Raghuram Rajan: Financial Sector Legislative Reforms Committee Report (FSLRC) - what to do and when?

Talk by Dr Raghuram Rajan, Governor of the Reserve Bank of India, at the First State Bank "Banking and Economic Conclave", Mumbai, 17 June 2014.

The Financial Sector Legislative Reforms Committee (FSLRC) Report is one of the most important, well researched, as well as well-publicized reports in Indian financial history. It not only lays out the functions of the financial sector and how it should be structured, but also how legislation and regulation governing it ought to look like. The authors of this report truly have to be commended for their national service. The report's influence will be felt for many years to come.

There is much to like and agree with in the report. In laying out the need for consumer protection, raising the issue of whether products sold are suitable for the target customer, and putting the onus on the financial institution to determine suitability, the report has forced regulators to review their consumer protection frameworks. We at the RBI are indeed engaged in such an exercise, informed by the valuable guidelines in the FSLRC report.

There is more of great value. The FSLRC's emphasis on the need for a clear monetary framework culminated in the Dr. Urjit Patel Committee report, which will guide our thinking in the years to come. Similarly, its focus on creating new institutions like the Financial Resolution Authority, which will help us resolve distressed financial institutions at minimum cost to the economy, is much needed.

I could go on. But I come here not to praise the FSLRC Report, but to debate some of it. I will argue that there are two fundamental areas of tension. One is the oversight of regulators. The FSLRC suggests laws that do not micromanage, giving regulators the freedom to fill in the details in consonance with the changing needs of the economy. At the same time, the FSLRC wants to check and balance the activities of regulators through judicial oversight. Too much of checks and balances could completely vitiate the flexibility afforded by rewriting laws. We need to find a proper balance, and the balance may vary with our level of development. I worry we have not thought through this fully.

The second area of tension is the appropriate size and scope of regulators. The FSLRC's recommendations seem somewhat schizophrenic here. On the one hand, it emphasizes synergies in bringing together some regulators into one entity. But in the process it suggests breaking up other regulators, with attendant loss of synergies. There is no discussion of the empirical magnitude of the synergies gained or synergies lost, which makes the recommendations seem faddish and impressionistic rather than based on deep analysis. Indeed, across the world, we see a variety of organizational structures in existence, suggesting that there is no one right structure. If so, there should be strong arguments for departing from the status quo, which the FSLRC does not provide.

Let me elaborate on these two issues.

The logic for regulation

The logic for regulation according to the FSLRC is to deal with market failure or, more colloquially, bad behaviour. The Commission talks about incomplete information or poor incentives as a reason for bad behaviour, but one of the most important reasons for the bad behaviour necessitating regulation is what economists call incomplete contracts; that is, the behaviour of the regulated entity (vis a vis customers, the public at large, the taxpayer, or the market) cannot be completely specified in contracts because it is too difficult to observe or verify in real time, or it can only be gauged across many contracts.

This means that while courts can enforce specific contracts, the regulator can sometimes do better. For instance, a bank may attract a lot of complaints from its credit card customers. While no single customer may think the case worth taking to court, and while no customer may be able to prove the bank was in the wrong, the large number of complaints will suggest to the regulator that the bank needs to shape up. By comparing the nature of the complaints it gets from this bank's customers with the complaints it gets from other banks, the regulator can gauge whether something is wrong with the bank and act. Similarly, if a particular product attracts a lot more complaints than other products, the regulator can ask the industry to modify the product appropriately, or even ban it.

A regulator may also have to prevent certain forms of contracting - such as the CDO squared and CDO cubed that emerged before the financial crisis. If the regulator thinks a certain kind of security will impose undue risks on the system, it can ban the security, even though it would have traded amongst consenting adults. While the regulator has no proof that the security will behave as it thinks, the regulator cannot wait till the risks occur, for it may be too late.

