Ch. 13: Fixed versus Floating: International Monetary Experience

Based on Feenstra and Taylor, Chapter 19 (KOM chapter 18). There are no lecture slides for this topic on the site yet; check against the lecture when it is given.

The previous chapters showed how exchange rates are determined in the short run (interest parity, IS-LM-FX) and the long run (PPP, monetary approach). This chapter asks the policy question: should a country fix its exchange rate or let it float? The answer is a cost-benefit comparison. Fixing brings gains from economic integration and can help with credibility, but it costs the country its monetary policy autonomy. The chapter then looks at fixed rate systems (many countries fixing together, cooperatively or not) and ends with the history of the international monetary system: the gold standard, the interwar period, Bretton Woods and the floating era, all read through the lens of the trilemma.

Opening case: Britain and the ERM

In 1990 the United Kingdom joined the Exchange Rate Mechanism (ERM), the European system of pegged rates centred on the German mark. In 1992 Germany raised interest rates to fight inflation after reunification. Britain was in recession and needed lower rates. Keeping the peg meant importing high German rates at the worst moment. On 16 September 1992 ("Black Wednesday") Britain left the ERM and the pound depreciated. The UK economy then recovered well. The episode is the textbook example of the cost of a peg: when the shock hitting home differs from the shock hitting the base country, the peg forces the wrong monetary policy on home.

Benefits of fixed exchange rates: economic integration

A floating exchange rate creates uncertainty and transaction costs for anyone who trades or invests across borders. A firm exporting to a country with a volatile currency faces exchange rate risk on its revenues; hedging costs money; converting currencies costs money. A credible fixed rate (or a common currency) lowers these costs and so encourages trade, investment and migration between the countries. The more two economies already trade with each other, the larger this gain.

• Trade gains. Empirical work by Andrew Rose found that pairs of countries sharing a currency trade much more with each other than similar pairs that do not (early estimates suggested roughly three times as much; later studies found smaller but still positive effects). Direct pegs show smaller effects than currency unions.
• Market integration. The benefit of fixing is increasing in the degree of economic integration between home and the base country: trade in goods and services, capital flows and labor flows. If Home sells half its exports to the base country, removing exchange rate risk with that partner matters a lot. If it sells 2%, it hardly matters.

Summary of the benefit side: the more integrated home is with the base country, the larger the efficiency gain from fixing.

Costs of fixed exchange rates: loss of monetary autonomy

Under a credible peg with capital mobility, uncovered interest parity forces i = i*: home's interest rate must equal the base country's rate. Home can no longer use monetary policy to stabilize its own output. Whether that matters depends on the type of shock.

Asymmetric shock

Suppose a negative demand shock hits Home only (for example a fall in consumer confidence at home). In the IS-LM-FX model the home IS curve shifts left.

• Floating: Home's central bank expands the money supply, LM shifts right (down), the interest rate falls, the home currency depreciates, net exports rise and output returns toward its original level. Monetary policy stabilizes the economy.
• Fixed: Home cannot cut i, because i must equal i*. Foreign is not hit, so the base country has no reason to cut its rate. Home's output stays low. This is the stabilization cost of the peg.

Symmetric shock

Suppose the same negative demand shock hits both Home and the base country. The base country cuts its interest rate to stabilize its own economy. Under the peg Home follows automatically: i falls with i*. Both economies get the stimulus they need, the exchange rate does not have to change, and the cost of the peg is small. Under a float Home would have done the same thing anyway.

Two further points from FT. First, the cost also depends on whether other adjustment mechanisms exist. If workers can move easily from the depressed home region to the booming base country (high labor mobility), or if a central fiscal authority transfers money to the depressed region (as in the United States between states), the loss of monetary policy hurts less. Second, the cost is smaller if home's monetary policy was not used well anyway; a central bank with no credibility may gain from tying its hands.

The symmetry-integration diagram

FT combine the benefit and the cost in one diagram. On the horizontal axis is the symmetry of shocks between home and the base country (more symmetric to the right). On the vertical axis is the degree of economic integration (more integrated upward).

• The FIX line slopes downward. Along it the benefit of fixing just equals the cost. The downward slope reflects a trade-off: if shocks are less symmetric (higher cost), a country needs more integration (higher benefit) to make fixing worthwhile.
• Above and to the right of the FIX line: integration and symmetry are high, benefits exceed costs, so fix.
• Below and to the left: costs exceed benefits, so float.

