What Sri Lanka can do to improve the credibility of its dollar soft-peg

[The Invisible]

ECONOMYNEXT - Sri Lanka’s Central Bank is forced to collect forex reserves to be able to pay off government debt when necessary and meet International Monetary Fund targets for the foreseeable future, which shows it has no choice but to operate a peg.

If that is the ground reality, the Central Bank has to shake off some of its practices that have destabilized the peg and undermined its credibility.

To avoid balance of payments crises forever and gain full control over the exchange rate, the Central Bank has to float the overnight interest rate.

However, if that is too radical, there are many things that it can do to somewhat limit the downside.

A currency peg breaks and weakens because of an incompatible interest rate, that is to say through various means, a central bank will print money via its discount windows, reverse repo auctions and most dangerously, through outright purchases of government securities to keep interest rates down, when credit demand picks up.

As of the second week of June, the Central Bank was conducting its monetary policy as best as it could, providing only overnight cash as close to the ceiling policy rate of 8.50 percent as possible through its reverse repo window and overnight auctions. It is easy to break the credibility of the peg, but it’s difficult to restore.

In Sri Lanka, because the Central Bank buys dollar inflows from the government, and the Treasury does not sell them in the market, the rupee goes downhill fast and it’s difficult to arrest it.

Reverse Repo Auction Money

Reverse repo money is a little more harmful than window money. The problem with all new money is that a commercial bank can use the new money to grant customer loans.

Because reverse repo auctions are conducted in the morning, and since banks get the money early, they can use it to fund the loan book.

The window money, on the other hand, is given late in the day and the window is only open for a short time. The window is really a lender of last resort, while a reverse repo auction is almost a lender of first resort since it is possible to fund a loan book with auction money.

As of the second week of June, the Central Bank had created Rs10-15 billion in cash shortage by getting rid of a part of its T-bill stock, and was funding money markets through overnight auctions.

It may seem a bit unfair to jump on the banks after the shortage was created, but the Central Bank should discourage banks from coming to its windows, because that is where all the trouble starts.

Easy Money

It is quite evident that some of Sri Lanka’s domestic banks are far too easy with going to the window or participating in a reverse repo auction. Key foreign banks on the other hand usually have a little excess cash.

The behaviour of the foreign banks show that it is eminently possible to run a bank without running to the discount window all the time. The real pressure on the currency (in the absence of any outright Treasury bill purchases by the central bank) is triggered by the banks that run to the window too often and not currency dealers.

If a bank borrows from the window and lends to an importer – say Bank of Ceylon or People’s Bank goes to the Central Bank’s window and gives a loan to the Ceylon Petroleum Corporation – the pressure will fall directly on the rupee. Even if they do not, and the money is spent domestically by its customer, a part of it will end up in the forex market through some other bank.

The Central Bank should, therefore, discourage borrowings from its windows and curtail reverse repo auctions to the minimum, or abandon auctions altogether and only give money from the window. In an ideal world, window cash should be limited to intraday liquidity.

NOP Limits

At the moment, there are limits on bank dollar holdings in its books. These Net Open Position Limits have limited value. The entire focus on dollar holdings is misplaced.

Just because George Soros, or any other investor, had billions of dollars, he cannot hit Sri Lanka’s rupee dollar exchange rate.

To hit the exchange rate, he has to get hold of some rupees first. Usually, a speculator will try to get ahold of some domestic currency by doing a currency swap. If there is only a limited amount of rupees available, the swap rate will hit the roof.

When a foreign speculator tries to hit a currency, they succeed because money comes from the discount window to do so and the rate does not go up too much.

This is why George Soros had to run from Hong Kong with his tail between his legs, as swap rates went up, while bigger countries collapsed. There is no true discount window at a currency board.

Once a dollar has been bought by a bank (for existing rupees), it matters little whether the dollar remains in the bank’s books or is sold to buy a car or Mysore dhal. Just because there are two trillion or whatever volume of dollars floating around the world, it does not make the rupee weak.

The rupee also cannot become weak if an exporter borrows (existing) rupees and keeps dollars abroad. Of course, interest rates can go up.

Only if the credit is financed by the window does a problem come.

But the real problem with narrow NOP limits is their reducing forex market depth. Let’s say on Tuesday there is big demand for dollars or inflows fall, or both. But because NOP limits are narrow, the market has no depth to absorb it. As a result, the rupee/dollar rate will tend to be excessively volatile and the Central Bank is forced to step in or operate a peg.

