The policy responses of authorities (both in the ERM and in countries whose currencies were linked to the ECU) to severe downward pressure on their exchange rates in the summer and early fall of 1992 ran the gamut from exchange market intervention to interest rate policy, to arranging syndicated loans from private markets, to imposing capital controls.
Official Intervention
One way of viewing an attack on a fixed exchange rate is as an assault on the central bank’s accumulated international reserve stock and on its access to international reserve credit. Speculators are buying international reserves from central banks at the fixed parity, expecting the value of those reserves to increase in terms of domestic currency. From that perspective, the supply of international reserves potentially available to the defending authorities in relation to the size of the resources available to the speculators becomes a crucial element in the defense of a currency.
ERM members had access to three sources of international reserves: their own stock of reserves, borrowing from international capital markets, and borrowing from official sources in the form of bilateral credits or loans from the multilateral very short-term financing facility (VSTFF). Each of these sources was tapped during the recent crisis.
Existing Stock of International Reserves
Whether a much larger stock of international reserves would have allowed authorities to resist the realignments and recourse to floating that actually took place, it is clear that official reserve holdings proved to be “small” in comparison with private sector positions. By the end of August 1992, the total stock of reserves (of foreign exchange) held by those ERM countries that were subsequently heavily involved in intervention was about $270 billion.13 Since these reserves were not pooled to ward off the sequential attacks on currencies, only the reserves of the defending central bank and its potential access to borrowed resources could be used to meet the demands of buyers of exchange.14 Those demands were massive. For example, the amounts involved in the September attack on the French franc were five times the amounts experienced in previous attacks: in earlier episodes, the amount of intervention required in one day might be $1.5 billion, whereas on certain days in September, it was $7-9 billion. Reports suggest that the Bundesbank alone spent about DM 92 billion on support purchases of EMS currencies during August and September.15
Borrowing from Private Markets
Several countries whose currencies came under attack resorted to the international syndicated loan markets, where sovereign loans can sometimes be arranged in a matter of days. For example, to signal its support for the ERM parity of the pound, the United Kingdom announced on September 3 a foreign currency borrowing program valued at ECU 10 billion. On September 18, Sweden engaged an ECU 8 billion syndicated loan and revolving credit line to replenish its currency reserves. The total borrowing by sovereign borrowers with weak currencies during and immediately after the crisis amounted to about $30 billion in September and October.16
However, the amount of foreign currency credit that the international banking system can extend even to a sound credit risk like a major ERM central bank is limited. A syndicate of domestic banks is constrained by large-exposure prudential rules and by limits on the amount of deutsche mark interbank lines it can arrange to finance its deutsche mark loan to its government. Such interbank lines are subject to Basle capital standards and are limited by prudential concerns, as is any direct sovereign lending by foreign banks. In addition, central banks are generally reluctant to become dependent on the domestic banking system for credit, preferring to borrow from official sources.
Borrowing from Official Sector
An important element of the ERM is the possibility of obtaining reserves through the VSTFF.17 This facility provides ECU credit for three and a half months, with the possibility of a further three-month extension. The ECUs borrowed are exchanged for the currencies of one or more member countries and used for exchange market intervention. Initially available only for compulsory intervention, the VSTFF was made accessible for intramarginal intervention in the Basle/Nyborg agreement of September 1987. Unlike compulsory intervention, for which VSTFF credit is automatically available in unlimited amounts, access to the VSTFF for intramarginal intervention is limited to twice the member’s debtor quota in the short-term monetary support mechanism. The permission of the central bank issuing the currency is still required for intramarginal intervention, but the agreement provided for a “presumption” that such permission will be forthcoming.18
Examination of the Bundesbank’s balance sheet indicates that claims of the Bundesbank on the European Monetary Cooperation Fund (EMCF) in connection with the EMS averaged about DM 27 billion in January-August 1992. The stock value of these claims increased to DM 52.7 billion in the second week of September and reached a peak of some DM 92 billion in the week ended September 23. Since Germany is the major creditor in the VSTFF, these numbers are probably a good indicator of the use made of this ERM financing facility.19
The creditor central bank, as well as the debtor central bank, may decide to limit the amount of credit for two reasons. First, both take a substantial exchange risk if the borrower finally devalues. Second, a significant volume of lending could well disrupt the monetary policy goals of the creditor bank.
