This issue of Banking Perspectives includes two articles on an important but generally overlooked subject: the interrelationship between bank liquidity regulation and the effectuation of monetary policy. The potential for liquidity regulation to reduce the efficacy of the Federal Reserve’s traditional approach to monetary policy is one of many underexamined costs of post-crisis regulation.
These articles also illustrate implicitly – and Paul Tucker, former deputy governor at the Bank of England, describes explicitly – another, broader consequence of post-crisis regulation: a significant growth in the depth and breadth of the power of central banks over economic activity. As Tucker states, “Since the financial crisis, the world’s central banks have gained much more power. ... More or less everywhere, central banks are now unambiguously part of the regulatory state as well as the fiscal state. Alongside activist judges, central bankers have, indeed, become the poster boys and girls of unelected power.”
Attempting to connect a few more dots, I focus in this column on the potential ramifications if the Federal Reserve were to adopt, without amendment, its recent proposal to integrate stress testing into existing minimum capital requirements while also proceeding to implement the so-called Basel IV accord. Together, these two actions would complete a quiet but revolutionary change in U.S. banking: for the largest banks that make up a majority of the industry’s assets, banks would no longer have any meaningful role in assessing their risk for capital purposes. Rather, the Federal Reserve, acting directly or indirectly via the Basel Committee, would model that risk for them. And these government-devised capital measures would drive credit allocation. The implications would be profound.
First, though, let’s start with how we get there. A key goal of the recent Federal Reserve’s stress capital buffer proposal is to establish a single capital standard as the binding constraint for all large banks. That proposed standard would require such banks to hold the following:
- A ratio of 4.5% equity capital to risk-weighted assets (measured under the Basel “standardized approach”); plus
- The higher of 2.5% equity capital to risk-weighted assets or a so-called stress capital buffer, reflecting the peak to trough losses that the bank would suffer under the Federal Reserve’s annual Comprehensive Capital Analysis and Review (CCAR) stress test; plus
- A global, systemically important bank (GSIB) surcharge, if applicable; plus
- A countercyclical capital buffer, if applicable; plus
- Four quarters of planned dividend payments.
While other capital requirements might remain in place, it is clearly expected (indeed, intended) that this new, combined requirement would be the binding constraint for all banks subject to it. The commendable goal of the proposal is to simplify capital planning for banks and pave the way for repeal of some of the other numerous current standards. But consider how that requirement would be built.
Basel standardized balance sheet. Before stress was applied, the assets of the bank would be risk-weighted to establish a starting capital ratio. Under Basel IV, those risk weights would be largely determined using models developed at the Basel Committee, known as standardized approaches. In particular, Basel IV includes a single Basel-devised model to measure all credit risk and a second Basel-devised model to measure all market risk.
Federal Reserve-devised stress scenarios and models. At that point, two standardized stress scenarios would be constructed by the Federal Reserve. It would then use its own models to determine how much revenue each bank would earn, and how much losses it would suffer, under the Fed-designed scenarios. (In reality, only one of the scenarios, the severely adverse, is likely to bind.) The Federal Reserve would then calculate which is higher as a percentage of risk-weighted assets: the stress losses as a percentage of capital, calculated by its own model under its own scenario, or 2.5%, a minimum capital buffer selected by the Basel Committee.
Basel and Federal Reserve-devised GSIB surcharge. The Federal Reserve determines the GSIB surcharge using a model designed by the Basel Committee and later amended by the Federal Reserve to be more stringent.
Federal Reserve-devised countercyclical buffer. There is no known model for determining the countercyclical capital buffer, and the Federal Reserve has not announced a clear standard for determining how it would be calibrated, so it will be the product largely if not solely of Federal Reserve discretion.
The remarkable result of this process is that at no point is a bank’s view of the risk of a loan or any other asset relevant to the capital it must hold against that asset.
The potential for liquidity regulation to reduce the efficacy of the Federal Reserve’s traditional approach to monetary policy is one of many underexamined costs of post-crisis regulation.
One could argue that this is nothing new. Basel I was a standardized approach to credit risk and governed for two decades, including up to and during the global financial crisis. It was only after the crisis that bank models were incorporated into risk-based capital, through Basel 2.5 and Basel III. (Of course, this history makes the return to standardized models in Basel IV rather inexplicable. It is an unusual variant on the so-called Pottery Barn Rule: “We broke it, we bought it – and plan to glue back together all the broken pieces, and then resell it.”) Furthermore, standardized measures are already the binding constraint for many large banks, as the so-called Collins Amendment to the Dodd-Frank Act imposes a minimum capital requirement based on the standardized approach, even for those banks running the advanced approaches under Basel 2.5 and III, and the banking agencies have chosen to impose similar standardized floors to various capital buffer requirements.
However, the credit allocation effects now would be more significant than they were pre-crisis, with Basel I, because the Federal Reserve has calibrated its new requirements to produce extremely high capital requirements by any absolute or historical standard. Research from The Clearing House demonstrates that the Federal Reserve’s CCAR stress tests are already causing banks to disfavor certain asset classes.
