## Monday, October 29, 2012

### Rethinking the Statutory Liquidity Ratio (SLR) in Indian banking

by Harsh Vardhan.

The CEO of a leading bank recently caused a flutter in the banking community by demanding the abolition of the Cash Reserve Ratio (CRR). RBI has promptly appointed a committee to look at this issue. The reserve ratios, CRR and SLR (Statutory Liquidity Reserve), are an important feature of Indian banking regulation. Alongside the debate about CRR, and new thinking about how monetary policy should be conducted, we should also review the SLR. SLR is a much bigger burden on the banking system and has no role in monetary policy.

### What is SLR?

SLR is the requirement imposed by the regulator on commercial banks that compels them to invest a percentage (currently 24%) of their Net Time and Demand Liabilities (NDTL) in approved government securities. Through this, today, 24% all the resources - deposits and borrowings - mobilised by commercial banks are invested in government securities. Currently bank deposits and borrowings are Rs.7 trillion which means that SLR places Rs.1.8 trillion into purchases of government securities. SLR creates a significant captive source of financing its borrowing program. This has three important implications:
1. SLR reduces the resources available for commercial lending by banks. Every rupee deployed in SLR is a rupee not invested in a private enterprise that needs capital. There is no free lunch: when capital given to the government, it comes at the cost of capital available to the private sector. Any reduction in the SLR (as in the CRR) will yield more capital for the Indian private sector. It is hence important to critically analyse both.
2. By creating a large captive source of deficit financing, SLR effectively subsidises government at the cost of savers and commercial borrowers. When a government has to borrow at a competitive rate in the market, the market exerts a check on irresponsible fiscal behavior of the government. When there is a large captive source of borrowing, the government is shielded from the pressures of the bond market and is more likely to engage in fiscal imprudence.
3. Such a large scale preemption of savings by the government through SLR fundamentally distorts the interest rate structure in the economy by artificially depressing the yield curve. This complicates the pricing of all assets in the economy.
If we want to "right-size" SLR we have to ask some important questions:
1. What is the rationale for imposing SLR?
2. What is the right level of SLR, that is consistent with this rationale and does not result in preemption of resources from the banking system?
3. Are there other conditions that need to be imposed on SLR so that it achieves the objectives?

### The rationale for SLR

What is the conceptual foundation for the regulator to impose SLR? The answer is: prudence. Banks raise public deposits with a promise to redeem them at par or more. To reduce the risk of the portfolio of the bank, the regulator ensures through SLR that at least some part is deployed in the safest assets available. But if prudence is the reason, what is the right level of such reserves that will ensure adequate prudence? Could it be that imposing a requirement as high as 24% is beyond prudence, and is actually a means for the government to preempt savings in the economy? It is hence important to ask the next question: What SLR do we need?

### What is the right level of the SLR?

Banks are in the business of taking risk. These risks are taken by deploying public deposits. The most potent weapon that the regulators have used against excessive risk taking is "risk capital" which the equity capital committed by the banks owners. In fact, the entire edifice of modern day bank regulation is based on provision of risk capital as a buffer against risk taking by banks. If we believe, as do most regulators, in risk capital as the buffer against risks, then it makes eminent sense for banks to hold this capital safely. This would logically lead us to conclude that prudence should demand that the bank's risk capital be held in very safe assets. In India, the risk capital requirement is 9% of risk assets which translates roughly to 6.5% of NDTL (given that the risk assets are typically 70% of NDTL). Therefore, the policy prescription should be: Banks must hold their entire risk capital in safe assets which should include both CRR and SLR.

Even if we assume the CRR is zero, this means that the theoretically right level of SLR would be around 6.5% of NDTL. If we scan the international landscape, this is the sort of number that we see in most countries. It is reasonable to argue that an SLR value above 6.5% of NDTL is motivated by pre-emption and not prudence. When the regulator prescribes a level of 24% for SLR, 6.5 percentage points are for prudence and the remaining 17.5 percentage points is really preemption by the government.

### The composition of SLR

The next important question about SLR is about its composition - what investments should qualify as SLR investments? Currently securities issued by the sovereign (Central and State Government bonds) are the only ones that are allowed as SLR investments. But if we accept prudence as the logic for SLR, then the regulation must make sure that these investments are as safe as they can be. This raises concerns about the rating threshold and of concentration risk. If Indian government securities are rated BBB and that of New Zealand government are AAA, it makes sense for banks to hold SLR in New Zealand Govt securities. Also, there should be limits on any individual issuer of securities, reflecting the standard risk management practice followed by any portfolio manager.

### The ideal SLR

Putting all the arguments above provides us an ideal construct of SLR as follows:
• SLR is imposed for the purpose of prudence and hence the operative principle is that banks should hold all the regulatory required risk capital in SLR
• The level of SLR should be consistent with the objective of prudence and anything over such a prudential level should be considered as preemption, which should be gradually eliminated.
• SLR should be invested in top rated securities available globally; furthermore there should be concentration limits on single security and issuer

### Dual limits structure for SLR

In the short term, it would be hard to come close to the ideal SLR outlined above. But there are some incremental changes that can be made without fundamentally altering the current framework that could provide banks with much greater flexibility. The regulator could prescribe 2 separate limits as follows:
• L1: is the minimum level of SLR that a bank would normally maintain
• L2: "core" SLR - a minimum below L1 that the banks can go down on SLR as long as the difference is only through repo arrangement on SLR with another bank
What does this mean? Let us assume that L1 is pegged at the currently prescribed level of 24%. We then define another limit, L2, which is closer to the prudential requirement of 6.5%. For simplicity, let us assume that L2 is set at 10%. This policy would demand that all banks maintain SLR at 24% but could go down this level upto 10% if and only if they enter into a repurchase agreement (repo) with another bank. Such a policy will mean that the banking system as a whole will continue to hold 24% SLR and so the government will continue to have access to this captive source of funding deficit. However, individual banks would be able to go down to lower levels if they have commercially viable opportunities to do so. Without diluting the overall investment by the banking system in government securities, it would provide significant flexibility to individual banks on commercial lending. In this respect, it is analogous to the idea of tradeable certificates for priority sector lending.