The exchange risk for the borrowing bank—whether bilateral or through the VSTFF—arises because, after a devaluation, it will book a capital loss if it has negative net reserves that must be repaid. The exchange risk for the creditor arises because the VSTFF denominates its claims and liabilities in ECU. In the September crisis, the Bundesbank probably suffered losses well in excess of DM 1 billion in its VSTFF lending to sustain the lira and the pound.
The second constraint arises if the lending central bank, say, the Bundesbank, insists on sterilizing the increase in deutsche mark base money created by the intervention with the borrowed deutsche mark in support of, say, sterling. Such sterilization is carried out by reducing its other assets: repurchase holdings, direct holdings of securities, Lombard loans, or holdings of other foreign currencies. Data indicate that about two-thirds of the net inflows were sterilized.20 Inflows as a result of the support purchases of EMS currencies of about DM 92 billion “presented quite exceptional challenges to the Bundesbank’s money market management.”21 Although the Bundesbank could have sterilized significantly more than it actually did, this would not have been without some disruption. For example, currency swaps of non-ERM currencies with the banking system were undertaken at rising deutsche mark interest rates, which was viewed as self-defeating. Ultimately, if the amounts that need to be sterilized become too large, the creditor bank faces the unattractive prospect of abandoning its interest rate policy—which, in turn, places a ceiling on the amount that the weak-currency central bank can borrow from the strong-currency one.22 The greater the difference in the strong-currency country—at the time of the exchange rate attack—between the interest rate requirements for defending the exchange rate and those for maintaining internal balance, the more binding this sterilization constraint is likely to be.
A related factor that can limit willingness to extend credit through the VSTFF is the number of currencies in the system that are simultaneously under siege. The Bundesbank’s (practical) limits are more likely to be hit when, for example, the Bank of Italy and the Bank of England and the Bank of France simultaneously appear as potential borrowers, than when just one is involved. The issue is whether any one central bank can reasonably be expected to match chips with the huge pile (from the private sector) that is now on the other side of the poker table. Indeed, for this reason interest in proposals for pooling the reserves of central banks continues.
Last, the effectiveness of exchange market intervention is conditioned by the flanking policies that are implemented with it. Such policies of course include interest rate policy. In this regard, private market participants have noted that the unequivocal strong joint statement of support for the existing parity by the Bundesbank and the Bank of France helped to relay the message that those betting on a devaluation of the French franc would be fighting two central banks rather than just one. Hence, the quality of exchange market intervention can be as significant as the quantity. On a broader level, it is merely another illustration of the proposition that the quality of monetary policy cooperation is itself an important element in the credibility of exchange rate parities.
Defensive Increases in Interest Rates
The classical prescription for a central bank facing an unwarranted run on its currency is to raise its own lending rate to banks in order to “squeeze” short sellers of funds denominated in its currency.23 The basic idea is to make short sellers pay dearly for the domestic funds they need to borrow for delivery under the forward contract. If the increase in lending rates filters through to other interest rates, it also serves to attract additional demand for assets denominated in the home currency. More fundamentally, the essence of the credibility of a fixed exchange rate is in convincing markets that the authorities will take whatever action is needed on interest rates to defend the exchange rate commitment—even if in so doing, they incur other costs. The rub, of course, is that circumstances can arise in which markets perceive the costs of that action as being so high as to raise doubts that the authorities will be both willing and able to sustain them. Four of the more important side effects of increases in interest rates in the weak-currency country should be mentioned.
First, the same increase in interest rates that serves to squeeze speculators can simultaneously squeeze the funding of securities and banking markets, which typically finance their positions through short-term rollover credit.24 Since central banks also carry the responsibility for maintaining the safety and soundness of the financial markets, they have to be careful about inducing an extreme punitive rise in central bank lending rates, except perhaps over the very short term. If some large banks are already laboring under heavy loan losses—or even worse, are on the brink of needing government assistance to remain solvent, rises in interest rates exacerbate their problems and may increase the fiscal deficit (if explicit or implicit government insurance of commercial bank liabilities is present).
Second, in situations where the household and corporate sectors have allowed their debt-to-income and debt-servicing positions to become unsustainably high, a sharp rise in interest rates makes it more difficult to reduce those ratios to more manageable levels—without simultaneously reducing spending to such a degree as to slow economic activity appreciably. Where the economy is already in recession, the problem takes on an added dimension because of the risk that a protracted period of high interest rates could force the pace of disinflation to become too rapid, and perhaps even push the economy into a debt-deflation cycle.
