Governor Elizabeth A. Duke
At the 2012 Bank Presidents Seminar, California Bankers Association, Santa Barbara, California
January 13, 2012
It's a pleasure to be here this morning to address the 2012 California Bankers Association Bank Presidents Seminar. Before joining the Federal Reserve, I spent most of my career in banking, much of it as president of a community bank. So I know how important these conferences can be as you try to meet the expectations of regulators, customers, employees, communities, and shareholders--and to do so safely and profitably--with limited resources. For more than three years now I have viewed banking from a different perspective, as a regulator whose primary concerns are the safety and soundness of the banking system and consumer protection. Three years may sound brief, but I think that the intensity of my tenure more than offsets its relatively short duration. I arrived at the Federal Reserve in 2008, just in time to serve on the front lines of the combat against the collapse of our financial system. I witnessed first-hand the damage that can result from reckless lending and weak risk-management practices.
As we all know, fallout from the crisis was not limited to those who engaged in the activities that were at the center of the problem. Indeed, as the crisis developed, conditions deteriorated so severely that many banks that had been considered financially strong, that had never made a single sub-prime loan, found themselves struggling for survival. This was particularly the case for banks in states like California that were most heavily impacted by a sharp deterioration in real estate markets and a significant increase in unemployment. Community bankers' efforts to address the asset quality problems that followed have been crucial to recovery. We are just now starting to reap the benefits of those efforts and are seeing some improvement in community banks in California and across the country.
But while the effects of the economic crisis are receding, bankers are now facing a wave of increased regulatory requirements. For the most part, the new regulations are directed at the largest institutions, whose failure would pose the greatest risk to the financial system, or at the lending practices that led to the crisis. Even so, the changes are so sweeping that many industry analysts have questioned whether the overall weight of regulation poses a threat to the future of the community bank model.
I do believe in the community bank model and its future. Indeed, I believe there is a real place for the customization and flexibility that community banks can exercise to meet the needs of local communities and small business customers. Still, the disproportionate cost of regulatory compliance for smaller institutions is real. Financial supervisors must be vigilant in efforts to maintain financial system stability and ensure that consumers are able to understand their financial product choices, no matter where they choose to bank. However, as we and other agencies craft regulations to implement the Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank Act) and adjust supervisory practices to meet these priorities, I think we must avoid a one-size-fits-all approach to supervision.
To this end, the Federal Reserve last year formed a subcommittee of the Board to oversee the supervision of community and small regional banking organizations. I chair the subcommittee and am joined in this effort by Governor Sarah Bloom Raskin, who complements my community banking background with her experience as a banking lawyer and a state bank supervisor. In addition, in 2010 the Federal Reserve formed the Community Depository Institution Advisory Council (CDIAC) with membership drawn from smaller banks, thrifts, and credit unions. Council members meet with the Federal Reserve Board twice a year to share their perspectives on the lending environment, regulatory issues, and the economy.1
Today, I plan to discuss with you some ideas and initiatives that are on our subcommittee agenda as we strive for balance between our supervisory responsibilities and the effect supervisory practices have on the cost of compliance and credit availability.
Before I begin, I should note that the views I will share today are my own and do not necessarily represent the views of my colleagues on the Federal Reserve Board. But I think it is also safe to say that my concern and commitment to get this right is shared by all of my colleagues on the Board of Governors and by the staff throughout the Federal Reserve System who do the really hard work every day.
Clarifying Supervisory Expectations
One of the most daunting aspects of any comprehensive financial regulatory reform legislation such as the Dodd-Frank Act is the sheer volume of new regulatory proposals and final regulations. I still remember the experience as a banker of reading through hundreds of pages of dense language, paying close attention to the footnotes, trying to determine whether a regulation even applied to my bank and, if it did, what was expected of us. That was before we even got around to figuring out how we were going to meet the requirements, let alone what compliance was going to cost. And now, as a regulator, I still read every rule, guidance, and proposal with the knowledge that the effort expended to understand new regulatory requirements is itself an additional burden.