There is good reason to believe this would end poorly for the U.S. economy. Governmental attempts at direct or indirect credit allocation have a dismal history. Economic growth would suffer, as a diversity in risk tolerance and judgment produce greater opportunities for businesses and individuals to obtain bank credit. Furthermore, because standardized risk measures and the Federal Reserve’s loss-forecasting models are necessarily crude and one-size-fits-all, relying on less data than the banks’ own models, much of the capital allocation they drive is likely to be misallocation. Lastly, systemic risk would increase as large banks were forced to concentrate in asset classes favored by the governmental models. (There is also the converse risk that non-banks would concentrate in asset classes disfavored by the governmental models; because those actors operate outside the purview of the Federal Reserve and do not qualify for lender-of-last-resort support, a collapse in prices for those assets would constitute its own systemic risk.)
Collectively, then, these proposals fail one of Paul Tucker’s proposed preconditions for the exercise of power by an independent agency: namely, that unelected regulators not have power to make “big choices on values and distributional issues.” The good news is that a few simple changes could allow the Federal Reserve to retain its proposed construct while preserving competition in risk management, innovation in risk analysis, and diversification of risk.
First, banks could be permitted to use their own loan forecasting models to determine stress losses for purposes of the stress capital buffer, subject to backtesting and Fed review of those models. Indeed, this approach is currently used by the Bank of England and the European Central Bank. It is worth noting how strange the Federal Reserve’s mono-model approach appears in light of broader developments in finance and technology. As I learned at a recent Federal Reserve Bank conference, artificial intelligence and machine learning are widely expected to revolutionize areas where large amounts of data are available to solve complex problems. Risk management in banking presents a classic case for this solution. Doing so would also address the concerns expressed by some policy makers about the systemic risk associated with the use of similar models by multiple financial firms. (A concern particularly applicable, I would note, with the mono-model approach the Fed has employed in stress testing.)
The good news is that a few simple changes could allow the Federal Reserve to retain its proposed construct while preserving competition in risk management, innovation in risk analysis, and diversification of risk.
Second, rather than relying on a single scenario, the Federal Reserve could incorporate multiple scenarios into stress testing – for example, by adding individually tailored, bank-developed scenarios to its own standardized scenarios. (It could also greatly improve its own scenarios by subjecting them to a notice and comment process.) Using multiple scenarios would eliminate the soundness risk that a given bank’s balance sheet would be structured to avoid a single, highly implausible stress scenario designed by the Federal Reserve – at considerable risk that such a balance sheet would be vulnerable to altogether different but more likely scenarios. It would also eliminate the systemic risk that all large banks would begin to have more similar balance sheets.
Third, outside the realm of stress testing, banks could be permitted to use their own models rather than standardized ones to establish their risk weights. The Federal Reserve allows the banks to model market risk (under Fed oversight) and could do the same for credit risk. The case for market risk appears even stronger, given the ongoing difficulties that the Basel Committee has encountered in trying to capture those highly complex risks through standardized measurement tools. Certainly, the current observation period for the Basel Committee’s recent (and still not finalized) changes to market risk capital measures can be used to determine how sophisticated and dynamic the Basel model is. Basel IV is not a law or treaty, and the Federal Reserve is free to depart from it; alternatively, there are ways to comply with Basel IV without incorporating it into the Federal Reserve’s new capital construct, which, like stress testing in general, has no Basel root.
Fourth, at least so long as it continues to insist that its stress tests be fundamentally countercyclical in design – tougher in good times, less so in bad – the Federal Reserve should publicly commit to a countercyclical capital buffer of zero. As a countercyclical tool, the stress capital buffer (SCB) and countercyclical buffer are duplicative, and if one is to be retained, the SCB is a much better tool. It is appropriately sensitive to banks that have more cyclically sensitive portfolios, while the Basel countercyclical buffer would require all banks to hold more capital irrespective of the composition of their portfolio. Such a public commitment would also serve the goal of providing predictability to bank investors by removing the potential for an any time/any amount/any reason capital charge.
Each of the proposed reforms has strong merits on its own. But the need for their adoption grows significantly when one considers the path we currently are on, and the need for a reinjection of private sector judgment into the process for allocating credit in this country. At the very least, the current path seems worthy of debate, certainly at the congressional and agency level. Because many of the steps toward standardization have been taken individually, however, there has not been an occasion to consider the wisdom of their collective force. Now seems like a good time.
(Note: Indeed, one wonders whether the Basel Committee insists on calling its recent amendments the “finalization of Basel III” rather than Basel IV in order to elide the fact that a fundamental change in approach has occurred. While global regulators have consistently and loudly expressed concern about variations in risk weighting among banks, they have been less forthright in acknowledging that their proposed solution is for regulators to enforce consistency by doing much of that analysis themselves).