By Wilfried Paus
At the wake of the EU implementation of Basel IV it appears that the modernization of financial risk measurement, which was initiated by the 2008 Basel II reforms, has been largely dismantled. Even more so, there remain voices that go far beyond the reduction of excessive variability in the capital requirements envisaged by the framework underlying the EU CRR3 legislation [BCBS 2017]. This article is intended as a plea to revive precise risk measurement and re-think risk management processes for effectiveness rather than letting the sector slide back further into the intellectual middle ages.
In October 2024, I was invited to a book event at Frankfurt’s Goethe-University organized by the Leibniz Institute for Financial Research SAFE. The central component of the program was the updated edition of the book “The Bankers’ New Clothes” co-authored by Anat Admati and Martin Hellwig [Admati/Hellwig 2024]. Both authors were present and accompanied by an equally distinguished panel of financial industry representatives[1].
The central message conveyed by both authors during the discussion was that, despite all regulatory reforms following the 2007-2009 financial crisis, banks remain inadequately capitalized for the financial risks they carry. It was accompanied by popular claims that bankers are oblivious to threats to the financial system and have a tendency to exploit its complexity to game relevant regulatory requirements to their own advantage. The other panelists bravely pushed back against these allegations. But after a good hour of debate, both sides still could not reconcile their differences.
What triggered me to write this article was the ultimate suggestion by both authors to increase the level of required capital for a financial institution to a fixed amount of its balance sheet of around 20%. In my view such a proposal would be nothing more than a fallback to the 1988 Basel I capital regime – at a 2.5 times higher rate (actually a lot more given that the scope would go beyond income earning assets). I believe that academics around the globe have more to offer than such blunt instruments.
But other stakeholders of the financial system should also revisit their role or, at minimum, rethink how they perform it. Let me share a few aspects to consider:
Financial institutions have a vital interest in adequately measuring their risks
Banks are in the business of trading money and thereby deliberately take risks. They have no interest in losing money irrespective of its root cause, i.e., whether these are due to position taking, result from flawed processes or are induced by uncontrollable external events.
Good measurement of all known (and unknown) risks enables better risk management through improved controls and more effective resource allocation. Overly complex models used in risk measurement, however, can mask otherwise visible risks and misdirect bank resources. If a substantial number of staff are deployed to optimize the outcome of models rather than improve their precision, this is a strong warning sign that risk management has moved out of focus. It is each financial institution’s responsibility to maintain the right balance here.
Risk management needs risk sensitive capital requirements
Despite the obvious weaknesses exposed in its early years, I disagree with statements (repeated prominently in the book event mentioned earlier) which dub Basel II as a failure. The Basel II framework and in particular its build-up fostered a far deeper understanding of banking risks and their dependencies than ever existed before. Banks were incentivized to develop their risk measurement and management tools. Simultaneously, associated control frameworks became mandatory (e.g., [FRB 2011]).
Under the impression of the Lehman and Euro crisis, Basel III rather made it worse by trying to make it better through partially adding more complexity alongside additional conservatism. The introduction of the advanced credit valuation adjustment risk capital charge for counterparty credit risk is one example here [cf. BCBS 2011, Ch. II A.].
Basel IV now has pulled the plug on several advanced approaches to quantify financial institutions profiles of certain risks. One prominent example is the decommissioning of Advanced Measurement Approach (AMA) models for operational risks. While Basel II enabled industry collaboration to understand the nature of these risks, pool loss data and align management processes, the new purely revenue-based Standardised Measurement Approach (SMA) [BCBS 2017 pp128] is certain to end these partnerships. This is a natural consequence given that the revised own funds requirement has nothing to do with the drivers of operational risks, forcing stakeholders to revert to less rigidly controlled internal models to understand their threat potential.
Drive for the right balance of transparency and complexity
The strive for precision in risk measurement is occasionally blurring the view on the object being quantified. Overly complex models tend to put off normal users and invite arbitrage. Commonly it is more beneficial to get the sensitivities than the absolute values of a risk metric right (unless, of course, the risk metric in question is not itself a sensitivity).
But the real challenge to model use and acceptance rests with undue conservatism often imposed by regulatory guidance. There is a natural reflex to mitigate model uncertainty with add-ons, some of them overshooting the target and paralyzing development. For instance, the EBA guidelines for IRBA parameters have rightfully introduced Margins of Conservatism for model weaknesses but propose their use in an additive manner. Moreover, rigid requirements on statistical significance force more and more low default portfolios into the credit risk standardized approaches and thus outside of a correct risk assessment [cf. EBA 2017 Ch. 4.4 & 9].
