Progress and Confusion

9. Macroprudential Policy Regimes: Definition and Institutional Implications

Olivier Blanchard, Kenneth Rogoff, and Raghuram Rajan
Published Date:
April 2016
  • ShareShare
Show Summary Details
Paul Tucker

Policymakers and economists spent decades debating the design of monetary institutions before the wave of central bank independence measures in the 1990s. We now need to turn our eyes to a fresh institution-building challenge: macroprudential regimes. Even for skeptics, this is an important endeavor as, wisely or unwisely, these regimes are increasingly springing up around the world.

There is not yet agreement, however, on the meaning of “macroprudential.” Too often the term is used as a synonym for financial stability policy more generally; we don’t need two words for one thing. It is important that the IMF, the Bank for International Settlements and the Financial Stability Board coalesce around a common definition that can become embedded in usage.

A Definition of Macroprudential Policy

I define macroprudential policy to be a regime under which policymakers can dynamically adjust regulatory parameters to maintain a desired degree of resilience in the financial system.

Policy instruments—requirements on firms, funds, or structures set out in rules, regulations and international accords—are not completely static under such regimes. Nor are they time-contingent. They are to be state-contingent, varying as needed with threats to stability. That does not mean that they vary a lot. The better the design and calibration of the base regulatory regime, the less cause there will be to vary them. But where necessary, they can be varied to sustain the financial system’s resilience.

In terms of the objective—maintaining the resilience of the system—this is not, therefore, a regime for fine-tuning the credit cycle. That would be too ambitious. It is hard to know whether temporarily raising, say, capital requirements for banks would tighten or relax the supply of credit in the short run. Any macroprudential measure will reveal not only the action itself but also information about the authorities’ views on the state of the financial system. In contrast to monetary policy, where the data on the economy are in the public domain, a prudential policymaker has lots of private information about vulnerabilities in individual financial institutions and the linkages among those institutions. If the market is surprised that the policymaker is concerned enough to act, credit conditions might tighten sharply if market participants conclude, on the basis of the information newly available to them, that the actions taken are insufficient. If, by contrast, the market has been ahead of the authorities in spotting a lurking threat to stability and so is relieved that the policymaker is finally waking up, credit conditions might even ease. There are many scenarios in-between.1

Hence my proposed focus is less on trying actively to manage credit conditions and rather more on aiming to sustain a desired degree of resilience in the system. That has the merit of concentrating on the big issue in this field. The greatest social cost of a vicious credit cycle occurs if the eventual bursting of the bubble causes the implosion of the financial system, cutting off the supply of essential services. It is an ambitious enough goal, but a vital one, to get policymakers to prevent the ceiling collapsing in future. Big picture, the ceiling can be reinforced even if the authorities’ actions reveal disturbing information about the financial weather.

Filling Out the Concept

How should we think about dynamic adjustment—state-contingent adjustment—of regulatory parameters in order to sustain a desired degree of systemic resilience? Implicitly there are three things going on here, which need to become more explicit.

First of all, I am assuming that society has, or should have, a desired degree of systemic resilience. Does it want to avoid a big crisis once every million years, every thousand years, every one hundred fifty years? It seems clear that seventy years, roughly the gap from the crisis of the 1920s–1930s to the 2007–2009 crisis, is more than the public of either North America or Europe is prepared to bear. But I doubt whether society wants to rule out a crisis for the next five million years, since that would most likely entail banning huge swaths of financial activity.

Second, the macroprudential authority is assumed to have a picture (or model) of the exposures and interconnectedness of the financial system. It uses that to gauge the prospective effects of possible shocks—causing a sharp rise in defaults or trading losses—on the system’s residual resilience, and thus whether actual resilience falls short of desired resilience. Of course, that picture of the structure of the system will be flawed to a greater or lesser extent, and so has to be kept under review.

The third thing implicit in this conception of prudential regimes is the riskiness of the world. If you like, abstractly, what is the underlying stochastic process generating unexpected losses to financial institutions—the first-round losses that the system’s structure transmits across firms, funds and markets?

