The European Banking Federation welcomes the solution proposed by the EBA in its draft RTS for the prudential treatment of software assets, based on a prudential amortisation.
The RTS clearly explains the arguments that support a change in the treatment of software assets in the EU banking regulation. These have been brought by the EBF to the attention of legislators during the regulatory reform debate and, among others, include:
1. The competitive disadvantage suffered by banks compared to less-regulated entities such as financial technology companies.
2. The uneven playing field with international competitor banks which do not account for software as an intangible asset and, therefore, are not required to deduct it from CET1 capital.
3. The need for EU banks to invest in software to keep up with the digital transformation that will enable them to compete in the global markets and to develop proper cybersecurity systems.
4. The fact that software assets have value for the institution, as well as for any potential acquirer, as the institution could not continue its functioning without its software.
The EBA’s proposal on the prudential amortization of software assets is the right one in the view of the EBF as:
- It is clear.
- It can be implemented relatively simply.
- It is easy to supervise.
- It is neutral with respect to accounting standards, thereby ensuring a level-playing field within the EU.
However, the proposed calibration of the solution with only a two-year prudential amortisation period is overly conservative, and does little to resolve the issues described above. The EBF argued, in its paper of 4 December 2019, that a six-year calibration period would largely tackle those problems while still maintaining a sufficiently prudent approach.
The EBF therefore supports the prudential amortisation approach but calls the EBA for a longer amortisation period for the reasons already elaborated in the EBF paper and reflected in the response to this consultation.
In general, the European banking industry highly supports a swift implementation of the new legal framework and encourages authorities to make it enforceable as soon as possible.
In the background of the current crisis and under the adoption of the quick fix to help banks finance the economy and sustain recovery, all stakeholders should do their best to ensure that the new rule can be enforced before the end of Q3 2020.
We regret noticing that, under the proposed RTS, even during the two years of the prudential amortization period, where the EBA recognises that software retains value, an increasing deduction still needs to be applied.
As already communicated to the EBA, banks’ investment in software is a key success factor not only in ensuring their business models remain competitive and resilient, but also in terms of ever-greater IT and cybersecurity risk. Furthermore, they contribute to banks’ profitability, which ultimately supports system-wide financial stability and reduces the likelihood of bank failure.
The COVID-19 crisis has evidenced the importance of a strong IT network for business continuity and resilience, with banks carrying on crisis monitoring, business as usual operations and customer support mainly by using distributed networks that have allowed their staff to work from home to an unprecedented extent. This has and will require continuing software investment.
The importance of software is reflected in CRR2 (Recital 27) and also recognised by bank supervisors:
- The currently applicable (CRR) prudential treatment of investments in software assets (in-full capital deduction) is counterproductive as it limits the capacity of EU banks to digitalise their operations. This aggravates the long-standing uneven playing field for EU banks vis-à-vis US and Swiss competitors, who benefit from full non-deduction of software assets under their prudential regimes, as well as vis-à-vis Bigtechs and Fintechs, who are typically not subject to capital requirements as they do not operate as credit institutions. As a result, in comparison to their direct competitors, EU banks are being held back by the currently applicable prudential regime.
- The change foreseen in CRR2 has the potential to correct this uneven playing field and gives EBA the opportunity to propose a meaningful exemption to capital deductions, in line with the letter and spirit of the Level 1 mandate and the legislators’ intention.
In a context characterized by a growing crisis and low interest rates, where future profitability will increasingly depend on the implementation of new business strategies based on digital services, the European banking industry strongly encourages the EBA to ensure that the exemption is indeed meaningful, not least by allowing a longer than two-year prudential amortisation period.
The proposal to consider only a two-year prudential amortisation period appears very conservative, when taking into account the following aspects:
- In the impact assessment analysis, the EBA refers to the fact that the average (accounting) amortization period of software in the EU is around six years. In comparison, we signal that a prudential amortisation period of two years is far too conservative (only one third of the average accounting amortisation period). In this regard, it should be considered that the accounting amortisation reflects the value in use of the software following already conservative accounting rules, verified by external auditors. The prudential amortisation should mirror this.
Our view is to calibrate the prudential amortisation period over the economic useful life of the intangible assets for at least 4 years, thus avoiding an excessive gap between the accounting and the prudential framework.
- EBA assessment of two years is driven by the expected time for a migration process of 1 to 3 years, which is based on limited data from recent resolution and M&A cases. According to the evidence that the EBF has received from its members, the range of migration period is quite longer than the EBA´s proposal.
