Foreign Exchange Hedging and the Interest Rate Defense

In the endgame of a fixed exchange rate regime, increases in interest rates to defend the currency may lead to an apparently perverse market response: further downward pressure on the exchange rate. This may result if a large proportion of investors’ foreign exchange exposure is dynamically hedged. This paper describes the trading operations involved in implementing dynamic hedges and the impact of these operations on central bank policy. The success of an interest rate defense hinges on the size and timing of the funding operations of those who are being squeezed relative to those engaged in dynamic hedging. [JEL F3]

Abstract

In the endgame of a fixed exchange rate regime, increases in interest rates to defend the currency may lead to an apparently perverse market response: further downward pressure on the exchange rate. This may result if a large proportion of investors’ foreign exchange exposure is dynamically hedged. This paper describes the trading operations involved in implementing dynamic hedges and the impact of these operations on central bank policy. The success of an interest rate defense hinges on the size and timing of the funding operations of those who are being squeezed relative to those engaged in dynamic hedging. [JEL F3]

Coincident with the internationalization of portfolios and the interlinking of money markets across currencies has been the expanded use of methods to hedge currency risk. While basic hedging instruments such as forward exchange contracts have a long history, the use of newer instruments such as exchange-traded options and futures contracts and over-the-counter (OTC) options and currency swaps has grown dramatically in the past decade. In addition, option pricing methods have been used in dynamic hedging strategies to construct tailor-made synthetic derivative products—a transplantation to currency markets of the portfolio insurance methods used to hedge equity market exposure.

The stock market crash of 1987 led to justifiable skepticism about the ability of mechanistic trading strategies like dynamic hedging actually to deliver the intended hedge protection when markets are illiquid.1 In addition, these strategies have been criticized for their tendency to exacerbate price trends. Such criticisms carry over to the use of dynamic hedging in currency markets, although currency markets are usually among the most liquid of financial markets.

In this paper, we examine the impact of dynamic hedging strategies on foreign exchange markets during those crisis periods when even the exchange markets can become illiquid. Though we place some emphasis on the well-known inability of these strategies to perform well for the hedger when a discontinuity in the exchange rate or an upsurge of volatility occurs, we are concerned primarily with the impact of hedging strategies on the efficacy of the classic central bank interest rate defense of a fixed exchange rate. It is generally believed that a central bank can defend an exchange rate if it is willing to raise short-term interest rates sufficiently high to squeeze speculators who are short in its currency. In the presence of dynamic hedging, however, mechanistic selling of the domestic currency may arise in the endgame of the interest rate defense, and this may overwhelm the credit lines available to the central bank for intervention in the exchange market before those squeezed by the interest rate increase start to buy. Thus, our ultimate focus is on market and central bank behavior in the crucial last moments of a fixed exchange rate, the boundary point toward which the collapsing system converges.

The paper is organized as follows. In Section I, we outline the growth of the foreign exchange markets in general, and the markets for currency derivatives in particular. In Section II, we consider the hedging operations of nonbanks and the techniques in general use. In Section III, we examine the theory and practice of dynamic hedging, and the hedging operations of banks in particular. In Section IV, we examine the mechanics of central bank currency intervention and the effect of interest rate defenses on market liquidity, focusing particularly on the response of dynamic hedging programs to interest rate increases. We also consider how the interaction between the timing of different trading programs—dynamic hedging versus closing positions to avoid a squeeze—and the credit lines of the central bank may force the central bank to abandon a fixed exchange rate if it is driven either to the limit of its credit lines with official lenders or commercial banks or to its self-imposed position limit before buyers of the currency arrive. Section V contains concluding remarks.

I. Markets for Foreign Exchange Products

The foreign exchange market is the largest financial market in the world, with average daily turnover in April 1992 estimated at $880 billion, compared with $620 billion in April 1989.2 The largest market segment is that for spot delivery—generally two days later—which accounted for just under half of the turnover in 1992, followed by the market for foreign exchange swaps, which accounted for 40 percent of turnover. The proportions of turnover resulting from outright forward deals, options, and futures were 7 percent, 5 percent, and 1 percent respectively.

The market is largely an inter-dealer market: 84 percent of transactions were made among financial institutions and other foreign exchange brokers and dealers in 1992. This characteristic is reflected in the average deal size, which for the U.S. dollar was approximately $6 million overall. Deals were relatively smaller in the spot market, in which the proportion of transactions with end users was higher, while in the derivatives markets deals tended to be larger. For example, the average size of an OTC deutsche mark-pound sterling option was $32 million.

The bulk of foreign exchange market activity still involves the U.S. dollar on one side of the transaction. The dollar was involved in 83 percent of all deals in 1992—including 72 percent of spot trades and 95 percent of swaps contracts—although this proportion had fallen since the previous survey conducted by the Bank for International Settlements (BIS) in 1989. Transactions between currencies in the European exchange rate mechanism (ERM) accounted for only 7 percent of aggregate turnover.

The 1992 survey results indicate how rapidly the use of financial derivatives has grown in recent years. While spot turnover increased only 14 percent between 1989 and 1992, forward transactions increased 60 percent, as did turnover in currency futures. Swaps trading increased 50 percent, and options trading increased by 124 percent.

The notional value of outstanding exchange-traded and OTC financial derivative contracts—including futures, forwards, forward rate agreements, swaps, options, caps, floors, and collars—grew from approximately $7.2 trillion at end-1989 to $17.6 trillion at end-1992.3 The 1992 notional amounts comprised $4.8 trillion in exchange-traded contracts, $4.7 trillion in swaps and $8.1 trillion in OTC options and forward contracts. By expanding the opportunities for borrowers and lenders to change the risk characteristics—such as maturity or currency denomination—of their portfolios, the growth in these markets has dramatically altered the nature of international finance and the behavior of market participants.

Exchange-traded derivative products—futures and options—are standardized, retail-sized products. Though they are retail in nature they are frequently used by the dealers in OTC markets to balance positions when credit lines with other financial institutions are filled or when wholesale counterparties are hard to find. Because the exchange’s clearinghouse is the counterparty to each contract and because positions are usually well collateralized through margin requirements, the evaluation of creditworthiness is less of an issue on organized exchanges than in the OTC market.4 The most actively traded financial derivatives on organized exchanges are futures on interest rates, primarily U.S. Treasury bond rates, Eurodollars, French government bonds, German Bunds, and Japanese government bonds.

The OTC markets in derivative products are concentrated in a small number of large banks and securities firms in the major financial centers. For example, bank holding companies with more than $10 billion in assets hold between 98 percent and 100 percent of all OTC derivative positions taken by U.S. banks.5 OTC contracts are often designed specifically for the needs of particular end users and therefore have tailor-made features such as maturity, currency denomination, and notional principal, and are frequently combined with other derivatives and sold as a package. Many OTC trades are inter-dealer trades in which dealers seek to balance their positions.