Shifting the FIX line

A country can also move in the diagram: if trade with the base country grows after the peg starts, it moves upward. This is the "endogeneity" argument for the euro: joining a currency union may itself create the integration and symmetry that justify it. FT apply the diagram to Europe versus the United States: US regions are well integrated and have a federal fiscal system; European countries were less integrated, with less labor mobility, so many economists argued that parts of Europe lay below the FIX line when the euro was created.

Other benefits of fixing

Fiscal discipline, seigniorage and the end of the inflation tax

A government that runs deficits can finance them by printing money (seigniorage). Under a peg it cannot print freely, because excessive money growth would drain reserves and break the peg. A peg therefore may impose fiscal discipline. The argument is weak in practice: governments with large deficits often simply break the peg, so the causality may run from discipline to peg rather than the other way around.

Nominal anchor and inflation credibility

Under PPP and a credible peg, home inflation equals base inflation plus the rate of depreciation, which is zero: π = π*. Pegging to a low-inflation country imports that country's credibility. Expected inflation falls, wage setters lower their demands, and inflation can fall with less unemployment than a domestic disinflation would need. This is why many high-inflation countries (for example in Latin America) used exchange rate based stabilization, and why the ERM tied countries to the Bundesbank.

Fixed exchange rates and foreign currency debt

Many emerging market countries cannot borrow abroad in their own currency. Their external liabilities are denominated in dollars or euros, while many of their assets are in local currency. Economists call the inability to borrow internationally in one's own currency original sin (Eichengreen and Hausmann). The result is a currency mismatch: a depreciation raises the home currency value of foreign currency liabilities more than it raises the value of foreign currency assets.

The wealth effect of a depreciation

Measure external wealth in home currency. Let A_F be home's foreign currency assets and L_F its foreign currency liabilities (both in home currency units). Assets and liabilities in home currency do not change value when E changes. A depreciation of %ΔE changes external wealth W by approximately:

ΔW = (A_F - L_F) × (ΔE/E)

If the country has more foreign currency liabilities than assets (a net short position in foreign currency), a depreciation reduces wealth.

The wealth loss hits firms and banks that borrowed in dollars, causing bankruptcies, falling investment and sometimes a banking crisis. A depreciation can then be contractionary: the negative wealth effect outweighs the usual positive expenditure switching effect on net exports. Fearing this, emerging market central banks are reluctant to let their currencies move much even when they officially float. Calvo and Reinhart called this fear of floating. In terms of the diagram, foreign currency debt raises the cost of floating and shifts the FIX line down.

FT illustrate with the 1997 Asian crisis: countries such as Indonesia, Thailand and Korea with large dollar debts saw their currencies collapse, and the resulting balance sheet losses deepened their recessions, while countries with smaller mismatches suffered less.

Fixed exchange rate systems: cooperative and noncooperative

So far one country pegged to a passive base. In a fixed exchange rate system several countries peg to each other. With N currencies there are only N - 1 independent exchange rates (the N - 1 problem, or redundancy problem), so one degree of freedom is left: someone has to set the system's monetary policy.

Reserve currency system

In a reserve currency system one center country (base) sets its monetary policy independently, and the other (noncenter) countries peg to it and hold its currency as reserves. Examples: the dollar in Bretton Woods, the German mark in the ERM. The center country keeps monetary autonomy; the noncenter countries must follow. This is an asymmetric system.

FT analyze it with two countries in IS-LM-FX. Suppose a negative demand shock hits the noncenter country (Home). Its preferred policy is lower i. Two cases:

• Noncooperative: the center (Foreign) ignores Home's problem and keeps i* unchanged. Home is stuck with a recession.
• Cooperative: the center agrees to lower i* somewhat. Home's i falls with it, so Home's output rises. The center's output rises a bit too (it was not hit by the shock), which it accepts as a cost. The joint outcome can be better for the pair than the noncooperative one if the center's loss is smaller than Home's gain.

The ERM 1992 is a noncooperative case: the Bundesbank set rates for German conditions after reunification, and Britain and Italy left.

Cooperative and noncooperative adjustments to exchange rates

A peg can be changed: a devaluation lowers the value of the home currency (E rises), a revaluation raises it. In an adjustable peg system, when Home devalues against the center, Home's net exports rise and the center's fall: a beggar-thy-neighbor policy. If all noncenter countries devalue at once against the center, they gain nothing against each other but the center country loses. Noncooperative unilateral devaluations, as in the 1930s, can lead to rounds of competitive devaluation. Cooperative realignments agreed by all members avoid this.