Big picture, we can think of the underlying risk process in the financial system as a whole as being at any time in one of three broad modes—normal, exuberant, or depressed. If that is right or helpful as a picture, then a very important policy question is whether or not to calibrate the base regulatory requirements designed to keep the system safe and sound—minimum capital requirements, minimum collateral requirements on derivatives transactions, and so forth—to exuberant states of the world. An argument against doing so is essentially ignorance and uncertainty. We do not know enough about the properties of the financial system to be confident about the effects on the supply of credit or of other financial services of calibrating the base regulatory requirements against the most vicious exuberant states of the world. This is a key moment in the argument. I am identifying a “prudent” approach to policy where selfconscious ignorance prompts policymakers to step back from calibrating the base regulatory regime to exuberant states of the world. That may be a mistake, but it is the choice made by international and national policymakers in the years following the 2007–2009 crisis.

If the regime is calibrated to a “normal” underlying risk-generation process, then we know that those regulatory requirements will be insufficient when the world moves into highly exuberant mode. In those circumstances, capital requirements or margin requirements or haircut requirements or whatever need to be changed in order to sustain the desired degree of resilience. Similarly, if the structure of the system changed in ways that made the propagation of first-round losses more contagious, the core regulatory parameters might need to be recalibrated, if only temporarily while deeper solutions were designed.

To be clear, this is not changing the goalposts. The goalposts stay fixed: the tolerance for crisis embedded in the desired degree of systemic resilience.

A Metaphor

A metaphor might help. In the spirit of Knut Wicksell, most monetary authorities think of themselves as moving around their policy interest rate in order to keep the actual short-term real rate (r) in line with estimates of the short-run equilibrium real rate (r*) as shocks occur to aggregate demand or supply. By analogy, macroprudential policy involves moving around, say, the actual capital requirement applied to intermediaries (K) in order to keep it in line with what is necessary to deliver the desired degree of resilience (K*).

The metaphor is of course imperfect in some respects. The suggestion is not that macroprudential policy be actively used, with frequent regulatory recalibrations. Unlike r*, K* is not entirely market determined but reflects the regime’s objective: how resilient the system should be. Nor is K the only instrument; leverage is not a uniquely useful measure of resilience. But I hope the metaphor helps to illuminate the thought of dynamic adjustment of regulatory instruments with a view to “neutralizing” the threats posed to stability by developments in the world.

Note that this conception of the purpose of macroprudential policy cuts through some of the current debates about its effectiveness. That work typically starts with questions about which among a range of instruments—whether tightening capital requirements or loan-to-value limits or other regulatory constraints—has the greatest and/or quickest effect on the rate of credit growth or asset-price appreciation. As already discussed, those questions are going to be incredibly hard to answer unless we can model the significance of the ex ante information asymmetries between the market and the authorities.

By contrast, the effectiveness of the conception of macroprudential policy outlined here is relatively straightforward. If, for a given balance sheet, core firms have to increase their tangible common equity by 10 percent, they will, broadly, be 10 percent more resilient against an unchanged risk environment. Even though the macroprudential intervention might reveal information about perceived threats to the system and so affect the price of risk in uncertain ways, the system itself is buttressed.

Instead, the hard part of the policymakers’ task is to judge the extent to which a regulatory lever must be tweaked to maintain the desired degree of system resilience. That underlines how important the choice of the base requirements is. The more demanding they are, the less likely it is that the macroprudential policymaker will need temporarily to increase them. But wherever they are set, circumstances will eventually arise where maintaining a static regime will bring on disaster.

The Design of Macroprudential Regimes

From that general conception of the purpose and operation of macroprudential regimes flow some precepts or pointers for their design.

The Scope of Macroprudential Policy Regimes

In the first place, although I have used banking as the core example, this is not, in fact, only about banking supervision. It is not just about maintaining the stability of the banking system in the sense of de jure banks. Regulatory arbitrage is endemic in finance. The industry is a shape-shifter.

If the authorities were to focus exclusively on maintaining a resilient banking sector, then systemic risks will turn up with a terrifying inevitability in some manifestation of shadow banking or somewhere else in the financial system. So the scope of the regime needs to be broad.

