06 January 2014

The 2013 EU-wide Transparency (?!) Exercise

On 16 December 2013, the European Banking Authority (EBA) published the outcome of the 2013 EU-wide transparency exercise. Indeed, there are lots of data about the covered 64 European banks from 21 countries of the European Economic Area (EEA): all together 730,000 data points including capital, Risk Weighted Assets (RWAs) and sovereign exposures. The aggregate total assets of the banks included account for 64% of total assets of the 21 EEA countries. As stated in the press release:
“Through this disclosure exercise, the EBA aims to promote greater understanding of capital positions and exposures of EU banks, thus contributing to market discipline and financial stability in the EU.”

It may sound great, but… what the heck transparency? Even if I looked into the data closely enough I did not become much smarter about the banks’ true capital positions than I was before. Drawing any definite conclusions at all apart from the fact that it’s a total mess with the reported numbers of the European banks seems mission impossible. I’m not jealous of anyone who is tasked with sorting it all out; the technical staff of the EBA and the ECB has a Gordian knot to solve.

As follows, I’d like to highlight quite a few issues with the disclosed data – for those who want or need to actually use them.

Crucial data not disclosed

Reportedly, the disclosure is part of the EBA’s effort for improving the understanding of banks’ internal models. However, some of the most crucial data are not disclosed on a bank-by-bank basis even if collected:
  • Internally estimated credit risk parameters (PD and LGD) that are used for the capital requirements’ calculation are not disclosed.
  • Realized default rates are not disclosed.
  • Realized loss rates are not disclosed.
Why not? I think it’s rather crucial information for the stakeholders to feel more comfortable about the banks’ internal models. Ok, these figures may easily be misinterpreted by the media and so one – but so are then the disclosed RWAs that are calculated based on these internally estimated risk parameters. The difference is that while looking at RWAs leaves those who are not familiar with the details, indifferent and contributes to creating false sense of confidence about the banks’ capital positions, comparisons of the risk parameters and the realized default- and loss rates across banks would not. So, good luck with digging in the banks’ Pillar 3 reports!

Further, it would have been helpful by the EBA, had they added crucial additional information that was there in the disclosed 2011 stress test data, such as banks’ total assets without any risk weightings. For the starters, one would like to compare the exposure figures with the total balance sheet volumes. When doing manual checks, I found that for some banks (notably out of the European banks identified as globally systemically important: Deutsche Bank, Societe Generale and Barklays) the reported exposure figures are a lot smaller than the amounts of assets in the balance sheets. Arguably, part of the assets and liabilities nets out, but still: if the reported exposure is just 50-60% of the balance sheet volume, one would like to know more how the netting is being done.

Misleading and outdated information on banks’ capital positions

Banks’ capital positions are reported in accordance with the CRD3 rules (Europe’s implementation of the so-called Basel 2.5). In addition, certain memo items are provided for the CRR / CRD4 (Europe’s implementation of the so-called Basel 3). By now CRD3 is more or less history; capital positions based on the CRD4 rules would be much more relevant. But as stated in the summary report of the exercise:
“In addition banks were requested to disclose some memo items relevant under the upcoming CRR/CRD4 regime. While these figures provide additional information, it is worth emphasising that they cannot be used for computing fully-loaded CRR/CRD4 capital ratios since the granularity of the information requested does not allow to do so.”

The main problem with the CRD3 figures is that the reported Core Tier 1 and Tier 1 Capital include amounts that might be described as the illusory capital:
  • Instruments that are not loss absorbing on a going concern, and should be phased out according to CRD4;
  • The share of the capital deductions that according to CRD3 can be made from the Tier 2 Capital (e.g. 50% of the IRB Provision shortfall);
  • Deferred tax assets that rely on future profitability as well as instruments that have been designed to inflate artificially the own funds of the institution (and as such, are not accepted in CRD4).
In other words, the reported capital positions and -ratios according to CRD3 rules are misleading and overestimated by definition.

At best, approximations can be made about the banks’ true capital positions. The bad news is that certain inconsistencies in the disclosed data across banks (see below) do not allow doing it without analyzing banks’ capital adequacy reports one-by-one.

Tricks with the transitional floors

From the “EU-wide Transparency Exercise Guidelines” document one can read that the banks were asked to calculate the RWA figures as reported to National Supervisory Authorities. This, among others, also means taking into account transitional floor requirements*.

So far so good: there is even a dedicated data field labeled as “RWA Transitional floors” in the disclosed dataset. Oddly enough, only six banks out of the 64 have a number greater than zero there. In reality more of them are reporting transitional floor capital requirement to their local supervisors. For example, all four Swedish banks included in the exercise locally have a transitional floor capital requirement, but no one is reporting a number for the transitional floors in the EBA exercise; so they look relatively better thanks to their low internally calculated risk weights.

As an illustration, for the Svenska Handelsbanken, the actual capital requirement is twice as high as the number reported in the EBA’s exercise. Even more strangely, for the Skandinaviska Enskilda Banke (SEB), at the time of first publishing on 16 December there was a RWA transitional floor of 38.3 billion euros included in the disclosed data for 30 June 2013, but now it’s zero. Probably SEB initially made a mistake and calculated the floor requirement according to the different method which ironically more correctly described the reality; later when comparing its results with the peers, it updated the number.

