Stress tests: why national CET1 calculations could vary

Author: Ben Bschor | Published: 24 Oct 2014
Email a friend

Please enter a maximum of 5 recipients. Use ; to separate more than one email address.

The EBA’s emphasis on consistent national data and results for its upcoming stress test will be examined on Sunday, when the ECB and EBA reveal how banks fared.

Banks in some EU countries are subjected to stricter rules than those in others when calculating common equity tier 1 (CET1).

This is due to transitional rules for the implementation of the Capital Requirements Directive (CRR). Those transitional rules can vary across jurisdictions and are largely taken into account for the purpose of the stress tests.

Headline figures
When the EBA and ECB release the results of the stress test and the comprehensive assessment on October 26, market participants will initially focus on three headline figures for banks: the CET1 ratio calculated for the asset quality review (AQR), a baseline scenario, and an adverse scenario.

These ratios of CET1 and risk-weighted-assets (RWA) are the key metrics that determine whether a bank fails or passes the stress test.

The AQR determines the quality of the banks’ assets. The baseline scenario assumes economic development in line with EC projections until 2016, including GDP growth of 1.5% in 2014, 2% in 2015, and 1.8% in 2016.

The adverse scenario assumes a deterioration in macroeconomic developments, including a decline of real GDP in the EU by 0.7% in 2014 and 1.5% in 2015, before rising by 0.1% in 2016. Employment falls significantly during that period.

Banks will pass the comprehensive assessment if their results show a CET1 ratio of at least 8% for the AQR and the baseline scenario, and of 5.5% for the adverse scenario.

But for the time being, banks in different EU member states do not necessarily calculate their CET1 capital on the same grounds, and it will take years before a level playing field is created.


  • EU member states decide to what extent certain assets are accounted for in regulatory capital during transitional periods
  • As a result CET1 ratios are not calculated on the same basis in all member states
  • This limits direct comparability of stress test results, although large improvements have been made over previous tests

Deferred tax assets
The most prominent example is the treatment of deferred tax assets (DTAs), where national rules are still applicable.

The exclusion of DTAs from CET1 is being phased out, but the CRR gives member states certain discretion about the speed of their inclusion.

Many EU members have chosen the most bank-friendly approach: a slow phase-in, starting at 10% in 2015 and increasing by 10% annually until reaching 100% from 2024. France, Germany, Ireland, Italy, the Netherlands, and Spain are among the countries that have chosen this path.

The UK, however, has skipped the transition entirely, with banks having to fully include DTAs in their calculations as of January 1 2014. In Denmark banks also need to fully take DTAs into account. And Luxembourg has opted for an accelerated phase-in of DTAs at a rate of 20% in 2014, 40% in 2015, and so on until reaching 100% from 2018.

No differences in AFS
But there is one notable exception of the rule that CRR national discretions are acknowledged in the tests: the treatment of available-for-sale securities (AFS).

In Germany, France, Ireland, Italy, Luxembourg and Spain losses from sovereign bonds held in the AFS portfolio are filtered out when calculating CET1.

This treatment will continue until the European fair value accounting rule IAS 39 is replaced. This was due in 2015 but has now been pushed back to at least 2017.

The Netherlands has chosen to remove the filter over the coming years. In 2014 the AFS filter has been reduced to 80%, and in 2015 and 2016 – that is, within the timeframe of the baseline and the adverse scenario – it will shrink further to 60% and 40% respectively.

UK’s regulator, the Prudential Regulatory Authority (PRA) announcend in December 2013 that in most cases a filter on AFS central government debt would be inappropriate because it would misstate the underlying solvency of the firm. The PRA said banks could still apply for permission to filter and each case would be considered individually.

But for the purpose of calculating capital in the stress test this member state discretion has been ignored: irrespective of the differences across countries the sovereign risk held as AFS is treated in the same way in the exercise.

This is significant, because the amount of sovereign debt held as AFS varies widely from bank to bank.

Viral Acharya, economics professor at the New York University Stern School of Business, and Sascha Steffen, associate professor of finance at the European School of Management and Technology in Berlin, compared data from banks in Spain, Greece, Ireland, Italy, and Portugal.

Their research, based on June 2013 figures, shows that within this group banks in Spain and Italy had by far the highest amount of sovereign bonds in their AFS portfolio.

Italian Intesa held €49.66 billion ($62.87 billion) as AFS (and €31.35 billion as HTM and €5.99 billion in trading), and Spanish Santander held €42.38 billion as AFS (and €22.33 billion HTM and €5.47 billion trading).

Unicredit (€39.45 billion AFS, €12.01 billion HTM and €6.56 billion trading) and BBVA (€21.93 billion AFS, €38.5 billion HTM and €4.95 billion trading) came in at third and fourth place respectively.

By comparison, Europe’s largest bank, HSBC, only has €350 million of sovereign bonds as AFS, and the second largest player, Deutsche Bank, has €1.31 billion. BNP Paribas’ €13.84 billion make it the second-largest holder of sovereign bonds as AFS among Europe’s ten biggest banks. But that’s still less than a third compared with Santander, Europe’s eight’s biggest bank.

This implies that some bank’s regulatory capital would look considerably stronger if sovereign debt held as available-for-sale was excluded from capital.

Martin Hellmich, professor for risk management at Frankfurt School of Finance & Management, says that capital ratios have never been directly comparable across EU countries – neither in past tests, nor in the 2014 exercise.

However, steps in the right direction to improve comparability have been taken. The AQR, the EBA single rulebook, the ECB’s direct supervision of significant banks and the introduction of joint supervisory teams as well as the introduction of leverage ratio requirements will all help the process. "But we are still far away from being able to directly compare the numbers," says Hellmich.

EBA and ECB acknowledge that there are differences in CET1 calculations and that transitional arrangements are taken into account for that purpose.

But the EBA also promised some clarity about this in the data as it will be published on Sunday. "In the legal framework competent authorities have the discretion, as per the CRR/CRD IV, to filter out unrealised losses," an EBA paper states. "Either choice will be clear in the transparency of the results."