Fixed, floating and managed systems, how a peg is defended by selling reserves or raising interest rates, the loss of monetary independence, why pegs break, devaluation against depreciation, the Marshall–Lerner condition and the J-curve. Through the Hong Kong dollar peg and the ERM exit of 1992.
Hong Kong fixed its dollar to the US dollar at HK$7.80 in 1983 and has held it ever since, within a band of 7.75 to 7.85. The rule is mechanical: if banks offer Hong Kong dollars at 7.85, the Hong Kong Monetary Authority must buy them with US dollars; if they demand Hong Kong dollars at 7.75, it must sell. Behind the promise sits an Exchange Fund of more than US$400bn.
In June 2025 the currency hit 7.85 for the first time in two years, as traders borrowed cheap Hong Kong dollars to buy higher-yielding US ones. Between 26 June and August the HKMA intervened seven times, buying HK$91bn. Each purchase drained cash from the banking system, local interest rates rose, and the trade stopped paying. In March 2026 it did the same again, buying more than HK$20bn in a month.
What does it take to fix an exchange rate, what does it cost, and why do most countries no longer try?
Source: Hong Kong Monetary Authority statements, 26 June 2025 to August 2025 and March 2026, as reported by Reuters, the South China Morning Post and The Standard.
A floating exchange rate is set by demand and supply in the foreign exchange market with no official target: the pound, the dollar, the euro against each other. A fixed exchange rate is set by the government or central bank at a declared value, or within a narrow band, and defended: Hong Kong, Denmark against the euro, the Gulf states. A managed float lies between: the currency floats but the authorities intervene to smooth or steer it, as China and India do.
A fixed rate is held in two ways. Intervention: when the currency is too weak, the central bank sells its foreign currency reserves and buys its own currency; when too strong, it does the reverse. Interest rates: raising them attracts capital and supports the currency. Both have limits: reserves run out, and interest rates set to defend a peg cannot also be set for the domestic economy.
Under a fixed rate a fall in the official value is a devaluation and a rise a revaluation; under a floating rate the market moves are depreciation and appreciation.
A devaluation improves the current account only if demand for exports and imports is elastic enough: the Marshall–Lerner condition, that the two price elasticities sum to more than one. Because volumes take time to respond, the balance usually worsens first and improves later: the J-curve.
Since 1983 the Hong Kong dollar has been fixed at HK$7.80 to the US dollar, never more than 7.85 or less than 7.75.
In summer 2025 traders borrowed cheap Hong Kong dollars, sold them for US dollars and pushed the rate to 7.85. The Hong Kong Monetary Authority stepped in seven times in two months.
What did it do?
At 7.85 the HKMA is obliged to buy every Hong Kong dollar offered to it, paying in US dollars from its US$400bn-plus Exchange Fund. Doing so drains Hong Kong dollars from the banks, so local interest rates rise, the borrow-and-sell trade stops paying, and the pressure fades. That is a fixed exchange rate at work, and it costs Hong Kong control of its own interest rates.
The diagram shows the market for **US dollars** in Hong Kong, priced in Hong Kong dollars. Demand for US dollars comes from importers and from capital leaving; supply from exporters and capital arriving. D1 and S1 clear at the 7.80 peg. The dotted lines are the band: 7.85 is the weak side for the Hong Kong dollar (more HK$ per US$), 7.75 the strong side.
Check: In the Hong Kong market for US dollars, who supplies US dollars?
US dollars are supplied by those who have them and want Hong Kong dollars: exporters paid in dollars and capital flowing in. Importers and outward investors demand dollars.
June 2025. Demand for US dollars jumps to **D2**: the carry trade. Left to float, the price would rise to 7.90, outside the band. At 7.85 buyers want US$90bn and the market supplies 70. The HKMA sells US$20bn from the Exchange Fund and takes Hong Kong dollars in return, which is supply curve **S2**. The rate holds at 7.85. Reserves fall by US$20bn.
Check: Traders borrow Hong Kong dollars to buy US dollars, shifting demand for US dollars to D2. Left alone the rate would reach 7.90. To hold 7.85 the HKMA must:
At 7.85 demand for dollars (90) exceeds supply (70). The HKMA supplies the missing US$20bn itself, buying HK$ with them.
Here is the cost. Every HK$ the HKMA buys leaves the banking system: the aggregate balance fell from over HK$170bn in May 2025 to under HK$60bn by April 2026. Scarcer HK$ means higher Hong Kong interest rates, which is what ends the outflow. But nobody chose those rates for Hong Kong’s economy. A fixed rate means importing the anchor country’s monetary policy: the impossible trinity.
Teacher tip: The impossible trinity: a country can have at most two of a fixed exchange rate, free capital movement and an independent monetary policy. Hong Kong chose the first two; the UK after 1992 chose the last two.