How come?

Well, for the starters, the implementation of the Basel I floors in EEA was a mess at the outset; different countries did it differently. (The issue is well explained in the economic commentaries of the Norges Bank: “The Basel I floor – transitional arrangement and backstop to the capital adequacy framework” (No. 8, 2012).)

So when discussing the methodology for the floor calculation back in 2011, the EBA decided to allow National Supervisory Authorities to choose between the two approaches (described in detail in the Methodical Note for the 2011 Capital Exercise):
  • Option 1 (floor based on the risk-weighted assets, as originally suggested by the Basel Committee): Transitional floor capital requirement = Max [(80%*(Capital Requirement based on Basel I) – (Total Minimum Own Funds)), 0]
  • Option 2 (the floor is defined as a floor for the regulatory capital – and not for risk-weighted assets.): Transitional floor capital requirement = Max [(80%* Capital Requirement based on Basel I) – (Total Own Funds), 0]
These two options have been allowed also this time. Note that the difference between them seems to be in just one word: “minimum”, “Total Minimum Own Funds” or “Total Own Funds.” The thing is that unlike Option 1, floor based on the Option 2 has no effect on banks with capital adequacy higher than 80 percent of the minimum under Basel I. In other words, such banks can attain very high capital adequacy by reducing their internal risk weights; they don’t have to bother about the floor requirements.

When continuing the example of Swedish banks: according to the Swedish rules they have been required to calculate transitional floor capital requirement based on the Option 1; however, for the purpose of the EBA Transparency Exercise they chose Option 2. At the same time, Norwegian banks have been required to calculate transitional floor capital requirement based on the Option 1 and so the DNB Bank did indeed report its number based on the Option 1.**

There is one more thing: while in the 2011 Exercise all banks in a country applied the same approach, now individual banks within a country may have used different approaches. (Take a look to the numbers of German banks, for example.)

Data issues

Whenever there are lots of data from different sources, there are also data issues. It just cannot be otherwise. Despite of the EBA’s instructions and guidelines, and technical Q&A-s, the banks appear to have reported some of the key figures differently. It can be (and often is) because of the different regulatory frameworks and accounting principles, but it may also be that the banks don’t yet have all data available in the requested format and/or have interpreted the data requirements not quite in the same way. Whatever the exact reasons, comparing banks based on the disclosed figures can easily result in misleading conclusions.

These are just some of the data issues that I noticed:
  • Different reporting of common equity before the deductions, and the deductions from the common equity. When I tried to correct the reported regulatory capital figures for the so-called illusory capital, I found that for one thing, banks (even in the same country) have reported intangible assets differently. For example, Nordea appears to have relatively large deductions from the reported figure of the “Common equity before deductions” when compared to its Swedish peer Swedbank (14.1% vs. 1.1% as of 30 June 2013). It so turned out that while the first one (Nordea) included goodwill in the common equity before deductions and thus, it was also included in the reported figure of “Deductions from common equity”, the second one (as a number of other banks in other countries) did not. So without looking into the banks’ individual reports, I couldn’t do the comparison of banks’ capital positions adjusted for the illusory capital that I thought would be more meaningful than the reported regulatory capital ratios.
  • Different level of detail in the reported figures for market risk. Not all banks have provided split between the traded debt instruments under the Internal Models approach. For example, there is no split for the RBS, for RZB, for BBVA and for some smaller German banks.
  • Different reporting of value adjustments and provisions. Relative to the (defaulted) exposures, banks are reporting very different levels of the value adjustments and provisions even if underlying risks could be assumed rather similar. For example, a number of banks do not report any provisions for the exposures that are not in default, while many others do; banks in the very same country often have different reporting approaches. As another example, some banks do not virtually do any specific provisions for the retail exposures not secured by the real estate; however, for the others such provisions amount to above 100% of the defaulted exposure amounts. Once again, no meaningful conclusion can be drawn based on the disclosed figures of the EBA’s transparency exercise.
In addition, it is well known that the reported loan-to-value (LTV) ratios are not easily comparable, that the principles for counting overdue days and recording defaults differ across countries and banks, and that the banks’ forbearance practices need a closer look. Let’s not even repeat the issues related to the banks’ internal risk estimates and the transitional floor capital requirements.

So much about transparency…

* Basel II was introduced in the European Economic Area (EEA) in 2007. To prevent banks' internal risk weights from reducing risk-weighted assets and thus banks' capital needs too much and too quickly, temporary, lower limits (referred to as the "Basel I floors" or as transitional floors) were set for how much capital could be reduced. These limits were set relative to the previous framework, Basel I, which had a fixed set of risk weights. Although the floor was originally intended to expire at the end of 2009, a Basel I floor of 80 percent of RWA was retained and continues to apply in Norway and most EU countries.
** Some adjustment factors have been ignored here for the sake of simplicity.

1 comment:

  1. About "The 2013 EU-wide Transparency and exercise" whatever you have shared here seems to me sound like very wonderful though very important as well. So happy to read about in such post. Thanks