So, in response to a suggestion that was made by one of our CDIAC members, we are now working to include, at the beginning of each regulatory proposal, final rule, or regulatory guidance, a statement outlining which banks are affected. In particular, when issuing supervisory letters, we try to state specifically if and how new guidance will apply to community banks. This way, banks won't waste resources on requirements that don't apply to them. Sometimes, this statement is relatively simple: for example, many provisions of the Dodd–Frank Act by statute apply only to the largest banks.
In cases where the rules apply to all banks, but expectations vary by bank size or the degree to which banks engage in specific activities, it may be more helpful to make distinctions throughout the regulation or guidance. For example, when we proposed interagency guidance covering incentive compensation, we tried to note under each provision the simplified expectations for community banks that did not make extensive use of incentive compensation.
Also helpful in clarifying supervisory expectations is the use of examples. Interagency guidance issued in 2009 covering workouts of commercial real estate (CRE) loans contained several examples of loan restructurings and how they should be classified.2 I think the examples helped bankers and examiners alike understand policymakers' expectations for the regulatory treatment of loan workouts. Still, even with the examples, we heard reports that the guidance was not being implemented consistently in the field. To further ensure consistency, we conducted extensive examiner training. After the training was completed, we conducted a review of hundreds of loan files from recent examinations. Our file review indicated that, with very few exceptions, the loans were classified according to the guidance.
I believe bankers come to fully understand supervisory expectations over time, as their experience grows. But when something radically changes the supervisory landscape such as new regulations, drastic changes in the economic environment, or reassignment of regulatory authority, more direct outreach might be needed.
For many years, the Reserve Banks have maintained local training and outreach programs for banks in their districts. For instance, the popular "Ask the Fed" calls conducted by the Federal Reserve Bank of St. Louis provide bankers with an opportunity to hear Federal Reserve staff discuss recent policy initiatives and to be updated on the most common problems that examiners see in the field. The calls also provide an informal setting for bankers to ask questions on issues of concern. In addition, consumer compliance webinars have provided a forum for the Federal Reserve to discuss emerging consumer compliance issues and to answer questions from webinar participants directly.3 These initiatives have been so successful that a planning group is studying options for leveraging these efforts in a coordinated, national initiative to improve communication of relevant information to community banks supervised by the Federal Reserve.
Despite all efforts to communicate supervisory expectations, there are inevitably disagreements between banks and their examiners about examination findings. The Federal Reserve Ombudsman actively works with banks that have concerns about their examinations or that wish to appeal an examination finding. The Ombudsman works independently from the bank supervision chain of command and has broad authority to mediate complaints, including the authority to refer matters to committees of the Board. The Ombudsman's office also has a follow-up program in place to protect the banks that contact it from retaliation. And the Ombudsman tracks the nature and source of complaints in order to identify potential systemic problems.4
Balancing Safety and Soundness with Credit Availability
Since 2008, many banks have seen their assets' quality criticized and their ratings downgraded. This is not surprising given the severity of the economic downturn and the effect it had on the quality of bank assets. However, some bankers complain that examiners also tightened their approach. Indeed, we repeatedly hear that fear of examiner criticism is one of the reasons banks hesitate to lend to small businesses. We take these concerns seriously. In response, we are actively communicating with examiners and have stepped up examiner training to ensure that supervision in the field is consistent with policy. At the same time, I think it is important to continually reevaluate our supervisory policies and procedures to ensure that the policies themselves are not unnecessarily constricting credit availability. Here are a few such policies that we are currently evaluating at the Federal Reserve.
