Each day, currencies worth 73 times U.S. daily output change hands

73× Foreign exchange traded per day, as a multiple of U.S. GDP per day

No market is bigger than the market for money itself. What sets the price it clears at, why does that price jump when a central bank moves, and what does a country give up when it decides to fix it?

Open with the number. It is the notebook’s check of the textbook’s comparison: more than 5 trillion dollars of currencies are traded on a typical day, and U.S. GDP is about 25 trillion dollars a year, or 68 billion a day. The ratio is 73. The textbook rounds daily GDP up to 70 billion and says “more than 75 times”. Against world GDP (the U.S. is about 15 percent of it) the ratio is still 11.

Why it matters: in the short run the exchange rate is set in this market, by investors who move money between countries in seconds. In the long run it is pinned down by something much slower, the prices of goods. The chapter is about those two horizons and the policy choices between them.

Ask the room: who has changed money abroad recently? Did you get more or fewer euros, yen or pesos for a dollar than a few years ago? Why do you think it changed?

Today: three questions

• What sets the exchange rate in the long run? Money, prices and the law of one price; why real exchange rates still drift.
• What moves it in the short run? Interest rates, a flatter IS curve, foreign rates as a demand shock and the policy trilemma.
• What happens when a country fixes it? Gold, Bretton Woods, currency crises and the euro crisis; then three policy choices.

These are the three acts of the deck and the three branches of the mind map on the title slide.

The headline result: in the long run, money and prices set the exchange rate; in the short run, interest rates move it. An open economy can have only two of a stable exchange rate, its own monetary policy and free capital flows, and the crises show what happens when that choice breaks.

The chapter builds on Chapter 8 (the quantity theory of money), Chapters 11 to 13 (the IS curve, monetary policy and AS/AD) and Chapter 19 (trade and the current account). Nothing in the short-run model has to be relearned: the open economy only adds a channel and a shock.

The long run: prices set the rate

Act I

The price of a dollar

How much has the price of a dollar, in yen and in euros, changed since fixed exchange rates ended in 1971?

Ask the room first. Collect two or three guesses: does a dollar buy more yen today than in 1971, or fewer? How about euros since the euro began in 1999?

Before the reveal, fix the units, because students mix them up all the time: E is the price of one dollar in foreign money (yen per dollar, euros per dollar). A rise in E means a dollar buys more: the dollar appreciates.

The dollar lost three quarters of its value against the yen by 1995

Data: Board of Governors of the Federal Reserve System, H.10 Foreign Exchange Rates, via FRED: yen per dollar (EXJPUS) and dollars per euro (EXUSEU, inverted), monthly. Chart inspired by Fig. 20.1 in Jones (2024), Macroeconomics, 6th ed.

How to read it: two lines on two vertical scales. The yen per dollar is on the left axis (hundreds of yen), euros per dollar on the right axis (about one euro). Both lines are the price of a dollar: up is a stronger dollar. Hover gives both values of the month you point at.

Do live:

1. Hover on January 1971: 358 yen. Then April 1995: 84 yen. The dollar bought less than a quarter as many yen.
2. Follow the euro line from 1999: up to 1.17 euros in October 2000, down to 0.63 in July 2008.
3. Go to the right end: 162 yen in July 2026, the most since 1986. The newest data go beyond the textbook.

Point at the scale: a fall from 358 to 84 is a fall of about 77 percent, which is why we say “three quarters”.

A rise in \(E\) is an appreciation; a fall is a depreciation

• From 358 yen in January 1971 to 84 in April 1995: the dollar depreciated by about three quarters.
• Against the euro it first appreciated (to 1.17 in October 2000), then depreciated (to 0.63 in July 2008).
• Since 2024 a dollar has bought 143 to 162 yen: the dollar is strong again against the yen.

The exchange rate is a price, and like any price it moves. The question is what moves it.

Numbers: the notebook’s Fig. 1 key numbers and “What to notice”. The textbook’s numbers are all in the data: 357 (358) yen in 1971, 87 in mid-1995, 137 in early 2023; 0.92 euros in 1999, 1.15 at the start of 2002, the low of 0.64 in 2008.

Units again: the Federal Reserve publishes dollars per euro; the book’s file turns it around so both lines are the price of a dollar.