While it may be fine to have narrow NOP limits when overall credit from the banking system slows down and excess liquidity builds up (strong side of the peg or in hard peg parlance also known as the strong side of the convertibility undertaking), it is not possible to prop up a currency when credit picks up (the peg or currency is at the weak side of the convertibility undertaking) while also targeting a fixed policy rate by printing unlimited amounts of money.

But it makes very good sense for NOP limits to be tied to a bank’s access to window cash. A bank that utilises the window for several days in a row should have its NOP limits cut or eliminated.

All of this will discourage banks from lending to customers without real deposits.

Discount Window Limits

It is not possible to guide an exchange rate (either to maintain a low inflation peg or to target a Real Effective Exchange Rate Index) without intervention. But, for any guiding to work, when the peg is on the weak side, the intervention has to be unsterilized, that is liquidity lost by dollar sales should not be replaced by the window.

Not even George Soros, who easily ‘broke the Bank of England’, could break a peg defended with unsterilized interventions. It is sterilized interventions that do not work. To be able to guide the exchange rate, sterilization of dollar sales has to be minimized to the lowest possible. Instead of NOP limits on the forex trading book, the central bank should slap limits on access to its discount windows. Let’s say, hypothetically, a bank is allowed to take printed money up to 5 percent of its capital at the ceiling policy rate, which was 8.50 percent from reverse repo windows or auctions. The next 5 percent of window cash should only be given at 9.0 percent. That way, banks can be encouraged to engage in more responsible lending.

It seems that there could be a daily liquidity variation of about 10 billion rupees in the banking system due to customers withdrawing cash. The first access limit should be set, keeping this in mind.

When such a limit is set, offenders like state banks who finance the Treasury overdraft with printed money should be a thing of the past.Several well-managed banks in Sri Lanka have internal limits on window borrowing. The Central Bank should examine these before setting the rules. It is a bit ironical that it is forex dealers of these banks, which do not pressure the currency, that get pulled up by the authorities because they usually have excess dollars and are blamed for ‘speculative’ trading. At first, fairly loose window access rules can be set, but they can be tightened over time.

Replace Term Repos with CB Securities

At the moment, the Central Bank’s net foreign assets are not enough to cover its domestic reserve money. And it has also run out of Treasury bills to sterilize dollar purchases, though it has some bills after the money printing binge in April. This column warned in December that running out of T-bills and sterilizing dollar purchases represented a danger to the rupee, based on the experience of the last crisis (Sri Lanka’s Central Bank should sell own securities in new credit cycle: Bellwether).

It said that terminating repo deals, which is also printing money, represented the latest danger to Sri Lanka’s economic stability. There was no expectation that the warning would crystalize into reality as soon as March. In March, out of 122 billion rupees of private credit, about 36.3 billion rupees was Central Bank credit, some or most of which came from terminated repo deals. One way to get over the problem is to get ‘special issue’ Treasury bills from the finance ministry like the Reserve Bank of India. But for reasons that will be explained later in the column, it is better to issue Central Bank securities.

Reserve Cover

The Central Bank is due to amend its governing monetary law soon. It is generally understood that ‘provisional advances’, a type of open-ended ‘window’ with no securities and no interest that is open to the Treasury, will be closed.

When amending the monetary law, the existing balance should be converted to three month bills at a discount equal to the market interest rate. Over time these bills can also be sold down (or directly swapped for dollars from the Treasury) to accumulate dollars.

At the moment, net foreign assets of the Central Bank is only about 90 percent of the monetary base.



When the provisional advances are sold down, the NFA cover will go up. New dollar purchases should be sterilized by 6-month Central Bank securities, which have not been issued for some years.

For the foreseeable future, no Central Bank profit transfers should be made until net foreign assets grow to be about 115 percent of the monetary base at least. Reserves should be kept as far as possible in US Dollars. Assets in Euro or Yuan should be avoided as much as possible.

If significant volumes of reserve assets are in dollar currencies, the reserve cover may need to be increased to about 120 or 130 percent of the monetary base to take into account any cross currency risks. The monetary law act can also be amended to prevent the dollar backing for the monetary base to be 100 percent at a minimum.

To settle government debt, a revolving sinking fund can be built by future profit transfers of the Central Bank. The Central Bank should get out of all public debt activities and let the Treasury manage its own affairs. Measuring foreign reserves against stocks like short-term debt or exports are meaningless. It is beyond the scope of this column to explain why.

Tiered Window Access

In order to avoid balance of payments troubles in the future, liquidity window operations should be carried out only against CB securities. As a result, banks that need liquidity would be forced to carry CB securities in their portfolio, creating demand for them.

Only overnight liquidity should be provided. No term reverse repo, or auctions should be conducted in the morning. If necessary, liquidity could also be provided against any US government securities that banks have. No Sri Lanka dollar bonds should be allowed as security for window operations under any circumstances.