Foreign exchange derivatives are important components of these markets, particularly the OTC markets. While the notional principal of outstanding exchange-traded foreign exchange derivatives at the end of 1992 was only $105 billion, there were $860 billion in currency swaps and $5.5 trillion in foreign exchange forwards and OTC options outstanding. In contrast, the notional principal of outstanding interest rate products was $4.4 trillion in exchange-traded contracts, $3.9 trillion in swaps, $634 billion in OTC options, and $2 trillion in forward rate agreements. Stock index derivatives totaled $245 billion.

The Markets for Currency Options

The market segment of particular interest in this paper is the market for options. The options market is divided into two parts: the market for exchange-traded options, and the OTC market. Currency options are listed on six exchanges in Europe and North America, but most trading takes place on the Philadelphia Stock Exchange (PHLX) and the Chicago Mercantile Exchange’s International Monetary Market (IMM). PHLX lists options on spot currency, while most of the contracts on the IMM are options on currency futures. Most listed options are available with a limited choice of maturities up to one year and have American exercise characteristics.6 In the OTC market, contract specifications are in principle negotiable, although there is considerable standardization. Furthermore, OTC options are options on spot currency rather than options on futures, and the European exercise convention is the norm. OTC options are also contracted in much higher amounts. While an individual deutsche mark contract on PHLX provides an option to buy or sell DM 62,500, options in the OTC market are written for amounts of at least $1 million.

The BIS data show that activity in the OTC market segment dominates total trading in currency options, accounting for 85 percent of turnover in April 1992. The U.S. dollar dominates the options markets: 98 percent of exchange trading and 74 percent of the OTC market’s turnover involved the dollar on one side of the transaction. Assuming that the currency composition of deutsche mark OTC options trading is representative of that for the other ERM currencies, only an estimated 10 percent of the OTC options market involves intra-ERM transactions. Moreover, two thirds of the banks’ options transactions, measured by notional principal, had other banks or dealers as counterparties.

While the BIS (1993) does not provide data on the maturity structure of the options market, it does provide it for forward contracts. These show that maturities are heavily concentrated in the near term: 64 percent of contracts mature within seven days, and only 1 percent have a maturity of longer than one year.

II. The Demand for Hedging in Foreign Exchange Markets

Open positions denominated in foreign currencies expose market participants to losses from exchange rate changes. Accounting for such risk is vital for portfolio managers with foreign currency exposure, corporations with foreign-currency-denominated assets or liabilities such as receivables or payables, or banks with currency exposure. These risks can be reduced by taking an offsetting position in the foreign currency. For example, a long position is hedged by shorting the currency in some fashion. This may consist of a spot sale with borrowing in the foreign currency to cover settlement, the purchase of a forward or future contract that locks in the level of the exchange rate for future payment, or the acquisition of a put option or the sale of a call option on the currency. Access to these instruments differs across types of hedgers: exchange-traded futures or options are retail products with little credit risk but with relatively high margin requirements; OTC products provided by banks and nonbank dealers are typically offered in much larger notional values and require a credit line from the bank to the customer along with a bank’s assessment of its exposure to a given client. Options generally provide a partial hedge. For example, a portfolio manager may buy a put option to ensure a floor to the domestic currency value of the foreign currency component of its portfolio, but the portfolio remains subject to risk of currency fluctuations while the portfolio value is above the floor.7

At the end of 1991, institutional investors—mutual funds, pension funds, and insurance companies—in OECD countries had total assets of approximately $11.7 trillion, compared with the assets of commercial banks, which totaled $19.6 trillion.8 The sizes of their foreign currency exposures in absolute terms and even in relation to their total assets can be quite large. For example, U.S. mutual funds and pension funds held $214 billion in foreign assets, or 5 percent of their combined end-1991 assets of $4.1 trillion. In contrast, U.K. mutual funds and pension funds invested $151 billion abroad—23 percent of their total assets. Institutional investors in Germany, Japan, and the Netherlands also invest sizable proportions of their assets abroad. More significant, perhaps, there are few restrictions on the foreign investments of institutional investors in industrial countries, and the trend appears to be toward relaxing those constraints that do exist. Banks, in contrast, often have well-defined position limits—either statutory or self-imposed—on their foreign exchange exposures.

Managers of pension funds, mutual funds, and bank trust accounts generally hedge their currency risk, often using dynamic hedging operations to create synthetic securities. For fixed-interest holdings of pension funds with obligations denominated in a given currency, the hedge reflects the desire by fund management to place a floor on the long-term value of foreign-currency-denominated holdings. For funds investing in foreign equities, the long-term reasons for establishing currency hedges are not as obvious because of the long-run tendency of exchange rates to conform with purchasing power parity. Nevertheless, in the short term—on a quarterly or annual basis—fund managers’ performance, and therefore their compensation, is often compared to a benchmark. Moreover, fund managers seek to protect short-term performance from significant declines to prevent an increase in redemptions. Similarly, for pension funds, underfunding of liabilities may force an injection of securities into the fund that tests the liquidity of the parent entity. For these reasons, fund managers are sensitive in the short term to exchange rate movements and will wish to hedge positions. In the simplest hedging operation, fund directors may establish currency risk targets or limits to which management must adhere by following agreed hedging strategies. To place an absolute ceiling on losses from currency fluctuations, fund directors may mandate the acquisition of a put option to cover the entire foreign exchange position of the fund.

If they are willing to bear more risk from volatility changes, fund directors may instruct management to replicate a put dynamically.9 This method has become typical for fund management. As indicated below, this buy-high, sell-low strategy will, ex post, have been less costly than an actual put if volatility declines, and more costly if volatility increases. Finally, many portfolio managers follow a constant-percentage portfolio insurance strategy: this is a buy-high, sell-low dynamic strategy that does not replicate a put option.10 Rather, it is driven entirely by price movements. For example, one realization of this strategy may aim at outperforming a 50-percent hedged position and would begin with a 50-percent hedge. A 1-percent move in the exchange rate would trigger an x-percent change in the hedge ratio. If the foreign currency appreciated by 10 percent, the hedge ratio would fall to 50 – 10x percent. Currency depreciations would be met with opposite adjustments in the hedge ratio. The strategy tends to work well when exchange rate changes come in trends but fails with a jump in volatility.11

Dynamic strategies are often implemented through cross-hedges—that is, a hedge may be implemented through shorting a currency whose exchange rate is highly correlated with the currency in which the fund holds securities. The purpose is to take advantage of greater liquidity in the exchange market or an interest rate premium in the currency used for the cross hedge.

Individual firms and portfolio managers ultimately must turn to banks to engage in foreign exchange hedging since banks are the principal dealers in the foreign exchange spot and derivatives markets. By taking the opposite side of a transaction undertaken by a customer, a bank will acquire foreign exchange exposure that it will then attempt to eliminate. For those exposures that do not net out in the course of a day’s trading with other customers—for example, currency or value-date mismatches in forward contract long and short positions or different features of options contracts—the bank must actively seek coverage by initiating its own transactions in the same OTC and exchange-traded derivatives markets or in the underlying markets.

