05 October 2014

Notes on ECB’s Comprehensive Bank Assessment (Part 6)

Everyone interested in the outcomes of the ECB’s Comprehensive Bank Assessment is probably keenly awaiting the results. Sources (whoever they are) have told Reuters that these will be announced on Sunday, Oct. 26. The date is officially not confirmed but anyways, according to the ECB, it will be in the second half of October. Reportedly, banks will be given their results about 48 hours before publication, to enable them to review the figures and prepare any response.

Some are making predictions of how many banks and which ones will fail. “Candidates” include but are not limited to: Germany’s Commerzbank, Austrian cooperative bank Volksbanken, politically-connected banks such as Germany's HSH Nordbank. Some want to see blood:
"Significant capital shortfalls and/or a significant number of banks not passing the test will be the criteria against which the assessment's credibility will be measured," Alain Laurin, associate managing director of Moody's Investors Service has been reported to have said back in July.

ECB’s officials are already pulling down expectations concerning the severity of the conclusions, while arguing that by the time results are published, banks will have done a lot already:


“Personally, I believe that European banks are in a better situation than markets believe. The comprehensive assessment will deliver transparency. [...] A number of banks have already started to address possible capital shortfalls. I think that is a good idea. It is better to go to the market before the outcome of the comprehensive assessment is available than to do so when all banks do together at the end of October.”
Danièle Nouy, Chair of the Supervisory Board of the Single Supervisory Mechanism, in an interview to Äripäev, conducted on 26 August 2014.

“While the exercise will only conclude next month [that is: October], we can already see signs that it has affected both the speed and quality of deleveraging. [...] This acceleration of the process suggests that, once the final results are known and residual uncertainty is removed, banks will be in a stronger position to resume new lending. “
Benoît Cœuré, Member of the Executive Board of the ECB, at IMF/Banka Slovenije high-level seminar on “Reinvigorating Credit Growth in Central, Eastern and Southern European Economies”, Portoroz, 26 September 2014.

So not so excited, guys: final conclusions, once disclosed, are probably going to be rather boring. Let’s see what we are going to get anyway.

Just to remind, complete methodology and process manual for the quantitative component of the comprehensive assessment consists of the three documents:
What is described in these three papers is ought to be the basis for conclusions, of course complemented by qualitative and other considerations.

While I have discussed the Asset Quality Review (AQR) and the EBA 2014 bank stress test earlier in this series (AQR in Part 3, and stress test in Parts 4 and 5), I haven’t so far digested into the last manual. This is what I did yesterday. The last manual is not the least one; it: a) outlines the stress test quality assurance process, and b) describes how AQR findings and stress test results are integrated (the so-called join-up). Namely, according to the ECB, join-up is what sets the comprehensive assessment apart from any other previous European exercises.

Comprehensive assessment stress test process is / was divided into four phases:
  • Phase 1 (May-June): Bank-led stress test
  • Phase 2 (July-August): Quality assurance
  • Phase 3 (September): Join-up
  • Phase 4 (September-October): Finalisation and disclosure
Note the order: while ECB’s Asset Quality Review was still ongoing, banks already did the stress test calculations; thereafter ECB and NCAs checked the quality of the stress tests, and only then, as the last step, the quality-assured stress test results were joined-up with the AQR results. Wouldn’t it have been more straightforward to use AQR-adjusted balance sheet, capital ratios etc. as inputs to the stress test instead of deriving final results from the stress test calculations that have been done based on not that reliable banks’ accounting figures? Shortage of time and certain other issues can be an excuse, of course, but now one cannot be sure if AQR findings are properly taken into account in stress testing. The current reverse order makes join-up incomprehensible even if the detailed methodology is published (the outcome will be combination of so many adjustments, limitations and omissions – see the discussion below).

Quality Assurance (QA)

As for quality assurance then the described process looks as rigorous as ever. How it is/was executed is yet another question to which an outsider cannot answer. The purpose of the QA exercise is/was to ensure that banks have applied the prescribed methodology and translated the impact of the baseline and adverse scenarios on their balance sheet in an appropriate manner. 