Beyond the textbook (the website version of Fig. 1 in the book chapter, all 25 currencies): with 1999 as the base, the dollar has lost value against the Swiss franc (index 55 in September 2026) and gained against the yen (138), the Mexican peso (181) and the Brazilian real (283). With 1971 as the base, the Swiss franc (20) and the yen (45) gained the most against the dollar.

Ask: which of these countries had low inflation? Hold the answer: the next two slides.

The law of one price ties the exchange rate to two price levels

\[EP = P^w \quad\Longrightarrow\quad \bar{E} = \frac{\bar{P}^w}{\bar{P}}\]

\[\bar{E} = \frac{\bar{M}^w}{\bar{M}}\times\frac{\bar{V}^w}{\bar{V}}\times\frac{\bar{Y}}{\bar{Y}^w}\]

The real exchange rate, \(RER \equiv EP/P^w\), the price of U.S. goods in foreign goods, equals 1 in the long run.

In the long run the exchange rate is the relative price of two piles of paper money.

The Model overview, first block. Assumption: arbitrage in goods. If a good is cheaper abroad, traders buy it there and ship it here until the gap closes. So once prices are written in one currency, a traded good costs the same everywhere: E P = P^w. Units: euros per dollar times dollars per good equals euros per good.

Apply it to the whole basket, so P and P^w are price levels, and use the quantity theory of Chapter 8 for each: P = M V / Y. Then the long-run exchange rate is the ratio of the two price levels, and each money supply enters with an elasticity of one. Doubling the number of dollars doubles U.S. prices and halves the value of the dollar (the notebook’s check: E from 0.92 to 0.46).

Intuition: a dollar is worth what it buys. If a basket costs 100 dollars here and 92 euros in Germany, a dollar is worth 0.92 euros.

The real exchange rate is the price of U.S. goods in foreign goods. The nominal rate is the price at which two currencies trade; the real rate is the price at which two countries’ goods trade. The law of one price sets the real rate to one in the long run.

Full derivation, with logs: the book chapter’s Model overview.

In growth rates, the currency with more inflation loses value

\[g_E \approx \pi^w - \pi \qquad\qquad g_{RER} \approx g_E + \pi - \pi^w\]

Inflation, U.S.	Inflation, Japan	yen per dollar, each year	after 25 years
4%	2%	about −2%	\((1.02/1.04)^{25} \approx 0.62\) of the start

Under PPP, the dollar should lose value against the yen at the yearly rate of the inflation gap.

Step 4 of the Model overview. Take logs of E = P^w / P and subtract last year’s version. The change in a log is approximately a growth rate, and the growth rate of a price level is inflation. This is the growth rule for a ratio: g of X/Z is g of X minus g of Z.

The exact yearly version is E_t / E_(t−1) = (1 + π^w) / (1 + π). The notebook simulates it: starting at 250 yen, 245.2 after one year and 153.9 after 25 years, a fall to 0.615 of the start, exactly (1.02/1.04)^25. The real exchange rate stays at 1.000 throughout.

With the quantity theory, inflation is money growth minus output growth, so the same equation says that a country that prints money faster ends up with a weaker currency.

The second equation, for the real exchange rate, is the one we use on the data in a few slides: when the real rate changes, either the nominal rate moved more than the inflation gap or prices moved while the nominal rate stood still.

The number of units in a currency means nothing by itself

A dollar buys 162 yen but less than one euro, so the yen must be a weak currency and the euro a strong one.

The level of \(E\) only reflects units. A new currency worth 1,000 old ones changes \(E\) a thousandfold and nothing real. What matters is how \(E\) changes, and what goods it buys.

A frequent confusion, also in the press. In September 2026 a dollar bought 1,359 Korean won but only 0.82 Swiss francs (the notebook’s numbers). That does not make the won “weak” in any useful sense. Korea could replace 1,000 old won by one new won tomorrow; E would fall a thousandfold, and every price in Korea would be divided by 1,000 too. The real exchange rate, and every real variable, would be untouched: the classical dichotomy of the long run.

What does carry meaning: the change in E over time (the dollar depreciated against the yen from 1971 to 1995), and the real exchange rate EP/P^w, which compares what goods cost.

Ask: if a country drops three zeros from its currency overnight, what happens to its living standards that day? (Nothing.)

Inflation in two countries

If the law of one price holds, what must be true of inflation in the United States and in Japan since 1971?

The dollar lost about three quarters of its yen value from 1971 to 1995. Was U.S. inflation higher or lower than Japan’s? By roughly how many points a year? Write down a guess.