First, is the interagency supervisory guidelines issued in 2006 that required additional risk monitoring of CRE lending whenever loans reached certain thresholds--100 percent of capital for loans secured by construction and land acquisition activities and 300 percent of capital for loans secured by total non-owner occupied commercial real estate, including construction and land acquisition activities. 5 Although these thresholds were never intended to be hard caps, we hear from banks that they are now widely regarded as such. Many bankers have told me that they manage their loan portfolios to stay below these thresholds and forego growth in these loan categories, even when promising, creditworthy lending prospects are available. These reports were so widespread that the Government Accountability Office (GAO) conducted a study on supervision of CRE lending and concluded that the guidance warranted clarification.6 Federal Reserve staff is working with the other agencies on ways to clarify the expectations of the interagency CRE guidance and reduce any unintended restrictions on sound lending.
Given the high level of losses experienced on construction and land development projects over the past several years and the number of failed banks with high concentrations of such lending, it seems prudent for banks to keep construction loan exposures at, or even well below, 100 percent of capital. However, losses on loans secured by existing non-owner occupied commercial real estate have been much more modest, suggesting that banks' efforts to maintain concentration ratios below 300 percent of capital may constrain lending for some creditworthy borrowers.
Moreover, to generate loan volume without increasing real estate lending, many banks are now targeting growth in commercial and industrial (C&I) loans, a type of lending with which they may have less expertise. In fact, banks have generally experienced higher loss rates on C&I loans than on commercial real estate secured loans (excluding construction loan losses), even through the crisis. So this portfolio shift has the potential to increase rather than decrease expected losses.
For my part, I believe we should retain and perhaps strengthen the 100 percent of capital guideline on construction lending and consider dropping or modifying the 300 percent guideline for non-owner occupied CRE loans. Analysis undertaken by Federal Reserve staff suggests that retaining only the 100 percent guideline on construction loans would still be adequate to identify the banks at greatest risk. And bankers would still have to manage their loan portfolios appropriately and consider the risks of concentrations. But this change would help to eliminate any perception of a cap on what, in many cases, could be prudent, secured lending.
One factor that contributes to regulatory uncertainty is the intersection of Generally Accepted Accounting Principles (GAAP) and regulatory requirements. Nowhere is this more evident than in the accounting for the allowance for loan and lease losses (ALLL). The current accounting standard requires provisions only to cover losses that have already been incurred. Conflicting views of the range of likely losses sometimes leads to a perception that the regulatory evaluation of the adequacy of ALLL levels involves something of a "black box." To further complicate things, the accounting standard generally requires estimation, using statistical analysis, of a bank's unique past loss patterns, but most community banks have neither the rich data nor the capability to perform such analysis. Federal Reserve staff is currently investigating whether there is any way to use available supervisory data to publish loss rate ranges that could be used as a starting point for any bank to calculate allowance amounts in a way that is simple to understand and not inconsistent with GAAP. Even if development of such a tool does not turn out to be feasible, staff is still working to amend its approach to clarify expectations, improve transparency, and heighten consistency.
Along the same lines, I hear enough feedback about troubled debt restructurings (TDRs) to lead me to believe that it could be helpful to issue guidance clarifying the regulatory and accounting treatment of TDRs, non-performing assets, and classified loans. Even though these designations have different definitions, time frames, and regulatory consequences, they often seem to be used interchangeably. Reportedly, some bankers are reluctant to offer modifications that would help struggling borrowers and enhance the potential for ultimate repayment because they are concerned that the loan would be classified as a TDR and remain classified even after performance under the modified terms is demonstrated.
Taking this a step further, now might be a good time to review the definitions and usages of asset classifications. Right now, the definition of a substandard loan encompasses a broad range of assets, many of which will never sustain a loss. Yet our classification and bank rating system gives the same weight to all loans in the same classification bucket, regardless of loss potential. As a result, when a bank starts to approach a level of classification that it believes would cause examiners to downgrade its CAMELS7 rating, the bank stops making loans that it believes will be classified even if there is no concern about actual losses from the loans.
