Strong, demanding supervision is needed more than ever - Andrea Enria (European Central Bank)
Keynote speech by Andrea Enria, Chair of the Supervisory Board of the ECB, 22nd Handelsblatt Annual Conference on Banking Supervision
Risk Models
We need to have expert judgement and some way of adjusting from the outcomes of the model and the way in which we look at risks.
Models are very useful instruments, I think. They are the best instruments we have to quantify things. But what they do is simple: they take the past, they crunch it into data and then they put the data into a machine and then try to project the future. They work very well when you are in a relatively stable structural environment and you have some sort of cyclical movements around a relatively stable trend. But at the moment we face geopolitical shifts, we have energy shocks, we have climate change, we have digitalisation. When we had the pandemic, we crunched data with our models to try to predict the amount of non-performing loans and we came out with humongous figures. Eventually, luckily enough, these did not materialise because there was a lot of public support and the economy rebounded much faster. Default rates were very small. But, by taking the data from the financial crisis and plugging them into the pandemic, we came out with the wrong answer. So, that for me was a lesson. We need to continue using models, of course – we use them in regulation. But we also need to have expert judgement and some way of adjusting from the outcomes of the model and the way in which we look at risks.
There is a perception that we are in some way increasing capital requirements by the back door through our internal model investigations leading to increased risk-weighted assets (RWAs) and that, in doing so, we are making capital requirements an unpredictable variable. In reality, there is no policy aimed at making internal models more costly and no intention to surprise banks with tighter requirements. Most of the impact of our work on internal models was accomplished through the targeted review of internal models, the final report on which was published in 2021. The overwhelming majority of internal model inspections that we have carried out since 2020, for instance in the area of credit risk, follow from requests submitted by banks for material model changes to be made in the context of the European Banking Authority (EBA) internal ratings-based repair programme, which the EBA started planning and communicating back in 2016. Under that programme, model changes can result in increased RWAs in those cases where the bank’s model specifications differed, to a material extent, from the requirements specified by the EBA. On an ongoing basis, we are required by law to only grant banks permission to use internal models if they meet the relevant requirements. With the implementation of the finalised Basel III reforms, which notably constrain the use of internal models under several dimensions, we should expect a slowdown in such supervisory activity. It is also up to banks to reduce the complexity of their internal model framework in those instances where, rather than achieving true risk sensitivity of capital requirements on material business lines, models are only used to achieve even very minor regulatory capital savings on non-core business.
Internal Controls
Internal controls are the backbone of the viability of a bank.
I don't know if you have read the report that Credit Suisse itself did after Archegos. It's a very candid report and is very interesting, in my view. It shows that it all boils down to internal controls and risk culture. That is a clear finding – they themselves identified it as a learning from that case. If you look at the last days of life of Credit Suisse, they delayed the publication of the annual report as requested by the SEC, and then the publication highlighted material deficiencies in internal controls. So, again internal controls are the backbone of the viability of a bank. That’s why we, as supervisors, do and need to continue to put a lot of emphasis on that.
Liquidity
Maybe tailor requirements for that specific business model rather than revise the global standard because of an outlier bank.
I believe that the liquidity coverage ratio is a good supervisory metric. I think the banks need to have liquidity buffers. This is what requires them to have it. It has worked. I am satisfied with the way in which it gives me an idea of where the liquidity of different banks is. You are right: there have been some rather fast outflows of deposits in some cases, so we need to study this and maybe see whether the calibration of certain factors is good.
领英推荐
But honestly, if I look at what happened, for instance, at Silicon Valley Bank: the rule is always a metric that we apply to all the banks. But then there are outlier business models. That model was very concentrated: almost exclusively uninsured large-ticket depositors; all homogeneous in terms of activity; all venture capital tech firms in the same region; also the size of the borrowing. There was also a bit of groupthink between the depositors, borrowers and bank management, so they went out very fast. I haven’t read about the testimony of Michael Barr in the newspapers today; I want to read it carefully and understand it better. But my impression is that, in those cases, it is?supervision?that needs to look at the business model and – if anything – maybe tailor requirements for that specific business model rather than revise the global standard because of an outlier bank.
The fast-paced normalisation of the monetary policy environment has in turn required us to brush up supervisory tools aimed at tackling risks of the banking business that had long not been included among the priorities. I am thinking of interest rate risk in the banking book and funding and liquidity risk.
Capital Management
It is true that in the United States, authorities took a decision?to decouple the rules applicable for the large systemically important banks from those for medium-sized and smaller banks, expanding in 2018 the scope of banks benefiting from lighter regulatory approaches. In the European Union, we traditionally implement the Basel standards more consistently across banks of all sizes.
However, that does not hold for the G-SIBs, the global systemically important banks, which are more comparable in terms of business model and which compete on global markets. EU G-SIBs are in the same ballpark as US G-SIBs in terms of capital requirements: the weighted average Common Equity Tier 1 (CET1) capital requirements for EU G-SIBs stood at 10.6% as at the fourth quarter of 2022. For the US G-SIBs, the CET1 capital requirements were a little higher at 11.1%. But I don’t want to make too much of simple comparisons – there are many differences in the types of risks that banks run, and we know for example that the risk weights attached to European banks’ assets tend to be lower than those applied in the United States.
#riskmanagement #creditrisk #riskmodeling #capitalmanagement #baseliii #ifrs9 #europeancentralbank #andreaenria
https://www.bankingsupervision.europa.eu/press/speeches/date/2023/html/ssm.sp230328~1797047d39.en.html