Third, in countries with both extremely weak fiscal fundamentals and a large share of short-term and floating-rate government debt, a large increase in interest rates can feed back quickly and powerfully to increase the government’s fiscal deficit. At some point, increases in interest rates may actually weaken the attractiveness of the domestic currency if market participants believe that they increase debt-servicing problems or if interest volatility presents asset holders with unacceptably high levels of market risk.
Fourth, high interest rates—maintained mainly to defend an exchange rate target—will often be viewed as having a high opportunity cost in terms of domestic economic activity, particularly where the economy has been in recession, where unemployment rates are high, where inflationary pressures are moderate and receding, and where the consensus forecast is for slow growth in the period ahead. The greater the current differences are between the domestic and the external requirements for monetary policy in the weak-currency country, the more likely it is that questions will be raised about the wisdom of “tying one’s hands” on monetary policy—whatever the long-term benefits of such a policy strategy. In this situation, increases in interest rates will be politically unpopular. In fact, it is precisely because this potential political fallout may limit the central bank’s flexibility to increase interest rates quickly, either in anticipation of or after an attack on the currency, that some observers believe that central bank independence is necessary for true interest rate credibility.25
In addition to the potential costs of making defensive increases in interest rates, central banks sometimes also have to contend with the possibility that such action may not always have its intended effect on speculators or on the exchange rate. If short sellers of the domestic currency have put their financing in place before the defensive increase in interest rates comes about, they will not be squeezed by the rate rise—unless the monetary authorities are prepared to hold rates high for a long period. Recent developments in derivative markets may also have weakened the power of the classical interest rate defense. In some dynamic hedging strategies, a rise in the spread between foreign and domestic interest rates actually mandates a forward or spot sale of the domestic currency.26 Thus, raising the interest rate in the classical manner could perversely induce sales of the currency being defended; the more widespread such dynamic trading strategies become, the more central banks would need to take this automated response into account.
Several of the constraints and concerns outlined above were evident during the recent currency turmoil.27 Sweden was a textbook example of the limits of an aggressive interest rate defense under conditions of a large fiscal deficit and significant financial fragility. The three-month siege of the krona began on August 21 with a sequence of increases that moved the Riksbank’s marginal lending rate from 13 percent to 500 percent on September 16. The pass-through to other interest rates was rapid. For example, the ask rates for interbank one-month funds jumped from 16 percent to 70 percent between September 7 and September 17. The large interest rate increases imposed more losses on an already troubled banking system, directly through rises in funding costs and indirectly by exacerbating the already existing debt-deflation problem. These losses were imposed in an environment of ongoing capital injections into the banking system and a large and mounting fiscal deficit. In the end, after failing to reach agreement with opposition parties on how to bring the fiscal deficit under control, the authorities had reluctantly to abandon their peg to the ECU and permit the krona to float.
In contrast to Sweden, the United Kingdom was a case where the authorities attempted to defend the currency with minimal use of the interest rate, opting instead to rely on massive intervention in the spot foreign exchange markets. Only when it became clear that intervention alone would not stem the tide did the monetary authorities resort to increases in interest rates. Here, the relevant background included two years of recession, excessive debt accumulation in the private sector, and a financial market structure under which any rise in the Bank of England’s minimum dealing rate spreads relatively quickly to the whole yield curve, including to retail markets and mortgage markets. After its minimum dealing rates had been raised from 12 percent to 15 percent on September 15, it became clear that these changes had no impact on the selling pressure: traders, apparently recognizing the high costs of maintaining this rate for long, continued to sell sterling, and the exchange rate did not budge from its floor value. The Bank of England then ceased the defense of the floor parity and dropped the minimum dealing rate back to 12 percent. By October 16, it had dropped the minimum dealing rate to 8 percent.
In Italy, it was apparently the link between interest rates and the fiscal deficit that led traders to question the sustainability of high interest rates. Money market interest rate rises have a large impact on Italian Government finance because of the total size of government debt and because large shares of the debt are in the form of treasury bills (BOTs, 29 percent) and floating-rate securities (CCTs, 48 percent); the rest of the debt is primarily medium- and long-term bonds (BTPs).28 Nonetheless the Bank of Italy did make significant use of the interest rate defense throughout the attack on the lira; it raised in steps its discount rate from 12 percent to 15 percent, allowing the interbank overnight rate to reach 36 percent on September 16.
