Rules vs discretion in macroprudential policies

Mario Quagliariello, Massimo Libertucci, 24 February 2010

a

A

Rules vs. discretion…

In monetary policy, discretion is essential to offset output fluctuations in Keynesian frameworks. Conversely, monetarists propose a tight, fixed rule to ensure price stability. The time-(in)consistency literature, launched bu Kydland and Prescott (1977), shows that discretion-based solutions would be the first-best in terms of agents’ utility, but they are not time-consistent. In fact, strategic responses of rational, utility-maximising agents lead to an ex-post sub-optimal arrangement; rules ensure that – ex-post – at least a second-best is achieved.

More generally, policy tools that are based on rules leave less room for policy error. Moreover, once in place, they act as an effective pre-commitment device. A rule-based approach, however, would require a very high degree of confidence that the predefined variables would always correctly perform as intended, without noisy signals. This is difficult to achieve for inflation targeting, much more so for identifying financial instability. Indeed, the adoption of a purely rule-based framework focusing on a macroeconomic indicator (as in either a simple rule such as Friedman’s or an inflation target framework) has faced several drawbacks including, for instance, the inability to face unexpected structural changes. A discretionary framework successfully addresses this issue, by allowing policymakers to actively learn from observing the interaction of relevant stakeholders.

Flexibility and adaptability of discretion do not come without costs. They entail limited predictability of decisions as well as a tendency towards forbearance, with an incentive for policymakers to postpone backfiring decisions, particularly if they are subject to some form of political pressure – including the pressure of public opinion.

Given the trade-off between ex-ante efficiency of discretion and ex-post efficiency of rules, hybrid regimes of constrained discretion have been extensively adopted in monetary policy. In particular, to mitigate the possible pitfalls generated by discretion, central banks may opt for a clearly stated, transparent and accountable decision making process (e.g., public “reaction function”).

…and macroprudential policies

The adoption of macroprudential policies has been advocated by the G20 Leaders and the Financial Stability Board. In very general terms, the new macroprudential approach envisages greater focus on the analysis of systemic risk and vulnerabilities that may emerge as a consequences of macro-factors (Borio 2009). In terms of regulation, the objective is to create a more disciplined and less procyclical financial system that better supports balanced, sustainable economic growth (FSB 2009).

The new macroprudential objectives should be pursued (also) through the introduction of countercyclical tools, such as capital buffers (GHOS 2010). The principle behind the functioning of the countercyclical buffers is straightforward: capital should be built up in good times, when risk is accumulated in banks’ portfolios, and depleted in bad times, when risk actually materialises. The main issue is therefore the identification of bad and good macroeconomic times and the definition of the policy response. This can be either left to discretion (i.e., to a purely case-by-case assessment by public authorities) or based on rules (i.e., linked to the dynamics of one or more predefined macroeconomic or financial variables). Similarly, the final goals of the policy – avoiding excessive credit imbalances or “just” smoothing the cycle – can be predetermined by rules or decided from time to time.

Under a discretion-based regime, authorities would require banks to build-up buffers in periods of economic exuberance. On the basis of a pre-announced set of variables, they should:

  • declare that a boom is in progress;
  • decide that the boom is a cause of excessive risk-taking;
  • impose capital savings for “rainy days”. In bad times, as discretionally assessed by the authorities themselves, banks would be allowed to release buffers.

This would reduce the risk of procyclicality, but may increase the risk of default at individual bank level. In fact, macroprudential capital buffers, targeting the state of health and risk conditions of the banking sector as a whole, would be independent of idiosyncratic solvency situations, that constitute the raison d'etre of microprudential capital requirement.

In practice, the authorities would be asked to continuously address the micro-level objective and the system-wide one. This is probably what makes the parallel with monetary policy not entirely satisfactory. In the case of macroprudential regulation and supervision the same policy instrument – bank capital – would be used to address both microeconomic purposes (reducing the probability of default of individual institutions) and macroeconomic goals (reducing system-wide risks and procyclicality), violating Tinbergen’s rule. Admittedly, micro- and macro-stability are compatible most of the time, but they may conflict in some circumstances.

