Views on improving the integrity of global capital markets
21 July 2016

Analyzing Banks’ Credit Risk: Expectations for New Accounting Guidance

The ongoing woes of Italian banks has produced sharp spikes in the levels of their non-performing loans (estimated to be a fifth of Italy’s GDP). With this situation coming relatively soon after the global financial and European sovereign debt crises, it serves as a reminder that the bank business model is susceptible to economic headwinds and that banking-related crises may be a recurrent characteristic—what some call the new norm.

At the same time, in today’s reporting environment, analyzing the credit risk of bank assets is hampered by the delayed write-down of loans, the incomparability of how loan carrying values are measured, and inconsistencies in the classification of loans as non-performing. For this reason, stakeholders, including investors, have been keenly waiting for the revised accounting guidance under both US GAAP and International Financial Reporting Standards (IFRS). The International Accounting Standard Board’s (IASB’s) expected credit loss (ECL) model is articulated in IFRS 9, issued in July 2014 (effective from 1 January 2018). The Financial Accounting Standards Board’s (FASB’s) current expected credit loss (CECL) model is contained in Accounting Standards Update Topic 326, issued in June 2016 (effective from 1 January 2020 for public entities with early adoption, allowed from 1 January 2019).

With the revised accounting guidance finally released, the question is whether investors will be better equipped to analyze and compare credit risk across banks. Three considerations can help to unpack this question:

  • The IASB and FASB revised models are viewed as an improvement and expected to provide more timely loan write-downs than today’s reporting, but both models also face some criticisms.
  • The impact (i.e., changes in provisioning levels) of adopting new guidance at an individual bank level remains a question.
  • The comparability challenges will most likely continue because of having multiple models.

Economic Relevance of New Models

At a top level, the difference between the FASB’s and the IASB’s models (CECL and ECL, respectively) is that in the FASB model, lifetime expected credit losses for loans are recognized, but the IASB model does the same only when there has been a significant deterioration in credit quality. Otherwise, the IASB model recognizes only a portion of lifetime expected losses (a 12-month measure).

Both of the models are expected to yield a more timely recognition of impaired financial assets than the current reporting requirements, and during the deliberation, on balance, there was general investor support for the change. In a 2013 Comment Letter about the proposed updates related to impairment, we reported on a survey we conducted that showed greater support for an expected loss approach (41%) relative to the currently required incurred loss approach (9%). In total, global survey respondents indicated a slight preference for the IASB ECL model over the FASB CECL model. Although preferences tended to reflect the location of respondents (i.e., US members supported the FASB model, whereas IFRS respondents supported the IASB model).

But these models were not exempt from criticisms. Member comments indicated that those supporting the FASB CECL model did so because of its seeming emphasis on prudence, whereas those with concerns indicated it recognized unrealistic and non-economic impairment charges up-front, lessening the relevance of impairment over net interest income ratio. Those supporting the IASB model, believed that it  better reflected the economics of the lending business, in which Day 1 expected losses are priced into the initial lending interest rate, and that a distinction in accounting for performing and non-performing loans is warranted. The principal objection to the IASB ECL model was the view that a 12-month expected loss recognition criterion is somewhat arbitrary and that this horizon did not necessarily have an economic basis. Some also expressed concern about its potential for sudden recognition losses and increase in loan loss allowance (described as the “cliff effect”) during challenging economic environments.

Question on Day 1 Impacts remains

An ex-ante quantification of the incremental impact of an expected credit loss model versus existing incurred loss is a tricky business and fraught with difficulties because it is hard to accurately and effectively undertake such a forecasting exercise of accounting effects before the actual effective date. The appropriate loan loss provision on Day 1 of adoption will depend on the economic environment, lending volume, asset class characteristics, and underlying borrower quality—all of which are variable factors. As an illustration, the following table shows that the level of loan allowances for a sample of large EU and US banks varies over time depending on the economic environment and particular bank asset class characteristics.

Loan Loss Allowance/Gross Loans at the Four Largest US and Four Largest EU Banks

IFRS ReportingUS GAAP Reporting
YearDeutscheHSBCBNPBarclaysJP MorganCitigroupBOAWells Fargo
20080.7%2.5%2.8%1.4%3.1%4.3%2.4%2.4%
20091.3%2.8%3.6%2.5%5.0%6.1%3.9%3.0%
20101.0%2.1%3.7%2.8%4.7%6.3%4.3%2.8%
20111.0%1.8%4.0%2.4%3.8%4.7%3.6%2.4%
4-year average (08-–11)1.0%2.3%3.6%2.3%4.1%5.3%3.6%2.6%
20121.2%1.6%4.1%2.2%3.0%3.9%2.6%2.0%
20131.5%1.3%3.7%1.5%2.2%3.1%1.9%1.8%
20141.3%1.1%3.7%1.1%1.9%2.5%1.7%1.4%
20151.2%0.9%3.5%1.1%1.6%2.1%1.4%1.3%
4-year average (12-–15)1.3%1.2%3.7%1.5%2.2%2.9%1.9%1.6%
 
Source: Banks’ Annual Reports and Bankscope Data.