In detail: the EBF gathered data referring to both distressed and non-distressed M&A cases, involving different types of software assets (core software assets, banking platforms, accounting tools etc.).
We noticed that, despite the different business choices of the acquiring entity in terms of amortization period of the specific software under consideration, the software asset was, or still is, in use for a far longer period than the one took for reference by the EBA in the draft RTS.
In different cases, the software assets were used, as a separate software asset, for four years after the acquisition date, and some of them are still in use as of now; in a different example the software asset was used for seven years, and even for a longer period in other cases. An example of “reverse integration” has been collected also, with the acquiring bank deciding to implement the core system of the acquired entity to its business units.
For confidentiality reasons, the EBF cannot provide, at this stage, more information referring to the aforementioned cases. However, we remain at your disposal for further engagement on a bilateral basis.
- Software assets would keep value for more than two years after banks are resolved (as it is the case for most European banks, as opposed to a liquidation regime):
• Capital requirements apply to banks in going-concern so it is not unreasonable to suppose that a bank that meets all capital and liquidity requirements at any point in time is highly unlikely to enter resolution from one day to another, as seems to be the conservative assumption that has led to set the limit to the partial non-deduction at two years.
• It seems that the EBA analysis does not take into account the prudential value of software when the resolution strategy is the bail-in or the bridge bank. This is relevant since the preferred resolution strategy for most of the European Banks is bail-in, and one of the consequences of applying an open bail-in strategy is that software´s useful life does not change during a resolution process.
- The evidence collected by the EBA points to two years’ worth of software value from the entry into resolution/insolvency/liquidation proceedings, not from the date of initial recognition or first use. Therefore, a healthy bank which started to use a software three years ago (amortised over six years from an accounting standpoint) will not recognise any value from a prudential standpoint, whereas, according to the EBA’s own evidence, it may use the software on average for an additional 2 years if it enters into resolution/insolvency/liquidation proceedings today.
Option B would be preferable for the following reasons:
- Banks have already all the necessary data for a quick implementation.
It would thus be easy to implement and therefore meet co-legislators’ intent to implement the new regime as quickly as possible.
- The operational constraints would be limited for EBA as well, given the prior validation of software accounting data by external auditors.
- It is also fundamentally desirable that both frameworks be aligned for consistency purposes.
The vast majority of the European banking industry expressed a preference for Option B. This option has been considered as the most conservative one and, in our opinion, should further support the request to extend the prudential amortisation period.
The opinion of the European banking industry of the cost/benefit analysis developed by the EBA is that the results presented in the Consultation paper (CP) are excessively optimistic.
In the baseline scenario, the EBA assumes the current negative impact on CET1 ratio, under the current framework based on software being fully deducted, to be 34.6 bp, over a restricted sample of banks able to provide sufficiently accurate and complete data in the 2018 QIS survey on software.
Applying Option B, EBA calculated a positive impact on CET1 ratio of 19.8 bp in 2018 (Figure 6 - page 39 of the CP), which would be equivalent to a gain of 57% of the activated software. In our opinion, this outcome overestimates the effect of option B.
The reason under this result could be the assumption EBA made on page 32, paragraph 11, of its RTS: “Note that due to data limitation, some assumptions were necessary to assess the impact of the different policy options:
‐ For the purpose of the impact assessment, the investments in software are assumed to be capitalised in full as of 31 December of each year.”
In a simple steady state software portfolio, in which each piece of software is activated with a useful life of five years on 31st December of each year, the gain in CET1 ratio is, according to our model (see attached) 51%. In this case the EBA result of 57% seems plausible given that currently software investments are rising. However, if an alternative, and more realistic assumption, is made (i.e. the activation of software is equally distributed over the year - e.g. 1/12 per month end) the same steady state software portfolio shows a gain in CET1 ratio of only 41%. According to our model, the EBA might have applied a best-case assumption, while a more realistic case assumption would consistently reduce the positive impact of the new legal framework.
Please find attached a simple EXCEL-Sheet running the above-mentioned software portfolios.
The Tab called “annually” contains three models based on year-end results: Model 1 (assuming that, according to EBA, Software is activated as the end of the year only), Model 2 (assuming that the Software is activated at midyear), Model 3 (assuming Model 1 but with a three-year prudential amortization period).
The Tab called “monthly” runs an exact model assuming a steady monthly activation of software.