Michael Hsu

Video: OCC Acting Comptroller Michael Hsu at the 2024 Just Economy Conference

Online Event Archive Recorded April 4, 2024

The Acting Comptroller of the Currency Michael Hsu provides remarks and answers questions from the National Community Reinvestment Coalition’s (NCRC) President and CEO Jesse Van Tol at NCRC’s Just Economy Conference.


NCRC video transcripts are produced by a third-party transcription service and may contain errors. They are lightly edited for style and clarity.

Hsu 00:14

Good morning. Okay, one more time. Good morning. So, thank you so much for having me back. I remember doing this last year and the walkup music was Metallica, which is a little bit more, it’s very, I remember those, I remember these things.

So, first of all, I’m deeply honored to join you today for the Just Economy Conference. And last year, I spoke about elevating fairness, as a top priority for the OCC. And I am proud that that commitment continues.

Today, I want to highlight several developments since we last met, as well as look ahead and share some thoughts on fairness as it relates to artificial intelligence and fraud. At this conference last March, I highlighted the OCC ‘s commitment to promoting safe, sound and fair bank overdraft protection programs. Shortly after an APR 2023. The OCC released guidance based on our supervisory experience to assist banks in managing the various risks associated with overdrafts. The guidance identified two practices; authorize positive settle negative and representment that may result in heightened risk exposure. It also highlights sound risk management and pro-consumer practices that banks can employ to strengthen their overdraft protection programs.

As we approach this first anniversary of the overdraft guidance, it’s worth reflecting on what progress has been made and what gaps remain. Since the OCC has heightened attention on overdraft began, the overdraft fees charged by OCC regulated banks in aggregate have fallen over 40% from $6.5 billion to $4 billion in 2023. Right that’s money that goes stays in the pockets of those who needed.

Supervisory data continue to show declines in overdrafts quarter over quarter. Among large banks several have outperformed their peers in terms of reducing fees and adopting pro-consumer features, such as grace periods, while others have reduced overdraft fees only modestly. I don’t have a way to show this but if you go if you’re online in my speech as a graphic which shows bank by bank those changes. Encourage you check that out.

Importantly, all OCC-supervised large banks have stopped assessing authorized positive settle, settle negative fees or insufficient fund fees and most have discontinued collecting sustained overdraft fees. OCC-regulated midsize and community banks have made or are in the process of making various pro-consumer changes to their overdraft protection programs. For instance, most have eliminated representative fees or started offering de minimis grace amounts or grace periods. This includes those community banks that derive an outsized amount of revenue from overdrafts. Progress with regards to representment has been more challenging due to the critical role played by the core processors. I had an opportunity to share community bankers’ concerns on this issue with representatives of the core processors and encourage them to continue to enhance their systems to facilitate transparency and pro-consumer practices. We understand the largest core processors are taking steps that will allow banks to identify and address representment practices. As those plans develop will continue to encourage banks and the core processors to take steps to protect and empower consumers.

Another core element of elevating fairness at the OCC is our continued effort to strengthen fair lending supervision. Last year I highlighted the OCC has comprehensive update to the fair lending booklet of the Comptroller’s handbook, which remains an important source for transparency and detail regarding the OCC supervisory practices. The OCC has continued to deploy an enhanced risk-based supervision process. This includes the fair lending risk assessment, as well as an annual statistical screening process that leverages hunted data. We run several screens on the data for each OCC-supervised bank, including for redlining. In the past year, we’ve developed new screening methods that can identify potentially discriminatory CRA assessment areas by using census and geographic data to detect when communities of color are located near but excluded from the boundaries of an assessment area. We remain focused on efforts to efficiently deploy supervisory resources to identify weaknesses and potential discrimination.

Internally, we’ve implemented a program to provide more advanced real-time support to fair lending reviews. Fair lending experts from our legal, policy and economics departments now work proactively with OCC supervisory offices to identify and address complex fair lending issues to enhance and expedite the examination process. This includes deploying staff skilled in the analysis of complex models to support examinations involving banks that had begun to incorporate advanced analytics such as machine learning and artificial intelligence into underwriting systems and fair lending risk management programs.

Ensuring that the federal banking system is free of discrimination is a part of our mission and critical to safeguarding trust in banks. That’s why we are committed to continuing to strengthen our supervision of fair lending, especially as banking practices and technologies change. Now, the OCC has also elevated fairness to the work of project REACH, which stands for the Roundtable for Economic Access and Change. Project REACH brings together leaders of banking, civil rights, community advocacy, business and technology to identify and reduce key barriers to financial inclusion for underserved minority communities. It’s a unique program with enormous potential. I’d like to provide an update on three project REACH focus areas, starting with credit invisibles. Tens of millions of consumers in America are unable to access credit because they lack a credit file and do not have a FICO score. Many of these so-called credit invisibles, pay their rent and utility bills on time, and are otherwise credit-worthy. Making them visible to the banking system helps get them onto the first rung of the economic ladder. As part of Project REACH several national banks have undertaken a pilot program to use alternative non-FICO data, primarily from deposit accounts to qualify consumers for first-time credit cards. As of October 31, 2023, over 110,000 accounts were established under this pilot. Participating banks have been monitoring key performance metrics to track customers credit progression after account opening, and the initial reports are promising. After 12 months, the average FICO score was 680. Recall this population initially was credit invisible and had no FICO score at all. The success of this pilot shows how fairness outcomes can be accelerated through collaboration and engagement. Because that success can also inspire. Some REACH participants have begun discussing tackling similar credit invisible challenges for veterans and justice-involved individuals.