If part of the constraint on using increases in interest rates to defend a fixed exchange rate arises because innocent bystanders as well as speculators are hit by the same higher cost of credit, can something be done to restrict the squeeze to the speculators—or at least to limit the spread of higher interest rates to the rest of the economy? The French defense provides an example of one effort in this direction—albeit one that can only be used for a short period and only when the banks themselves are prepared to cooperate. Although initially reluctant to raise interest rates in the early stages of the attack, the French authorities increased rates as the attack against the franc intensified. Banks with normal commercial requirements received relatively cheap funding that could be passed through to customers to avoid a rise in the base rate (about 15 percent of bank credit depends on the base rate, which governs loans primarily to small businesses). Thus, many regular customers could obtain near normal interest rates, while others had to pay higher rates. Lending was distinguished not by the nationality of the buyer but by the class of business. This temporary credit allocation mechanism had some success in raising the cost of credit for those market participants who did not have either access to appropriate paper (for discounting) or preferential access to the Bank of France’s repurchase facility.29 It may also have helped to share the burden of the cost of interest rate adjustment in a way that was less politically unpopular than a more uniform increase in rates would have been. But it is clearly a tactic that can only be used as a very short-term instrument of crisis management, since going beyond that would unduly strain the banking system and, more fundamentally, would risk the efficiency gains from following a market orientation to the allocation of credit.
Under a fixed exchange rate arrangement, the costs associated with asking weak-currency countries to adopt large increases in interest rates are normally diminished by having the strong-currency country share the adjustment burden by reducing its own interest rates. In the run-up to, and during the ERM crisis itself, this course of action was severely constrained by the Bundesbank’s assessment that it would be incompatible with controlling inflationary pressures within Germany.30 In short, the level of interest rates that many other ERM countries deemed appropriate for dealing with their domestic economic situations was different (lower) than the level of rates that the anchor country in the ERM deemed appropriate to its own domestic economic conditions and responsibilities. Markets are sensitive to such conflicts over the appropriate course of monetary policy in an area of fixed exchange rates, and may have reasoned that the anchor country would have preferred a realignment of the deutsche mark against many other ERM currencies to a large reduction in German interest rates. That of course, plus the cumulative losses of competitiveness in some ERM countries with relatively high inflation rates and the constraints on interest rate increases in some weak-currency countries, created the appearance of a “one-way bet” for speculators that merely fueled exchange market pressures.
Useful information is also to be gained from the countries whose currencies were not attacked during the crisis—particularly the Netherlands and Belgium-Luxembourg. These countries have maintained close links with the deutsche mark and have consistently refrained from any efforts to decouple their monetary policies from that of the anchor country. In so doing, they have apparently been successful in convincing markets that their commitment to the exchange rate parity is paramount among monetary policy objectives. Even though high interest rates have some elements of vulnerability (e.g., the high debt/GNP ratio and the relatively large fiscal deficit in Belgium), markets have presumably decided that, if necessary, those costs would and could be absorbed to protect the longer-term benefits from exchange rate stability. Because these countries are relatively small, any decoupling option (of either interest rates or exchange rates) may prove less attractive. In any case, it is easy to see why these countries might argue that taking early action on interest rates at the first sign of capital market pressure (as well as supporting rates with intramarginal intervention along the lines of the Basle/Nyborg agreement) is the right preventive medicine for avoiding crisis-induced realignments in the ERM.
Capital Controls
As is well known, three ERM countries (Spain, Portugal, and Ireland) reluctantly resorted either to capital controls or to a tightening of existing regulations during the exchange market crisis (see Annex VI for a more detailed description of these measures). A sudden imposition of capital controls can almost always reduce the short-term pressures that a central bank faces in an attack on its currency. The controls operate by breaking the speculator’s (or hedger’s) chain of finance to the domestic banking system—thereby inhibiting his ability to fund a redenomination of his currency position. Over time, channels of avoidance typically grow, and the initial slow drain of reserves accelerates (unless the fundamentals that initially motivated the attack are addressed). But the most damaging effects show up in the longer term. Since investors have had the experience of seeing their liquid investments transformed by controls into illiquid claims, it will be much harder in the future to attract portfolio investment inflows and to assure market participants that open capital markets are an essential element in a country’s long-term outward-looking policy strategy. In most cases, investors will demand in the aftermath of controls a risk premium on the country’s liabilities beyond the usual markup for depreciation risk. For this reason countries faced with massive and unrelenting capital market pressures often conclude that it is less costly over the long run either to realign the parity or to resort to temporary floating than to impose capital controls.31