Check: When the HKMA buys Hong Kong dollars with its reserves, cash in the banking system falls and Hong Kong interest rates rise. This shows that a fixed exchange rate:
Intervention changes the money supply automatically, and rates move to close the gap with the anchor currency. Hong Kong imports US monetary policy.
The strong side. In spring 2025 money poured into Hong Kong and demand for US dollars fell to **D3**. The rate would have gone to 7.70. At 7.75 the market offered US$50bn and wanted 30: the HKMA bought the US$20bn surplus with new Hong Kong dollars. Reserves rose, cash flooded the banks and local rates fell towards zero, which is what set up the carry trade a month later.
Check: Capital now flows into Hong Kong and demand for US dollars falls to D3. The rate would fall to 7.70. At the 7.75 strong side the HKMA:
At 7.75 the market supplies more dollars than it demands. The HKMA absorbs them, adding to reserves and to HK$ in circulation, which pushes local rates down.
Why pegs break. Hong Kong’s holds because reserves cover every dollar in issue; most do not. On 16 September 1992 the UK spent reserves and raised Bank Rate to 15% to hold the pound in the ERM; the selling went on, and it left that evening. Once markets think a rate is wrong, defending it costs more than devaluing.
Check: In September 1992 the UK left the European Exchange Rate Mechanism after spending billions of reserves and raising interest rates to 15% in a day without stopping the sale of pounds. Why do pegs break?
A peg is a promise backed by reserves. If markets believe the rate is wrong, selling can exceed what reserves and rate rises can absorb, and the cost of defending it exceeds the cost of giving up.
Suppose a country does devalue. The J-curve shows its current account. At first the balance **worsens**: contracts are fixed, so the same imports cost more and the same exports earn less. Over a year or so volumes respond and the balance improves past where it started, provided the Marshall–Lerner condition holds: PED for exports plus PED for imports greater than one.
Check: A country devalues by 10%. Its price elasticity of demand for exports is 0.6 and for imports 0.5. According to the Marshall–Lerner condition, the current account will:
What matters is the sum. Above one, the extra export volume and the lost import volume outweigh the worse terms of trade.
The label says the HKMA buys US dollars. At 7.85 there is excess demand for US dollars; buying more would push the rate past the ceiling. To defend the weak side it sells US dollars from its reserves and buys Hong Kong dollars.
Explain how a central bank maintains a fixed exchange rate, and evaluate whether a fixed exchange rate system is preferable to a floating one.
A fixed exchange rate is one the authorities hold at a declared value or within a narrow band, such as the Hong Kong dollar at HK$7.80 to the US dollar within 7.75 to 7.85. A floating rate is set by demand and supply with no target; a managed float is a floating rate the central bank steers. Under a fixed rate a change in the official value is a devaluation or revaluation; under a float the market produces a depreciation or appreciation.
When capital flows out, demand for foreign currency rises and the home currency would fall below the band. On a diagram of the market for US dollars, demand shifts right and the free-market rate lies above the ceiling. At the ceiling demand for dollars exceeds supply, and the central bank fills the gap by selling dollars from its reserves and buying its own currency; the HKMA did this seven times in summer 2025, buying HK$91bn. It can also raise interest rates to attract capital back. Both have costs: reserves fall, and buying the home currency drains it from the banks, so interest rates rise whether or not the domestic economy needs them. A fixed rate therefore means importing the anchor country’s monetary policy, which is the impossible trinity: a fixed rate, free capital flows and an independent monetary policy cannot all be had.
A fixed rate gives traders and investors certainty, imposes discipline on inflation and, when credible, deters speculation. A floating rate absorbs shocks, lets interest rates serve the home economy and needs no reserves, at the cost of volatility and occasional overshooting.
**However,** the case depends on the economy. A small, open economy trading heavily with the anchor country and lacking a credible central bank of its own, such as Hong Kong or Denmark, gains most from fixing. A large economy with its own inflation target, such as the UK, loses more from surrendering monetary policy; its experience in September 1992, when reserves and a 15% interest rate failed to hold the pound in the ERM, shows how a peg can collapse once markets doubt it. And if a fixed rate has to be devalued, the current account improves only if the Marshall–Lerner condition holds, and after a J-curve in which it first worsens. For most large economies a float with an independent inflation target is preferable; for small open economies with large reserves, a fixed rate can be worth its cost.
Hong Kong Monetary Authority interventions: HKMA statements and press reports, 26 June to August 2025 and March 2026. The quantities on the diagram are for an illustrative market, chosen so the arithmetic is clear. Diagrams drawn with Rough.js and perfect-freehand, both MIT licensed.