IFRS9 Critical Issues and Possible Solutions - The IRB (Internal Ratings Based) Approach Risk-Enterprise Analytics

Transcription

1 IFRS9 Critical Issues and Possible Solutions - The IRB (Internal Ratings Based) Approach Risk-Enterprise Analytics

2 Introduction The stated objective of IFRS9 is to establish principles for the financial reporting of financial assets and financial liabilities that will present relevant and useful information to users of financial statements for their assessment of the amounts, timing and uncertainty of an entity s future cash flows Essentially, IFRS9 is therefore about accounting for a firm s financial assets and liabilities in a manner that is consistent with the fundamental concepts of accounting that have been in place since accounting first began. IFRS9 is not however intended to fundamentally change an institution s existing practices and processes with regard to the assessment, quantification and surveillance of its credit risk exposures. IFRS9 also clearly emphasises throughout the intended use of reasonable and supportable information that is available without undue cost and effort. It does not, therefore, intend nor envisage - the adoption of dubious conceptual approaches that are reliant upon assumptions derived from high-level analysis of data that have questionable representativeness or pertinence to a financial institution s actual assets and liabilities. Equally, IFRS9 does not necessitate nor encourage an adaptation of unproven approaches that are both time-consuming and costly to implement and (in worst cases) potentially may produce spurious prospective ECLs (expected credit losses) that are dubious in nature. Moreover, while different regulatory requirements (such as Basel and IFRS9) are aimed at different aspects of good credit risk management, all expect institutions to effectively understand, measure, monitor and manage the credit risks to which they are exposed. Analytical consistency (and the verifiable demonstration of that consistency) between assumptions used to manage the business in practice, derive Basel calculations, produce IFRS 9 calculations, and other credit-related calculations are essential and we expect regulators to insist upon analytical consistency going forward after the expiration of any initial transition concessions. Consequently, the adoption of divergent methodologies that lack analytical consistency merely stores up problems for the future in terms of costs and resources that would best be avoided by following a coherent approach outset. The Advantages of an IRB (Internal Ratings Based approach) Irrespective of whether such approaches have been formally recognised for Basel capital allocation purposes, most lending institutions already employ IRB approaches for running their businesses, determining lending criteria, and quantifying and surveilling credit risk. Such methodologies reflect sound credit risk practices built up, enhanced and validated over many years. The extension of the application of existing IRB methodologies for IFRS9 reporting purposes avoids the critical issues and problems outlined above.

3 In particular, an IRB approach leverages and builds on the analysis already performed as a matter of course for normal everyday business operations as well as (in many cases) for Baselcompliance purposes. Importantly, such an approach retains analytical consistency with an institution s existing risk management framework and practices. Furthermore, an IRB-based approach is more efficient and avoids the costs (but questionable benefits) of investing in incompatible alternative approaches to ECL computations, A robust IRB (internal ratings based) approach already means that internal ratings are assigned and reviewed taking into consideration: past, current and prospective economic conditions; a prospective view of the borrower s credit quality; and market information including external indicators and market prices. Indeed, it is standard best practice within an IRB approach: To incorporate the firm s past performance within financial analysis To evaluate the extent to which past financial performance is expected to change going forward To consider the business profile of the borrower and how that is likely to change prospectively To incorporate past and forecast economic and sector-specific conditions To utilise external indicators (such as third-party ratings, market prices, and spreads) to refine the analysis (particularly when such indicators suggest different credit outcomes than those derived from the application of the institution s core methodology). Therefore, by commencing with the existing IRB outcome, a sound (and analytically consistent) IFRS9 methodology should focus on: the borrower s latest TTC PD rating; the institution s past observed default experience; prospective in-house forecasts of GDP (or sector GDP) and should partition the TTC PD rating into its prospective PIT PD rates as a function of these inputs. In summary, the advantages of determining IFRS9 ECL calculations directly from an institution s normal IRB-based outcomes are: (a) analytical consistency (with other measures of credit) from outset; (b) much greater efficiency in terms of time and cost; and (c) the avoidance of dubious ECL computations (arising from inconsistent models) The IRB-Based Approach to IFRS9 An Overview Step 1: Updating the Asset s Credit Score The IRB-based approach commences by reviewing the borrower s credit score or rating in accordance with the firm s usual business practices. Step 2: Reviewing an Asset s IFRS 9 ECL Category The next stage is to determine the ECL category of the borrower for IFRS 9 purposes. This is facilitated by comparing the borrower s latest score / rating with that at the outset or initial recognition (as illustrated below).