Since there would only be a finite volume of CB securities in issue, which banks can use to get liquidity, the Central Bank will not be able to provide an unlimited amount of liquidity and create balance of payments troubles (to sterilize interventions until large volumes of forex reserves are depleted).

For this arrangement to work well, policy rates have to be ideally floated. If not, the policy corridor should at least be widened from the current 7.25 percent and 8.50 percent to 5.00 and 10.00.

It must be understood that without a full float of the rates, it is not possible to absolutely fix the exchange rate. However, like in Singapore, it will be possible to control the exchange rate over the longer term.

Interest Rates

In seems increasingly likely that the 6.25 percent overnight rates seen in 2014 may not come back. But it is not necessary to have a rate cut to boost growth. When the credit cycle turns, economic activity picks up. Several factors will drive economic activity in the short term.

Business confidence is one. There seems to be some business confidence emerging after people meet Mangala Samaraweera and Eran Wickramaratne. On the negative side, a falling rupee will push up prices and kill purchasing power and economic activity.

Even in well-managed countries, interest rates have been moving up, but growth is also there. Singapore’s short-term rates have been steadily moving up. Singapore also does not have a policy rate. It manages to grow without ‘rate cuts’.

The three month Singapore swap offer rate (SOR), a key benchmark in the country, has steadily moved up over the past five years.

It can also be seen that the overnight volatility is very high due to the lack of a policy rate, but it does not lead to a hysterical rise in longer-term rates and may even be inverted at times.

The Singapore Interbank Offered Rate has also been climbing. If rates do not fall back to 2014 levels in Singapore, with much better fiscal policy, may be there can be no expectations of rates falling to 6.0 percent levels in Sri Lanka either. Sri Lanka has several reasons for high interest rates: the main one is currency depreciation and inflation and the second one is the high reserve ratio.

It is possible to operate without a reserve ratio if necessary or with a reserve of about two percent, once more responsible lending and liquidity management develops at domestic banks. The third is the practice of selling excess dollars coming to the Treasury to the Central Bank, instead of to the market.

Credit Pick Up

The good news is that credit seems to be picking up. There can be no BOP problems if the economy is in the doldrums, absent capital flight. Unfortunately, there seems to be some.

If economic activity and credit picks up, the Central Bank will not be able to collect forex reserves at the rate it has been doing until January 2018. Already some UNP politicians, including Ravi Karunanayake, who helped create the 2015 BOP crisis with Arjuna Mahendran, is making noises about high interest rates.

Careful thought should be given to cutting rates, or narrowing the policy corridor. For one thing the US ‘normalization’ will have unforeseen effects on capital flows. US rates will also hit pegs with much more credibility than Sri Lanka’s highly unstable peg.

The real pressure on the currency (in the absence of any outright Treasury bill purchases by the central bank) is triggered by the banks that run to the window too often and not currency dealers

If there is a peg, and its credibility is broken and the Central Bank tries to float the currency to avoid interventions, it is usual for the fall to overshoot 5 to 10 percent before the float takes hold and exporters are comfortable to sell.

If the currency floats fairly freely, it will also fall, when the US Dollar strengthens. All it takes is a couple of term reverse repo auctions to break the credibility of the peg.

But it is much more difficult to restore credibility. So care must be taken. For that reason, the Central Bank, at the first opportunity, should appreciate the currency and create a ‘buffer’ in the exchange rate, in case there is capital flight from time to time, when the US continues to raise rates. This has to be done even if tiered limits are set for window access. It also is a must if a bastardized peg in the form of a ‘flexible exchange rate’ is followed.

The Big NO-NO

The other big danger to the peg is the practice of selling Treasury dollars to the Central Bank. If the guided peg breaks and there is an attempt to float, this is an absolute no-no, even if liquidity is short. All Treasury dollars should be sold in the market for old rupees, not to the Central Bank for new rupees.

Excessively high interest rates are needed to correct currency pressure, because of this practice. Even after the most recent hike, the US rate is 2.0 percent. Sri Lanka’s is a thumping 8.5 percent.

That should be more than enough to maintain a fair degree of credibility in the peg, if reverse repo auctions are abandoned and Treasury dollars are sold in the market.

There are many costs of currency instability, even in the short term.

The business of importers is disrupted. Exporters also waste time in speculating, instead of making longer-term plans. They may borrow to keep dollars out, crowding out other businesses. Everybody’s business plans and costs budgets are disrupted, a bit like regime uncertainty.

The Central Bank is also having a tough time with President Sirisena taking pot shots at Prime Minister Wickremesinghe.