FMP 9: Foreign Exchange Markets

A dealer publishes two prices. The bid is the rate at which the dealer buys the base currency from the client, and the ask, also called the offer, is the rate at which the dealer sells it. So a client buys at the ask and sells at the bid, and the gap pays the dealer for standing ready to trade.

Spot rates are shown to four or five decimal places, and that precision is not decoration, because the whole spread on a large corporate trade sits in the last two or three digits. On January 16, 2020, EURUSD was quoted at one point bid 1.11347 and ask 1.11354, giving a spread of 0.00007 U.S. dollars per euro.

Forward rates use the same base currency as spot rates, and dealers quote forward points rather than the outright number. For most pairs a point is one ten thousandth of a unit of the quote currency, so the figure is divided by 10,000. Japanese yen points are the exception, being multiplied by 1/100.

Take EURUSD on January 16, 2020, spot bid 1.11347 and ask 1.11354, with the three-month forward quoted bid 63.2 and ask 63.5. The points give a forward bid of 1.11347 + 0.00632 = 1.11979 and an ask of 1.11354 + 0.00635 = 1.11989. Euros were dearer three months out than spot because of the gap between dollar and euro interest rates, not any view about the euro.

The points carry a spread of their own, which stacks on the spot spread. Here it is 0.3, being 63.5 less 63.2, or 0.00003, and adding the spot spread of 0.00007 gives a forward spread of 0.00010. Extend the maturity and the effect grows: the ten-year forward came to 1.11347 + 0.22160 = 1.33507 on the bid side and 1.11354 + 0.22960 = 1.34314 on the ask, a spread of 0.00807. Long-dated forwards tie up more dealer capital and carry years of counterparty credit exposure.

Points can be negative. USDCAD in July 2018 traded with the Canadian rate below the U.S. rate, putting the forward rate, in Canadian dollars per U.S. dollar, below the spot rate. Screens sometimes drop the minus signs, so traders compare the figures: ask points below bid points means the points are subtracted, because a forward spread must be wider than the spot spread.

A forward transaction in which two parties agree to exchange currencies on a single future date is an outright transaction. An FX swap is assembled differently, with currency changing hands on two dates, typically bought spot and sold back forward. Both legs are agreed at once, so both rates are known from the outset.

What an FX swap delivers is funding. A U.S. company needing euros can borrow dollars at home, buy 1 million EUR spot, and simultaneously agree to sell 1 million EUR one month forward. The European operation has its euros, and interest is paid in the domestic currency. On January 16, 2020 the one-month forward bid for EUR was 21.1 points, so more dollars come back in a month than the same euros would raise today, cutting the funding cost.

The mirror case is a company holding foreign currency it will spend later while preferring to earn interest at home: it sells spot and buys back forward. A currency swap, or cross-currency swap, is a separate instrument, exchanging principal plus an interest stream in one currency for principal plus an interest stream in another.

Source: BIS Triennial Central Bank Survey, 2019, ordered here from largest to smallest; shares computed here.

Foreign exchange futures trade actively worldwide, and the CME Group lists contracts on the dollar against many currencies. Exchange-traded quotes obey one rule without exception: the U.S. dollar is always the quote currency, because the exchange treats a foreign currency as an asset priced in dollars. A six-month USDCAD forward quote of 1.3000 becomes a futures quote of 0.7692 U.S. dollars per Canadian dollar, which is 1 divided by 1.3000.

Contract sizes are fixed by the exchange. Popular CME Group contracts cover 62,500 GBP, 100,000 AUD, 100,000 CAD, 125,000 CHF, 125,000 EUR and 12.5 million JPY. Maturity months include the next three calendar months plus March, June, September and December for the following 20 months, and several cross rates are listed too.

Long and short currency positions

A currency position is always a position in the base currency of the pair used to describe it. A long position gains when the quoted number rises, because each unit of the base currency then commands more of the quote currency, and a short position gains when it falls. One economic position has two correct descriptions, and the labels swap when the pair is inverted. A long Canadian dollar futures position at 0.7692 profits when the Canadian dollar strengthens; in USDCAD terms that same position is short and profits when the number falls.

Describe the exposure before naming any position: state which currency must be delivered and which received, choose a quote direction, then attach the labels. A British importer owing South African rand is short rand. Buying rand forward creates the offsetting long rand position, and in a GBPZAR quote, counting rand per pound, that hedge is a sale of GBPZAR forward.

Firms must quantify their exposures to exchange rates at future dates before judging whether those exposures are tolerable. Three categories are separated: transaction risk, translation risk and economic risk. What each one touches matters more than its definition.