The broader point is that a lot of regulatory action stems from the regulator exercising sound judgment based on years of experience. In doing so, it fills in the gaps in laws, contracts, and even regulations. Not everything the regulator does can be proven in a court of law. Courts do not interfere in the specific decisions of a corporate board - using the business judgment rule, they do not second guess business decisions, and only pull up boards when there is a violation of the legal process of arriving at a decision. In the same way, there are a range of regulatory decisions where regulatory judgment should not be second guessed.

The danger of excessive legal oversight

Yet one reading of the FSLRC is that almost everything the regulator does, not just the framing of regulation or the process by which decisions are reached but also the exercise of regulatory judgment as well as policy decisions, is to be subject to legal appeal. For that, it wants to create a Financial Sector Appellate Tribunal. The intent is to place more checks and balances on regulatory actions. Note that the process by which the regulator reached a decision, as well as the conformity of the decision with basic principles such as natural justice, can already be challenged through a writ petition in High Court. Even now, some regulatory decisions can be appealed to the central government. But how much checking and balancing is enough? Do we want even policy decisions to be appealable? Can legal oversight become excessive?

There are three dangers we have to guard against. The first is to ask tribunals to make judgments that they simply do not have the capability, experience, or information to make, and where precise evidence may be lacking. If we attempt to do this, we will undermine the very purpose of a regulator. Of course, one could trust the good sense of the tribunal to follow a "regulatory judgment" rule and not intervene in a broad array of matters, but does this not imply a double standard - we trust the tribunal's judgment but not that of the regulator. More likely, though, past experience suggests that entities like to justify their existence, and if set up, a tribunal will intervene more than necessary.

The second danger is that easing the appellate process will invite appeal. In a developed country with well-established regulations, a case history of judgments, and speedy delivery of justice, this would not be a problem. In India, where the financial system is developing and many new regulations have to be framed (more so if we move to a principle based approach for legislation), and where the tribunals will have a significant amount of learning to do, the encouragement to appeal could paralyze the system and create distortions, as needed regulations are held up and participants exploit loopholes.

Finally, in every country, a healthy respect for the regulator serves to keep participants on the straight and narrow. In a developing country, where private behaviour is less constrained by norms or institutions, this is especially important. But to the extent that private parties with their high priced lawyers can check the regulator, that healthy respect dissipates. So the final danger is that the regulator could become a paper tiger, and lose its power of influencing good behaviour, even in areas that are not subject to judicial review.

Am I arguing that no checks and balances are needed? Certainly not! But there are already checks and balances in place, including review by constitutional courts like high courts through writ petitions. Senior officers of the regulator are appointed, and can be removed, by the government. The FSLRC recommends an annual report to parliament, as well as regular discussions with parliamentarians. These are good suggestions, which would add to oversight.

Some could argue that SEBI is already under the Securities Appellate Tribunal, so why not bring other regulators under a tribunal? So long as the Tribunal only questions administrative decisions such as the size and proportionality of penalties, I do not see a problem. But if it goes beyond, and starts entertaining questions about policy, the functioning of a regulator like the RBI, which has to constantly make judgments intended to minimize systemic risk, will be greatly impaired. Indeed, because of the tendency of any new organization to overreach to justify its existence, one should be careful about tying the financial regulator with further judicial oversight. Better to revisit these issues a few years from now when both regulation and oversight mechanisms are better developed.

Finally, we do understand that if the regulator wants to be trusted, it has to display the greatest competence and integrity. The RBI, despite the general deterioration in the probity of public institutions, has maintained a reputation for integrity. We cannot be complacent about this and have to work on maintaining a culture, as well as service conditions, that encourage integrity. We also have to work continuously on upgrading our capabilities so we match the fast pace of change in the financial sector.

Regulatory architecture

Another area where there are tensions in the FSLRC's reasoning is on regulatory architecture. Let us take the suggestion to merge all regulation of trading under a new Unified Financial Agency. So the Forward Markets Commission, as well as the bond regulation activities currently undertaken by the RBI, would move under a new roof, as would SEBI. But this assumes that the central synergy is the fact that the instrument is traded. But could other synergies exist? And how important are they?