Despite these difficulties in quantifying aggregate incremental impact, there are some helpful reference resources that educate on the likely impacts:

  • IFRS Foundation IFRS 9 Investor Outreach Slides based on an October 2015 presentation by IASB Board Member Sue Lloyd. These slides provide insightful empirical data points, including anticipated, hypothetical impacts on regulatory capital on the basis of current regulatory requirements.
  • A 2016 Deloitte Global Bank Survey of 91 global banks showed that 60% could not quantify the effects of IFRS 9. Of those that could (40%), loan loss provisions are expected to increase by 25% across all loan asset classes on initial application of IFRS 9.

Furthermore, in January 2016, the European Banking Authority announced that it is undertaking an exercise to assess the impact of IFRS 9 on EU banks.

Because the FASB CECL model has just been recently issued, it may take some time before similar education resources on the effects of the FASB approach will be available.

Comparability Challenges

Despite prior concerted efforts to develop a joint standard, by issuing differing revised financial asset impairment guidance, the FASB and IASB missed an opportunity to converge their requirements and enable global banks to provide comparable financial asset impairment information. Unfortunately, the accompanying disclosures for these revised standards will also not enable investors to reconcile the FASB and IASB models.

The comparability challenges of these two different models will be exacerbated by their different adoption dates (2018 versus 2020). Different impairment models will likely need to be applied by reporting entities during the transition period, especially given that both models allow early adoption. As a result, during the transition period, investors will need to be able to compare (within and between entities) the incurred loss approach with the IASB ECL or FASB CECL models.

Another source of interpretation complexity will arise from the fact that prudential regulators (e.g., Basel committee) require a different version of expected credit loss for the purposes of determining regulatory capital. Hence, within the corporate report, there could be information in regard to different types of expected loss information (i.e., incurred loss versus IASB/FASB expected loss models versus Basel expected loss model).

An expected loss methodology necessitates the application of forward-looking, macroeconomic inputs and estimates, such as the probability of default (PD), loss-given-default (LGD) across asset classes, and migration matrices of PD and LGD across time. These multiple and complex management judgments around applied inputs could result in subjective judgments and thus inconsistency and incomparability of the reported impairment amounts.

In addition, the IASB and FASB models both have the following specific features that require significant management judgment:

  • IASB ECL (Judging significant deterioration in credit quality): IFRS 9 requires a three stage categorization of financial assets across a continuum of credit risk as follows: Performing (Stage 1), Underperforming (Stage 2), and Non-performing (Stage 3). The judgment on significant deterioration in credit (i.e., move from Stage 1 to Stage 2) in a manner that is consistent and comparable across banks and faithfully representational could prove challenging.
  • FASB CECL: Estimating the lifetime expected losses under the FASB model will be quite a difficult judgment with potential for subjectivity and incomparability of management judgments made across banks.

Overall, investors will face the challenge of understanding how management has made such decisions and whether these judgments are comparable between reporting banks. The standard setters have enhanced disclosures to help investors better understand management judgments.

Main Takeaway

In sum, although there was a missed opportunity to converge the guidance, the revised guidance from the IASB and FASB has several upsides. First, it is anticipated that there will be a timely and less pro-cyclical write-down of loans relative to current reporting requirements. Second, it will encourage a better integration of credit risk management and accounting information systems within financial institutions. But the adoption of these revised accounting standards will also introduce additional layers of measurement complexity that will require comprehensive disclosures before investors can meaningfully apply the accounting information as effective signals and measures of the relative credit quality of different banks. In other words, investors will likely have access to a richer information set than today, but they will not be exempt from the challenge of comparing credit risk across entities.

Notes

Impetus for Revising Accounting Standards

The 2008 global financial crisis was a watershed moment because it provided the impetus for reforming the accounting for financial instruments including loans. During the crisis, there was a widespread acknowledgment among stakeholders, including the G–20 that the existing accounting framework needed to change. There have been several data points during the past few years that affirm the shortcomings of the incurred loss method. For example, the 2014 European Central Bank (ECB) asset quality review of 130 EU banks required a gross downward adjustment of carrying values on balance sheets to the tune of €47.5 billion, and the bulk of these were attributable to additional loan provisioning (€42.9 billion). Furthermore, a 2014 CFA Institute report  showed that the delayed impairment of loan assets, low profitability of US and EU banks, and increased investor risk aversion toward them all contributed to their low price-to-book ratios (P/B) during the financial crisis.

The global financial crisis resulted in a number of initiatives aimed at improve bank financial reporting:

  • Accounting Standards Revised Guidance: The IASB and FASB have respectively revised their financial asset impairments guidance.
  • The 2014 European Central Bank’s (ECB) asset quality review assessed the adequacy of loan provisioning and categorization of loans as non-performing exposures. An additional €142 billion of loans were characterized as non-performing by the ECB due asset quality review.
  • Financial Stability Board (FSB) Enhanced Disclosure Task Force (EDTF) recommended voluntary disclosures that improve the risk reporting of large banks, including disclosures on credit risk management. There has been an uptake of EDTF recommendations by a number of large global banks.

