Introduction
In a recent Brookings paper, former Federal Reserve Gov. Daniel Tarullo discussed several potential policy changes to the stress testing regime. Each of the proposals would reduce the transparency of the bank capital regime and inject unwarranted and counterproductive uncertainty into the bank capital process. In this post, we discuss the issue of transparency in the stress tests and comment on the policy changes reviewed in the paper.
Transparent Regulation is a Hallmark of Good Government and Good Capital Policy
In his paper, Daniel Tarullo takes a dim view of transparency in the stress testing process and equates a call for transparency with efforts to “lobby for less dynamism in stress testing.” This claim only seeks to divert attention away from the simple fact that reduced transparency in government regulation is harmful to everyone. It is often said that “sunlight is the best disinfectant” and that a government that operates in the light of day is subjected to vigorous oversight that limits the exercise of arbitrary authority. The notion that the stress tests and financial stability stand to benefit from less and not more transparency stands in clear opposition to decades of experience and common standards of good governance.
There is an unstated assumption embedded in the call for less transparency that if banks respond to the design of the stress tests, then they are “cheating.” This viewpoint portrays a deep, serious, and dangerous misunderstanding of how and why banks respond to capital regulation. Bank capital requirements are regulatory pronouncements about risk. Bank assets that are riskier require greater capital to withstand potential losses. Banks respond to capital requirements by trading off the economic benefit of holding the asset with the risk incurred by holding it. If an asset’s risk – either real or as perceived by regulators through capital requirements- increases, then banks have an incentive to reduce their holdings. When a bank engages in this practice it is typically considered good risk management rather than cheating.
Moreover, this is exactly the dynamic that is created by existing risk-based capital requirements, or so-called “point-in-time” requirements. If a regulator’s assessment of housing risk results in an increase in housing-related regulatory capital requirements, as has happened in the context of the recent Basel 3 Endgame capital proposal, banks should re-evaluate the tradeoff between the risk and return of holding housing assets. This is how banks manage their risk while allocating bank funding to its most productive use. Indeed, this very fact coupled with the fact that Basel 3 Endgame’s proposed changes are fully transparent has generated quite a lot of useful public comment that will positively impact the final regulation. Much the same would be achieved if the stress tests were subject to the same degree of public transparency.
Oddly, there appears to be a belief that adapting to capital regulation in the context of “point-in-time” requirements is fine, but doing so in the context of dynamic capital requirements via stress tests is not. This is incorrect. Banks should respond to capital requirements and risk whether they be static or dynamic. Imagine that a particular coastal area becomes more prone to flooding because of recent beach erosion but insurance companies are not allowed to raise flood insurance premiums and home builders are required to continue building in the area. This would be crazy. Obviously, as conditions change, affected parties judiciously managing their risk should adapt. The notion that this basic truth is reversed in the case of bank capital stress tests is without any logical basis.
More Opacity in Stress Testing Will Not Improve the Stress Tests
Several of the policy options discussed in the paper would increase the opacity of the stress testing regime. Adding more opacity to the stress tests will not improve their ability to measure capital adequacy.
Consider the policy option of including “second round effects” in the stress tests. Second round effects encompass a wide array of highly speculative economic impacts that could, hypothetically, be triggered by significant losses in the banking sector. One such example is the idea that significant bank losses would precipitate a wave of asset “fire sales” by banks that would then further depress asset prices and further erode bank capital.
Such impacts are an interesting theoretical possibility but are far – far as in lightyears away– from being settled economic science. Economic modeling is quite unlike Newtonian physics. There are no clear and predictable relationships that are observed regularly with high precision that are backed by clear and (nearly) incontrovertible theories. Our ability to predict and model such “second round” impacts is little more than slightly informed guesswork not rooted in any definitive empirical or theoretical body of research. Rather, the inclusion of any such effects would be mere story telling by the unchecked dictate of Federal Reserve staff without any independent oversight on their modeling. Frankly, economists have quite enough trouble accurately estimating the immediate and direct (first round) impact of economic stress on bank capital levels. The idea that we should further complicate that exercise by including fanciful tales about what may or may not happen because of second or third round effects is a dangerous idea that would threaten the validity and empirical rigor of the stress tests.
What would be achieved by adding such speculative impacts to the stress tests? More regulatory uncertainty and unchecked discretion thereby allowing regulators to unilaterally determine capital levels without having to explain or justify the rationale for their determinations. If capital is perceived as being too low, just go ahead and fire up the staff’s “fire sale scenario machine” and presto-change-o, new and higher bank capital!
The paper discusses similar policy initiatives along the lines of allowing Federal Reserve staff to make frequent and undisclosed changes to the underlying statistical loss models that drive the stress testing process. Imagine such an approach being applied in any other regulatory context. For example, imagine that the government could simply decide to change the rule by which personal or corporate income taxes are determined without revealing these changes to the public. Or imagine that automobile safety standards were changed by the government without informing auto manufacturers of these changes so they could adapt to the new standards? These examples make clear that allowing for undisclosed and unchecked changes to the bank capital regime would only create chaos and frustrate banks’ ability to understand the regulatory landscape, which would impair their ability to channel resources to their highest and best economic use.
Decoupling the Stress Tests from Capital Requirements: Unhelpful Policy Ping Pong
Another policy option discussed in the paper, suggests that the link between the stress tests and capital requirements, that was implemented by the Federal Reserve in 2020 through the stress capital buffer (SCB) rule, should be abandoned. This policy proposal is highly problematic for two reasons.
First, such a change would result in the sort of policy ping pong that typifies bad policymaking. Prior to the SCB rulemaking, the stress tests did not have a transparent and well-defined impact on bank capital requirements. At the time, then Governor Tarullo touted the SCB as a means of simplifying and streamlining the bank capital regime while clearly linking capital requirements to the results of the stress tests. Undoing all of that now would simply create more uncertainty in the regulatory process. Surely, ripping the stress tests out of the bank capital framework now would only invite some future policymaker to graft them back on. An ongoing and torturous cycle of – this year they’re in and next year they’re out until they are back in again would wreak havoc on the bank regulatory regime without offering any benefit. Year-to-year gyrations in capital rules generated by needless policy vacillation wouldn’t do anything but create a confused and ineffective regulatory system.
Second, de-coupling the stress tests from capital requirements would only invite the sort of secretive “shadow regulation” that endangers the transparent and equitable administration of government. If the stress tests are not part of the formal capital regime that are governed by strict transparency standards, the stress tests could be used as a supervisory cudgel to achieve capital levels that could not otherwise be achieved through consistent and transparent capital regulation. This possibility would be even more likely if the stress tests could be changed and adjusted as regulators see fit without any public oversight. Open and transparent regulation is important to ensure that regulation is always applied consistently regardless of the prevailing winds of the moment.
Conclusion
Transparency in all forms of regulation and government administration is a long held and widely accepted standard of good governance. There is no compelling argument to abandon this transparency principle in the case of bank capital requirements and the stress tests. Banks must understand the capital rules to which they are subject so that they can appropriately allocate funding to its highest and best use for the economy. A regulatory capital regime that purposefully obfuscates the rules will only reduce the ability of banks to effectively allocate capital while creating counterproductive and dangerous policy uncertainty.