For instance, in forward trading where a real commodity is delivered, regulatory oversight over the real markets for the commodity where price is discovered, as well as over warehouses where the commodity is delivered, may be important sources of regulatory synergy. Should the FMC be subsumed under the Unified Financial Agency or would it be better off having stronger links to the ministries overseeing the real commodities? I think the answer needs more investigation.

Similarly, is the regulation of bond trading more synergistic with the regulation of other debt products such as bank loans and with the operation of monetary policy (which requires bond trading) than with other forms of trading? Once again, I am not sure we have a compelling answer in the FSLRC report. My personal view is that moving the regulation of bond trading at this time would severely hamper the development of the government bond market, including the process of making bonds more liquid across the spectrum, a process which the RBI is engaged in.

The FSLRC also seems to be inconsistent in its emphasis on synergies and regulatory uniformity. It proposes all regulation of trading should move under one roof, all regulation of consumer protection should move under another roof, but the regulation of credit should be balkanized - banks should continue to be regulated by the RBI but the regulation of the quasi-bank NBFCs should move to the Unified Financial Agency, a regulatory behemoth that would combine supervision of trading as well as credit. This balkanization would hamper regulatory uniformity, the supervision of credit growth, and the conduct of monetary policy.

More broadly, the FSLRC seems to have a somewhat idealistic view of the benefits of reorganization. It seems to believe that once activities are combined in an organization, synergies can be fully exploited while if they exist in separate organizations, synergies will not be exploited. I too shared such a view, but I now believe it is too extreme. Silos within a large bureaucratic regulator may prevent synergies from being exploited, while frequent inter-regulatory meetings can allow regulators to capture many of available synergies between their activities. Indeed, one particularly useful proposal by the FSLRC is to put the Financial Sector Development Council on a firmer footing. It is a good venue for inter-regulatory cooperation, and its benefits are further augmented by personal interactions. For instance, Chairman SEBI and I try to get together once every month to note and resolve issues.

At the same time, while negotiations and cooperation between regulators can overcome organizational barriers, it is not wise to give a regulator a responsibility and leave the tools for exercising that responsibility in other hands. The RBI has responsibility for managing the internal and external value of the rupee, and more broadly, for macroeconomic stability. As a number of multilateral agencies and academics have recognized, the ability to shape capital inflows is now a recognized part of the macro-prudential tool kit. But by taking away control over internal capital inflows from the RBI, isn't the FSLRC taking away an important tool from the RBI?

If it ain't broke...

Lest all this sound like an unthinking defence of regulatory turf, let me add that there are places where the RBI could give up powers. For instance, if the government wants to manage its own debt, there is no reason for the RBI to stand in the way. I don't believe the government suffers any less from conflicts of interest in debt management (unlike the views of the FSLRC), but the RBI could well carry out the government's instructions without any loss in welfare. I imagine, however, that the government will depend on deputations from the RBI for a while for advice.

Instead, think of my remarks as an attempt to draw out the important and undoubted benefits of the FSLRC report, while eschewing grand schemes with dubious chances of success. Undoubtedly our laws need reform, but that is no reason to try entirely new approaches to legislation, overlaid on entirely new regulatory structures, complemented by entirely new oversight over regulation. Undoubtedly, we have had, and will have, periods when regulators have not gotten along with each other. But is that a reason to merge some organizations and break up others, perhaps ensuring dysfunctionality along many other dimensions? After all, there is no single regulatory architecture that has emerged with distinction from the crisis. Instead, different regulatory architectures have succeeded or failed based on the circumstances of the country and the quality of the regulator.

Undoubtedly, we have also had occasions when regulators have exceeded their remit or been high-handed. But is that a reason to subject their every action to judicial second-guessing? Is there a reason we need more checks and balances, or are we trying to solve a problem that does not exist.

As the Chinese would say, let us recognize the value of crossing the river by feeling each stone before we put our weight on it. Let us not take a blind jump hoping that a stone will be there to support us when we land. Or in American, if it ain't broke, don't fix it!