1. The statement that SLR decreases resources available for private investment is only correct at the most superficial level - if you take 24 away from 100, you get 76.

If we look just one level deeper, this statement breaks down. Banks collect deposits from retails customers, and some corporate customers. If you asked the majority of depositors in India what kind of risk they want to take in their deposits, the answer would be *none*. Hence, it is good argument that all retail deposits should *only* be lent to zero risk, in other words, the Government of India. (The ratings argument is just silly - in a post 2008 world, would any person believe that a rating by some agency should be logic enough to put money away from our control - the Government of India - into the Government of New Zealand? What was Greece rated in 2007?). Under that argument, what should SLR be? ~ 75%. Why this argument? Well, when you put *your* money into a bank, what are you looking for - that the private sector be well funded, or that your money should be safe?

The other argument - that there is something called "Risk Capital" that is a magic wand that saves the banking system is not credible. It would need us to believe that a set of fancy calculations defend us from the kind of failures that banks have faced over the last 80 years. Simple back testing shows that neither Lehman nor Bear would have surived 2008 even with the *new* capital requirements. When banks go pear shaped, they do it because they engage in behaviour that is so completely dumb that it seems close to criminal. No amount of "Risk Capital" would help that. Hence, the requirement that a certain amount of deposits be held in a "liquid" investment is generally accepted. Why bring Risk Capital into the discussion at all? It's not about protecting the equity of the Bank as a corporation, it is about protecting the interests of the depositors.

As a technical point, I don't understand the need for the suggestion to use Repos to maintin SLR - that is already legal today. If YES bank wanted to hold no Government Securities at all and satisfied it's requirements only by the method of repos with qualified participants, the RBI does not - under current regulation - even need to know. In fact, if you held all your SLR requirements by doing the reverse repo that the RBI offers on an overnight basis, it would still be perfectly legal.

2. While the argument for prudence makes a lot of economic sense, I doubt if it is feasible in India of today. Can any country where at least 60% population needs some form of government support and where there is a great need to invest in infrastructure, ever be fiscally efficient? In other words, in a country with so much needs, is an AAA rating possible? If not, then with a low rating why would an intelligent investor park money in Indian g-secs?

Perhaps a way forward is to phase out SLR over the next few decades. As the instances of extreme poverty falls, the governments then should be made to raise money from the "real" bond markets.

3. You said
If Indian government securities are rated BBB and that of New Zealand government are AAA, it makes sense for banks to hold SLR in New Zealand Govt securities.

But everybody knows that rating agencies are dumb and/or corrupt. You can see that that the rating agencies gave AAA rating to mortgage bonds which then was reduced to Junk within a span of 3 months. Spain enjoys better rating than India? Why? Simply because credit rating agencies are dumb and/or corrupt.

Why to allow such credit rating agencies to decide "safe" assets in Indian Banking? In fact, if they would have, then Indian banking would have encountered losses.
Also, you are exposing to exchange rate risk if you allow investing in other sovereign bonds.

4. I totally agree with your point stated above.. but even the CEO you mentioned said that CRR should be abolished or apprehension prevailing about SLR, why PSU banks are investing in SLR more than required (In given economics scenario).. One argument can be that they are not allowed to invest in high rated foreign securities( or other lucrative avenues) but do you really think that many european countries which are having higher rating than us are safe investment.... For the argument sake we can agree that SLR should be reduced..... But for that to happen we have to ensure that Indian Banks should be capable enough to Invest or Lend the remaining proactively with good risk management practices...

5. There are several problems with this approach.
1. Linking the required level of prudence in liquidity ratios prescription to risk capital is meaningless as the min. capital requirement is for safety whereas SLR is for liquidity which aims at ensuring not only liquidity but also for its signalling effect that the bank is liquid to meet any eventuality. Hence, it has to be as high as possible without compromising on other two basic tenets of banking viz. safety and profitability. Further, no amount of risk capital can save the bank if it takes huge and unacceptable risks and is less liquid.
2. The pre-emption theory has no basis as banks are maintaining more SLR assets than 23%.
3. Sovereign ratings by rating agencies have become so unreliable that if the suggestion that banks can subscribe to securities of better rated countries has little merit.

6. it is 23% and not 25%. SLR serves the purpose of prudence and allows GOI/State Governments, some of them, the credit worthy ones- to raise capital for their uses which is just as well and may be better than allowing banks to lend that to Vjay malyas!

1. a) Lots of countries get good safety in banks with much, much lower levels of SLR.

b) If prudence was the goal, you'd want to buy safe government bonds, i.e. those of AAA sovereigns. A near-speculative grade issuer would not make the grade.

c) Achieving safe banks and sensible banks by prohibiting private credit is like achieving road safety by banning driving. The PURPOSE of banks is to lend money to high quality projects in the private sector.

Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticism is perfectly okay; uncivilised language is not. We delete any comment which is spam, has personal attacks against anyone, or uses foul language. We delete any comment which does not contribute to the intellectual discussion about the blog article in question.

LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.