4 We would generally recommend adoption of the following rules for classification (although institutions are of course free to follow different rules): A low credit threshold of 5.50 (equivalent to BBB- on the Rating Agency scale) i.e. any borrower currently scored at 5.50 is designated as category 1 for IFRS 9 purposes (with ECL calculated on the 12-month ECL basis) irrespective of any changes in its rating / score since initial recognition. An impaired threshold of 9.75 (equivalent to CC or worse on the Rating Agency scale) i.e. any borrower scored at 9.75 or worse is automatically designated as category 3 for IFRS 9 computation purposes. An absolute increase of 5% or more per annum in the annual (through-the-cycle) PD rate (from original recognition of the asset or liability) leads to a category 2 designation for ECL purposes. Step 3: Determining Analytically Consistent Prospective PIT PD Rates for the IFRS 9 Exercise The second component of the IFRS9 framework derives PIT (point-in-time) PD rates for use in the subsequent computation of expected credit losses (ECL) for individual assets and liabilities. Based on our analysis, GDP measures (and their derivatives), combined with observed past default experience appear to facilitate the best predictor of prospective default experience (relative to the assigned TTC PD rates). Such a relationship also makes intuitive sense given the importance of the economy to credit risk. Therefore, prospective PIT rates are derived as a function of (a) the current TTC (through-thecycle) 1-year PD rates by rating category / score, (b) the firm s observed past default experience by (TTC) scoring category, (c) the current borrower profile at the calculation date (i.e. the

5 number of borrowers by each rating / scoring category, and (d) GDP growth rates (historical and projected). The first set of inputs (as illustrated below) for this exercise are: The associated TTC (through the cycle) or LRA (long run average) default rates associated with each rating category / score together with the expected minimum and maximum default rates for the same categories. The firm s current issuer profile The firm s observed default experience by rating category / score When an institution has a fully (statistically) credible volume of in-house defaults then that experience should be used as a basis for the above inputs. However, for many asset classes (e.g. large corporates and specialised finance) the in-house experience will often lack full statistical credibility. In such cases, it is appropriate to first map the in-house model to (for example) rating agency methodologies and we can then provide rates for input based on that wider global experience. (For partially credible in-house default experiences, a credibility-weighted average of the in-house default experience and the wider experience is the correct approach to employ for these inputs).

6 The second set of inputs (as illustrated below) relate to observed past GDP values together with forecast GDP values. The optimal GDP-based metric can vary by country and we supply and recommend deployment of that metric which (based on past observations) has been most predictive of prospective default experience (on a normalised, aggregate basis) for the country concerned. The IRB-based approach first derives normalised aggregate default rates as a function of the above inputs using the current borrower profile for the purposes of normalisation. Next, the methodology derives rating or score-specific predicted PIT PD rates as a function of (a) the derived NADRs (normalised aggregate default rates), (b) the TTC 1-year PD rates for each score, and (c) the input minimum and maximum PIT PD rates by each score. As a by-product, the methodology also determines analytically consistent prospective adjustment factors for later use in determining prospective LGD (loss given default) values in connection with ECL (expected credit loss) computations for individual assets / liabilities. This exercise is typically carried out just prior to conducting the instrument-specific ECL calculations. Step 4: ECL Calculations

7 The methodology then determines (as required under IFRS 9) the expected credit loss (ECL) for each asset / liability using the analytically consistent prospective PIT PD values and the asset s specific characteristics. The initial inputs are illustrated below: Current Year is the year of calculation Current Score is the latest TTC PD Score Current Economic Value of Assets and Current Volatility are the current values for the borrower that would be deployed in the LGD methodology EIR is the contractual interest rate on the loan unless the facility is classified as impaired (in which case the EIR represents the IRR internal rate of return calculated at first recognition of impairment to equate prospective contractual cashflows to the fair value of the asset at the recognition date) The second band of inputs (as illustrated below) required are those that have been derived for the portfolio as a whole (in step 3) i.e. the prospective PIT PD values for the exercise (and associated adjustment factors for prospective LGD computations)

8 The EVA points are used for determining for computing prospective LGD values. The PIT PD rates (and the TTC PD Rate) are those that correspond to the borrower s current PD score (and were derived during step 3). The final set of required inputs (as illustrated below) relate to the contractual cash flows of the asset / liability is assessed. Timing relates to when the contractual payment (shown in amount) is due and is expressed in years measured from the effective ECL calculation date. Amount is the contractual payment. Exposure at Default is the expected EAD prior to payment (or receipt) of the corresponding contractual payment. Prospective LGD values are also derived by calling the LGD model for each future cashflow period. The ECL (expected credit losses) for the asset / liability are then calculated in accordance with the IFRS 9 requirements (as illustrated below)

9 Contact us EMEA Risk-Enterprise Limited Risk-Enterprise copyright Disclaimer The information contained in this brochure is both of a specific and general nature and is provided for your guidance and assistance. Whilst any recommendations are intended to advise or direct you on the specific actions believed to be required by your business, neither the Document s authors nor Risk-Enterprise Limited will accept liability or responsibility for any loss or damage suffered by you or another party as a result of reliance being placed on the information or guidance supplied. If you wish to unsubscribe please let us know by .