If, then, macroprudential policy cannot just be about banking supervision, it must also be about the parts of the financial system that typically fall under the jurisdiction of securities regulators. That requires either reform of the mandates and mind-sets of securities regulators or, alternatively, a reshaping of the regulatory architecture to give one authority general macroprudential jurisdiction over the financial system as a whole.

Rules versus Constrained Discretion

Further, state-contingent adjustment of regulatory parameters is plainly not about writing detailed rules that are put out for extensive consultation in the usual way (in the US jargon, “notice-and-comment” rule-making). By the time any planned temporary adjustment had gone through such a process, it might be too late. Macroprudential policy, as conceived here, involves the exercise of discretion, and so needs to be carefully constrained to serve society’s agreed-upon purposes.

Credible Commitment: An Independent Policymaker with Instrument but Not Goal Independence

Macroprudential regimes plainly face a problem of credible commitment. Dynamically adjusting the regulatory parameters to strengthen the system during a boom is likely to be unpopular. If the tools were in the hands of politicians, they might well decline to deploy them because, despite their better instincts, they would know that the credit or asset-price boom was making the electing public feel good, and so more likely to reelect them.

There might also be a strict time-inconsistency problem in the narrow sense of a social planner with unchanged preferences departing from a long-run optimal plan because it can improve on the plan in a single period. This question is underresearched.

The problem of credible commitment makes the case for a macroprudential regime of the kind I have described being in the hands of an independent institution, insulated from day-to-day politics. That institution need not necessarily be a jurisdiction’s central bank, but it might be. I return to that below.

Whoever the policymakers are, they are going to be mighty powerful. That entails some constraints on institutional design if the regime is to enjoy democratic legitimacy.

First—and of course this echoes work on monetary policy a few decades ago—the goals should not be set by the independent unelected policymakers themselves. The objective should be decided by the people’s elected representatives, after due public debate. That means more than elected politicians specifying a vague objective such as “preserve financial stability.” It means that elected politicians also need to set or bless the standard of resilience for the system as a whole.

That is not going to be easy because, on the face of it, it amounts to asking politicians a question that many would not want to answer: “What is your tolerance for a crisis?” And yet a standard of resilience, and so a tolerance for crisis, is implicit in the minimum capital standards applied to banks and other financial institutions.

Taking into account the differences in other parts of system, that implicit standard should be applied elsewhere in the broadest sense, as follows: given the nature of the risks to stability from sector X, what requirement is needed to ensure that it is no more likely to bring down the ceiling than banking? That would entail making assumptions about the structure of each sector, its vulnerability to risk, and the wider systemic consequences of distress within the sector for the provision of core financial services. Technocrats and researchers need to find ways of having these debates with elected policymakers.

No First-Order Distributional Choices

Where a policy regime is delegated to an independent institution, insulated from day-to-day politics, that should not entail society delegating first-order distributional choices. The stress is on choices. The suggestion is not that these regimes cannot have distributional effects. Such effects might be foreseen by the politicians who are doing the delegating, whether or not they are expected to average out to zero over time.

If that design constraint is accepted, it poses some concrete questions for macroprudential regimes. For example, is it okay for independent agencies to change maximum loan-to-value (LTV) or loan-to-income ratios without the usual process of formal consultation? When I was in office, my view and Mervin King’s view was that that would probably go a bit too far because there would inevitably be underserving losers whose voice had not been heard. I thought that the Bank of England found a neat solution to this a year or so ago (after my period of office): with mortgage credit rising quite rapidly, the BoE put a limit on the percentage of any bank’s portfolio that could be accounted for by high-LTV mortgages. In other words, it did not ban the writing of or borrowing on high-LTV mortgages but instead focused on what it judged was required for the resilience of the system.

That’s just one type of instrument. A lot of work would be needed if the constraint “no big distributional choices” were to be accepted. For each potential macroprudential instrument, it would be necessary to debate whether policy decisions would turn on the authority making a big distributional choice. If that is not debated in advance and decided by elected politicians, it will come back to bite these new policy regimes.