Because of internal risk-control operations and regulations on foreign exchange risk, banks are active in using dynamic hedging techniques. Typically, they will hedge the net exposure to exchange rate changes acquired through transactions with clients, but they may leverage exchange risk when trading for proprietary accounts.

Regulation on banks’ net foreign exchange positions varies widely across industrial countries.12 In some countries, such as the United States, banks’ exposures and internal controls are monitored on a regular basis, although there are no specified limits. Elsewhere, as in, for example, Germany, Japan, and the United Kingdom, guidelines or stronger constraints limit open positions to a specified ratio to total capital. Banks’ internal risk management controls include the separation of dealing operations—in which buy/sell orders are taken—and back-office activities where contracts are confirmed and settled, the imposition of open position limits on the dealing book, and limits on the extension of credit to individual counterparties.

A bank that writes an option becomes exposed to the possibility that the option will be exercised and it will have to buy or sell foreign currency (depending upon whether it has written a put or a call). The simplest hedge in this case would be to acquire a perfectly offsetting contract. For a bank that maintains a large options book, many of its options contracts will indeed offset each other. However, to hedge the remaining options exposure, banks will generally turn to the more liquid underlying markets and hedge their exposures by creating synthetic options. Dynamic hedging strategies provide a simple means by which complicated options books can be hedged by constructing synthetic options.

As the discussion in Section I indicates, trading in options is only a small part of the foreign exchange market. Most of banks’ foreign exchange exposure comes from dealing in the spot and forward (including swaps) markets. As with their options-based exposures, banks will actively hedge their net exposure arising from these other transactions. Moreover, not all options (or other) transactions entered into by banks are derived from hedging operations. Unlike transactions in the underlying markets, options provide tools for taking positions in the volatility of spot exchange rates or exchange rate futures, instead of or in addition to speculating on the future direction of these underlying assets. Banks both sell packages of options to their customers that allow them to choose their own degree of exposure to the level, direction of change, and volatility of the underlying asset, and enter into transactions with other dealers to do the same for their own account.

III. Mechanics of Option Pricing and Dynamic Hedging

Banks and other large institutions active in the foreign exchange market use dynamic hedging techniques extensively to reduce the risks they incur from their foreign exchange trading. Dynamic hedging involves, in its essence, taking positions in domestic and foreign currency loans (or bonds) that together replicate a put option to hedge a long position in foreign exchange. This section explains how this hedging portfolio is constructed and how it is affected by changes in the foreign interest rate. The key implication is that under fairly general conditions an increase in the foreign interest rate will result in an increase in the short position in foreign currency.

Pricing Put Options on Foreign Currency

Because option pricing theory is at the heart of dynamic hedging, it is helpful at this point to review the basic option pricing formula for foreign exchange—the Garman/Kohlhagen formula—before describing how dynamic hedging works.13 Although banks and other wholesale traders may use more sophisticated pricing methods that account for varying interest rates and exchange rate volatility, the Garman/Kohlhagen formula is in general operational use by pension fund and other portfolio managers and it is pedagogically useful for illustrating the management of risk in a bank’s foreign exchange book.14

Suppose that a customer buys a European put option to deliver deutsche mark for dollars after T periods for an exercise price of $X per deutsche mark. The value of the put option, Pt, is:

Pt=[1N(d1)]exp[rDMT]S+[1N(d2)]exp[r$T]X(1)

where rDM and r$ are the (constant) risk-free instantaneous deutsche mark and dollar interest rates, S is the current dollar/deutsche mark spot exchange rate, and X is the exercise or strike exchange rate of the option.15 N(d1) is the value of the normal distribution function evaluated at the argument

d1={ln(S/X)+(r$rDM+σ2/2)T}/σT,(2)

where σ is the (constant) instantaneous standard deviation or volatility of the exchange rate S. Finally, d2=d1σT. The put price, or premium, is graphed against the spot exchange rate in Figure 1. The premium is a downward-sloping, convex, function of the exchange rate and lies closer to the option’s intrinsic value, max [0,X – S] (depicted as the dashed 45° line, which coincides with the horizontal axis to the right of X) the shorter the time to maturity. Note that it is possible for the price of the option to be less than its intrinsic value for deep-in-the-money puts.

Figure 1.
Figure 1.

Currency Put Option Premium and Delta

Citation: IMF Staff Papers 1995, 003; 10.5089/9781451973396.024.A002

Implementing the Dynamic Hedge

The put pricing formula is determined by finding the short position in deutsche mark loans and the long position in dollar loans such that a portfolio with these positions and also short a put is riskless with respect to small exchange rate movements.

Thus, an investor that wants to hedge its exposure to fluctuations in the dollar/deutsche mark exchange rate can either hedge a long deutsche mark position by buying a put option, or use equation (1) to determine positions in deutsche mark and dollar loans that mimic the value of a put—that is, to create a synthetic put. The basic security in the first half of the formula is a loan promising to deliver one deutsche mark in T periods—this has a deutsche mark present value of exp(–rDMT) and a dollar value of exp(–rDMT)S. The coefficient –[1 – N(d1)] indicates that the mimicking portfolio should consist of a short position of a fraction of such a deutsche mark loan—that is, a short deutsche mark position. Similarly, the dollar position is long a fraction [1 – N(d2)] of a loan promising to pay X dollars in T periods with a present dollar value of exp(–r$T)X. However, since d1 and d2 constantly move with the exchange rate, the interest rate differential, and the standard deviation projected for exchange rate movements, the positions must be adjusted constantly—hence the term dynamic hedging—to maintain the equivalence of the position to a put option.16

The foreign exchange exposure of the bank that sells the put is to the possibility of having to buy deutsche mark at the exercise price at date T. Under the assumptions behind the pricing formula, it is not necessary to hedge the total face value of the contract prior to the exercise date. How much of the face value to hedge, which in turn determines the hedge ratio, is determined by the option’s delta, the change in the value of the option with respect to a movement in the exchange rate. From the pricing formula developed above, the delta of a currency put option is –[1 – N(d1)]exp(–rDMT). Thus, a rise in the dollar value of the deutsche mark makes it less likely that the option will be exercised and reduces the value of the put. The put delta takes values between –1, for a deep-in-the-money option that would almost certainly be exercised, to 0, for a deep-out-of-the-money option that would never be exercised (see Figure 1). The negative of delta, therefore, provides a proxy for the probability of exercise. Delta multiplied by the number of units of foreign currency provides an estimate of the expected foreign exchange that is sold short at any point in time to hedge against possible exercise of the option.