There was an interagency proposal issued in 20058 that would have established a two-step process for classifying commercial loans. The first step would be to evaluate the loan (the probability of default) and the second step involved estimating the potential loss (the loss given default). Many larger banks already use this approach in their own internal loan classification systems. On the other hand, many commenters at the time, including those from both small and large banks, objected to the proposal for a variety of reasons and the agencies decided to table it for the time being. Still, as I went back and reviewed the proposal, it struck me that some form of such guidance could be helpful in responding to the complaint that I hear over and over that loan classifications often ignore the loss protection provided by guarantors or collateral. And I think it might allow more accurate measurement of credit risk in all loan portfolios, not just commercial loans. So we are continuing to study the idea and the effect it might have on asset quality ratings if it was applied to different portfolios.
I also believe that when conditions warrant a change in CAMELS ratings, the update should take place as quickly as possible. It is always important to promptly identify problem banks and make use of all supervisory tools to foster their recovery. But I believe that as a bank stabilizes and demonstrates improvement, we should ensure that examiners move just as promptly to assign ratings that reflect the improved financial and managerial condition of the bank and free it from restrictions that could delay the bank's return to prudent lending activity.
Encouraging Creative Supervisory Approaches to Emerging Problems
So far, I have talked mainly about making adjustments to conventional supervisory approaches. In some cases, however, I think it might make sense to challenge some of our traditional thinking.
The first area that I think could benefit from a fresh approach is our requirement for the active marketing for sale of properties acquired in foreclosure, often called real estate owned, or REO, properties. While existing statutes and regulations do not prohibit financial institutions from renting REO properties, supervisory guidance encourages sales as the primary disposition tool. The problem with this requirement in the current environment is that having banks, servicers, Fannie Mae, Freddie Mac, and the Federal Housing Administration all following the same approach with thousands of properties on the market at the same time may actually be exacerbating the slump in housing prices.
Our research suggests that if lenders were permitted in some cases to rent residential REO property rather than sell it at fire-sale prices, it could better balance rental and owner-occupied markets and thus help housing prices stabilize sooner. Banks would have the opportunity to offset carrying costs and potentially increase their ultimate recovery. And the net result of removing some properties from the distress sale inventory could ultimately lead to higher recoveries for all holders of REO. Given current market conditions, banks could still divest property within statutory time frames, but with better results for themselves, surrounding property owners, and the economy. We are in the process now of exploring ways to clarify our guidance regarding rental of residential REO properties.
Let me turn now to a supervisory innovation that, I believe, will prove quite valuable in ensuring the stability of our financial system: stress testing. At the height of the crisis, the Federal Reserve, working with other banking agencies, conducted stress tests on the largest financial institutions and published the results. Banks that did not appear to have sufficient capital to withstand adverse conditions were required to raise additional capital. Last year, we reviewed the largest banks' capital plans in light of stress test results and used the information to guide approvals of planned capital distributions. The Dodd-Frank Act requires annual supervisory stress tests for institutions with assets of at least $50 billion and internal stress tests for institutions with assets of $10 billion or more. There are no current bank-wide stress testing requirements for banks with assets less than $10 billion. The only expectations for smaller banks are those contained in existing guidance, such as for interest rate risk or for commercial real estate concentrations.
As strongly as I believe in the concept of stress testing for capital adequacy, I don't believe that it needs to become a complicated, expensive, and burdensome process for smaller institutions with traditional business models. In many cases, smaller institutions are already incorporating the impact of adverse outcomes or stressful events into their existing risk management and business decision-making process.
The capital stress testing framework that we have developed is a dynamic way of looking at potential threats to capital in the context of the company's ability to replace capital through its earnings power, the current level of capital, and plans for capital distribution. For traditional community banks, the primary threat to capital is the risk of loan losses. We now have data on losses by loan category for two periods of severe financial stress, the recent financial crisis and the savings and loan crisis.