Another key focus of project REACH has been revitalizing minority depository institutions or MDIs. 26 banks have signed a pledge to support MDIs with over $500 million in financial support and technical assistance to help those institutions grow and expand in a safe, sound and sustainable way. Participants created an MDI FinTech playbook and an online portal, through which MDI employees can receive specialized training and information to help them better serve their communities. In addition, several large banks have assigned senior executives to spend a year working on-site with MDIs to share best practices, and strengthen operational and managerial aspects of successfully running a bank. MDI feedback has been overwhelmingly positive and all participants have noted the mutual benefits of collaborating as partners and building relationships, not just relying on transactions.

Project REACH participants have also helped to expand access to housing for low- and moderate-income communities. This includes homeownership on tribal lands, which has distinct challenges. To address those complexities, REACH participants hosted a national webinar for bankers to demystify the process and support opportunities for home finance in Indian Country. The group also sponsored deal-making sessions among tribal leaders and banking officials that focused on affordable housing. Project REACH also helped facilitate several banks launches of special purpose credit programs focused on providing mortgage financing for minority communities as permitted under the Equal Credit Opportunity Act. Another group of REACH participants has worked on expanding the supply of affordable housing through accessory dwelling units or ADUs. Focusing on California or zoning laws for ADUs had recently changed. They developed a white paper and piloted a loan pool to facilitate ADU financing. These other project reach efforts have sought to tap into the creative energies of bankers, civil rights and community organizers and technology leaders to collaborate and craft novel solutions to key fairness challenges. I’m proud of what Project REACH has been able to accomplish and looking forward to what REACH participants will do in the future.

Looking ahead. Fairness is not self-enforcing. ie It doesn’t happen on its own. Achieving fairness requires work. We must be just as vigilant about ensuring fairness with new technologies and emerging issues as we are with traditional banking products and services like overdrafts and mortgages.

Today, two emerging areas standout; AI and fraud. From a fairness perspective, AI holds both promise and peril. AI has the potential to reduce bias and enable fair access to credit and banking services in ways that humans have been challenged to do. We should welcome this and create spaces to safely explore and develop that potential. At the same time, however, AI has the potential to perpetuate and exacerbate the biases, discrimination and unfairness that are deeply embedded in the data feeding AI systems. The challenge is that there can be strong echoes of race, gender and other characteristics in large datasets, even when race data are removed. This means that even if a data set were quote, unquote, “colorblind,” the redundant encodings and other variables would likely reflect and reveal race as a factor. To guard against this, banks need to have appropriate oversight and governance of the models they use. This includes monitoring for fair lending impacts as credit models are developed, validated, monitored and tested.

The intersection of AI and fraud also warrants close attention. The Federal Trade Commission estimates consumer fraud losses in 2023 exceeded $10 billion, a 14% increase over 2022. The rise can be attributed to increasingly sophisticated digital tools used by fraudsters, the accessibility of AI tools is making it easier to generate deep fakes to clone voices and to otherwise defraud individuals and businesses. I am concerned about a potential explosion in AI-driven fraud, with elders and members of vulnerable communities being disproportionately preyed upon. To successfully fight AI-driven fraud will likely require AI-driven solutions. Banks and AI companies are best positioned to deliver those. In the UK, the payment systems regulator recently issued a new reimbursement requirement for authorized push payment fraud, where criminals trick an individual to send money to a fraudulent account. Under the new requirement, the customers bank and the receiving bank must reimburse the customer for losses with the reimbursement costs split 50-50 between each bank. In addition to making the consumer hole, the requirement incentivizes both banks to develop and implement more effective fraud controls.

This liability model deserves greater discussion and debate here in the US, as noted by some consumer advocacy groups who are probably in the room now. In cases where AI plays a role in the fraud, splitting the liability between the customers bank, the receiving bank, and the AI platform would be a good starting point for consideration. Such a liability split would protect consumers while creating strong incentives for those best positioned to develop effective defenses. In addition, the gap between large and community banks when it comes to fraud data and AI capabilities must be addressed. Large banks can develop and train anti-fraud models using their own internal data. Community banks cannot. Large banks can also invest in AI talent and IT systems in ways that are impractical for community banks. A significant part of the solution lies again, with the core processors upon which most community banks rely. The OCC has recently deepened its engagement with the core processors and intends to expand this engagement going forward in the provision of financial services to communities fairness, rest as much with the core processors as it does with community banks.

