP data and the average long-term DR (default rate) of a given portfolio. If the share of Stage 2 in the Bank is lower than the long-term average for the polish banking sector in a given portfolio (or three times DR), then the Bank classifies exposures into the Stage 2 until the average is reached, where the credits are moved in the order corresponding to their distance from Stage 2 in based on the other 2 criteria mentioned before. This approach ensures, to the minimum extent required, the consistency of the shares of Stages in the Bank with the average share in the banking sector. The described solution replaced the quantile regression used so far in order to statistically identify significant changes in risk. Each of the three criteria described is applied separately.

The other significant changes to the models concerned the consistency of segmentation for other models from NDD (for the recovery rate / RR / and exposures at default / EAD / models) and the calculation of these risk parameters into NDD time series.

However, compared to the end of 2021, the bank did not introduce any significant changes in portfolio quality forecasting and continues to use trend analysis for retail portfolios and quantitative / expert analysis for the other portfolios. In particular, due to the instability of internal and external conditions, the probability of the pessimistic scenario (50%) is still high.

In total, the changes described above did not have a significant impact on the level of expected credit losses on the date of implementation (end of April). The amount of impairment losses for the Bank decreased by PLN 3 million. Changes in the default definition resulted in decrease in the level of assets classified in Stage 3 by PLN 147 million gross carrying amount, mainly due to the implementation of a relative threshold (1%) in the calculation of days past due.

In 2022, the Bank selected customers operating in higher-risk industries and increased PD on them by 100%, resulting in a PLN 197 million increase in expected credit losses in the working capital portfolio. This impact was taken into account for loans with a total gross carrying amount of PLN 15 832 million. The analysis of industries took into account the indirect impact of the armed conflict in Ukraine, the marked deceleration in domestic demand and investment, the burden of interest costs resulting from loans and advances (due to the high level of NBP interest rates) and the demand of individual branches of industrial processing. Adjusted industries with the largest share in the Bank’s loan portfolio are, by PKD division, as follows: 49 land transport and pipeline transport, 42 civil engineering works, 55 accommodation, 41 construction works for the erection of buildings, 77 rental and leasing, 23 manufacture of other non-metallic mineral products, 24 manufacture of metals.

The industry analysis took into account the indirect impact of the armed conflict in Ukraine.

Sensitivity analysis of ECL in established changes of PD and RR/LGD parameters

The table below presents the results of the ECL sensitivity analysis for the assumed changes in PD and RR/LGD parameters carried out separately for exposures subject to individual and group analysis. For the exposures included in the Bank analysis, the PD and recovery rate (1-RR=LGD) increase and decrease by 1% and 5% scenario were presented compared to the values used to calculate the expected credit loss as of date 31 December 2022. For the exposures analyzed individually, the estimated impact is presented as a reduction of recoveries from collaterals included in the debt collection scenario by 10%.

Changes in impairment allowances level (ECL) in different scenarios of changing the influencing parameters for the calculation of write-offs (in millions of zlotys).

According to par. 5.5.10 IFRS 9 exposures that are considered as low risk credit exposures at the reporting date may remain in Stage 1, regardless of the scale of the relative credit deterioration from the initial recognition. According to par. B.5.5.22 of IFRS 9, the credit risk of a financial instrument is considered low when:

• the financial instrument has a low risk of default,
• the borrower has a strong capacity to meet its contractual cash flow obligations in the near term
• adverse changes in the economic and business conditions in the long term may, but will not necessarily, reduce the ability of the borrower to fulfil its contractual cash flow obligations.

The Group applies a low credit risk criterion for three portfolios: exposures to banks, exposures to local government units and exposures to the State Treasury and the National Bank of Poland.

Financial assets for which at the balance sheet date the Group will identify a significant increase in credit risk from the initial recognition are classified in Stage 2. The Group recognizes that for a given asset a significant increase in credit risk has been identified if a quantitative or qualitative criterion is met, in particular if contractual payments are more than 30 days past due, where the occurrence of a given criterion is verified at the exposure level.

