Top Bank Regulator Takes on ‘Drive Fast, Crash’ Risk Culture

The Wall Street Journal’s Risk and Compliance Journal spoke with Acting Comptroller of the Currency Michael Hsu—one of the nation’s top banking regulators and a leading voice on risk and on banks being “too big to manage"—about the past year and the challenges looming for the industry.

This interview has been edited for length and clarity.

WSJ: We saw several high-profile bank failures in the past year. Where are we now?

Hsu: The national banking system is sound. I can say that fairly confidently, in part because, following those failures from earlier this year, banks and supervisors really focused on those risk areas to make sure that there are proper controls. There’s enough resilience, financial resources, et cetera, so that we can ensure that we’ve got a sound system for everyone.

WSJ: You have raised the problem of “too big to manage" and potentially taking action against banks with persistent weaknesses. Are you in a place now where you can talk about potential future actions?

Hsu: Internally, we have these discussions a lot. I’m not really prepared to have discussions externally. Part of the logic of laying out that framework is just to put the onus of proving whether a bank is too big to manage or not on the bank.

If there’s kind of a framework that has escalating consequences, it should be able to fix those things. And if it can’t, then that is kind of de facto evidence that [a bank] may be too big to manage and we’ll take action accordingly.

WSJ: The business side can see risk managers as a drag when things are going well. Do you think, given some of the shocks in 2023, they have more of a voice?

Hsu: If you go through the post mortems of risk management failures—Credit Suisse with Archegos, JPMorgan Chase with the London Whale, the ’08 crisis—if you really dig into those reports, somebody in the institution knew. It wasn’t necessarily a lack of knowledge. The credibility of those who knew that was disproportionate to those who are making the money and those who are running the businesses. What we want to do as supervisors is to make sure that there’s a balance.

These banks are not monoliths. There’s lots of different groups who exist within them, and it’s good to recognize that there are people and processes and systems to try to make that all work. And we try to make sure that we’re really providing that necessary support for folks to do their jobs.

WSJ: We’re in a new era in terms of banking technology. Some people, for example, have savings accounts on their phones that they could empty in seconds. Do we need more friction in banking to avoid a repeat of, for example, the SVB situation?

Hsu: There’s a couple ways to approach this issue. The macro view on it is that things are speeding up—transactions, the velocity of money movement and of information in general. The information environment and the ability to act has sped things up, and that’s something that we all need to grapple with. We need to take our time to really think about that deliberatively to say, “OK, where does this really impact things, and what do we do about it."

At the micro level, this does have implications for liquidity and for bank liquidity, and I think we saw that with Silicon Valley Bank and with Signature. There’s been a stronger emphasis and focus on really short-term liquidity. We’ve put a lot of attention on this to make sure banks [can] generate the cash necessary to survive faster bank runs.

WSJ: Securities and Exchange Commission Chair Gary Gensler has talked about artificial intelligence as a potential major risk to the financial system. Are you worried about AI?

Hsu: I do worry about AI from a number of different angles. I don’t think we can zero in on just one of those risks. I think the right way to approach it is with a wide aperture because it’s so nascent and developing. Every week is like a new week in the AI space.

There are the financial stability risks that Gensler has highlighted. We also have to think about a whole host of other ways in which AI is being used directly and indirectly within the financial services space. Something that’s a little closer to home is going to be related to microcredit decision-making, and discrimination and bias related to that and fraud detection.

Banks have been doing machine learning for quite some time. How they use that, and how they expand that to other uses, and uses of Gen AI really are questions that are live, that we have to think through. On top of that, I go back to operational [issues]. A lot of these operational processes are no longer held just solely within the bank. There’s almost like a supply chain of these things where the bank relies on someone, the third party relies on someone else. And so we really have to start thinking with a bit more of a supply-chain mindset in terms of operational resilience.

WSJ: What is your outlook on these issues?

Hsu: What I’m optimistic about is if we can get these things right, then it really enables the financial system to just go. It enables banks to focus on lending, it instills confidence and trust. That’s why we put so much time and energy into this. I have a saying, “The better a car’s brakes, the faster you can drive it safely."

Some people want to sequence it like “drive fast, crash, build brakes." No. It’s better to do these simultaneously, because [with] stronger controls you can drive confidently at higher speeds. And that’s ultimately what we want.

Write to Richard Vanderford at Richard.Vanderford@wsj.com