Waiting for Basel? Next steps for Canada’s bank capital regime - Remarks by Jeremy Rudin to the C. D. Howe Institute, Toronto, Ontario, April 6, 2017

Quote

Until recently, the Committee had been working to complete the reform agenda, but those negotiations have stalled, and it is not clear when they will resume in earnest.

In our view, there are important issues that still need to be addressed. Our strong preference is for the Committee to get back to work on those issues, and to reach an agreement that all member jurisdictions will support and implement.

At the same time, Canada is preparing to move ahead on its own, with a Canada-specific plan for improving our capital regime. If we conclude that the Basel process is unlikely to restart in the near future, OSFI will put that plan into action in a measured way.

Introduction

My thanks go to the C. D. Howe Institute for organizing today’s event.

Canada has reached a significant juncture in the evolution of our approach to bank regulation. In the years following the global financial crisis, agreements reached by an international body — the Basel Committee on Banking Supervision — have been, by far, the most important influence on our regulatory agenda.

Now, that influence may be waning.

Basel III — A stalled international agreement

Most of you will know that in the wake of the financial crisis, the Basel Committee undertook a program of far-reaching reforms to the international standards for bank capital requirements. The agreement that the Committee reached in 2009 — commonly referred to as Basel III — delivered substantial improvements in the design and calibration of minimum capital requirements; improvements that were felt both globally and here in Canada.

However, that agreement also left some significant unfinished business.

Until recently, the Committee had been working to complete the reform agenda, but those negotiations have stalled, and it is not clear when they will resume in earnest.

In our view, there are important issues that still need to be addressed. Our strong preference is for the Committee to get back to work on those issues, and to reach an agreement that all member jurisdictions will support and implement.

At the same time, Canada is preparing to move ahead on its own, with a Canada-specific plan for improving our capital regime. If we conclude that the Basel process is unlikely to restart in the near future, OSFI will put that plan into action in a measured way.

Moreover, that plan will recognize that an effective capital regime includes much more than the minimum requirements that have been the primary focus of attention at Basel in recent years.

Building a stronger capital regime that responds to lessons learned from the financial crisis

At OSFI, we learned some valuable lessons about the banking system, and the regulations that protect it, from the global experience during the financial crisis.

This experience points to three areas where the previous international bank capital regime — known as Basel II — did not perform well.

First, an effective capital regime must establish minimum requirements for banks’ ability to absorb losses that will protect their solvency if they experience severe stress. Clearly, in many countries this was not achieved under Basel II.

Second, an effective capital regime should support confidence in the banking system. Yet, we saw during the crisis that confidence in the solvency of some major banks dissolved, even when those banks were reporting strong capital ratios as measured under Basel II.

Third, an effective capital regime provides healthy incentives for banks in both good times and bad. In good times, an effective regime discourages banks from taking excessive risk by requiring higher capital for riskier assets. In bad times, an effective regime encourages banks to maintain adequate capital without resorting to fire sales of assets, or drastic reductions in lending.

As it stands today, the international capital regime does a much better job at protecting the solvency of banks than did its predecessor. Under Basel III, the quantity and quality of capital that banks are obliged to hold is much enhanced compared to Basel II. This means that banks are much better protected against major losses than they were before the crisis, improving the overall resilience of the global banking system.

This is no small accomplishment.

More work is needed to improve the incentives that the bank capital regime provides

We have made less progress, however, on improving the incentives that the capital regime provides to banks. We want a system that requires banks to hold more loss absorbing capacity if the assets that it owns are riskier. This both discourages banks from loading up on risk, and ensures that they have enough capacity to absorb the losses that they could plausibly experience.

The tools that we use to achieve these incentives are risk weights. The capital regime requires banks to classify their assets according to defined categories of risk. Each category is then assigned a weight that reflects the risk.

There are two approaches to measuring risk weights in the Basel system.