A writer of a put option may, therefore, hedge the option dynamically according to the prescriptions of the put pricing formula. First, it must establish the portfolio that mimics the value of the option, for example, by shorting [1 – N(d1])]exp(–rDMT) deutsche mark spot for dollars and buying [1 – N(d2)]exp(-r$T)X in U.S. Treasury bills. As the exchange rate fluctuates, the now-hedged writer of the option must adjust the short deutsche mark and long dollar positions according to the formula to continue to mimic the option. Typically, the adjustments will not be continuous; instead, to avoid transactions costs, adjustments to the mimicking portfolio will be made as part of a regular rebalancing exercise.17

Among other assumptions, the put pricing formula is based on the assumption that exchange rate volatility will remain constant during the life of the contract. Because volatility typically is not constant, the mimicking portfolio will never perfectly track the actual option’s value—gains or losses relative to the initial option premium will always occur—and so the portfolio must constantly be adjusted to changes in volatilities as measured, frequently, by implied volatilities in options prices. If volatility jumps above the value implicit in the price of the actual put option, the writer of the put who engages in dynamic hedging will take a loss and the buyer of the put will gain. It is well-known that strategies to create synthetic options to hedge actual options through the use of dynamic trading, designed to be delta neutral, can be used to take positions on volatility in underlying prices and in interest rates.18

The loss to the writer is immediately apparent if the portfolio is marked to market. A volatility increase will, ceteris paribus, increase the value of the actual option (a liability) and leave unchanged the value of the hedging portfolio (the supposedly balancing asset). Alternatively, if the option value is not marked to market, the loss will be booked through the dynamic adjustment of deutsche mark and dollar positions until the exercise date. According to the hedging strategy, a rise in the exchange rate will cause the writer of the put to reduce the short deutsche mark position: the writer of the option will buy deutsche mark when the deutsche mark appreciates and sell when it depreciates. This “buy dear-sell cheap” strategy generates a foreseeable loss to the writer of the put for which the writer is compensated by the put premium. If volatility jumps, however, the premium will be insufficient to cover the now greater-than-expected realized loss on these hedging trades.

To hedge the risks acquired from their OTC options transactions with other dealers, banks generally construct a dynamic hedge by purchasing or selling currency in the spot market to close the currency exposure, and entering into a swap contract to shift the exposure to coincide with the maturity date of the option. Indeed, such transactions have become part of banks’ normal operating procedures (Walmsley (1992)).

In order to monitor its overall exposure, a bank must have a method to break down each option in its book into its implied foreign exchange position. It can then determine its global net position in each currency by adding its net position from trading in other foreign exchange products to its net position implied in its options book, and then hedge the combined exposure. The foreign exchange equivalent into which a bank will decompose its options will depend on the currency options pricing formula used by the bank, but it will usually be based on delta hedging methods. The bank calculates the delta for all the contracts it has written or bought and multiplies these by the face values of the contracts. These are then added up for each currency to estimate the expected net foreign currency delivery requirement. For European-style options, in which exercise is only possible at maturity, the hedge portfolio will include futures or forward contracts that offset these amounts, while for American-style options, the hedge will include cash positions because the exercise date is uncertain. Because the management of the foreign exchange book is global, the amounts required to hedge the options will be netted against spot and forward net positions.

For example, suppose that the global position in the currency option book of a bank making a market in derivatives is short one OTC European deutsche mark put option that allows the holder to sell DM 1 for $X at time T, and long one European put option to sell DM 1 for French francs at T*. If the bank uses the Garman/Kohlhagen formula, its deutsche mark position from its options book is:

article image

In these formulas, d1 and d1* are defined as above with the appropriate volatilities and exercise prices substituted for each option. If the bank is also long deutsche mark in its forward and spot trading, it can determine its global foreign exchange exposure in deutsche mark by adding these three quantities. The bank can then hedge the foreign exchange risk by taking the opposite position in the forward market. Because the implied delivery dates across its deutsche mark contracts may differ, this still leaves the bank with an interest rate risk that can be hedged through appropriate deutsche mark forwards or swaps.

Response of the Put Option Delta to Changes in Foreign Interest Rates

The previous discussion identified the option delta as a crucial determinant of the hedge portfolio. Because the discussion to follow focuses on the response of investors with dynamic hedging programs in place to increases in the interest rate in a foreign country whose currency is under attack, this subsection examines the effect of changes in rDM on the put option delta.19 In most circumstances, delta increases in magnitude as the foreign interest rate increases (i.e., the size of the short deutsche mark position in the hedge portfolio increases when rDM rises). However, since the derivative is of ambiguous sign, Figures 24 show how the slope varies as the values of other parameters are changed.20 Figure 2 shows that for common values of the volatility parameter the relationship is always negative. However, if the exchange rate’s annualized volatility falls below 5 percent, the curve slopes upward after rDM rises above approximately 45 percent. Figure 3 shows that as the intrinsic value of the option, X – S, rises, a similar reversal in slope is possible. Indeed, for deep-in-the-money options, delta is everywhere increasing in rDM. Finally, Figure 4 shows that as the time to maturity of the option increases, further increases in rDM can lead to an increase in delta.

Figure 2.
Figure 2.

Sensitivity of Put Delta to the Foreign Interest Rate and to Volatility

Citation: IMF Staff Papers 1995, 003; 10.5089/9781451973396.024.A002

Figure 3.
Figure 3.

Sensitivity of Put Delta to the Foreign Interest Rate and to the Intrinsic Value

Citation: IMF Staff Papers 1995, 003; 10.5089/9781451973396.024.A002

Figure 4.
Figure 4.

Sensitivity of Put Delta to the Foreign Interest Rate and to the Contract Maturity

Citation: IMF Staff Papers 1995, 003; 10.5089/9781451973396.024.A002

The inference that can be drawn from these figures is that, for commonly observed parameter values, delta declines with increases in the foreign interest rate. However, for options that have long terms to maturity or that are deep-in-the-money or were written in periods of low volatility—or a combination of these characteristics—it is possible that for a sufficiently large increase in rDM, the relationship between delta and rDM could actually become positive. Subsequent increases in rDM would then lower the short deutsche mark position held to hedge the short put position. Note however, that except for multiyear options—which are extremely rare—it is highly unlikely that any increase in rDM would actually lead to a decline in the initial short deutsche mark position.

IV. Hedging in a Crisis

Dynamic hedging strategies are not an entirely new activity—stop-loss trading has always been triggered by price movements beyond a certain threshold. Dynamic hedging simply mechanizes this response. To the extent, however, that the technique has been adopted by large segments of the financial intermediation industry and can be implemented more rapidly than previous techniques, dynamic hedging strategies have added to trading volume and have accentuated price movements by contributing to momentary illiquidity. In this section, we consider how the widespread use of dynamic hedging techniques may interact with central bank exchange rate and liquidity policies to undermine a defense of a fixed exchange rate system.