Other examples include the reservation to allow model-based approaches to transaction monitoring applied for the identification of money-laundering activities [BaFin 2024, Ch. 5.5.1]. This regulation deprives banks of concentrating their investigations of suspicious payments patterns to those with a high success rate. Instead, it mandates broad-brush rule-based investigations which produce an overwhelming share of false positive alerts.
We need more daring policy makers to accept the occasional errors in bank’s forecasts and estimates to ensure that, on balance, the results will be closer to reality.
Supervisory model understanding is key
By nature of their role, policy makers commonly don’t have enough insight on what banks are doing every day to manage their portfolios. This may yield incomplete or erroneous conclusions (see, e.g., [Paus 2024] in reference to [Faccia / Hünnekes / Köhler-Ulbrich 2024]). But supervisory authorities have both the power and qualifications to deep dive into such processes and the models integrated within them.
Unfortunately, supervisory audits often still tend to focus on regulatory rules and waste valuable time on interpretating the letter of the law rather than understanding the risk subject. Again, regulation can blur the perspective here, e.g., EBA Stress Test 2025 will conservatively disallow the reduction of operational risk capital in an economic downturn [cf. EBA 2024 Ch. 5.5] even though the aforementioned SMA was exactly designed to have this sensitivity.
Supervisors need encouragement to deep dive into the reality of day-to-day risk management when inspecting an institution and try to develop an understanding of what triggers internal activities and what does not.
Risk Management needs fresh academic ideas paired with easier access to data
The objectives or mandates of financial institutions, regulators and supervisory authorities bear the risk of obscuring the bigger picture of how make the financial system more transparent and robust. Consultancy firms can help to connect the dots but are also subject to dependency structures with their clients. For unbiased, high-quality ideas from the outside, however, no-one is better placed for this than academia.
But a common obstacle for more tangible research is limited access to bank’s internal data as banks have reputational risk concerns of leakage of information on their clients. To bridge the divide between those two legitimate interests, data pooling initiatives can be a catalyst. But this requires a robust legal framework to enable the data provider sharing information without risking breaches of the abundant data protection legislation. Such a framework can again only be delivered by policy makers.
We all need to change and help each other
None of the points I made above are genuinely new. But so was the conclusion of the distinguished panel mentioned in the introduction. My conclusion is that all stakeholders mentioned above need to take a step back for a deeper look at what drives their counterparts and revisit some of their own paradigms.
To put it bluntly: to shape the financial system for the better requires modest bankers, daring policy makers, inquisitive supervisors and open-minded academics. Whichever camp you belong to: please make the first move.
References
Admati, A./Hellwig M. [2024]: The Banker’s New Clothes,2nd ed., Princeton University Press, Princeton 2024
Basel Committee on Banking Supervision [2011]: Basel III: A global regulatory framework for more resilient banks and banking systems, Bank for International Settlements, Basel, 2011
Basel Committee on Banking Supervision [2017]: Basel III: Finalizing post-crisis reforms, Bank for International Settlements, Basel, 2017
Board of Governors of the Federal Reserve System [2011]: SR 11-7: Guidance on Model Risk Management, Washington, D.C. 2011
Federal Financial Supervisory Authority [2024]: Interpretation and application notes on the Money Laundering Act, Bonn 2024
European Banking Authority [2017]: EBA/GL/2017/16: Guidelines on PD estimation, LGD estimation and the treatment of defaulted exposure, London, 2017.
European Banking Authority [2024]: 2025 EU-Wide Stress Test – Methodological Note, Paris, 2017.
Faccia, D. / Hünnekes, F. / Köhler-Ulbrich, P. [2024]: What drives banks’ credit standards? An analysis based on a large bank-firm panel, Frankfurt am Main 2024, European Central Bank Working Paper Series No. 2902.
Paus, W. [2024]: What drives banks’ credit standards? An analysis based on a large bank-firm panel – Discussion, Frankfurt am Main 2024, FIRM Research Conference
[1] Consisting of Martin Blessing (Chairman of the Board of Directors of Danske Bank), Simon Gleeson (Clifford Chance) and as moderator Mark Whitehouse (Bloomberg).