One is the standardized approach which, as its name implies, prescribes standard asset classes and their associated risk weights.

The other is the internal ratings based (or IRB) approach which allows authorized banks to use their own risk models as input into the risk weights.

In Canada, our small and mid-sized banks use the standardized approach. Our largest banks are authorized to use the internal ratings based approach.

What we see, both globally and domestically, is that neither of these approaches is doing a good enough job of calibrating risk weights in proportion to risk.

The underlying problem in each approach is the mirror image of the problem in the other. In the standardized approach, the problem manifests itself in risk weights that do not vary enough from bank to bank. In the internal ratings based approach, risk weights vary too much from bank to bank.

Weighing in on the issue of unwarranted variability of risk weighted assets

How did we arrive at this situation?

As I noted, the standardized approach groups all assets into a few simple categories, each of which is assigned a single, unique risk weight.

For Canadian purposes, at least, some of those categories are too broad for a single risk weight to accurately reflect the underlying risk. For example, there is a single risk weight for all residential mortgages.

As a result, the standardized approach places the same risk weight — and therefore the same capital requirement — on a bank that specializes in mortgages on investment properties as it does on a bank that specializes in mortgages on owner-occupied properties.

Yet, we know that mortgages on investment properties are riskier than mortgages on owner-occupied properties. In that sense, the standardized risk weights do not vary enough across banks.

If we turn to the internal ratings based approach, we find that risk weights vary too much across banks. This is seen most clearly when banks using the internal ratings based method are asked to determine the risk weight that they would assign to a common, specified portfolio of assets.

The Basel Committee undertook such a study and the results were telling.

In this study, a group of globally active banks using internal ratings were asked to assign risk weights to a model portfolio of assets. One bank assigned a risk weight to the portfolio that was 40 per cent lower than the median risk weight assigned by all other banks in the study. That would mean this bank would hold 40 per cent less capital than the other banks for the same level of risk.

This is the type of outlier result that should not occur.

This degree of variation in risk weights fails against all three objectives of a sound capital regime.

First, the outlier portfolios lack enough protection against loss. Second, this sort of variation undermines confidence in the banking system as the measured capital ratios of the outlier banks are not representative of the true capitalization of those banks. Third, a regime that tolerates this degree of variability does not provide healthy incentives to banks to manage their risks. Rather, it provides an incentive to try to manipulate the system to reduce capital requirements. This is the last thing we want banks to be working on.

When a bank is authorized by its regulator to use a model to calculate risk weights, it must apply that model responsibly. At OSFI, we authorize a bank to use the internal ratings based approach where we believe the outcome will be a more accurate risk weight than the standardized approach. It is also our view that the effort a bank must put into building and maintaining its model creates a degree of discipline and knowledge that should make it a better risk manager.

Allowing a bank to use its own model should not, however, give it free reign. We impose a system of controls on the range of outcomes that the banks' models can produce. This aligns with our overall approach to working with information that we obtain from banks: trust, but verify.

We think that this balanced approach can be the basis of an international consensus that will better discipline the use of internal models. We need to find a set of constraints that will substantially diminish the unwarranted variability in risk weights across banks, without losing the benefits for risk management that can come from the use of internal models.

This is the important work at hand for the Basel Committee.

While we remain hopeful that the Basel Committee will reach an agreement that Canada can support, the possibility remains that negotiations could be stalled for some time. Finding consensus among the 27 jurisdictions represented on the Basel Committee — jurisdictions that differ along legal, economic and cultural lines — is always a challenge. Now that some are questioning the value of multilateralism more generally, the challenge is even greater.

A retreat from international capital work is not a retreat from our domestic agenda

What would we do if we eventually came to the conclusion that the Basel process was stalled indefinitely?

We would not stand still. Rather, we would address the outstanding issues in Basel III in a way that works for Canada. We would do so with an eye to addressing the inappropriate incentives that arise when assets that pose very different risks are given the same risk weight.