Quantitative criteria

Taking into account the requirements of the standard, the Group defined two basic characteristics of the quantitative model:

• the measure on the basis of which risk change assessment is made,
• the materiality threshold of the measure, above which the Bank recognizes that there has been a significant increase in credit risk.

The measure, on the basis of which risk change assessment is made, was set by the Group as the ratio of:

• current credit risk assessment defined as lifetime PD in the horizon from the reporting date to the maturity date determined on the basis of the characteristics effective as at the reporting date,
• the original credit risk assessment defined as lifetime PD in the period from the reporting date to the maturity date determined on the basis of the characteristics applicable as at the date of initial recognition.

The assessment of significance of credit risk deterioration is carried out by comparing the observed measure with the threshold above which the Group considers that a significant deterioration in credit risk occurred.

The allocation threshold is designated as the reporting date at the single exposure level by a statistical model based, among others, on information on the credit risk assessment as of the date of the initial recognition, the time from the date of the initial recognition of the exposure and historical price volatility.

The tables below present the arithmetic average values of the risk change measure (*) as at 31 December 2022 and 31 December 2021 determined for the most significant portfolios covered by the quantitative model.

(*) The measure on the basis of which the risk change is assessed is determined by the Bank as the ratio of:

• current credit risk assessment defined as lifetime PD in the horizon from the reporting date to the maturity date, determined on the basis of the characteristics applicable as at the reporting date,
• original credit risk assessment defined as lifetime PD in the period from the reporting date to the maturity date, determined on the basis of the characteristics valid at the date of initial recognition.

The change in the average values of the risk change measure is a consequence of the complete reconstruction of the TL model and the conditions for including exposures in this model.

Qualitative criteria

As a result of the monitoring process carried out by the Group, the qualitative criteria for the allocation to Stage 2 are identified, such as:

• the amount of arrears simultaneously above the set materiality threshold (PLN 400 for retail exposures and PLN 2 000 for non-retail exposures) and the relative threshold of 1% for over 30 days up to 90 days inclusive,
• a delay in repayment over 90 days,
• occurrence of forbearance status,
• exposure is on the Watchlist.

In addition to the above, for individual monitoring the Group has defined a number of specific quality criteria for various types of portfolios, such as, inter alia, changes in the internal rating, changes in supervisory classes for selected segments (eg specialized financing), warning signals identified in the monitoring system and credit risk management or the results of individual monitoring.

In the case of granting credit holidays under the Act on crowdfunding for business ventures and assistance to borrowers of July 14, 2022, the Group applies an approach consistent with regulatory guidelines in this regard. Granting credit holidays does not result in automatic reclassification to Stage 2. However, such reclassification is performed if the deterioration of credit risk is affected by additional factors indicating the debtor’s problems. During the credit holidays, the Group suspends the counting of overdue days.

Financial assets for which at the balance sheet date the Group has identified occurrences of the default event are classified in Stage 3.

The Group recognizes that for a given asset a default was identified if at least one of the following occurred:

• amount of arrears simultaneously above the set materiality threshold (PLN 400 for retail exposures and PLN 2 000 for non-retail exposures) and the relative threshold of 1% for over 90 days,
• exposure during the restructuring process,
• other qualitative impairment trigger.

For SME and corporate segments, default is identified at the customer level, whereas for the retail segment at the customer/product group level. The criterion of days and amounts of delays is also defined at the level of identification.

The Group applies a six-month quarantine period effective from the moment all defaults cease to exist.

At the end of 2022, the Group additionally included CHF mortgage loans in Stage 3 in accordance with the principles presented in the Note 46.3.

Forecast of risk parameters

Based on significant inertia of retail portfolios, a trend analysis of historical default rates have been applied. Based on the history of realized default rates for portfolios of retail exposures, trends were estimated, which were then used for future projections. For non-retail portfolios projections are based on expert judgment of the economic conditions applied to the long term average through the cycle parameters. The analysis for non-retail portfolios consists of the following steps: an expert evaluation of the forecasted economic conditions based on Group’s projections and studies carried out by the Central Statistical Office in Poland (GUS), translation of this evaluation onto quantitative measure at the scale 0-100% indicating the phase of the economic cycle (e.g. 75% represents situation where in the past 75% of observation situation is better and in 25% is worse), finally getting the corresponding quantile of the historical default rates and use of it as the forecast for first year. For the second year forecast assumes the linear convergence to average through the cycle parameters which is assumed to take place in the fifth year (which mirrors few years long credit cycles). In 2022, the history of DR default rates was updated in accordance with NDD.