One way to test whether community banks have enough capital to withstand stress would be to apply the level of losses generally experienced in a stressed environment to bank loan portfolios to estimate potential losses. In such an analysis, a bank with concentrations in a lending category that could experience high losses during times of economic stress, such as construction lending or credit card loans, would have a high potential loss rate applied to those portfolios and thus require more capital than a bank with a more conservative portfolio. The rough estimates based on average losses by loan category could be further refined by looking at the individual bank loan classifications. We could similarly study the impact of stressful conditions on pre-provision net revenue, but I don't believe these fluctuations will have nearly as strong an impact on stressed capital as loan losses would have. And this analysis could be accomplished in smaller banks using existing call report data and supervisory asset classifications.
There are two points here. First, stress testing does not necessarily have to involve additional burden on the banks. Second, and perhaps most importantly, I hope that bankers do not feel that they have to spend scarce resources trying to conform to stress testing expectations that apply only to larger institutions.
Regulatory Burden and Mortgage Lending
I've focused most of my remarks today on policy ideas the Federal Reserve has the authority to pursue in order to reduce regulatory burden. One area that imposes especially heavy costs of regulatory burden on community banks, but for which the Federal Reserve no longer has rule-writing authority, is mortgage lending.
Community banks have long been a source of funding for mortgages that didn't fit the underwriting criteria to qualify for purchase by the government-sponsored enterprises (GSEs). These loans typically are held on balance sheet rather than securitized. Consequently, community banks retain 100 percent of the credit risk of these loans and have no incentive to make loans without regard to the consumer's ability to pay. But because community banks hold the loans on balance sheet, they may charge higher rates than those for prime GSE loans or include balloon payments to account for the liquidity and interest rate risk of holding the loans in portfolio. Further, community banks with small portfolios do not realize economies-of-scale so their costs are higher. And keeping up with regulatory change can be difficult, especially for banks that rely on purchased software and for loan officers who do not specialize in mortgage lending, but rather make the occasional mortgage loan to satisfy a customer need.
Traditional community lending is totally different than sub-prime lending. But it is difficult to make that distinction if regulatory requirements use the interest rate on the loan as a proxy to identify sub-prime loans. No one can argue with the need for stronger regulation to prevent the lending abuses that led to the current foreclosure crisis. However, I think it would also be unfortunate if the laws and regulations put in place to require other lenders to adopt the same responsible practices long used by community banks are so complicated and expensive that they have the unintended effect of forcing some community banks to leave the market.
It is difficult to think about crafting a regulatory regime that is simplified for smaller lenders that retain 100 percent of the risk they take, but I think it is important to try. The Federal Reserve had to think through a lot of these issues as we drafted proposals for escrow requirements and the definition of qualified mortgages. The Dodd-Frank Act made a start by crafting exemptions for banks in rural or underserved areas, but I think broader exemptions could be warranted for the thousands of smaller banks that make loans in small metropolitan areas or suburban areas. The Dodd-Frank Act transferred rule-writing responsibility for most mortgage-related statutes from the Federal Reserve to the Consumer Financial Protection Bureau (CFPB). So the CFPB will be finalizing the rules for which we issued initial proposals. The CFPB has already established an office for outreach to community banks and I have discussed this issue with them. I urge you to continue to explain how the burden of regulation, including regulations covering mortgage lending, impacts the ability of community banks to meet the credit needs of their communities.
The bottom line is that doing the necessary work of protecting our financial system and its customers comes with a cost. But that doesn't mean that we shouldn't continue to try to at least limit the burden. I hope you have noticed the number of times I have referred to reports from bankers or concerns expressed by borrowers. Even though I have experience living with regulatory burden, I still need regular feedback from bankers and consumers of banking services to understand the impact of regulation--both its effectiveness in achieving intended results and in limiting any unintended consequences. The Federal Reserve makes every effort to identify opportunities to reduce burden or improve credit availability. To assist us in this effort, I hope you will continue to point out such opportunities as you see them.