In conclusion, I want to re-emphasize how integral fairness is to the OCC mission and Bethany read this out earlier but I want to repeat it. Quote: ‘To ensure that national banks and Federal Savings associations operate in a safe and sound manner, provide fair access to financial services, treat customers fairly, and comply with applicable laws and regulations.’ We’ve made a lot of progress in elevating fairness in the federal banking system since I spoke to you last and rest assured we will continue to focus our efforts on fairness going forward. Thank you.

Van Tol 14:59

‘Fairness is not automatic,’ I love that. I love that quote. And, Comptroller, one of the things that we have been concerned about, you know, we have the new CRA rule, which I’m gonna ask you more about in a minute. But CRA was really passed in response to redlining. And yet, in a sense, the laws that more directly address redlining, ECOA, and other things, are not fully integrated in the review. There’s sort of a retroactive integration. If there’s a fair lending issue, it can result in a downgrade. But we’ve taken a look at the last dozen or so banks to settle with the Department of Justice for redlining, and actually, 10 of the 12 had passing CRA grades, two of them had outstanding CRA grades. How, as you think about the implementation of the new CRA rule, and some of the other moving pieces and with respect to regulatory policy, how do these things come together and be coordinated?

Hsu 16:09

So I can’t say much about the implementation of the rule. What I will say is that, Jesse, you’re absolutely right. You know, the CRA statute passed in 1977, that was was passed as part of a package of other legislation to address these deep, deep discrimination, discrimination, and redlining. And we are fully committed at the OCC to carrying out the statutory purpose of CRA, which is to address that; to make sure we need to fully assess how banks meet the credit needs of everyone in their community, including and especially low- and moderate-income neighborhoods. So that’s really critical. We’ve remained fully committed to that we’ll use all the tools available to us to do that.

Van Tol 16:53

Shift in the bank merger reform, you’ve talked a lot about this recently, issued some proposals, the FDIC recently issued a proposal on bank merger policy, really thinking about the steps that you have taken. What’s critical to us is this notion that every merger has some type of adverse effect. You know, could be product loss, could be job loss, could be branch closures could just be the fact that you know, one plus one equals slightly less than two in the aggregate lending. Talk about your views for how the public interest is served and evaluated in the bank merger process.

Hsu 17:44

Sure, sure. So I think there’s a popular conception that all bank mergers are like that. That’s not the case. That is sometimes the case. And I think what we’re trying to do by updating the frameworks by which we analyze is to make sure that when that is the case, they don’t get approved. I do want to clarify that there aren’t, mergers can be pro-community, they can be pro-competition, they can be pro-financial stability. So I want to be careful that we don’t paint all mergers as bad, or all mergers as good. They’re not also all good. There are some out there who advocate that. That’s not true. We need to be we need to be very discerning about the mergers and part of this update. The more discerning we are, the more transparent we are. We need assist a banking system that can be dynamic and pro-community. That’s that’s the objective. That’s what we need to drive towards. We can’t just freeze the banking system in place. That’s not good, long term for anybody. So I think we share the objective and like the devils in the details, we have a policy statement, we put it out for comment. And we really I want to echo something that Bethany had said, please give us your comments. I know there’s some comment fatigue. Now there’s a lot of things to comment on. We take them very, very seriously. And the more we hear from you, the better we can set those bars, set that yardstick so that we see good mergers.

Van Tol 19:10

Mike, you’ve also talked about and expressed concerns about the resolveability of non-Gsib- the globally systemically important banks, banks, what concerns do you have there? What risks do you see in those kinds of mergers?

Hsu 19:30

So no bank should be too big to fail, period. I want to tell a really quick story. Back in 2008, I was at the SEC at the time, and in September, Lehman failed. Lehman filed for bankruptcy. I was there. I was part of that team overseeing that. And it created huge uncertainty and the impacts of that loss on hundreds of millions of people who had nothing to do with that. It was felt by everybody. And people lost their savings, people lost homes. It’s just the blast zone of problems from that failure you had a lot of it was just not right? It was unjust. Two days later, AIG got $85 billion loan from the Fed. Later on that fall, we get a $40 billion equity injection from Treasury Department. We required further bailouts. Because the fear was that if AIG did this, we had the same problem, that even more people would feel more pain. Both of those outcomes are unjust. That is the core of too big to fail problem. Dodd Frank gets passed. Dodd Frank then mandates never, that’s not going to happen again. There needs to be living wills, resolution plans, large those kinds of bank need to be resolvable. So that framework gets put in place. I had a role in helping to put some of that into place for the largest G sibs. Now the issue is, in my opinion, all large banks should have elements of that, so that we can avoid that in the future. We have, we have reliable, we can be confident in those protections for all large banks in the future. And that’s part of what the proposals and things that we’re working on both at the OCC and across the federal banking agencies to make sure we avoid that.

Van Tol 21:23

Thank you. I also want to thank Comptroller Hsu for your leadership on CRA. And no we we have a complication with respect to the final rule. But we also know that much of the banking industry is preparing nonetheless to implement the rule and we’re excited for what that brings. So thank you for your leadership and another NCRC round of applause for the Comptroller.

Hsu 21:52

Thanks for having me, Jesse.

Print Friendly, PDF & Email
Scroll to Top