Depending on the element checked (credit risk, market risk, securitisations, net interest income or other pre-provision profit – the focus is/was on those areas of banks’ stress test results that may materially underestimate the capital impact of the stress), QA is supposed to have included:
  • Consistency checks with EBA methodology;
  • Comparison of starting point parameters such as PD and LGD estimates with the outcomes of AQR;
  • Parameter consistency check within the bank;
  • Comparison of bottom-up stress test results provided by the banks with the results of the top-down model owned by the ECB;
  • Comparison of bottom-up stress test results provided by the banks with the results of other banks within the same market;
  • Comparison of bottom-up stress test results provided by the bank with the results of other banks across the Single Supervisory Mechanism (SSM).
A traffic light threshold approach (or Red/Amber/Green, the RAG, approach) has been applied:
  • Green results: ok;
  • Amber results: “comply or explain” for banks;
  • Red: non-compliant, do it again.
QA is ought to have resulted in an updated bottom-up stress test result that addresses concerns raised during the QA. In terms of the comprehensive assessment, these results are/were intermediate and served as the starting point for the join-up; join-up both changed starting points and evolution, and potentially triggered further discussions between the ECB, the national competent authorities (NCAs) and individual banks.

Above I already made my comment about the somewhat illogical order of things.

I wonder what the ECB’s top-down models and -estimates are that serve as benchmarks to the banks’ own calculations. Given the extensive usage of such top-down models, that’s a key issue as far as quality assured stress test results are concerned.

Another issue is level playing field for banks that use ECB’s benchmark parameters vs banks that have their own internal models. Can the result of the QA for banks with stress testing models on average be below the mean result of the ECB model? Maybe banks with more sophisticated models indeed have portfolios that are of superior quality to those of banks without such models, but how big can the difference possibly be? We will see from the published numbers what banks and regulators are thinking. 

 “Join-up”

As I read the ECB Comprehensive Assessment Stress Test Manual (August 2014), there will be two sets of stress test results for each bank participating in the comprehensive assessment: 1) the so-called “pure” stress test results, and 2) stress test results that are “informed” by the AQR.

Depending on the AQR findings, AQR adjustments to a bank’s preliminary numbers can have been multiple, and applied both to the 2013 year-end figures and to the baseline and adverse scenario projections, incl.:
  • Changes in the fair value classification (e.g. HTM vs. AFS vs. fair value) of assets;
  • Corrections of the identified asset segment misclassifications (e.g. if a bank has classified CRE exposures as corporate);
  • Reclassification of exposures from “performing” to “non-performing” based on the EBA’s simplified NPE definition;
  • Dealing with any identified issues in relation to legal costs;
  • Material adjustments resulting from the level 3 fair value exposures review of derivative pricing models;
  • Taking into account findings relating to CVA identified in the AQR;
  • Adjustments for the cases where issues have been found in relation to deconsolidation or consolidation of assets (during AQR, a number of cases were found where banks had deconsolidated securitised portfolios from the prudential perimeter but not from the accounting perimeter; these portfolios should have been reconsolidated for the stress test).
… And most probably I have forgotten something from the list.

Depending on the specific issue, there are four “methods” for making the adjustments:
  • Method 1: Centrally-led (i.e. conducted by the ECB and the NCAs) adjustment: adjustment via direct use of AQR-built models;
  • Method 2: Centrally-led adjustment: simple scaling of stress test result through direct adjustment of position value;
  • Method 3: Bank-led adjustment;
  • Method 4: No adjustment required.
Wow, that looks impressive – all the abbreviations and everything. Sounds a bit ironical? So it is. Just as a side note: during the comprehensive assessment process alone a number of new abbreviations has been introduced: AQR, DIV, QA, PP&A, RAG, … These are in addition to the many that are more or less common in banking and/or financial supervisory parlance, e.g. CRE, NCA, SME, CVA, SSM, HTM, AFS, NPE, PD, LGD, EL, IBNR, IRB, RWA, CET1, SPV, DTAs, CRR, CRD IV ... I wonder how long it takes for a “normal person” to start orienting in this alphabet soup. I wonder if it all has become overly complex and inconsistent… Can all the decision makers in the ECB possibly understand this stuff? Let’s hope so... Taking deposits and lending money (basic banking) is not such a complex business; yet accounting and regulatory reporting “to make believe” most definitely is an art.

Going back to the topic, the intentions of a proper stress testing exercise may even be fair and honest, but the reality strikes. A number of join-up issues have been highlighted in the ECB’s Stress Test Manual itself, including but not limited to the following. Sentences in italic are quotes from the manual; […] indicates omitted parts.

