Significant Increase in Credit Risk According to IFRS 9: Implications for Financial Institutions

The aims of accounting standard setting is substantially different to those of bank regulation. Financial reporting follows as a general purpose to provide information to those outside the firm to support decision usefulness. In contrast to this prudential bank regulation seeks to decrease the frequency and cost of bank failures and to protect the financial system as a whole by limiting the frequency and cost of systemic crises [1].


Introduction
The aims of accounting standard setting is substantially different to those of bank regulation. Financial reporting follows as a general purpose to provide information to those outside the firm to support decision usefulness. In contrast to this prudential bank regulation seeks to decrease the frequency and cost of bank failures and to protect the financial system as a whole by limiting the frequency and cost of systemic crises [1].
These differing objectives are at the center of a standing debate over loan loss accounting as reflected in recent high profile proposals: Impairment of Financial Assets-The Expected Loss Model (EFRAG), Basel Committee on Banking Supervision, and Guidance on accounting for expected credit losses. In response to the financial crisis and these proposals the IASB has introduced a new standard IFRS 9 on impairment [2], which requires a three-step approach, which in general replaces the current incurred impairment model with a new expected loss model. The new model is very much influenced by prudent bank regulation and the aim to reduce income smoothing [3].
The new model requires three stages of impairment considering changes in credit quality since initial recognition ( Figure 1).
In the first stage financial assets are included, which are exposed to a low credit risk. A significant increase in credit risk could not be measured since the initial recognition. Even if no impairment was incurred, for these assets the 12-month ECL is calculated. Interest revenue is calculated based on the gross carrying amount [3,4].
12-month ECL is the expected credit losses which are caused by default events that are possible within 12 months after the reporting date [2]. It is often misunderstood, as it is not the expected cash shortfalls over the 12-month period but the entire loss on an asset weighted by the probability that the loss will occur in the next 12 months. This stage is summarized as the under-performing state. In the second stage financial instruments are included which were exposed to a significant increase in credit risk since initial recognition and no objective evidence of impairment is provided. For these assets, lifetime ECL are recognized, but interest revenue is still calculated on the gross carrying amount of the asset. Lifetime ECL are the expected credit losses that result from all possible default events over the expected life of the financial instrument. Expected credit losses are the weighted average credit losses with the probability of default ('PD') as the weight [1].
The third stage includes financial assets for which objective evidence of impairment at the reporting date has taken place. For these assets, lifetime ECL is recognized and interest revenue is calculated on the net carrying amount (that is, net of credit allowance) [2].
In this three-stage approach the IASB has implemented a general change from the incurred to expected loss model. According to the Abstract This study explores the literature about definitions and concepts when a significant increase in credit risk is achieved. In response to the financial crisis the IASB has introduced a new standard (IFRS 9) on impairment, which requires a three-step approach, which in general replaces the current incurred impairment model with a new expected loss model. This research paper summarizes alternative impairment models and particuarly focus on the significant detreortaion criteria, which is a corner stone of the new IFRS 9 impiamrnet model. The expected loss model is not completely new within the accounting literature. The study provides early insights into implemention of IFRS 9 on impairment, as IFRS 9 will become applicable 2018. It is also relevant for regulators, as it becomes obvious due to the non-existance of a dominant approach; the question arises if the regulator should provide more guidance to avoid that all companies purse completely different model resulting in decreasing comparability for investors.
EFRAG (European Financial Reporting Advisory Group) and FEE (Federation of European Accountants), the issue with the incurred loss model is that impairment losses (and resulting write-downs in the reported value of financial assets) can only be recognized when there is evidence that they took place (-have been incurred). It is not permitted to reporting entities to subjectively consider expected losses.
The assumption is that prudent recognition of loan losses could have potentially decreased the cyclical moves in the recent financial crisis [5].
However according to Bushman and Landsman [5] it is not only a model change, but it also shows that regulation of financial reporting in light of the recent financial crisis needs to develop towards the prudent regulatory model. The reality is that the regulation of corporate reporting is just one piece of a larger regulatory configuration and that forces are at play that would "subjugate accounting standard-setting to broader regulatory demands" (p. 260). IFRS 9 illustrate this point.

Literature Review
This literature review starts to provide an overview of the different impairment models existing: Incurred Loss Model, Expected Loss Model, Fair-Value based Model and Dynamic Provisioning. The selection of the methods is based on an analysis of the EFRAG and the FEE, who provided a summary of the dominant models. This is followed by a detailed literature review on the increase of credit risk.

Incurred loss model
Under the incurred loss model, investments are regarded as impaired if there is no longer reasonable assurance that the future cash flows related to them will be either collected in their entirety or when due. Entities search for evidence of situations that would indicate impairment, such triggering events are seen when the entity: • Is experiencing notable financial distress.
• Has defaulted on or is past due on making interest or principal payments.
• Is likely to undergo a major financial reorganization or enter bankruptcy or • Is in a market that is experiencing significant negative economic change.
According to Leventis et al. [6] many researchers have concluded that loan loss provisioning is used as a tool for earnings management. However according to an empirical study about commercial banks they can draw an opposite conclusion that the introduction of IAS 39 in the past has mitigated earnings management behavior.

Expected loss model
Under an expected loss impairment model, estimates of future cash flows used to determine the present value of the investment are made on a continuous basis and do not rely on a triggering event to occur. Even though there may be no objective evidence that an impairment loss has been incurred, revised cash flow projections may indicate changes in credit risk. Under the expected loss model, these revised expected cash flows are discounted at the same effective interest rate used when the instrument was first acquired, therefore retaining a costbased measurement. Calculating the Impairment cost is the same as the Incurred Loss Model.

