How Do Exchange Rates Work? What Sets the Price of a Currency

A country must decide how much it lets its currency move, and the choices sit on a spectrum:

• •Free floating, where the rate is set entirely by supply and demand in the market, as for the dollar, euro, pound and yen, with central banks intervening rarely
• •Managed float, where a central bank intervenes to smooth movements or resist a trend without committing to a target
• •Pegged, where the rate is fixed against another currency or a basket and defended by buying and selling reserves, as the Hong Kong dollar has been against the American dollar since 1983
• •Currency board, a hard version of a peg in which the domestic money supply is fully backed by foreign reserves
• •Dollarisation or adopting another currency entirely, as Ecuador, El Salvador and Panama have done, which imports the other country's monetary policy wholesale
• •Monetary union, where a group of countries share one currency and one central bank, as in the euro area

The constraint governing all of this is one of the more useful results in economics. A country can have any two of three things and never all three: a fixed exchange rate, free movement of capital across its borders, and an independent monetary policy. The logic is straightforward. If capital can move freely and you set an interest rate different from the anchor country's, money floods in or out chasing the difference, which forces the exchange rate to move, and defending the fix means abandoning your interest rate. So a country picks: the euro area has fixed rates internally and free capital movement and no national monetary policy; China long maintained a managed rate and independent policy by restricting capital flows; Britain and the United States have independent policy and free capital and accept whatever the rate does. Attempts to have all three end in currency crises, the most famous being Britain's exit from the European exchange rate mechanism in September 1992, when the government raised interest rates twice in a day and spent billions of reserves before conceding.

Several forces operate on different timescales. In the short term, interest rate differentials dominate, since capital chases yield and a currency whose central bank is expected to raise rates appreciates on the expectation rather than on the event. Risk sentiment matters almost as much, with the dollar, yen and Swiss franc strengthening during crises because investors move into what they consider safe regardless of fundamentals. In the medium term, inflation differences and current account balances matter, since a country with persistently higher inflation should see its currency decline. In the long term, the theory is purchasing power parity, which says rates should adjust so that a basket of goods costs the same everywhere, and which the Economist's index comparing the price of a hamburger across countries illustrates. Purchasing power parity works poorly over years and reasonably over decades, and its failures are systematic, since prices of things that cannot be traded, such as haircuts and rent, are genuinely lower in poorer countries and always will be.

A falling currency is neither good nor bad in itself, and it redistributes. Exporters gain, since their goods become cheaper abroad, and the tourism industry gains for the same reason. Importers lose, and since almost everything contains imported components, the cost feeds into domestic prices, which is imported inflation. Anyone holding debt denominated in a foreign currency loses badly, which is the mechanism that turns currency crises into debt crises in developing economies and is why the accumulation of dollar-denominated borrowing is watched closely. Savers holding domestic assets lose purchasing power internationally. A rising currency reverses all of this, and is often unwelcome to a government because it makes exports uncompetitive, which is why the phrase currency manipulation exists and why countries have periodically been accused of holding their rates down deliberately to support exporters.

Exchange rate forecasting is notoriously bad, and there is a well-known result behind it: a 1983 paper by Richard Meese and Kenneth Rogoff found that no economic model outperformed a simple assumption that tomorrow's rate will equal today's, over horizons up to a year, and decades of subsequent work have improved on this only marginally. The explanation is that rates are asset prices, so they incorporate expectations, and therefore move on the difference between what happens and what was already expected, which is unpredictable by definition. The scale of financial flows relative to trade flows amplifies this, since the market is driven by portfolio decisions rather than by the underlying exchange of goods. A related consequence is that a currency can stay far from any fundamental value for years, which is why carry trades, borrowing in a low-interest currency to invest in a high-interest one, are profitable for long periods and then lose everything at once when the rate moves.

Most currency trading is financial rather than commercial, which is why rates behave like asset prices and move on expectations. A country can have at most two of a fixed rate, free capital movement and an independent monetary policy, and attempts at all three end in crises. Interest rate differentials and risk sentiment drive short-term movements, inflation and trade balances the medium term, and purchasing power parity only over decades. Falling currencies help exporters, raise import prices and severely damage anyone holding foreign-currency debt.