CURRENCY boards are making a comeback. Several countries have established them in an attempt to stabilize their currencies and control inflation. Although they limit the freedom to set fiscal and monetary policy, currency boards have proven effective, and the trade-off may sometimes be worth it.
Currency boards, a system of fixed exchange rates that was common in colonial territories during the first half of the twentieth century, fell into disuse when colonial regimes were dismantled. In recent years, however, interest has revived in currency boards as a means of stabilizing currencies and bringing order to economic conditions generally, when more conventional arrangements have proved wanting.
The first of the recent reintroductions of a currency board was in Hong Kong in 1983. Although highly successful, Hong Kong’s currency board remained something of an anomaly until Argentina set up a similar arrangement in 1991, under very different circumstances. In the fall of 1991, the dissolution of the former Soviet Union sparked more debate about the usefulness of currency boards and whether they could bring financial stability to the Baltic countries, Russia, and the other countries of the former Soviet Union (BRO countries) during the early stages of transition. In 1992, Estonia, the first of these countries to introduce a separate currency, began operating a currency board; in 1994, Lithuania followed with a similar arrangement.
Currency boards have proven unusually effective methods of pegging the exchange rate, but they do not come without a price. Governments that operate currency boards must accept restrictions on the way they conduct fiscal, and especially monetary, policies. These restrictions mean that currency boards are not for everyone, but they can be very effective in countries suffering from hyperinflation, or in new countries in which financial stability is not yet well established.
Currency boards are obliged to supply—on demand and without limit—the foreign currency to which the domestic currency is pegged in exchange for local currency (bank notes and, when applicable, reserve deposits of commercial banks held with the currency board); and vice versa. The currency board sets the exchange rate for these transactions.
What do you back? The simplest arrangement is for currency boards to back only bank notes, as in Hong Kong, whose currency board, known as the Exchange Fund, originally had no direct role in commercial bank clearing operations and, therefore, no other base-money liabilities. Commercial bank clearing is carried out by the Hongkong & Shanghai Bank (HKSB). Since July 1988, however, the HKSB has been required to hold balances with the Exchange Fund, in amounts determined by the authorities, which the clearing accounts held with the HKSB by other banks must mirror. The Exchange Fund is thus now linked to the clearing accounts of commercial banks.
Before Argentina and Estonia introduced currency boards, their central banks were already involved in commercial bank clearing arrangements because interbank settlements were carried out through the central bank. Whereas Argentina’s commercial interbank transactions were highly efficient, in Estonia, the only means of clearing was through accounts with the central bank. In neither case would it have made sense to undermine the security of accounts with the central bank by denying their convertibility, thus, the guarantee of foreign exchange applies to both bank notes and deposits with the central bank.
How much backing? To instill confidence that it can honor the pledge of convertibility, a currency board should start with foreign exchange reserves sufficient to back a substantial proportion of the outstanding value of relevant liabilities at the chosen exchange rate. When confidence is lacking, this proportion needs to be at least 100 percent. Ideally, the central bank should also hold surplus reserves in order to deal with contingencies such as bank failures. In 1983, the Hong Kong Exchange Fund had net reserves (gross reserves net of borrowing) in excess of what was required to honor convertibility. Argentina’s Convertibility Law requires full backing for the monetary base in terms of gross reserves. So long as reserve liabilities are sufficiently long term or easy to refinance, a sufficiency of gross reserves can be considered adequate backing. Uncertain economic conditions in Estonia following independence in 1991 provided a strong case for full backing in net as well as gross terms. However, Estonia was fortunate—the Bank of England returned gold deposits made by the country before World War II. Thus, when Estonia’s currency board was launched in June 1992, it had 90 percent cover, which quickly rose to over 100 percent when more gold and other moneys were restituted.
Which currency? The choice of currency to which to peg will depend, among other things, on whether the currency is strong or weak, as well as on trade ties. For both Hong Kong and Argentina, the obvious choice was the US dollar, the dominant unit of account for trade and other financial transactions. In Estonia, although the Finnish markka or Swedish krona might have recommended themselves from the point of view of trade, the deutsche mark was chosen because of its strength.