Transaction risk attaches to receivables and payables, amounts of foreign currency genuinely paid or received on identifiable dates. A British company buying goods from South African suppliers and settling in rand is exposed to GBPZAR, the number of rand per pound, and if the rand strengthens each invoice costs more pounds. Suppose it also sells into Portugal and prices its goods in euros. It now holds a euro receivable and is exposed to EURGBP, where a weaker euro converts into fewer pounds.

Outright forward transactions lock both, and the direction of each trade follows from the direction of the cash flow. Rand must be bought, so rand are bought forward and the rate paid to suppliers is fixed today. Euros will be received, so euros are sold forward. A forward fixes a rate, not a good outcome: if the rand weakens afterwards, the company still settles at the contracted rate.

Amounts and timing are seldom known precisely. A forward sized for a receivable that arrives late leaves the company delivering currency it has not collected, and one sized for a receivable that shrinks leaves it over-hedged. Treasuries hedge a conservative share of a forecast exposure, roll contracts as timing firms up, or use options.

Translation risk shows up on a reporting date rather than a payment date. Assets and liabilities denominated in a foreign currency must be restated into the reporting currency when financial statements are prepared, producing gains and losses owing nothing to operating performance. Earnings held in foreign subsidiaries rather than repatriated are translated too, usually at the fiscal year average exchange rate.

The difference from transaction risk governs the hedging decision: transaction risk moves cash, while translation risk does not, although it can move reported earnings considerably.

Hedging translation exposure only makes sense for one future date, since covering the same assets at both the one-year and the two-year reporting date counts the first year twice, which is over-hedging. Even then it deserves scrutiny, because forward contracts settle in cash: accounting risk has been swapped for cash flow risk. The case is strongest when there is a real plan to sell the assets, retire the liabilities, or repatriate income at a known time.

Financing is often the better answer. Funding assets in a country with borrowings raised there leaves gains and losses on the assets offset by losses and gains on the liabilities: the company above could fund its GBP 10 million facility with GBP 10 million of sterling borrowings and carry no net translation exposure.

Economic risk is the risk that future cash flows are affected by exchange rate movements, and it can exist where no foreign currency contract does. A U.S. firm selling software in Brazil and denominating the price in dollars has no transaction risk on those sales, yet its economic risk is considerable. Should the real decline against the dollar, customers in Brazil find the software more expensive, and either demand falls or the firm cuts its dollar price.

Home markets offer no shelter. A U.K. firm with no production or sales overseas can still be damaged when exchange rate movements give a foreign competitor fresh reason to expand there. Economic risk resists measurement, because the answer depends on how customers, competitors and prices respond over years, so it belongs in strategic analysis rather than a hedging programme.

Why many exposures are safer than one

A multinational is exposed to a long list of currencies, and the length of that list is itself protection. Movements in different exchange rates are not perfectly correlated, so a portfolio of exposures is less volatile than a single exposure of the same size.

Treasurers frequently prefer options to forward contracts, because an option gives downside protection against an adverse exchange rate movement while leaving the firm free to benefit from a favourable one. Buying an option on each currency is expensive, since each is priced off one currency and none of the diversification is credited back. The cheaper alternative is one option, bought in the over the counter market, on the whole portfolio: long 100,000 units of currency A, long 200,000 units of currency B and short 75,000 units of currency C, as one underlying. That is a basket option, priced off the volatility of the basket.

The saving and the residual risk share a source. A basket option protects the aggregate and not the individual legs, so the firm keeps cross-currency risk: currency A can move against it while currency B moves in its favour, leaving the basket near its strike and the option worthless although one exposure lost money. Protection depends on correlations behaving, and they are not stable. Timing is the same problem again. A multinational exposed in every month can trade monthly maturities, or a cheaper option written on the average exchange rate for the year, known as an Asian option, and the two combine into an Asian basket option.

Exchange rates are set by supply and demand, as the prices of all financial assets are, and many interrelated factors push on both sides. No method predicts a future exchange rate with precision; what these variables explain is the direction of pressure.

The balance of payments records the value of what a country exports less the value of what it imports, and trade creates flows of currency. When exports from Country A to Country B increase, exporters convert foreign currency revenue into their own currency, demand for the currency of Country A rises, and it strengthens relative to that of Country B. Rising imports work the other way, because importers must buy the currency of Country B.

There are equilibrating forces here, since a stronger currency in Country A makes its exports dearer abroad and demand for them falls. Commodity exporters show the link plainly. Canada is an oil exporting nation, so the Canadian dollar is influenced by the price of crude oil. It traded above the U.S. dollar through 2011-2014, when crude prices were high, and weakened once the oil price declined.

Monetary policy is the second lever. A central bank expanding the money supply puts more currency behind an unchanged quantity of goods. If Country A expands its money supply by 25% and Country B leaves its own alone, the currency of Country A tends to decline by 25% against that of Country B, all else equal.