Table below shows 12-month PD forecasts used in the calculation of expected credit losses in baseline scenario. For retail portfolios the parameters are weighted with the gross carrying amount limited to PLN 2 million at the loan level and at the customer level for SME loans. For non-retail, the parameters are weighted with the gross carrying amount limited to PLN 20 million at the client level.

Scenarios definition

The PD parameters presented in the previous section refer to the baseline scenario of portfolio quality development. They reflect the assumption of a moderate economic slowdown amid persistent high inflation and interest rates (GDP growth by about 5%, average annual inflation of about 14% and WIBOR 3M at the end of the year over 7%). The assumptions for the remaining scenarios and the weights assigned to them are presented below.

In the applied approach the Group used 3 scenario of evolution of quality of the portfolio: baseline (presented above), upward (assuming positive change in the credit quality of the portfolio in the next years compared to the baseline) and downward (assuming negative change in the credit quality of the portfolio in the next years compared to the baseline). The baseline scenario has the probability of occurrence of 45%, upward of 5% and downward of 50%. High probability of downward scenario reflects Group’s expert judgment of the possibility of realization of some risks the economy of Poland faces and their significant impact on credit portfolio with regard to:

• the development and impact of the war in Ukraine on the Polish economy and the related disturbances in the supply of raw materials,
• maintaining or increasing interest rates, which translates into a high increase in the burden on certain customer groups,
• inflation remaining at record levels and its impact on some customer groups, in particular in terms of the increase in prices of energy and energy raw materials,
• greater than expected economic slowdown due to growing cost pressure on entrepreneurs,
• and possible mutations of COVID 19 and subsequent waves of the pandemic.

Individually the risk of these scenarios is equal or below 50% in the Group’s view but their number implies high risk of occurrence of one of them.

The diversified nature of the observed threats and the breakdown of the dependencies between the parameters of the quality of the loan portfolio and the macroeconomic variables indicated in the previous section means that it is impossible to formulate scenarios in the form of extreme changes in macroeconomic factors. Therefore, the Group applied an alternative approach in which the PD change scenarios are determined based on the historical variability of the DR. The downward scenario is assigned values corresponding to the high past observations, and similarly to the upward scenario, the values corresponding to the low past observations are assigned. This translates into the following 12-month PD forecasts for 2023.

The Group also carried out analysis confirming the lack of dependence of the recovery rates for non-performing exposures (RR parameter) on the economic situation. Therefore, the same recovery rates are assumed in each of the scenarios.

The subsidiaries of the Bank determine expected credit losses according to IFRS 9. Due to their characteristics and portfolios the scenarios used in the calculation of expected credit losses is not fully aligned.

Sensitivity analysis regarding the forecast of the macroeconomic situation

The Group estimates probability weighted expected credit losses taking into account 3 macro-economic scenarios: baseline (occurring with a probability of 45%), upward (assuming positive change of the quality of the portfolio in the next years compared to the baseline, occurring with a probability of 5%) and downward (assuming worsening of the quality of the portfolio in the next years compared to the baseline that could occur with a probability of 50%).

The changes in expected credit losses presented in the table below for exposures without impairment were designated as the difference between the expected credit losses calculated for a specific macroeconomic scenario and expected credit losses calculated taking into account all scenarios macroeconomic factors weighted with the probability of their realization (in accordance with IFRS 9).

The tables below present the changes in impairment allowances and gross carrying amount of financial assets not measured at fair value through profit or loss by class of financial assets:

The total value of undiscounted expected credit losses at the time of initial recognition of financial assets purchased or originated credit impaired in the period ended 31 December 2022 amounted to PLN 56 263 thousand.

The total value of undiscounted expected credit losses at the time of initial recognition of financial assets purchased or originated credit impaired in the period ended 31 December 2021 amounted to PLN 11 026 thousand.