Fair-value based model
In a fair value based impairment approach, an entity would book impairment losses on a financial asset held at amortized cost to the extent that the fair value of that asset is less than its carrying amount. Fair value would include credit and noncredit related changes in fair value i.e. using market based values and discount rates. It is envisaged that a fair value-based impairment approach would incorporate a -trigger of some sort, since automatic adjustments in fair value movements would seem contrary to carrying a financial asset at amortized costs. This model would imply to be very volatile and clearly pro-cyclical, although not as much as a full fair value model.

Dynamic provisioning
The dynamic provisioning model contrary to the incurred loss model follows the main objective to improve the financial soundness of banks. Dynamic provisioning applies an ant cyclical approach, .i.e. in "good times" a loan reserve is set-up so it will not face insolvency due to charge-offs and provisions in "bad times" [7,8].

Discussion of Models Expected loss-model versus incurred loss-model
One of the main differences between the expected loss and the incurred loss model is that it is more subjective in nature, as it is based significantly on the estimate of cash flow expectations prepared by the reporting entity. This process of analyzing future cash flow streams is inherently subjective and represents potential to become susceptible for earnings management [4]. "Too little too late" was the summary often mentioned as part of the exposure draft discussion of IFRS 9. Critics of the accounting standard argue that the recognition after identification of evidence, such as a counterparty failing to meet its contractual obligations, is much too late because the expenses in the income statement for impairments then accumulate in economic downturns when losses materialize. This provisioning regime therefore increases pro-cyclicality. In good times, when lending is already at a high level, banks are not required to set aside buffers for expected losses, and thus overstate the economic value of the loan portfolio and understate losses in the income statement. As a result, lending can be expanded beyond the amount that would be possible under a different accounting regime. In economic downturns high credit losses occur, but the lack of available provisions increases the losses reported in banks' income statements, which reduces capital and may force banks to recapitalize or reduce lending and sell assets. Hence, provisions set aside in good times could serve as a buffer against risk; one that alleviates the impact of these effects and reduces the likelihood of banks becoming insolvent. Therefore some safeguards need to be built into the process such as disclosures of methods applied and periodical back testing and immediate reflection of the results of the back testing in the models applied for the future.

Expected loss model
Significant deterioration: A main new trigger within IFRS 9 to fall into stage 2 is a significant increase in credit risk. As expected, IFRS 9 does not provide detailed guidance what should be criteria's, as this represents typical principal-based accounting increment. Contrary to that according to IFRS 9 indicators are provided according to those financial institutions can conclude that a 12 month ECL is only required: • Low credit risk.
• Borrower is financial sound to meet is obligations.
• Adverse macroeconomic changes will not have an impact on the borrower's capacity to fulfill its obligations.
However the Basel SCRAVL (Sound credit risk assessment and valuation for loans) document limits this so called low credit expedient [9,10]. "The Committee expects that the low credit ingredient would be used by banks only in rare and appropriate circumstances, since the Committee views lending activities as the core of the bank's business" According to IFRS 9 the assessment of significant credit risk deterioration should include a multifactor and holistic approach. Whether one factor needs to be weighted more than another one needs judgment. IFRS 9 to simplify this judgment provides a list of nonexhaustive factors, which are summarized as followed: • Significant change in internal price.
• Other changes in the rates of terms of an existing financial instrument.
• Credit default swap price.
• Duration that the fair value is less than the amortized cost.
• Other market information related to the borrower.
• Significant change in the credit rating.
• Internal credit rating downgrade.
• Significant change in the value of the collateral.
The assessment of whether there has been a significant increase in credit risk is based on an increase in the probability of a default occurring since initial recognition. Under the Standard, an entity may use various approaches to assess whether credit risk has increased significantly, however it needs to be ensured that the approach is applied consistently.
In the literature different approaches are therefore discussed for significant deterioration [7].
• Non-portfolio based approach.
Portfolio-based approach defines conditions for significant credit risk decrease depending upon the quality of the portfolio. A low risk portfolio from inception would imply a relative low change in the credit risk to comply with the significant deterioration criteria. While a high risk portfolio would encounter a relative higher change in credit risk to have a significant deterioration.
In a non-portfolio based approach, regardless of the basis common criteria would be applied. What is often mentioned is a downgrade indicator for instance a three notch downgrade.

Implications for financial institutions:
According to Basel Draft Guidance banks are requested to develop systems and processes to generate all available and sound information to attain a high quality, robust and consistent implementation of the IFRS 9 expected credit loss model. Financial institutions as Basel recommends planning high upfront investments, as quality would far outweigh associated costs from a long-term perspective [10].

Conclusion
This study explores the literature about definitions and concepts when a significant increase in credit risk is achieved. With IFRS 9 prudent accounting has been a strong influencer for the change from an incurred to an expected loss model. This research paper summarized alternative impairment models and particularly focus on the significant deterioration criteria, which is a cornerstone of the new IFRS 9 impairment model. The expected loss model is not completely new within the accounting literature. The study provides early insights into implementation of IFRS 9 on impairment, as IFRS 9 will become applicable 2018. The new model will have a material impact on the financial institutions bank systems and processes. Moreover, it will also have a tightening effect on earnings management [9].
Based on the findings, the question arises if the regulator should provide more guidance to avoid that all companies purse completely different model resulting in decreasing comparabity for regulators and investors.