Let students use the equation of two slides back: g_E ≈ π^w − π. A fall from 358 to 84 yen over 24 years is a fall of about 6 percent a year. If PPP explained all of it, U.S. inflation would have been about 6 points higher than Japan’s every year.

Collect guesses: higher or lower, and by how much. Many students remember Japan’s deflation and guess a large gap.

U.S. inflation ran above Japan’s in most years since 1971

Data: World Bank, World Development Indicators, consumer-price inflation (FP.CPI.TOTL.ZG); the U.S. and Japanese rows match FRED’s FPCPITOTLZGUSA and FPCPITOTLZGJPN; World Bank country groups. Chart inspired by Fig. 20.2 in Jones (2024), Macroeconomics, 6th ed.

How to read it: annual consumer-price inflation in percent. The menu picks a group of countries (“Textbook: U.S. and Japan” first); a click on a name in the legend hides or shows it. The vertical axis stops at 100 percent, so hyperinflations are cut, and the title says so.

Do live:

1. Hover on 1974: Japan’s inflation jumped to 23 percent after the first oil shock. Then 1980: U.S. inflation peaked at 13.5 percent.
2. From the late 1990s, Japan had deflation in many years; U.S. inflation stayed positive except in 2009.
3. Pick “G7” and add Germany; then pick “Europe & Central Asia” and add Switzerland. Over 1971–2024, Switzerland averaged 2.2 percent and Germany 2.7 percent, against 4.0 for the U.S. The low-inflation countries are the ones whose currencies gained most against the dollar in Fig. 1.

Data caveat: add Argentina and you get only seven points, 2018–2024. Its official price index was discredited in 2007–2015, and the World Bank series starts after that.

The direction fits the law of one price; the size does not

• The prediction: U.S. inflation was higher. Over 1971–2024 it averaged 4.0% a year, against 2.4% in Japan.
• But from 1971 to 1995 the gap was only 0.93 points a year (5.67% vs 4.74%). The dollar fell about 6% a year.
• After 1995 the gap widened to 2.10 points a year, yet the dollar gained against the yen.

Inflation gaps point the right way over decades, but they cannot explain the size or the timing of the moves.

Reveal the prediction first: yes, U.S. inflation was higher, so the dollar should depreciate against the yen. The notebook’s key numbers: 1971–1995, U.S. 5.67 percent and Japan 4.74 percent a year, a gap of 0.93 points; 1996–2024, 2.53 and 0.43 percent, a gap of 2.10 points; 1971–2024, 3.98 and 2.43, a gap of 1.56 points.

The textbook says U.S. inflation was higher “by only about 1 point a year” before 1995; the data agree. A 1-point gap over 24 years accounts for a fall of about a fifth, not three quarters.

After 1995 the puzzle flips: the inflation gap doubled, which should have weakened the dollar faster, but the dollar went from 84 to 162 yen.

So something other than relative price levels moved the yen. The next figure measures it: the real exchange rate.

The real exchange rate

If the law of one price held exactly, Japan’s real exchange rate would stay at 100. What do you expect the data to show from 1971 to today?

Recall the definition from Japan’s point of view: the price of Japanese goods in U.S. goods, (P^Japan / E) / P^US, as an index with 1971 = 100. If PPP held, the nominal rate would offset the inflation gap exactly, and the line would be flat.

Ask: flat, rising, or falling? And if it moves, should it come back?

Japan’s real exchange rate rose 2.5-fold to 1995, then fell back

Data: Groningen Growth and Development Centre, Penn World Table 11.0 (Feenstra, Inklaar and Timmer 2015), price level of GDP (pl_gdpo), Japan relative to the U.S.; our estimate from EXJPUS and World Bank consumer-price inflation via FRED. Chart inspired by Fig. 20.3 in Jones (2024), Macroeconomics, 6th ed.

How to read it: two measures of Japan’s real exchange rate against the U.S., both with 1971 = 100. The solid line is the textbook’s measure from the Penn World Table, which compares the price of everything each economy produces (GDP prices). The dashed line is our own estimate, built from the data of Figs. 1 and 2: the inflation rates multiplied up into price levels, divided by the exchange rate.

Do live:

1. Hover on 1995: the Penn index peaks at 245; our consumer-price estimate at 286.
2. Follow both lines down after 1995. In 2022–2023 the Penn index fell below its 1971 level (91 in 2023).