Another drawback of discretion is that warnings and recommendations may have adverse effects, with the warnings possibly turning into self-fulfilling prophecies for example. This might imply significant political pressure on macroprudential authorities.

On the other hand, under rule-based frameworks any policy reaction would be left to pre-defined automatic mechanisms and triggers. This would avoid time-inconsistency; moreover, authorities would be forced to identify the (still subjective) optimal balance between the micro and macro objectives once for all. Unfortunately, the design of the set of rules may be extremely difficult, particularly for a brand new policy, which should be applied world-wide. Moreover, rules are subject to Goodhart’s law: the informational content of a targeted variable would be vanished once it becomes part of a (publicly known) rule.

So, can a combination of rules and discretion be a viable solution also for macroprudential supervision? The Bank of England (2009) proposes a constrained discretion macroprudential regime: this would be largely discretional to allow policymakers to adapt as they learn, but still systematic, transparent and accountable thanks to the predetermined constraints. These may assume, for instance, the form of pre-defined objectives (ideally, a numerical target), decision-making frameworks (detailing the analyses underpinning decision making), accountability measures (public report, parliamentary scrutiny, etc).

A new proposal

Looking at pros and cons of the two extreme options (pure discretion and hard rules), we agree with the idea that a constrained discretion can be a viable compromise. Our favourite setting is however slightly different, with greater emphasis on rules. A predefined rule would determine the policy reaction (e.g., in terms of request for building up / depletion of capital buffers) to changing economic conditions. The set of variables may be different across jurisdictions, but should be announced clearly and transparently ex-ante. The relevant authority would have the option to override this rule, allowing the banking system to increase and decrease the buffer as appropriate, in the light of a broader range of information on the state of health of the economy. However, this would represent an exception, to be used in rare circumstances and properly explained to market participants. In our view, a more extensive use of discretion may determine competition in laxity across jurisdictions and reduction in cross-country comparability. In sum, rules would be the everyday framework, while discretion would represent an extreme resort.

We also believe that such a system can work better if the macroprudential tools are put in the hands of a regional or global macroprudential authority – less prone to external pressures – rather than domestic ones. Certainly, the actual effectiveness of such a system will depend crucially on the independence, reputation and accountability of the macroprudential authority, as well as on its ability to enforce its recommendations. The final outcome will also be affected by the interaction between the micro- and the macro-prudential authorities, since they will handle the same tool. Should this institutional setting be considered too ambitious, a rigorous peer review process, as suggested also by the Financial Stability Board (2010), would ensure that national authorities either comply with their own pre-commitment or explain possible deviations to a group of peer authorities.

Authors' note: The opinions expressed are those of the authors and do not necessarily reflect those of the Bank of Italy.

References

Bank of England (2009), “The Role of Macroprudential Policy”, November.

Borio, Claudio (2009), “The Macroprudential Approach to Regulation and Supervision”, VoxEU.org, 14 April.

Financial Stability Board (2009), “Improving Financial Regulation - Report of the Financial Stability Board to G20 Leaders”, September.

Financial Stability Board (2010), “FSB Framework for Strengthening Adherence to International Standards”, January.

Financial Stability Forum (2009), “Addressing Procyclicality in the Financial System”, April.

Group of Central Bank Governors and Heads of Supervision (2010), “Group of Central Bank Governors and Heads of Supervision reinforces Basel Committee reform package," Press Release, 11 January.

Kydland, Finn E and Edward C Prescott (1977), “Rules Rather than Discretion: The Inconsistency of Optimal Plans,” Journal of Political Economy, 85(3): 473-492, June.

Topics: Global governance, Monetary policy
Tags: checks and balances, Macroprudential policies, monetary policy

Massimo Libertucci
Economist, Regulation and Supervisory Policies Department, Banca d’Italia
Mario Quagliariello
Principal Bank Sector Analyst at the European Banking Authority (EBA)