The meeting on 7 December focussed on the final unresolved issue: the output floor for capital requirements calculated using internal models. The GHOS set this at 72.5%. Bundesbank President Jens Weidmann and BaFin President Felix Hufeld represent Germany in the GHOS. "The conclusion of the Basel III reform package is important because it means that, eleven years on from the outbreak of the financial crisis, another major lesson has finally been learned. It also eliminates the regulatory uncertainty that was weighing on banks," Bundesbank President Jens Weidmann explained. "The output floor we have now set is hardly the outcome Germany had been hoping for, but it’s a compromise that all the stakeholders can live with," BaFin President Felix Hufeld remarked, adding, "What mattered to us was that the global banking regulation regime does not depart from the principle of risk sensitivity and that it continues to permit the use of internal models."

Limiting deviations

The main thrust of the revised Basel rules is to curb the unintentionally large deviations in the capital requirements which banks calculate using their internal models. For banks which use their own internal models, the output floor will cap the potential capital saving these models offer at 27.5%, compared with the figure produced by a standardised approach.

"The new rules present stiff challenges which institutions will now need to get to grips with. But banks will now have nine years to gradually acclimatise to the new requirements, and that’s doable,” Hufeld noted.

Successful negotiation by BaFin and the Bundesbank

BaFin and the Bundesbank firmly advocated the idea that the credit risk standardised approach should better capture actual risk content and that banks should still be able to calculate capital requirements for many portfolios using their own internal models. One successful outcome of the negotiations is that verifiably small losses in real estate lending business, which is of particular importance to the German market, can be recognised as having a risk-mitigating effect (known as the “hard test”).

"All the members of the two bodies – the GHOS and the Basel Committee – have committed to push for full and timely implementation of every element of the Basel III package in their respective countries. That was crucial for gaining our approval for the reforms," Bundesbank President Jens Weidmann said.

Interview
BaFin President Felix Hufeld: "I am not overjoyed, but the outcome is workable"

BaFin President Felix Hufeld, who, together with Bundesbank President Jens Weidmann, represents Germany in the GHOS, explained in an interview the main aspects of the agreement and what it means for German banks and the market as a whole.

Mr Hufeld, after a long struggle you did manage to reach an agreement in the GHOS after all. Could you tell us about the details of the compromise?

The majority of the revisions to the Basel III Framework were settled back at the end of 2016 in the decisions of the Basel Committee in Santiago. Agreements were reached there regarding constraints on the use of internal models and new standardised approaches for credit risk, operational risk and derivative risks. In addition, global systemically important banks are expected to have a higher leverage ratio. For all of these issues we had already managed to find compromises in the negotiations that we are quite satisfied with.

The main sticking point was the calibration of a measure to limit the variability of the models that banks use to calculate their risks. There were differing views on how such an output floor should function and its consequences for banks' capital resources, which led to delays in reaching a consensus. The agreement now provides for an output floor of 72.5%, based on the use of standardised approaches.

Are you content with this outcome?

I am not overjoyed, but the outcome is workable, and is a very important milestone towards maintaining a global standard in banking regulation. I am satisfied with the outcomes of the negotiations regarding real estate, bonds and exposures to corporates and to retail clients, which are of particular importance for Germany. Above all else, though, the key point is that global banking regulation has not departed from the principle of risk sensitivity and still permits modelling. A de facto abolition of this principle would have crossed the red line for me. Modelling risks based on loss history and forecasts forces the banks to develop a very detailed understanding of their risks. We supervisors need this knowledge, too, in order to understand and assess the capital requirements of systemically important institutions and thus the financial system as a whole.

I am not trying to sing the praises of modelling. We had to prevent it from being overused and limit undesirable variability in the results. We hope that this will be achieved, among other things, via the input floor, which limits the probability of default and the loss given default, and the output floor I mentioned earlier. And then we come to the things for which I would have preferred a different outcome: on the basis of the effects that we know about so far, the calibration of the output floor is on the boundaries of what is acceptable. Even if it only affects banks using models, it will demand a great deal of certain institutions, including those that operate portfolios with impressively good default histories and that thus have justifiably low risk weights for these portfolios.

Why was this Basel III compromise so important?

The capital and liquidity rules are intended to help stabilise the markets. It would be a cause for considerable concern if, due to a single crude risk measure, banks were only able to succeed in the balancing act of meeting regulatory capital requirements on the one hand and owners' expectations on the other by means of high-risk new business.

Together with the Deutsche Bundesbank, we are responsible for the stability of our banking market. As the integrated financial supervisor, we also see the overarching effects on the financial market as a whole. We are at a point now where we need to weigh up the extent to which additional capital requirements stabilise the markets before they start having the opposite effect. If capital costs are too high, this can place an undue burden on the banking industry. In particular here in Germany, we need the banks, because our economic system, which is primarily formed of Mittelstand companies, is financed more by bank loans than via the capital markets. In other countries, this is not the case.

Speaking of other countries, I'm sure the problems that banks face in other countries are no smaller than those in Germany. So why was it still so difficult to reach an agreement?

Because the problems the banks face are of a very different nature. The business models and risk structures are different. And since the current set of rules provides for options and discretion, regulation is not applied in such a uniform manner that all parties entered into the negotiations with equal preconditions. For example, we Europeans implemented the mother of all output floors – the Basel I floor, which was intended to prevent the capital requirements from dropping too low through the use of models – in a different way from other jurisdictions. Over many years, this led to more favourable capital requirements in Europe. In countries in which the model outputs already have tighter lower limits, the increase in capital requirements due to the new regulation is, of course, low to non-existent.

When are the new rules supposed to enter into force?

The rules are intended to apply from 2022. The output floor itself will then be phased-in over five years, until it reaches its full level of 72.5% on 1 January 2027. In other words: from today, banks have nine years to gradually adjust to the new requirements. They should make use of this time.

Is this the end of the process, or are there still important things to be done?

There are indeed still some things to be done. The framework itself is a giant mosaic. Not every little stone fits in perfectly; some repair work will be needed. Moreover, we will need to take greater account of how different regulatory requirements interact – including those from accounting rules – so that we are able to fully assess the overall effect on bank balance sheets. This is far from trivial, and will require a certain amount of further work on our part.