The two measures differ in level because they compare different baskets, but they move together year by year.

Prices alone predicted 269 yen per dollar in 1995; the market said 94

• The prediction: the line is far from flat. Japanese goods became 2.5 times as expensive relative to U.S. goods by 1995.
• Starting from 348 yen (the 1971 average), the inflation gap of 1972–1995 predicts 269 yen in 1995. The actual average was 94.
• Since 1995 the real exchange rate has fallen back, below its 1971 level by 2022.

Real exchange rates can move a long way, for decades. Something beyond money must drive them.

Numbers: the notebook’s “law of one price, 1971–1995” key numbers. Apply the yearly growth factors (1 + π^Japan)/(1 + π^US) of 1972–1995 to the 1971 average of 348 yen per dollar, and you get 269 yen in 1995. The actual average was 94. The yen’s real exchange rate rose a lot: the textbook’s “apparent contradiction”.

Beyond Japan (the website version in the book chapter): Germany (index 163) and Switzerland (262) also rose until 1995. By 2023 Japan (91) and Germany (87) were back below their 1971 level, while Switzerland stayed at 150.

Data lesson: Bermuda’s Penn price level in 2000 is more than 1,000 times the U.S. one, an obvious error in the table. Check a surprising number before you build on it.

Next: one reason the real rate drifts, the Balassa-Samuelson effect.

Fast productivity growth in factories makes a whole country dearer

\[P_TA_T = P_NA_N = W \quad\Longrightarrow\quad P_N = P_T\,\frac{A_T}{A_N}\]

\[RER = \left(\frac{A_T/A_N}{A^w_T/A^w_N}\right)^{1-\theta}\]

With \(\theta = 1/2\) and a productivity edge in traded goods of 3 points a year: \(1.015^{24} \approx 1.43\), a real appreciation of 43% in 24 years.

Only traded goods obey the law of one price. Haircuts follow local wages, so rich countries look expensive.

The Balassa-Samuelson block of the Model overview, which writes the textbook’s verbal explanation as a small model in our own notation.

Assumptions: two sectors, traded (cars, phones) and nontraded (haircuts, restaurant meals), with output per worker A_T and A_N. Workers move between sectors, so both pay the same wage W, and competition makes the wage equal the value of what a worker produces. The law of one price holds for traded goods only. The price level is an average of the two prices with weight θ on traded goods.

Step 1: a haircut costs more where the barber could have earned more in a factory. Step 2: substitute into the price level. Step 3: the traded parts cancel in E P / P^w, and what remains is relative productivity.

Intuition: productivity grows fast in factories, wages rise across the whole economy; factory goods do not get dearer, because they compete with the world, but services do. A country whose traded-goods productivity grows faster than abroad sees its real exchange rate rise. Japan caught up fast until 1995.

The number is the notebook’s check: a gap of 3 points a year, θ = 1/2, 24 years: times 1.43.

The price of a burger

Should a Big Mac cost the same number of dollars in Oslo, Tokyo and Mumbai? If not, where should it be dearest?

Ask for a ranking, not only yes or no. Remind the room that a Big Mac is partly traded inputs (beef, buns, the machines) and partly local labor and rent, which cannot be shipped.

Note for India: the index there uses the chicken Maharaja Mac, since McDonald’s in India sells no beef. Always look at what a data series measures.

Big Macs cost more in richer countries

Data: The Economist, The Big Mac index, open data (release of July 2026); Penn World Table 11.0, real GDP per person in 2023. Chart inspired by Table 20.1 and the Balassa-Samuelson discussion in Chapter 20 of Jones (2024), Macroeconomics, 6th ed.

How to read it: each dot is an economy in the July 2026 release. Across: GDP per person in 2023, on a ratio scale. Up: the price of a Big Mac in dollars at the market exchange rate. The line is fitted to the log of GDP, so it is straight on this axis. Hover for each country; hover on the line to read its equation.

Do live:

1. Point at the slope: about 2 dollars more for every tenfold rise in GDP per person.
2. The cheap end: Indonesia, India, the Philippines. The dear end: Switzerland, Uruguay, Norway.
3. Look for Japan and Taiwan: rich, yet cheap Big Macs. For Japan this is the weak yen of 2026, the fall of the real exchange rate in the last figure.

