- Source: The challenge of taking macroprudential decisions: who will press which button(s)? Jaime Caruana of the BIS, September 24, 2010
3. How much macroprudential policy should be embedded in the regulation?
Or, to put the question differently, how far can macroprudential policy rely on rules embedded in regulation as opposed to discretion in supervisory decisions? “As much as possible, but no more”, I would say. Given uncertainty regarding low-probability events entailing high costs, it is difficult to press the button, particularly in good times. That said, an internationally agreed framework can help the authorities to exercise and to justify discretion. As an illustration, let me focus on measures to make prudential standards less procyclical (the time dimension of the macroprudential approach). Given the limited time available, I omit those that address common exposures and interconnections as sources of systemic risk (the cross-sectional dimension).
The Basel Committee on Banking Supervision seeks to embed in regulation the macroprudential principle of increasing capital buffers in good times that can be drawn down during periods of stress. Two capital buffers have been endorsed by the Governors and Heads of Supervision.
First, there is a proposal to conserve capital. 5 This conservation buffer above the minimum requirement is purely rule-based, as it is set as a fixed proportion of risk-weighted assets. The buffer can be run down during a period of stress, lessening pressure to restrict credit. But its primary objective is to ensure that banks that incur losses and thereby, with realistic accounting, approach the minimum do not pay out capital, which would further deplete their reserves. During the crisis, most of the banks continued to make distributions at the accustomed, blue-sky rate, paying dividends and bonuses and repurchasing shares. This buffer is best thought of as a microprudential tool with macroprudential implications, since it would leave the system more resilient as a downturn deepened.
Second, there is an additional countercyclical buffer. 6 This tool is much more based on discretion. Through similar restrictions on dividend payments, banks would be constrained to accumulate this buffer during periods of very rapid credit growth in order to mitigate the build-up of systemic risk. Supervisors would then release the buffer as strains materialised in order to absorb losses. The ratio of credit to GDP would serve as a common reference for the build-up phase – indicating when to push the button. But it would not establish a hard-and-fast rule. Authorities would be allowed to rely on a broader set of indicators, including, for instance, asset prices. And a set of principles would guide its release, based on indicators of emerging strains. Given the uncertainties involved, it has not proved possible to reach agreement on a simple rule. Inescapably, judgment remains critical in this case.
The operation of this countercyclical buffer requires a decision to act. Let me offer a perspective on the very real challenges involved by harking back to my last appearance here at the Chicago Fed and recalling what was on the Bank of Spain’s dashboard when we decided to require dynamic provisioning based on through-the-cycle parameters.
The context was that of a new currency union. The virtual euro had replaced the peseta at the beginning of 1999. Interest rates in Spain had converged to European levels well before then. There was a strong prospect that the macroeconomy would be more stable, supporting the servicing of higher debt. The interest rate sensitive sector of the economy was responding vigorously and real estate prices were recovering from lows reached in an earlier cycle. By 1999–2000, mortgage credit was growing at 15% per annum and house prices were rising at about 10% per annum. With inflation above the European average, and even with real growth at 5%, rapid credit growth amid keen competition in the financial sector was a cause for concern. 7 Because of the euro, the interest rate was not an available tool.
Macroprudential policy was the only possible option.
Contrary to the view that the authorities will inevitably identify a potential problem too late, the Bank of Spain identified the build-up of risks very early: the peak in property prices did not occur until 2007. Early identification is very good from the perspective of building buffers, but it poses a number of problems from the perspective of signalling and communication. When the authorities respond to risks that do not materialise for years, the inevitable controversy and criticism about the measures taken can undermine their effectiveness, which partly depends on acceptance of their rationale. Thus, timing the call is difficult, and from the communication perspective there are risks of making it either too early or too late. Dynamic provisioning did make banks stronger than would otherwise have been the case, and there is some suggestive evidence that it moderated the credit boom as well. Whether the Bank of Spain’s response was proportional to the challenge, or more should have been done, remains a matter of debate.
Regardless of the final outcome, this was not an easy call. Banks were not pleased to report reduced profits as they built up their statistical loan loss provisions. Provisioning that cost around 20% of operating profits and that was not required of international peers did not go down easily. In addition, there were technical challenges from the accounting profession, incorrectly concerned with artificial profit smoothing. It will never be an easy call: uncertainty will be high, and therefore explaining the trade-offs to set reasonable expectations will be key. An internationally agreed framework for the countercyclical buffer can only help the authorities. But the onus to push the button remains on them.