Inflation acts on exchange rates through prices. If USDCAD is 1.2500, the Canadian dollar price of goods in Canada should sit 25% above the U.S. dollar price of the same goods, otherwise a theoretical arbitrage opportunity appears. A product costing USD 100 in the United States and CAD 130 in Canada earns an arbitrageur CAD 5 per unit, since USD 100 converts into CAD 125; price it at CAD 120 and the trade reverses. Transportation costs and tariffs stop most such trades.

Applied across time it becomes purchasing power parity. Take inflation of 3% per year in the United States and 1% per year in Switzerland, with USDCHF starting at 1.05, so a basket costing 100 USD costs 105 CHF. After one year it costs USD 103, being 1.03 X 100, and CHF 106.05, being 1.01 X 105, so the rate becomes 106.05 divided by 103, or 1.0296. The franc has strengthened by roughly 2%, matching the inflation gap.

Real interest rates and nominal interest rates

Nominal interest rates are the rates quoted in the market, stating the return earned in units of the currency itself: 4% per year means 100 grows to 104 in one year. What those units buy depends on prices. A basket costing 100 at the start of the year costs 103 at the end if inflation is 3%, so an investor who deposited the 100 can buy 104 divided by 103, or 1.0097 baskets. Real purchasing power has risen by 0.97%, which is the real interest rate.

Both rates can turn negative. Nominal rates on the Swedish krona, the Japanese yen, the Danish krone, the euro and the Swiss franc were negative after the 2007-2008 financial crisis, and roughly USD 13 trillion of government bonds offered yields below zero in 2019. The objection that rates cannot fall below zero because cash pays zero underestimates the cost of storing cash, so negative rates do not necessarily present arbitrage opportunities.

Purchasing power parity is approximately right over long horizons and badly wrong over short ones. Covered interest parity links the forward rate, the spot rate and two interest rates by an arbitrage between instruments that all trade today, so it holds much more precisely.

A trader holds 100 GBP and wants U.S. dollars in T years. One route stays in sterling: invest at the sterling risk-free rate and at the same moment sell the resulting sterling forward at the T-year GBPUSD forward rate F. The other converts at once, so 100 GBP becomes 100S dollars at the spot rate S, invested at the dollar risk-free rate. Neither route leaves uncertainty about the dollars received, so without arbitrage they must give the same result.

Suppose F sits above that level. An arbitrageur borrows 100S USD at the dollar rate, converts to 100 GBP, invests at the sterling rate, and sells the proceeds forward at F, so the dollars arriving exceed the dollars owed. Below parity the steps reverse. Because real desks borrow and lend at different rates, the argument fixes a narrow range of forward rates, not one value.

Now nail the direction down, because this is where the result is most often applied backwards. For a pair written XXXYYY the quote counts units of YYY per one unit of XXX. When the risk-free rate for XXX is higher, XXX is weaker forward, so fewer units of YYY buy one unit of XXX. When it is lower, XXX is stronger forward. The currency paying the higher rate trades at a forward discount, which cancels its interest advantage.

Read backwards, the relationship recovers an interest rate differential. EURUSD mid rates on January 16, 2020 were 1.113505 spot and 1.11984 for three months, the averages of the quotes given earlier. Rearranging with T at 0.25 gives a ratio of one plus the dollar rate to one plus the euro rate of 1.0229. Market rates agreed: three-month interbank borrowing was about 1.84% in USD against about -0.39% in euros, and 1.0184 divided by 0.9961 is 1.0224.

Covered interest parity is about forward exchange rates and is enforced by arbitrage. Uncovered interest parity is about exchange rates themselves, and nothing enforces it. Its claim is that, once expected exchange rate movements are allowed for, every currency should offer an investor the same return. Take currency X with a risk-free rate of 2% and currency Y paying 6%: for the two to be equally attractive, currency Y should be expected to weaken by about 4% relative to currency X.

Potential violations are everywhere. The interest rate in USD in 2020 was much higher than the rate in EUR, yet the U.S. economy was considered much stronger than many European economies, and few market participants read that as a sign of a stronger euro. If both parity conditions held, the forward rate would equal the expected future spot rate.

The carry trade

A carry trade is the position taken by an investor who does not believe uncovered interest parity. Borrow in a low interest rate currency, convert at spot, and invest in a high interest rate currency, unhedged. Parity says the expected profit is zero, because the high-rate currency should depreciate by the interest advantage. The trade earns money whenever that depreciation fails to arrive.

Two features make the exposure worse than the arithmetic suggests. Returns arrive as a long run of small gains interrupted by rare large losses, because low-rate funding currencies tend to appreciate sharply when investors retreat from risk, which is when the position is hardest to exit. Leverage does the rest.