Data caveat: Venezuela is dropped. Its GDP per person in the table, about 160 dollars a year, is not plausible; the price data broke down in its hyperinflation.

The law of one price fails for burgers, in the direction the theory predicts

• In July 2022 a Big Mac cost from $2.34 in South Africa to $6.26 in Norway ($5.15 in the U.S.).
• The rich-country premium is the Balassa-Samuelson effect: the burger includes the cook’s wage and the rent.
• Whole price levels show it too: GDP cost 69% of the U.S. price in Japan, 52% in China and 25% in India in 2023.

A cheap Big Mac is mostly a cheap cook, not a mispriced currency.

Numbers: the textbook’s Table 20.1 is the July 2022 release, and every dollar price matches the notebook’s table. The one exception, the euro area, is rounding: the textbook divides 4.65 euros by a rounded 0.98 euros per dollar.

By the burger standard in 2022, Norway’s krone was 21.6 percent dearer than the U.S. and China’s yuan 30.9 percent cheaper.

The price levels in the third bullet are the Penn World Table levels (U.S. = 100) from the book chapter: Switzerland 101, Japan 69, China 52, India 25 in 2023. Poorer countries have lower price levels, as the model predicts.

The cloud is wide, so income is far from the whole story; exchange-rate swings (Japan in 2026) move dollar prices too.

A cheap burger does not prove a cheap currency

China’s Big Mac costs 30.9% less than the U.S. one, so the yuan is undervalued by 30.9% and must rise.

Part of the gap is the Balassa-Samuelson effect: in a poorer country, local labor and rent are cheap, so the price level is low even when the currency is fairly priced.

This myth appears in news stories every time the index is released. The Economist itself publishes a GDP-adjusted version of the index for this reason.

The correction uses the scatter: at China’s income, the fitted line predicts a cheaper burger than in the U.S. The right comparison is with countries of similar income, not with the U.S. Even then, the law of one price holds only for traded goods, and a Big Mac is mostly not traded.

The serious version of the claim, that China held its currency down in the 2000s, rests on other evidence (large current account surpluses and reserve accumulation behind capital controls), which Act II’s trilemma helps to read.

The short run: interest rates move the rate

Act II

In the long run goods set the exchange rate; in the short run, money markets do

	Long run	Short run
Nominal rate \(E\)	relative price levels (quantity theory)	set in currency markets; moves with the interest rate
Real rate \(EP/P^w\)	equal to 1 (law of one price)	moves with every surprise in \(E\): prices are sticky

Shipping goods takes months; moving money takes seconds. So in the short run the price of a currency follows the return on holding it.

This is the textbook’s Table 20.2, in our own words, and the hinge of the chapter.

Long run: the left column is Act I. Short run: inflation is sticky (the short-run model of Chapters 9 to 13), so over weeks P and P^w barely move. By equation (5), g_RER ≈ g_E + π − π^w ≈ g_E: any surprise move in the nominal rate is a move in the real rate.

Recall the hook: foreign exchange traded per day is 73 times U.S. daily GDP. Most of it is financial, not trade in goods. What do these investors care about? The return on holding each currency: the interest rate, and what they expect the currency to do.

Uncovered interest parity: two ways to invest a dollar must pay the same

\[1 + i_t = (1+i^w)\,\frac{E_t}{E^e_{t+1}}\]

\[\ln E_t \approx \ln E^e_{t+n} + n\,(i_t - i^w)\]

A U.S. rate 1 point higher for 5 years, with the anchor unchanged: the dollar jumps about 5% today (4.85% exactly).

A higher interest rate makes a currency stronger at once, by more the longer the gap is expected to last.

The UIP block of the Model overview. The textbook argues in words that higher U.S. rates make the dollar stronger; UIP is the standard equation behind that argument.

Assumptions: sticky prices; free financial flows; investors care only about the expected return; the expected future rate is anchored by PPP in the long run.

Step 1: kept at home, a dollar grows to 1 + i. Sent abroad, it becomes E euros, grows to E(1 + i^w), and is changed back at the expected rate E^e: E(1 + i^w)/Ee dollars. Money flows into the better route until both pay the same: the notebook checks 1.0600 at home and 1.0600 abroad.

Step 2: solve for today’s rate and take logs; for a gap expected to last n years, the effect is n times the gap. Notebook: a gap lasting 1 year gives a jump of 0.95 percent; 5 years, 4.85 percent (the log approximation says 1 and 5).