The time available for the comprehensive assessment exercise is being referred to as a constraint. For example, consider these quotes from the manual (stresses added):
  • “The AQR may identify cases where exposures are not currently considered impaired but are very likely to be in the near future. These future losses are identified and noted but are not included in the AQR-adjusted CET1 ratio, nor are they produced in time to be shared with banks and factored into the bank-led stress test results.”
  • “Misclassifications to asset segments identified in the preliminary findings of the credit file review in the AQR should not be communicated to banks for inclusion in the stress test as they will not be complete and will be at different stages for different banks.”
  • “There will be no adjustments made to RWA or IRB provisioning shortfall for adjustments made to P&L items or reserves. […] …the level of data required to carry out the join-up and QA adjustment process would be dramatically increased which is neither feasible in the time frame, nor worthwhile considering the second-order impact.” Maybe not worthwhile, but some NCA’s seem to have had objections in this regard…
Several references are being made to avoiding issues with the Market Abuse Directive:
  • “All issues with derivative pricing models highlighted by the AQR should be communicated to the banks applying the comprehensive approach. […] For the avoidance of doubt – the bank should not be informed of the AQR quantification of the impact – only the nature of the issue to avoid issues with the Market Abuse Directive.”
  • The same story with the AQR findings relating to CVA: “The impact of the finding on the December 2013 position should not be communicated (so that it does not create issues with respect to the Market Abuse Directive). The impacts should therefore be forward-looking only, and not impact the December 2013 CET1 ratio.”
As it appears, the regulators are in an odd position: they see the models of the banks that at least in some respects are playing zero sum game against each other, and have to not just judge on the banks’ positions, but also incorporate their judgment into the stress testing exercise. 

That the ECB’s staff now knows more about the banks’ positions than the banks by themselves seems also be true for credit risk: “The experience of collating loan tape data in the AQR is that in many instances the AQR exercise has access to substantially more granular data than the “bottom up” analysis that the banks themselves will have performed when conducting the stress test.”

Further, the need to avoid consequences for market disclosure by the banks is being implied as a limiting factor to the join-up:
“For the avoidance of doubt, the correction of asset segment misclassifications should have no direct impact on the December 2013 available capital. Furthermore, as changes to accounting metrics are not being communicated as part of this exercise this is expected to have no consequences for market disclosure by the banks. Findings could have an impact on RWA, but – based on the level of adjustments in the DIV process found – this is considered to be of secondary importance.”
Secondary importance or not, an outsider cannot know.

More effects of the AQR findings are not at all taken into account while being referred to as immaterial:
  • “Where deviation in EL suggested by comparison of unadjusted with adjusted AQR parameters is less than or equal to a pre-defined quantitative threshold, it will be judged as immaterial and need not be investigated further.” What’s the referred pre-defined quantitative threshold, by the way?
  • “It should be noted that any second order indirect impacts of accrual accounted credit risk adjustments will not be reflected as part of the join-up (e.g. interest income implications of changes in the NPE projection).”
  • “The join-up approach for IBNR provisions will involve adjusting the year-end 2013 starting balance sheet for these findings; however, projected IBNR flows will not be adjusted.”

Once the AQR adjustments are made, the capital shortfall is calculated with respect to the CET1 ratio. At this crucial point, the comprehensive assessment is not free from the national discretion (stresses added):
  • “AQR: 8% CET1 on the AQR adjusted 2013 balance sheet (using national discretionary definitions of CET1 under CRD IV/CRR and national discretion for the phasing-in of deductions from CET1 integrated with a common approach for the application of prudential filters for sovereign assets in the AFS portfolio);
  • Stress test – baseline: 8% CET1 on the AQR adjusted baseline 2014-2016 year-end balance sheets (using national discretionary definitions of CET1 under CRD IV/CRR and national discretion for the phasing-in of deductions from CET1 integrated with a common approach for the application of prudential filters for sovereign assets in the AFS portfolio);
  • Stress test – adverse: 5.5% CET1 on the AQR adjusted adverse 2014-2016 year-end balance sheets (using national discretionary definitions of CET1 under CRD IV/CRR and national discretion for the phasing-in of deductions from CET1 integrated with a common approach for the application of prudential filters for sovereign assets in the AFS portfolio).”
Only the approach to the application of prudential filters for sovereign assets in the AFS portfolio is common – and this because of the earlier market pressure.

In short: when considering the above, I can only say that credibility of the join-up outcome remains a big question mark.

Disclosure of the assessment results

As part of the disclosure package, the ECB has promised to publish:
There will be lots of data points published, both those that are useful and those that are adding to the confusion. Let’s look at the relevant figures once templates are filled with numbers.

No comments:

Post a Comment