When a fixed exchange rate regime moves toward a crisis, speculation against the currency is generally channeled through forward sales of the currency to the banking system. Some margin is required by counterparty banks, but this can be leveraged up by a factor of ten or more by the speculator. In a crisis, these sales will generally not be matched by other customers’ forward purchases of the currency. The central bank defending the currency may intervene with forward purchases, but the extent of such an operation is limited by the unwillingness of a central bank to risk large capital losses on negative net foreign exchange positions and by limits on credit lines to the central bank made available by the major dealing banks.21 Once the central bank ceases to buy its currency in the forward market, banks must balance their forward purchases with spot sales of the currency (to balance the net currency position) and by currency swaps (to balance maturities).

Once again, during a crisis, the central bank will be the most important buyer on the spot market through its intervention to maintain the fixed exchange rate. At the same time, it provides its currency through the discount window to the banks that need to sell currency in order to match their forward and spot foreign exchange positions as discussed in the previous paragraph. By providing liquidity to banks through this kind of facility, the central bank is effectively financing the attack on its own reserves. To settle its spot transactions, the central bank must deliver its own foreign exchange reserves or draw down lines of credit from other central banks or multilateral entities. As its short foreign exchange position mounts during the intervention, the central bank must act by raising the discount rate. This increases the cost to speculators who speculate against the currency by borrowing from the central bank. The central bank may also impose a squeeze on short sellers by channeling available credit away from identified speculators.

This final operation is the classic interest rate defense of a fixed exchange rate. It works through a liquidity effect in the money market—domestic credit grows less rapidly than central bank net reserves decline, thereby producing a decline in the supply of the domestic settlement medium. If large short positions in the currency are due for settlement, holders of short positions may sell foreign exchange to the central bank rather than face the high interest costs of rolling over overnight loans in the weak currency. The costs to holders of short positions are further accentuated if in addition they are caught in a squeeze so that they have to pay more than the discount rate to obtain funds.

The market’s acquisition of foreign exchange from the central bank does not arise exclusively from forward sales by nonbank speculators. Speculators and hedgers may also buy put options on the weak currency from the banks. Again, in a crisis, the banking system will likely be unable to find nonbank sellers of puts to balance these positions.22 To hedge, the bank that writes the put may create a long position in a synthetic put by selling the weak currency forward, by selling on the futures market, or by selling spot and entering a swap contract. Any of these operations will trigger a spot sale of the weak currency to the central bank as described above.

A common hedging strategy employed by customers is the implementation of a range forward, depicted in Figure 5.23 An investor holding Italian government BTPs (Buoni del Tesoro Poliennali—medium- and long-term bonds), for example, who invested in these bonds because of their relatively high interest rate, might buy a put option and sell a call with a higher exercise price chosen so that the revenue from the sale of the call equals the price of the put. In such a strategy, the investor bears the risk of exchange rate depreciation in the range, X¯X, but is protected from very large depreciations, as demonstrated by the payoff function RF. The bank that sells the range forward is short a put option and long a call option with similar strike prices. The bank may choose to hedge these exposures by creating synthetic options. The hedge portfolio for both of these option positions requires a short lira position.24

Figure 5.
Figure 5.

Range Forward Payoff Diagram

Citation: IMF Staff Papers 1995, 003; 10.5089/9781451973396.024.A002

The Effect of Interest Rate Changes on Dynamic Hedging

Once a central bank raises interest rates in defense of the fixed exchange rate, hedging operations may trigger further sales of the currency rather than the purchases anticipated from the squeeze. This result follows from the relation between interest rate movements and the hedging portfolio in equation (1).

Intuitively, the interest rate differential between the two currencies reflects the expected rate of depreciation of the exchange rate plus a risk premium. Unless volatility increases or attitudes toward risk change, a rise in the differential between deutsche mark and dollar interest rates means that the deutsche mark is expected to depreciate more rapidly against the dollar—that is, the hedge ratio increases.25

With an unchanged current exchange rate, exercise price, and exchange rate volatility, the put option is much more likely to finish in the money when the foreign interest rate jumps upward.26 That the option is more likely to be exercised means that it provides a higher effective hedge to a portfolio manager covering a deutsche mark exposure. The manager of the bank’s portfolio who uses a synthetic put in a dynamic hedging operation must likewise provide an increased hedge ratio in response to the greater probability that the option will be exercised. This means that he must short sell more deutsche mark so that his synthetic put continues to mimic an actual put. Taken to an extreme, if deutsche mark interest rates rise so high that, according to the underlying theory, it is almost certain that a put option will be exercised, the put then provides the equivalent of a 100-percent hedge ratio. The bank’s portfolio manager using a synthetic put must similarly sell sufficient deutsche mark to cover his entire deutsche mark position to provide the same coverage as an actual put.

How important will the dynamic hedging response be? Figure 3 provides some indication of the response of dynamic hedging programs during the final days of a managed or fixed exchange rate regime. In the days leading up to the collapse of an exchange rate band regime the gradual depreciation in the spot exchange rate will have a significant effect on the hedge ratio, necessitating a gradual increase in the short foreign currency position. However, in the final hours or minutes of such a regime or of an absolutely fixed exchange rate, the use of large interest rate increases to defend the fixed exchange rate can result in increases in the hedge ratio of a similar magnitude. What makes this effect important is that in a fixed exchange rate regime, or a quasi-fixed system such as the ERM, the boundary values for the nominal exchange rate become focal points for speculation about the direction of change of the exchange rate. Not only will more investors begin to hedge their exposures as the risk of realignment or a change in the parity increases, but the options that are written for customers will all tend to have similar exercise prices, so that they will tend to react similarly to changes in foreign interest rate. Moreover, as the spot foreign exchange rate falls toward the lower boundary value—a natural value for the put option’s exercise price—the option’s delta will become more sensitive to changes in the foreign interest rate (Figure 3).

In the United Kingdom on September 16, 1992, the Bank of England increased the base lending rate twice, from 10 percent to 12 percent and then again to 15 percent (effective the next day).27 The one-month London interbank offer rate (LIBOR) increased from 10.4 percent at the end of the previous day to 28.9 percent at the end of trading on September 16. According to equation (1), such an interest rate increase would have resulted in a 22 percent decrease in the delta (or increase in the hedge ratio) of an at-the-money put, from –0.54 to –0.66—a larger change than would have been obtained from a 1.5 percent depreciation at the initial interest rate.28 In the Swedish market, the increase in the marginal lending rate from 75 percent to 500 percent on September 16 led to an increase in the one-month Stockholm interbank offer rate (STIBOR) from 25 percent to 70 percent. An increase of this magnitude implies a 14 percent decrease in delta.29 On November 19, the eve of the Swedish devaluation, the one-month STIBOR rate rose from 13.9 percent to 28 percent, implying an 18 percent increase in the hedge ratio.30