The trilemma model of Fig. 7 uses a jump of 5 percent per point: a gap expected to last about 5 years.

With sticky inflation a higher nominal rate is also a higher real rate (Fisher), so the same chain runs from R: R up, E up, the real exchange rate up.

After the jump, the currency is expected to slide back

The Fed raised its rate, so investors expect the dollar to keep rising from here.

UIP says the reverse: \(g^e_E \approx i^w - i_t\). With higher U.S. rates, the dollar jumps now and must then be expected to lose value by the interest gap, so holding dollars pays no more than holding euros.

Step 3 of the UIP block. Rearrange UIP as E^e/E = (1 + i^w)/(1 + i). If U.S. bonds pay more, the dollar must be expected to depreciate from now on, by exactly the gap. The only way to have both a higher rate and an expected depreciation is an immediate jump up, then a slide back toward the anchor.

This is the same logic as overshooting in Dornbusch’s famous model, which the textbook does not cover. Mention it as a pointer for curious students.

Caveat for later (Devil’s advocate): in the data, high-interest currencies often do not depreciate as UIP predicts; that is why “carry trades” can pay for a while. UIP is a benchmark, not a law of nature.

Net exports add a second channel to the IS curve

\[\frac{NX_t}{\bar{Y}_t} = \bar{a}_{nx} - \bar{b}_{nx}\,(R_t - \bar{R}^w)\]

\[\widetilde{Y}\!_t = \bar{a} - \bar{b}\,(R_t - \bar{r}) \qquad \bar{b} \equiv \bar{b}_i + \bar{b}_{nx}\] \[\bar{a} \equiv \bar{a}_c + \bar{a}_i + \bar{a}_g + \bar{a}_{nx} - 1 + \bar{b}_{nx}(\bar{R}^w - \bar{r})\]

Same form as the closed-economy IS curve: only the meaning of \(\bar{a}\) and \(\bar{b}\) is richer.

The open-economy IS block. Assumption 1: when U.S. goods are expensive (a high real exchange rate), exports are low and imports high. Assumption 2: by UIP, the real exchange rate rises with the U.S. real rate and falls with the foreign one. Together: net exports fall with the gap R − R^w. With b_nx = 0 we are back in Chapter 11.

Step 1, the trick: write R − R^w = (R − r̄) − (R^w − r̄). Now trade depends on the same gap as investment, plus a term that only the rest of the world can move.

Step 2 and 3: divide spending by potential output, substitute the shares, subtract one. The slope parameter b̄ now adds the trade channel to the investment channel; the shock ā now contains a demand shock from abroad, b_nx (R^w − r̄).

Why trade reacts like investment: national saving is used at home or lent abroad, S = I + NX (Chapter 19). A higher U.S. rate makes both uses less attractive at once.

The notebook’s baseline: b_i = 1 and b_nx = 1, so b̄ = 2, the value of Chapters 11 and 13, split half and half. The shares C 60 + I 19 + G 20 + NX 1 − 100 = 0: no shock at home.

Raising interest rates in an open economy

The Fed raises the real rate from \(\bar{r} = 5\%\) to 6%. In a closed economy with \(\bar{b}_i = 1\), output falls by 1%. With trade, \(\bar{b}_{nx} = 1\), how far does output fall? Is the IS curve steeper or flatter?

Give students a minute with the IS curve of the previous slide. The answer is a number: Ỹ’ = −(b_i + b_nx)(R’ − r̄).

A common wrong answer: “trade does not matter for interest rates”. Ask: what does a higher U.S. rate do to the dollar (UIP), and what does a stronger dollar do to exports?

A higher interest rate now cuts net exports as well as investment

Model figure: the open-economy IS curve \(\widetilde{Y} = -(\bar{b}_i + \bar{b}_{nx})(R - \bar{r})\) with \(\bar{b}_i = 1\) and \(\bar{r} = 5\%\); dashed: no trade channel. Adapted from Fig. 20.4 in Jones (2024), Macroeconomics, 6th ed.

How to read it: short-run output Ỹ across, the real interest rate R up. The solid line is the open-economy IS curve; the dashed gray line is the IS curve without the trade channel, the closed economy of Chapter 11. The Fed raises R from r̄ = 5 to R’ = 6 percent: the economy moves along the curve from A to B. The title splits the fall in output into investment and net exports. The slider sets b_nx.