In principle, backing should consist of external reserves, in the form of (relatively short-term) interest-bearing assets, in the currency to which the domestic currency is pegged. This is to ensure that movements in asset prices do not undermine the backing and to provide interest, either to add to the capital of the central bank or currency board, or to provide income to the government. Thus, although the bulk of Estonia’s reserves was in gold when the currency board was first set up, this was rapidly converted into interest-bearing deutsche mark assets. The reserves of both Hong Kong’s Exchange Fund and of Argentina’s central bank are mostly held as interest-bearing US dollar assets.
Setting the exchange rate. The exchange rate selected should ideally neither undervalue nor overvalue the currency unduly, or represent a significant departure from the prevailing market rate. Currency boards are most likely to be introduced to arrest a sharp depreciation. In this case, the speculative attack that triggered the depreciation is likely to have corrected most or all of any prior overvaluation. The exchange rate prevailing just before Argentina and Hong Kong set up their currency boards was not felt to pose any threat to competitiveness. In the case of Estonia, the ruble had depreciated so much after independence that there was concern that it was undervalued. On balance, however, it was felt that the advantages of an undervalued currency for trading would outweigh the disadvantages of high inflation.
Who has access? In Hong Kong, only banks with the authority to issue notes have the right to convert Hong Kong dollar cash into US dollars at the Exchange Fund. Individuals and enterprises must rely on the competitiveness of the interbank currency market to ensure they get similar rates to those offered by the Exchange Fund. In Argentina, all comers are, in principle, entitled to exchange pesos but, in practice, only banks can deal with the central bank. In Estonia, initially, to build confidence, the general public could exchange their bank notes at the central bank. Now, however, only banks have practical access to the Bank of Estonia.
The robustness currency boards confer on exchange rates, in comparison with conventional pegs, does not come without a price. The operation of such an arrangement entails a number of restrictions—more than are entailed by a peg—on the conduct of fiscal and monetary policy.
Government finances. One of the most important conditions is that there be no lending by the currency board to the government. For the link between reserves and base money to be credible, the assets used as backing must be external. Modest compromises are nonetheless sometimes made. For example, Argentina allows—in principle—a portion of the backing for base money to be in the form of dollar-denominated Argentinian Government debt, although the central bank is prohibited from extending credit directly to the public sector. If domestic assets are employed, they should never be denominated in local currency; this would undermine the central bank’s ability to honor its guarantee.
Because the government cannot obtain financing from the central bank, it must turn to commercial banks or bond issues to meet domestic financing needs. If commercial banks are to have any resources left to lend to the private sector, or if the appetite for bonds is limited, the domestic government borrowing requirement must be fairly modest—ideally, zero. It is notable that the Government of Hong Kong has traditionally run a surplus. Argentina’s public deficit, very large for many years, was brought under control before the currency board was established. In Estonia, an important factor in the success of the kroon, the new currency, was the introduction, on the day of the reform, of measures to reduce the ex ante government deficit for 1992 from 4 percent of GDP to zero. The domestic borrowing requirement has remained at zero.
Open market operations. With a pure currency board arrangement, the authorities have no freedom to influence interest rates through open market operations or by sterilizing the monetary effect of changes in external reserves. By forgoing control over interest rates, the authorities gain a more robust exchange rate. While this is the ideal situation, departures from it may sometimes be warranted by the need to smooth out temporary fluctuations in interest rates. It remains essential, however, that there never be an attempt to offset the liquidity effects of sustained capital inflows or outflows. In Hong Kong, occasional intervention to smooth fluctuations in interest rates has occurred. Originally, this was done indirectly, through the foreign exchange market. In 1990, however, the Exchange Fund began issuing bills to enable it to conduct more conventional open-market operations. With the introduction in 1993 of the Liquidity Adjustment Facility—a discount window—the Exchange Fund has taken another step away from being a pure currency board and toward being a conventional central bank.