The Design of Multiple-Mission Central Banks

Finally, what does this mean for central banks?

Clearly, central banks are reasonable institutions to have these powers, but it then makes them multi-mission authorities. We have known for decades, both in theory and even more in practice, that there is a problem with institutions having multiple missions, because they have incentives to concentrate their efforts on the one that is most salient to the people and most observable to the outside world, which of course many people would think of as being monetary policy. So how on earth are these institutions, central banks, to handle multiple missions?

A few things can be said. First, I think that if a central bank has macroprudential responsibilities alongside monetary policy responsibilities, it should have separate committees. The committees would sensibly have overlapping membership in order to harness the benefits of housing them in the same agency, but ideally each would have a majority of members who were on that committee only.

Now that is what was done in the UK in the architectural reforms of 2012, and there is in fact a separatemicroprudential policy committee as well. During the period when the new architecture was being designed and debated in the Westminster Parliament and beyond, a recurrent question was, why have three committees? My answer was, so that at each meeting, there is a majority of people in the room who are responsible for only that committee’s mission and so will make sure it is always a serious meeting. However preoccupied those on two or three committees are with other matters, they will be incentivized to step up to the plate of the meeting they are in, because there’s a bunch of people in the room who do only that.

A second imperative to help central banks in this new area is to make policy systematic, with the reasons for policy choices sufficiently transparent for the guiding principles to be observable and so capable of being subject to public scrutiny and debate. To be less transparent or systematic than in monetary policy could be damaging to a multi-mission central bank.

At first sight, this is a big deal. A problem that has plagued supervision and regulation is that not only are the outturns difficult to judge, but the outputs have sometimes been impossible to observe. I think the development since the crisis of systematic stress testing of firms’ resilience potentially changes that in profound ways. The political economy significance of stress testing is that it is done once a year and is highly public: the scenario is public and the results are public. So politicians can have the Fed into Congress, the Bank of England into the Westminster Parliament, the ECB into the European Parliament, and say this scenario seemed a bit silly, too strong, too weak; the results seem implausible given the scenario, and so on. And, of course, that would be informed by masses of debate around the stress tests, in the way that monetary policy decisions are surrounded by debate.

Central banks should be in the business of encouraging debate about their stress tests. Debate, research and criticism will help them build and improve these regimes over the next decade or so.

Interactions with Monetary Policy and Central Bank Balance Sheet Operations

I have been describing a setup in which monetary policymakers and macroprudential policymakers can cooperate, even coordinate, but have separate objectives.

In some ways this bypasses suggestions that monetary policy should be the preferred instrument for addressing financial stability risks on the grounds that, as Jeremy Stein has pointed out, monetary policy gets into all the cracks, affecting all asset prices. But in fact, it gets into all of the cracks relevant to financial stability only under autarky, with capital controls and all the transactions within the economy denominated in the local currency. It does not get into all the cracks if households, businesses and intermediaries can borrow from abroad in foreign currency, as they can more or less everywhere. Time and again over the past half century we have seen stability problems rooted in, or at least brought to a head by, external or foreign currency indebtedness. That being so, even if societies were prepared to divert monetary policy from the goal of nominal stability, it could not substitute for a policy of maintaining the resilience of the financial system against the complete range of potential threats.

The approach outlined here allows the deployment of monetary policy instruments, and central bank balance sheet policy more generally, to return to being parsimonious once macroeconomic normality is regained. In the face of a stability-threatening boom, macroprudential policy action would be preferred over interventions based on selling parts of portfolios of private-sector securities to drive up credit spreads.

That leaves open the possibility of the authorities intervening in specific markets in a severe economic downturn when monetary policy has reached the zero lower bound. But that opens up a broader set of questions, not addressed here, about how to frame the proper role of central banks with respect to monetary/fiscal coordination.


See Paul Tucker, “Banking Reform and Macro-prudential Regulation: Implications for Banks’ Capital Structure and Credit Conditions,” speech delivered at the SUERF/Bank of Finland Conference, “Banking after Regulatory Reform: Business As Usual,” Helsinki, June 13, 2013, Bank of England,

    Other Resources Citing This Publication