How long will it take for bank regulators to ask AI regulators for a little favour?

Sir, Michael Skapinker writes: “Why did no one speak up “inside Sil­icon Val­ley Bank before it col­lapsed? People thought speak­ing up would leave them vulnerable to vic­tim­isa­tion.” “Listen and you might learn something” FT July 17.

Thought, or knew? 

What should a risk manager know about the risk of informing the board that according to revised models that included the interest rate risk (duration), SVB’ risk weighted capital/equity requirements would increase substantially? 

Would a bank supervisor like to go on record informing his superiors, the regulators, that because of IRR, the 0% risk weighting of long-term US Treasury bonds made no sense?

Sir, instead of exposing humans to victimisation, it seems precisely the moment that we could make great use of artificial intelligence. 

E.g., I asked ChatGPT – OpenAI: “Should regulators and supervisors be aware of risks with US Treasury long-term bonds? 

It answered: “Yes, they should be aware of the duration risk and interest rate risk associated with long-term Treasury bonds held by banks”

But of course, AI could be vulnerable to victimisation too.

I asked Open AI:” Can those who become an Artificial Intelligence regulator, make you or any other AI participants agree with all they want you to agree with?” 

It answered: “Regulators aim to address ethical considerations, potential risks, and ensure responsible AI practices… AI systems don't possess independent consciousness or the ability to willingly agree or disagree with regulations. Their behavior is determined by their programming and the data they have been trained on.”

Sir, you know much to well, that for more than two decades I’ve been vociferating, as much as I can, my criticism against the risk weighted bank capital requirements. Clearly when someone does want to hear, he does not hear. For instance, as Upton Sinclair Jr. explained it: “It is difficult to get a man to understand something, when his salary depends on his not understanding it.”

Therefore, when I heard about OpenAI, I asked it a series of questions. From the answers you must agree I found an ally. I just wonder how long it will take for bank regulators to shut it up. “We can’t have someone questioning the risk weights of 0% government – 100% citizens. Can we?”

How long will it take to FT to really listen to some of its faithful subscribers?

PS. US Treasury long-term bonds still carry a 0% risk weight.

https://subprimeregulations.blogspot.com/2023/04/chatgpt-openai-should-bank.html

https://perkurowski.blogspot.com/2023/07/ai-chatgpt-open-ai-artificial.html

https://subprimeregulations.blogspot.com/2023/06/my-questions-to-chatgpt-openai-about.html

@PerKurowski

Require banks to hold capital against the unexpected, making sure those not able to manage perceived risks fail, fast

Sir, Nobuchika Mori, commissioner of Japan’s Financial Services Agency writes: “Regulators have made the global financial system more resilient by major regulatory reforms. Banks now have much bigger capital and liquidity buffers. Resolution frameworks have been strengthened. Derivatives markets are being made safer, while toxic forms of shadow banking have been detoxified.” “A holistic approach to future-proofing the financial system” May 26.

Holy moly! How come bank regulators have not been nominated for a Nobel Prize? I mean, if Bob Dylan and Juan Manuel Santos could each get one, and if what the author says is true, these also sound to be worthy one or two.

But no! Nobuchika Mori then goes on describing some pending issues that would not make them worthy of a prize, on the contrary, more worthy of being sacked.

He writes: “It is possible that funds are not being allocated in ways that foster economic growth. Credit should be provided in ways that enhance productivity, revitalise industry, foster innovations and create new businesses.”

Indeed! But should not the purpose of the banks have been defined before regulating these?

He writes: “We need supervision to establish whether regulation offers perverse incentives to banks to accumulate excessive risks.”

More than supervision what is needed are regulators that understand that with their current capital requirements, banks will accumulate too large exposures against what is ex ante perceived as safe (but that ex post could be risky) and too little exposures to what is perceived as risky, like those “small and medium-size enterprises” those who according to the author could “help to revitalise local communities”

“An Explanatory Note on the Basel II IRB (internal ratings-based) Risk Weight Functions” expresses: 

“The model [is] portfolio invariant and so the capital required for any given loan does only depend on the risk of that loan and must not depend on the portfolio it is added to.”

The explicit reason for that mindboggling simplification was: 

“This characteristic has been deemed vital in order to make the new IRB framework applicable to a wider range of countries and institutions. Taking into account the actual portfolio composition when determining capital for each loan - as is done in more advanced credit portfolio models - would have been a too complex task for most banks and supervisors alike.”

But here Nobuchika Mori wants the supervisors to take on even a more active and complex role. That would not be helpful. That implies assigning the supervisors too much capability and, explicitly, too much power.

Much better it is to get rid of all risk weighting setting a reasonable leverage ratio, like 10% capital for all assets (sovereigns included), to cover in part for what is unexpected… all in the knowledge that the faster banks that cannot manage perceived risks fail, the better for all.

The Questions!

@PerKurowski