Open market operations are also conducted in Argentina, but are normally designed merely to smooth out intramonthly fluctuations in the demand for cash. Until the Mexican crisis, outstanding swap operations did not typically exceed 1 percent of the monetary base, although, since then, efforts to contain interest rates have involved a much more extensive use of swaps and rediscounts. In Estonia, though, there are no open market operations or discount window facilities. Given the unpromising economic conditions prevailing when the currency board was introduced, the new system had to be watertight. The introduction, in 1993, of Bank of Estonia bills was intended to enhance the ability of banks to undertake collateralized lending to each other rather than to facilitate open market operations.
While smoothing fluctuations of interest rates through open market operations need not necessarily compromise currency board operations, allowing interest rates to be controlled administratively is definitely inadvisable. Such control would quickly lead to the system’s collapse because it would interfere with the mechanism through which banks bring equilibrium to the distribution of monetary assets and liabilities. If interest rates were too low, for example, banks would be subject to excessive withdrawals—and could collapse. Such controls are, therefore, absent from all three financial systems considered here.
Lender of last resort. Limitations on their ability to influence interest rates also constrain the scope of monetary authorities operating a currency board to act as guarantor of the banking system. Assistance for weak banks should be confined to foreign exchange reserves in excess of the currency board’s requirements and should be extended only when the banking system as a whole is threatened. To maintain the integrity of the currency board, the Bank of Estonia was divided—in an accounting sense—into two separate departments: the Issue Department for the currency board, and the Banking Department for other central banking functions, including emergency loans to the banking system. Foreign exchange in excess of what is needed for backing is transferred from the Issue Department to the Banking Department.
The surplus of foreign exchange in Hong Kong means that the banking system implicitly has very strong backing. A bank collapse threatening the whole banking system has yet to occur, and the treatment of BCCHK (Bank of Credit and Commerce International’s Hong Kong subsidiary) shows that lesser cases cannot expect help. In Argentina, following the implementation of the Convertibility Law, banks were put on notice that they would no longer be able to count on automatic assistance from the central bank. Steps were taken to strengthen capital adequacy requirements. Reserve requirements were kept high, to ensure adequate liquidity to meet potential deposit withdrawals. In 1995, when large deposit withdrawals occurred in the wake of the Mexico crisis, reserve requirements were lowered so that banks could draw on their reserves. But the scale of the banks’ problems was so large that the Argentine authorities had to soften their policy toward lending to banks, which were then also offered swaps and rediscounts to ease their liquidity position. In Estonia, lending to commercial banks occurred only in 1992 and 1994, and was confined to the surplus resources of the Banking Department.
Role of interest rates
Even though, with a currency board arrangement, the exchange rate for cash or base-money liabilities is officially pegged, there is still some scope for market rates to deviate slightly from the official rate. In practice, such deviations rarely exceed 1 percent of parity. Interest rates thus bear by far the largest burden of adjustment to market forces. In principle, with a currency board, interest rates are set at the discretion of the commercial banks—there is no “official” rate. Because of the possibilities for arbitrage, interest rates are unlikely to deviate by large amounts for very long from those applicable to the peg currency, except insofar as there is a risk premium reflecting either the probability of default or of a change in the exchange rate. Nonetheless, changes in demand for local currency will tend to result in significant deviations in the short run, until capital flows can bring cash balances to their new equilibria.
Interest rates in Hong Kong have frequently deviated, both up and down, from US rates, sometimes significantly, but there is a tendency for Hong Kong rates to follow US dollar rates over the long run (Chart 1). The exchange rate peg was briefly put to the test after the Mexico crisis—speculators attacked the Hong Kong dollar during January 11-13, 1995, putting downward pressure on the market exchange rate. In principle, the normal operation of the currency board would have triggered an offsetting rise in interest rates of sufficient magnitude to avert further speculation, but the authorities decided to magnify the effect on interest rates by draining liquidity from the market. Their strategy was the opposite of that of the Argentine authorities at this time, whose room for maneuver proved more limited. After peaking at 12 percent on January 13, Hong Kong interest rates returned to normal, and the speculative attack ended.