Industry sources indicate that indeed when there is an increase in the interest rate spread with no movement in the exchange rate, the forward rate discount will trigger a sell-off in the currency through dynamic hedging. During the ERM crisis of September 1992, for example, industry sources estimate that dynamic hedging sales to adjust positions because of increases in interest rate spreads, exchange rate movements, and increases in volatility accounted for 20–30 percent of the selling in the crisis.31 It apparently was a major factor in the lira market one week after the first devaluation and also in the Swedish krona market later in 1992. Up to 10 percent of the sales were due to increases in interest rate spreads. In the case of the United Kingdom, on September 16, 1992, the dramatic increase in forward discounts triggered sales of pounds. When interest rates rose and nothing happened to the exchange rates, the selling programs were turned on. The lack of movement (appreciation) in the exchange rate meant that the forward rate fell further below the floor. Thus, the full force of programmed sales triggered by interest rate movements was not offset by exchange rate improvement. Another source of the sales volumes at this moment was the rising perceived volatility resulting from the suddenly larger movement of the forward rate below the floor. The effect of dynamic hedging sales may also have been a source of some of the selling pressure observed on August 12, 1994, when the Italian lira depreciated sharply after the Banca D’ltalia raised the discount rate by 50 basis points, though the consensus view is that markets reacted mostly in the belief that the interest rate increases were fiscally unsustainable.

In an exchange crisis, therefore, a large defensive rise in the interest rate aimed at imposing a squeeze on speculators will instantaneously trigger hedging programs to order sales of the weak currency.32 The experiments conducted using the actual data on interest rates and historical volatility suggest that the selling triggered by dynamic hedging programs during an interest rate defense can be significant. The existence of a large amount of such programs in the market would undermine the use of an interest rate defense of a weak currency—the moment that a central bank raises interest rates, it might face an avalanche of sales of its currency rather than the purchases of the squeezed shorts that it had anticipated. In effect, the hedging programs make the hedgers insensitive to the added costs of funding their weak currency sales.

If the central bank has a credit line limit in foreign exchange or a self-imposed negative net reserve position, the upsurge of selling brought about by the interest rate increase might cause a sudden jump to its limit and force it to cease intervention in defense of the exchange rate. Whether this counterintuitive result occurs depends on the weight of these mechanistic traders relative to those caught in the short squeeze.

In one scenario, the hedging operation may in any case far exceed the amount of the weak currency demanded by those caught in the squeeze. In this case, the timing of the hedging sales—the prearranged rule for awakening the selling programs—relative to the time at which those caught in the short squeeze appear on the market is immaterial to the survival of the fixed exchange rate. Dominance by the mechanistic hedgers will defeat the interest rate defense.

In the scenario in which the amounts of these opposite transactions are roughly balanced or even where those caught in the short squeeze dominate, the timing of transactions is key. If the selling programs switch on instantly, but the buying operations to cover short positions occur with some lag, the central banks’ net short foreign exchange limit may be exceeded prior to the appearance of the buyers of its currency, causing the abandonment of the fixed exchange rate. Buyers might have appeared by the end of the day to offset the sellers, but the initial selling may unnerve the central bank and force devaluation. The devaluation will ratify the actions of both the sellers and those caught in the squeeze who hesitated. Sellers will have sold prior to the devaluation of the exchange rate, and those caught in the squeeze can buy back into the weak currency at a lower price. If the central bank simultaneously relaxes the high interest rates, overnight borrowing will cease to be a problem for those caught short, and the squeeze will be suspended.

V. Conclusion

In their impact on the viability of the interest rate defense of a fixed exchange rate, dynamic hedging programs can be interpreted as a new wrinkle on an old phenomenon. Skeptical participants in the foreign exchange market have sometimes interpreted a defensive increase in the interest rate as the last rearguard action before the abandonment of a fixed rate. In this light, the suddenly higher interest rate differential signals only the extent of the impending depreciation of the exchange rate and certainly not a drastic and extended tightening of liquidity in the weak currency’s money markets. Interpreting the interest rate increase in this way dictates that a speculative selling program should be begun. Dynamic hedging programs automatically place this interpretation on an interest rate increase; thus, they are a mechanization of the previously informal skepticism that occasionally arose about exchange rate defenses. To the extent that such programs are present in generating large selling volumes, they signal a major shift toward skepticism about the strength of the central bank’s adherence to the policy of defending the exchange rate, thereby undermining the efficacy of a previously useful defensive tool.

The scenario that we depict here is a technical story about the character of minute-by-minute trading in the death throes of a fixed exchange rate. A dramatic interest rate increase in a last ditch defense triggers dramatic selling pressure. If this technical feature of the market is important in the last moments of a fixed exchange rate, it is necessary to implement a defense operation that takes it into account. For example, it is often argued that a resolute defense of a fixed exchange rate regime requires that at an early date interest rates be raised gradually, though ultimately to high levels.33 Such a policy would trigger daily selling of the currency by dynamic hedgers, but not in quantities that would overwhelm the central bank’s net reserve limits before those caught short in the currency would appear as buyers. Thus, raising rates gradually in an interest rate defense may immunize the central bank against being pushed beyond its position limits.

REFERENCES

  • Bank of England, Derivatives: Report of an Internal Working Group (London: Bank of England, April 1993).

  • Bank for International Settlements, Recent Developments in International Interbank Relations (Basle: BIS, October 1992).

  • Bank for International Settlements, Central Bank Survey of Foreign Exchange Market Activity in April 1992 (Basle: Bank for International Settlements, March 1993).

    • Search Google Scholar
    • Export Citation
  • Brady Commission, Report of the Presidential Task Force on Market Mechanisms (Washington: Government Printing Office, January 1988).

  • Chiang, Raymond, and John Okunev, “An Alternative Formulation on the Pricing of Foreign Currency Options,” Journal of Futures Markets, Vol. 13 (December 1993), pp. 9037.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cookson, Richard, “Moving in the Right Direction,” Risk, Vol. 6, No. 10 (October 1993), pp. 2226.

  • DeRosa, David F., Options on Foreign Exchange (Chicago, Illinois: Probus Publishing Company, 1992).

  • Germany, Deutsche Bundesbank, “Off-Balance-Sheet Activities of German Banks,” Monthly Report of the Deutsche Bundesbank, Vol. 45, No. 10 (October 1993), pp. 4567.

    • Search Google Scholar
    • Export Citation
  • Dumas, Bernard, L., Peter Jennergren, and Bertil Näslund, “Realignment Risk and Currency Option Pricing in Target Zones,” NBER Working Paper No. 4458 (Cambridge, Massachusetts: National Bureau of Economic Research, September 1993).

    • Search Google Scholar
    • Export Citation
  • Feiger, George, and Bertrand Jacquillat, “Currency Option Bonds, Puts and Calls on Spot Exchange and the Hedging of Contingent Foreign Earnings,” Journal of Finance, Vol. 34 (December 1979), pp. 112939.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garber, Peter M., and Michael G. Spencer, “Foreign Exchange Hedging with Synthetic Options and the Interest Rate Defense of a Fixed Exchange Rate Regime,” IMF Working Paper 94/151 (Washington: International Monetary Fund, December 1994).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garman, Mark B., and Steven W. Kohlhagen, “Foreign Currency Option Values,” Journal of International Money and Finance, Vol. 2 (December 1983) pp. 23137.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gennotte, Gerard, and Hayne Leland, “Market Liquidity, Hedging and Crashes,” American Economic Review, Vol. 80 (December 1990), pp. 9991021.