And we would examine the variation in risk weights across Canadian banks that use the internal ratings based approach. We would want to make sure that the problem of unwarranted variation in risk weights that we see internationally, does not take root in Canada.

We recognize that moving ahead on our own would not give us a completely free hand. We are mindful of preserving a level playing field domestically, and for our banks that compete internationally. We are prepared to impose higher standards on Canadian banks than the international minimums when it is necessary to bolster safety and soundness. We have done this in the past, we are doing it to some extent in the present, and we are prepared to do so in the future. But these departures from a level playing field need to meet a stringent test.

The way forward to ensuring buffers play their intended role in the capital regime

To this point I have spoken mainly about the rules for setting minimum capital requirements for banks. However, there is much more to effective capital regulation than setting the minimum requirements.

We need to consider all of the elements of the capital regime, and to pay attention to how these play out over time.

For example, an effective regime will always encourage, and sometimes require, banks to hold additional capital — or buffers — above the minimum capital requirements.

It is tempting to conclude that any step that increases capital buffers will make the banking system more stable. After all, one of the lessons of the crisis was that banks needed to be better protected from insolvency.

But as I noted earlier, that is not the only lesson for bank capital that emerged from the crisis. We also want the capital regime to foster confidence in the banking system in bad times. That will not be the case if banks cannot allow their capital ratios to dip into the buffer range without triggering a severe loss of confidence.

Of course, a bank that is using up its capital needs to recapitalize. If this goes on long enough, or happens quickly enough, dramatic measures may be required. But for a bank that starts with capital well above its regulatory minimum, we all need to see the idea of using some of the bank’s capital buffer as the normal first step in the process of recapitalization. Otherwise, we will have inadvertently designed a capital regime that does not foster confidence in the system, and potentially undermines it.

This also underlines the importance of having a system for setting risk weights that has integrity; and that is seen to have integrity. Otherwise, a handful of faltering outliers could undermine confidence in all banks.

Not only should the capital regime foster confidence in bad times, it should also provide appropriate incentives to banks in bad times. In particular, the capital regime should avoid adding any artificial incentive for banks to rapidly de-lever during an economic downturn, either by selling assets at fire sale prices or sharply restricting new lending.

A capital regime that avoids this perverse incentive is a regime that establishes buffers that are meant to be used in bad times. As regulators and supervisors, we need to set out the circumstances under which we would allow, if not encourage, banks to let their capital ratios decline into — but not fall below — the buffer range. Having done that, we need to ensure that banks understand those circumstances, and that market participants understand them as well. Failing that, the capital regime could become part of the problem during a downturn, when it needs instead to be part of the solution.

Ideally, we would have an international consensus on the dynamic role of capital buffers; a consensus that was clear to banks and to market participants. This may be too much to ask in the current circumstances. We need to be prepared to build a clear plan for the use of capital buffers by Canadian banks, and to build a new and better understanding about the role of capital buffers in Canada.

Conclusion

As we move ahead to the next steps for Canada's bank capital regime, let us remind ourselves that we have a strong regime that we can be proud of. We have a strong regime that protects the interests of depositors and other creditors to the banks and fosters confidence in our banking system.

One of the reasons that we have a strong regime is that we have not shied away from improving it. As we look ahead, our top priorities for improvement are twofold: better aligning risk weights with underlying risks, and ensuring that capital buffers can be effectively used when they are needed.

We believe these objectives will be mutually reinforcing. Improving the ability of risk weights to capture the risk that a bank holds will not only improve the resilience of individual banks, it will improve confidence in the overall capital regime. That increased confidence will, in turn, improve the likelihood that the capital regime will act as intended in times of stress.

Our preferred approach would be to move forward on both of these priorities in partnership with our international peers on the Basel Committee. At the same time, we are preparing our own plan for improving the capital regime in Canada, while we watch to see if the international community will be willing to re-engage around the Basel table.