Chart 1Interest rate differentials
In Argentina, local interest rates tended to remain above equivalent US dollar rates during 1991-92, partly on account of strong demand for money resulting from economic growth and also, possibly, reflecting residual concerns about the viability of the currency board arrangement. These concerns came to a head in November 1992, when there was a speculative attack against the peso. The authorities let the currency board arrangement translate a loss of reserves and contraction of the monetary base of some 3 percent into market interest rates of up to 85 percent. The speculative attack petered out very quickly, and interest rates returned to normal. The Argentine peso was again subject to speculative pressure following the Mexican devaluation of December 1994. This time, however, the authorities were more concerned about the ability of the banking system to withstand such high interest rates. At first, they resisted the rise in market rates by cutting reserve requirements and offering rediscounts. The reserve loss continued until the authorities allowed interest rates to rise—on March 9, 1995, rates touched 70 percent. This action, combined with a package of tough fiscal measures, the announcement of financial support from the IMF, and support from the international financial community, restored confidence and put an end to the speculative attack. This experience demonstrates the difficulty of choosing between protecting the banking system, on the one hand, and the exchange rate, on the other.
In Estonia, interest rates for deposits and loans in the banking system have been very high since currency reform, but this reflects primarily the risk of default. The risk of default is as great for bank deposits as for loans; this is why both rates have tended to be above equivalent deutsche mark rates. Interest rates on Bank of Estonia certificates of deposit free of credit risk have been generally very close to deutsche mark rates, so realignment does not appear to be a concern. Mexico’s crisis appears to have had few direct repercussions on Estonia’s currency, although reserve flows into Estonia dried up over the first quarter of 1995. Market interest rates showed no reaction.
Inflation and competitiveness
While interest rates tend to follow those of the peg currency over the medium term, the inflation rate may not. The peg currency is usually that of a mature industrial economy, where productivity (and, therefore, income) in the traded-goods sector tends to grow relatively slowly. With nontraded goods (property and domestic services) in inelastic supply, their relative prices may rise faster in developing than in developed countries. This helps to explain why, since 1983, inflation has almost consistently been higher in Hong Kong than in the United States, despite the peg to the US dollar (Chart 2).
Chart 2Consumer price inflation, 1991-95
Relatively fast productivity growth from very low levels may also explain the high inflation rate in Estonia, despite the peg to the deutsche mark. The explanation of consumer price inflation in Argentina is less clear. It may have more to do with increased confidence resulting from Argentina’s successful currency-board arrangement, leading to more consumption and investment, thereby putting upward pressure on prices of nontraded goods. Argentina’s inflation rate has fallen steadily, however, and had largely converged on the US rate by 1994. These arguments should not, however, tempt policymakers to view inflationary developments with complacency. Because the credibility of currency boards depends, in part, on relinquishing the devaluation option, more attention than usual must be paid to unraveling wage and price rigidities. While the currency board itself may be proof against their influence, the real economy will not be.
Benefits outweigh constraints
There is little doubt that the currency board arrangements introduced by Argentina, Estonia, and Hong Kong, all of which were in difficult, albeit different, circumstances have served them well. The contrast between Argentina’s economic performance before and after 1991 is no less striking than the contrast between Estonia’s performance and that of most of the other BRO countries since 1992. The various reforms and departures from the rules of the game that have been witnessed in these countries attest to the constraining nature of currency board arrangements. But, for countries that have chosen to operate currency boards to help establish or maintain financial stability, it is important not to relax these constraints prematurely, because it is to them that currency boards owe their effectiveness.
This article is based on the author’s “Currency Boards: Issues and Experiences,” an IMF Paper onPolicy Analysis and Assessment (PPAA/94/18). September 1994.