    • Search Google Scholar
    • Export Citation
  • Grabbe, J. Orlin, “The Pricing of Call and Put Options on Foreign Exchange,” Journal of International Money and Finance, Vol. 2 (December 1983), pp. 23953.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Grossman, Sanford J., “An Analysis of the Implications for Stock and Futures Price Volatility of Program Trading and Dynamic Hedging Strategies,” Journal of Business, Vol. 61 (July 1988), pp. 27598.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Group of Ten, International Capital Movements and Foreign Exchange Markets (Rome: Group of Ten, April 1993).

  • Group of Thirty, Derivatives: Practices and Principles (Washington: Group of Thirty, July 1993).

  • Hull, John C., Options, Futures, and Other Derivative Securities (Englewood Cliffs, New Jersey: Prentice-Hall, 2nd. ed., 1993).

  • International Monetary Fund (1993a), International Capital Markets: Part I. Exchange Rate Management and International Capital Flows, World Economic and Financial Surveys (Washington: IMF, April 1993).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, (1993b), International Capital Markets: Part II. Systemic Issues in International Finance, World Economic and Financial Surveys (Washington: IMF, August 1993).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, International Capital Markets: Developments, Prospects and Policy Issues, World Economic and Financial Surveys (Washington: IMF, September 1994).

    • Search Google Scholar
    • Export Citation
  • Kat, Harry M., “Portfolio Insurance: A Comparison of Alternative Strategies,” Journal of Financial Engineering, Vol. 2, No. 4 (1993), pp. 41542.

    • Search Google Scholar
    • Export Citation
  • Kroner, Kenneth F., and Jahangir Sultan, “Time-Varying Distributions and Dynamic Hedging with Foreign Currency Futures,” Journal of Financial and Quantitative Analysis, Vol. 28 (December 1993), pp. 53551.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Leong, Kenneth, “Solving the Mystery,” Risk, Vol. 4, No. 1 (December 1990– January 1991), pp. 6871.

  • Melino, Angelo, and Stuart M. Turnbull, “Pricing Foreign Currency Options with Stochastic Volatility,” Journal of Econometrics, Vol. 45 (July–August 1990), pp. 23965.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Naik, Vasanttilak, “Option Valuation and Hedging Strategies with Jumps in the Volatility of Asset Returns,” Journal of Finance, Vol. 98 (December 1993), pp. 196984.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perraudin, William R., and Bent E. Sorenson, “Foreign Exchange Option Pricing in a Continuous Time Arbitrage Pricing Model with Stochastic Volatility and Jumps” (unpublished; Cambridge: Cambridge University, June 1992).

    • Search Google Scholar
    • Export Citation
  • United States, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, and Office of the Comptroller of the Currency, “Derivative Product Activities of Commercial Banks,” Joint Study Conducted in Response to Questions Posed by Senator Riegle on Derivative Products (unpublished; Washington: January 27, 1993).

    • Search Google Scholar
    • Export Citation
  • United States, Commodity Futures Trading Commission, OTC Derivative Markets and their Regulation (Washington: CFTC, October 1993).

  • United States, General Accounting Office, Financial Derivatives: Actions Needed to Protect the Financial System (Washington: GAO, May 1994).

    • Search Google Scholar
    • Export Citation
  • United States, Securities and Exchange Commission, The October 1987 Market Break (Washington: Government Printing Office, February 1988).

    • Search Google Scholar
    • Export Citation
  • Walmsley, Julian, The Foreign Exchange and Money Markets Guide (New York: Wiley, 1992).

*

Peter M. Garber is Professor of Economics at Brown University and holds a Ph.D, in economics from the University of Chicago. Michael G. Spencer is an Economist in the Capital Markets and Financial Studies Division of the IMF’s Research Department. He holds a Ph.D, in economics from Queen’s University. This paper has benefited from the comments of participants at the NBER, CEPR, Banca d’Italia Conference on the Microstructure of Foreign Exchange Markets held in Perugia, Italy, July 1–2, 1994, and at seminars held at the Bank of Japan, Princeton University, and Virginia Polytechnic Institute. The authors are grateful also for comments from Philippe Jorion, Paolo Kind, Richard Lyons, John Montgomery, Victor Ng, David Ordoobadi, and Thierry Pujol.

1

See, for example, the Brady Commission (1988) and U.S. Securities and Exchange Commission (1988) reports on the 1987 stock market break. The Brady Commission estimated that by the end of trading on October 16, only approximately one third of the portfolio insurance-motivated sales of stock and stock futures had been executed. Grossman (1988) anticipated this problem. Gennotte and Leland (1990) model the relationship between hedging operations and market liquidity and show how a relatively small volume of transactions initiated by hedgers can lead to a large price change.

3

These estimates are derived in U.S. General Accounting Office (1994). The notional value of a contract is the nominal amount used as a base to calculate a transfer of payments according to a contractual formula. For example, an interest rate swap may have a notional value of $10 million. This notional value is not delivered as principal. Rather, counterparties deliver or receive the net between fixed interest on $10 million and floating interest on $10 million, so the claims that counterparties might have on each other are far smaller than the notional value.

4

OTC derivatives dominate exchange traded products with limited liquidity such as longer-dated contracts or out-of-the-money options.

6

That is, the option can be exercised at any time prior to maturity. Under the European exercise convention, the option may only be exercised at maturity.

7

In addition, portfolios will be subject to basis risk when the security underlying the hedge instrument is not identical to the security whose return is being hedged so that the returns on the two securities are not perfectly correlated. A hedge constructed with a related, but not identical, instrument to the one whose value is being hedged is called a cross hedge.

8

See International Monetary Fund (1993a) for a discussion of the foreign holdings of institutional investors in industrial countries.

9

Using real put contracts to hedge long positions is not entirely free of volatility risk, of course, since changes in volatility can result in losses when put contracts are rolled over if the maturity of the contracts is shorter than the horizon of the hedging operation.

10

This strategy is referred to by Leland, O’Brien, and Rubinstein and Associates as a perpetual protection policy.

11

A constant percentage portfolio insurance strategy has an advantage over an option replication strategy in that at the end of the period a renewal of the hedge does not require a large trading operation. For an option replication strategy, at expiration the portfolio is either 100-percent hedged or completely unhedged. Renewal of the strategy for another period then requires a large jump in the hedge ratio.

12

See International Monetary Fund (1993a) for a discussion of the regulatory and internal constraints on banks’ foreign exchange trading.