The gold standard

Under the gold standard (classical period roughly 1870 to 1914) each central bank fixed the price of its currency in terms of gold and stood ready to buy and sell gold at that price. Because each currency had a fixed gold price, all exchange rates between them were fixed. Example: if one ounce of gold costs $20.67 in the US and £4.25 in Britain, the mint parity exchange rate is 20.67/4.25 = about $4.86 per pound.

The gold standard is a symmetric system: no country is the center, because every country pegs to gold rather than to another currency. The common monetary policy is set by the world supply of gold.

Price-specie-flow mechanism (Hume)

David Hume described how the system adjusts automatically. Suppose Britain runs a trade deficit. Gold flows out of Britain to pay for imports, the British money supply falls and British prices fall; in the surplus country gold flows in and prices rise. British goods become cheaper, the deficit shrinks, and the gold flows stop. The rules of the game asked central banks to reinforce this process (raise rates when losing gold, lower them when gaining gold), but in practice they often sterilized gold flows.

Strengths and weaknesses

• Strength: a nominal anchor; long-run price stability because money is tied to a physical commodity.
• Weakness: the price level depends on gold discoveries and on the world demand for gold, which can create deflation when gold is scarce (for example in the 1870s to 1890s) and inflation after big discoveries.
• Weakness: no monetary autonomy to fight recessions.

Interwar period

The gold standard broke down in World War I. Countries tried to restore it in the 1920s (Britain returned at the prewar parity in 1925, overvaluing the pound). During the Great Depression countries left gold one by one; research (Eichengreen and Sachs) shows that countries that left gold earlier recovered earlier, because they regained monetary autonomy and could reflate. The 1930s also saw competitive devaluations, trade barriers and capital controls: the noncooperative outcome.

Bretton Woods (1944 to 1973)

At Bretton Woods (New Hampshire, 1944) the Allies designed a new system to avoid the problems of the 1930s:

• The US dollar was pegged to gold at $35 per ounce, and the US promised to convert dollars held by foreign central banks into gold.
• Other currencies were pegged to the dollar: a gold exchange standard and a reserve currency system with the US as center.
• Pegs were adjustable: a country in "fundamental disequilibrium" could devalue with IMF approval.
• Capital controls were allowed and widely used, which let countries keep some monetary autonomy despite the pegs.
• The International Monetary Fund (IMF) was created to lend reserves to countries with temporary balance of payments problems; the World Bank to finance reconstruction and development.

Why Bretton Woods collapsed

In the 1960s US monetary and fiscal policy became expansionary (Vietnam War, Great Society programs), US inflation rose, and other countries had to import that inflation to keep their pegs. Germany and Japan, with low inflation preferences, resisted. At the same time the stock of dollars held abroad grew much larger than US gold reserves. This is the Triffin dilemma: the world needed more dollar reserves to grow, but the more dollars there were, the less credible was the promise to convert them into gold. Capital controls also became leaky, so speculative attacks became easier. In August 1971 President Nixon closed the gold window (ended dollar convertibility into gold). The Smithsonian Agreement of December 1971 tried to restore pegs at new rates, but by 1973 the major currencies were floating.

In the language of this chapter, Bretton Woods failed because the center country set policy for itself (noncooperative) while the noncenter countries wanted a different policy: the shocks and preferences had become too asymmetric.

The trilemma and its history

A country can have at most two of the following three: (1) a fixed exchange rate, (2) full capital mobility, (3) monetary policy autonomy. With a peg and capital mobility, UIP forces i = i*, so there is no autonomy. To keep autonomy under a peg, a country must close capital markets; to keep autonomy with open capital markets, it must float.

Obstfeld, Shambaugh and Taylor measure how closely interest rates of pegging countries follow base country rates. Under the gold standard and in the modern era with open capital markets, pegged countries' rates move almost one for one with base rates. Under Bretton Woods, when capital controls were common, the link was much weaker: controls really did buy autonomy. Floating countries' rates move much less with base rates in all periods. The data therefore support the trilemma.

The broad pattern of history: as capital mobility rose (gold standard) countries chose pegs and gave up autonomy; when democratic pressures for full employment grew (interwar, Bretton Woods) countries wanted autonomy and closed capital markets; when capital markets reopened from the 1970s, the big economies floated. Many developing countries today use intermediate regimes and partial controls.