Post-Crisis Trends

There has been a sense during the past two years that banks have transitioned from the effects of the global financial and European sovereign debt crises, which adversely affected the credit quality of loan assets held by these banks. The profile of profitability from 2008 to 2015 for eight of the largest US and the European Union banks (see the following two tables) shows that bank profitability remains volatile, anemic and below the pre-crisis levels.

But the impairment data trends show that there has also been a reduction in the cost of credit risk (i.e., loan impairment charge/net interest income) for these banks during recent years.

Return on Equity at the Four Largest US and Four Largest EU Banks

IFRSUS GAAP
YearDeutscheHSBCBNPBarclaysJP MorganCitigroupBOAWells Fargo
2008-12.2%6.8%7.5%15.0%4.2%-39.5%2.9%3.8%
200913.1%5.1%9.2%20.7%7.5%-1.0%2.9%12.0%
20104.6%9.5%12.1%8.4%10.3%6.6%-1.1%10.6%
20117.9%11.4%9.0%6.7%10.8%6.3%0.7%12.5%
4-year average (08-–11)3.3%8.2%9.4%12.7%8.2%-6.9%1.4%9.7%
20120.5%8.8%8.5%-0.4%10.9%4.1%1.9%13.3%
20131.2%9.3%6.0%2.0%8.5%6.7%-3.5%13.0%
20142.3%7.4%0.5%1.3%9.4%3.5%2.0%12.7%
2015-10.0%7.6%7.0%0.9%9.9%7.8%6.2%12.0%
4-year average (12–-15)-1.5%8.3%5.5%1.0%9.7%5.5%1.6%12.8%
 
Source: Banks’ Annual Reports and Bankscope Data.

Cost of Credit Risk (Impairment Charge/Net Interest Income) at the Four Largest US and Four Largest EU Banks

IFRSUS GAAP
YearDeutscheHSBCBNPBarclaysJP MorganCitigroupBOAWells Fargo
20089%58%27%42%54%63%59%60%
200921%61%36%62%63%79%103%47%
20108%34%20%45%33%46%55%35%
201111%28%15%31%16%24%30%19%
4-year average (08–-11)12.1%45.4%24.4%44.8%41.3%53.1%61.9%40.1%
201211%22%17%30%8%23%20%16%
201314%17%19%26%1%16%8%5%
20148%12%18%18%7%14%6%3%
20156%11%16%17%9%15%8%5%
4-year average (12–-15)9.5%15.3%17.7%22.9%6.0%17.1%10.5%7.6%
 
Source: Banks’Annual Reports and Bankscope Data.

 

The following table highlights non-performing loans. Getting this data confirmed the inconsistencies in the classification of loans as non-performing. These inconsistencies limit the comparability of the non-performing loans ratio and its effectiveness as a signal of relative credit risk.

Non-Performing Loans at the Four Largest US and Four Largest EU Banks

IFRSUS GAAP
YearDeutscheHSBCBNPBarclaysJP MorganCitigroupBOAWells Fargo
20081.4%2.6%3.7%3.4%0.6%2.7%1.7%0.8%
20092.8%3.3%5.5%5.2%2.7%5.2%3.6%3.5%
20101.5%2.9%5.9%5.5%3.1%5.5%4.5%4.6%
20112.3%4.3%6.3%4.8%2.7%5.3%4.8%4.3%
4-year average (08–-11)2.0%3.3%5.4%4.7%2.3%4.7%3.6%3.3%
20122.6%3.8%6.5%6.4%2.8%4.9%4.2%4.1%
20132.7%4.6%6.4%5.2%2.3%3.4%1.9%3.4%
20142.3%3.9%6.0%3.8%1.9%2.6%1.4%3.0%
20151.9%3.5%5.5%3.3%1.4%2.0%1.0%2.6%
4-year average (12-–15)2.3%3.9%6.1%4.7%2.1%3.2%2.1%3.3%
 
Source: Banks’ Annual Reports and Bankscope Data.

If you liked this post, consider subscribing to Market Integrity Insights.


Image Credit: ©iStockphoto.com/kevinjeon00

Tags: , , ,

About the Author(s)
Vincent Papa, PhD, CPA, FSA, CFA

Vincent Papa, PhD, CPA, FSA Credential, CFA, was the director of financial reporting policy at CFA Institute. He was responsible for representing the interests of CFA Institute on financial reporting and on wider corporate reporting developments to major accounting standard setting bodies, enhanced reporting initiatives, and key stakeholders. He is a member of ESMA’s consultative working group for the Corporate Reporting Standing Committee, EFRAG user panel, and a former member of the IFRS Advisory Council, Capital Markets Advisory Committee, and Financial Stability Board Enhanced Disclosure Task Force. Prior to joining CFA Institute, he served in investment analysis, management consulting, and auditing roles.

Leave a Reply

Your email address will not be published. Required fields are marked *



By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.

Close