13

See Garman and Kohlhagen (1983) for the development of this formula. For pricing formulas taking account of stochastic volatility, see Chiang and Okunev (1993), Kroner and Sultan (1993), Melino and Turnbull (1990), Naik (1993), and Perraudin and Sorenson (1992). Dumas, Jennergren, and Näslund (1993) derive options pricing formulas for currencies restricted by target zones as in the ERM. However, as the data in Section I indicate, most OTC and exchange-traded options contracts are written for dollar exchange rates.

14

Most exchange-traded currency options, except those traded on the Philadelphia Stock Exchange, are options on futures, for which the Garman/Kohlhagen formula for spot exchange options is inapplicable. In the standard formula for pricing options on futures, the foreign interest rate does not appear. The effects of foreign interest rate changes are felt through their impact on the futures price. Moreover, in the OTC market it is more common to price options with respect to the forward exchange rate than to the spot rate (see DeRosa (1992)).

15

This equation is identical in form to the Merton adaptation of the Black-Scholes put formula for a stock that pays a continuous, constant dividend. This formula is constructed on the assumption that the percentage change in the price of the underlying security, in this case the dollar-deutsche mark exchange rate, follows a Wiener process, and that the instantaneous interest rates in both countries and the standard deviation of the percentage exchange rate change are fixed parameters for the life of the option. Such a simple formula does not exist for American put options; these must be evaluated by numerical methods (see DeRosa (1992)).

16

Note that the ability to maintain a dynamic hedge depends critically on the existence of a liquid spot foreign exchange market in which the rebalancing trades can be executed. If, as happens during a crisis, markets become illiquid, investors that rely on dynamic hedging may not be able to adjust their portfolios and will be exposed to further exchange rate changes.

17

Since the hedge is not adjusted continuously, the bank will incur losses between rebalancing exercises. Leong (1991) argues that the option premium charged by a bank will, in equilibrium, equal the expected value of this “hedge slippage.”

18

See, for example, Cookson (1993) or DeRosa (1992).

19

For similar discussion of the effect on delta of changes in the other parameters (S, X, r$, σ, and T) the reader is referred to Garber and Spencer (1994).

20

These figures are drawn by varying one or more of the parameters, as indicated, from the following baseline: r$ = rDM = 0.03, S = X = 1, T = 1 month, σ = 0.15.

21

The ability of the central bank to enter into forward contracts with its own nationally chartered banks is circumscribed by credit line limits imposed by banks elsewhere on these banks.

22

Even if nonbank sellers of puts exist somewhere in the financial system, the selling bank seeking cover may not find them through the banking system. In a crisis, gross trading volumes surge, thereby causing many banks to reach their credit ceilings with other banks. As the banking system becomes illiquid in this way, transactions that passed through the banking system on a credit basis now must seek a cash market. To hedge, the selling bank will place an order to buy a put onto the organized currency options market, where credit risk is not an issue, and will find the potential seller in this market. As the crisis progresses and more interbank credit lines fill, volume will tend to move to the more secure organized exchanges.

23

We are grateful to Paolo Kind for suggesting this example.

24

Intuitively, a short put position is a contingent long position in the foreign currency (here, the lira), as is a long call position. Hence, both are hedged by going short in lire.

25

This is the only interpretation of an interest rate increase that is allowed by the options pricing model. The construction of the differential equations behind options pricing, however sophisticated, requires assumptions about the covariances of the macroeconomic forcing variables in the model—exchange rate changes, interest rates in both countries, and exchange rate volatility. The Garman-Kohlhagen model comes out of macroeconomic world in which interest rates and volatility are constant, and therefore exchange rates are exogenous and are uncorrelated with these parameters. These assumptions are required in order to transfer the Black-Scholes equity option pricing model to the currency market. While such assumptions may be reasonable in the context of an option on a single stock, they do not emanate from any known macroeconomic model. As a result, in establishing the delta hedge, the Garman-Kohlhagen model takes the exchange rate and its volatility as momentarily given. Through interest parity the increased foreign interest rate must result from either an increased expected rate of depreciation or a higher risk premium. Since volatility is fixed, the latter possibility is ruled out.

26

A central bank squeeze generally operates through overnight interest rates, which are not the interest rates used to value longer-dated options. Nevertheless, in a squeeze, a jump in overnight rates will usually have a strong impact on short-and medium-term interest rates, which are relevant to option pricing.

27

See International Monetary Fund (1993a) and Group of Ten (1993) for descriptions of the European currency crisis of 1992–93.

28

On September 15, the one-month LIBOR rate for dollars closed at 3.0625 percent and the historical volatility of the $/£ exchange rate, estimated over the previous month, was 15.8 percent per annum. As Figure 2 shows, if the banks sold options with volatilities higher than their historical levels, which during a speculative attack is very likely, the change in the hedge ratio may have been smaller.

29

Based on a historical volatility, calculated over the previous month, of 6.08 percent.

30

With an estimated volatility of 12.9 percent.

31

These estimates were obtained during confidential interviews with market participants in November 1992 (see International Monetary Fund (1993a)). A more precise estimate of the importance of dynamic hedging on September 16 would require access to confidential information (e.g., which firms were most active in the spot and derivatives markets, how much of their trading was due to dynamic hedging programs, detailed minute-by-minute transactions data). The Securities and Exchange Commission and the Brady Commission had access to such information for their analyses of the 1987 stock market crash. They concluded that portfolio insurance trading played a key role in driving stock prices down during October 16-20, 1987. Analysis of intra-day trading behavior showed that immediately after large price declines, portfolio insurance programs initiated large sales in both the futures and stock markets. Declines in futures prices were transmitted to the stock market almost immediately by index arbitrage programs. The Brady Commission (1988) estimated that between $60 billion and $90billion in equity assets were under portfolio insurance programs at the time of the crash, and that the typical program called for a 20-percent decline in the stock portfolio in response to a 10 percent decline in prices. The report estimated that portfolio insurance programs accounted for more than 40 percent of S&P 500 futures volume on the Chicago Mercantile Exchange and 23 percent of the large-trade volume on the New York Stock Exchange on October 19, 1987.

32

Who is actually squeezed in such a defense? All borrowers in the weak currency whose debts are due for settlement or rollover soon (after two days) will find that their costs and risks have suddenly jumped as they now have to pay high and volatile yields to the money market scalpers that are unleashed by the squeeze. This group could conceivably include even those who have constructed synthetic puts if they have established their short currency position by borrowing on overnight rollover credit, as Richard Lyons has pointed out to us. Typically, however, a synthetic option is constructed by establishing a short forward position whose expiration date coincides with the expiration date of the option. If many of the existing hedges were constructed within a one- or three-month period before the speculative attack and with a relatively long maturity, they would have locked in longer-term finance and the position would be immune from a short squeeze.

33

“Early” is relative to the time of outbreak of the next speculative attack. How to recognize when an attack will come in order to implement this early defense is problematic.

IMF Staff papers, Volume 42 No. 3
Author: International Monetary Fund. Research Dept.