Do live:

1. Start at b_nx = 1: Ỹ’ = −2.0 percent (investment −1.0, net exports −1.0).
2. Move to 0: −1.0, the closed economy, and the two curves coincide.
3. Move to 2: −3.0 percent. The curve turns flatter and B moves left.

Point at A: every curve passes through it, because at R = r̄ there is no gap to react to.

Trade doubles the punch of a rate rise: output falls 2%, not 1%

• The prediction: output falls by 2%, 1 point through investment and 1 through net exports.
• The chain: \(\uparrow R \Rightarrow \uparrow E \Rightarrow \uparrow EP/P^w \Rightarrow \downarrow NX/\bar{Y}\). The IS curve, slope \(-1/\bar{b}\), is flatter.
• The more net exports react, the more powerful monetary policy is: with \(\bar{b}_{nx} = 2\), a 1-point rise costs 3% of output.

In an open economy, monetary policy works through the exchange rate as well as through investment.

Reveal the prediction first: −2 percent, and a flatter curve. Numbers: the notebook’s slope check (−2.0 open, −1.0 closed) and the slider steps (b_nx = 0, 0.5, 1, 1.5, 2 give −1.0, −1.5, −2.0, −2.5, −3.0).

The “Your turn” of the book chapter: with b_i = 1, which b_nx makes a 1-point rise lower output by 3 percent? Answer: 2.

For the policy students: this is why central banks of small open economies watch the exchange rate so closely. Much of their policy works through it, and a currency that appreciates “on its own” tightens policy without any decision.

When the European Central Bank raises its rate

The ECB raises its real rate 2 points above \(\bar{r}\) for nine periods; the Fed only follows its usual rule. What happens to U.S. output and inflation in period 1? And in period 10, when the ECB’s rate returns to normal?

Hints if needed: by UIP, a higher foreign rate makes the dollar depreciate. What does a cheaper dollar do to U.S. net exports? And in the IS curve, the foreign rate sits inside ā.

Ask for signs first, then numbers. Most students get the boom; few predict the recession at the end.

A rise in foreign interest rates is a boom, then a slump, at home

Model figure: AS/AD with \(\bar{b} = 2\), \(\bar{m} = 1/2\), \(\bar{\nu} = 1\), \(\bar{\pi} = 2\%\), and \(\bar{R}^w - \bar{r} = 2\) in periods 1–9. Adapted from Fig. 20.5 in Jones (2024), Macroeconomics, 6th ed.

How to read it: short-run output Ỹ across, inflation π up. The original AS curve and the AS curve of the chosen period; the original AD curve and the shifted one, AD1 (the one that holds in that period in full color). The path so far carries the textbook’s letters A, B, C, D; the path still to come is gray. The slider picks the period; the title gives π and Ỹ.

Do live:

1. Period 1: the economy jumps from A to B. π = 3.00 percent, Ỹ = +1.00.
2. Periods 2, 4, 8, 9: inflation rises toward 4 percent (3.50, 3.88, 3.99, 4.00), the Fed raises U.S. rates, output returns to potential at C.
3. Period 10: the foreign rate is back at r̄, AD shifts back, and the economy drops to D: π = 3.00, Ỹ = −1.00.
4. Periods 12 and 20: down the AD curve back to A (π 2.25, then 2.00).

Point: the AS curves of periods 8, 9 and 10 are almost the same line, which is why the textbook draws C and D on one curve.

Half of the shock goes into output, half into inflation, and the end costs a recession

• The prediction, period 1: \(\bar{a} = \bar{b}_{nx}(\bar{R}^w - \bar{r}) = 2\), so \(\widetilde{Y}\!_1 = 2/(1 + \bar{b}\bar{m}\bar{\nu})\), which is +1%, and \(\pi_1 = 3\%\).
• At \(C\) output is back at potential, but inflation is near \(\bar{\pi} + \bar{a}/(\bar{b}\bar{m}) = 4\%\).
• Period 10: the shock ends and output falls to −1%. Inflation then halves its gap each period.

The U.S. did nothing, yet it boomed and slumped: interest rates abroad are a demand shock at home.

Reveal the prediction first. The AS/AD block of the Model overview: substitute the policy rule into the IS curve to get AD; put the Phillips curve in; solve. With b̄ m̄ = 1 and ν̄ = 1, the denominator is 2, so half of the shock goes into output. The notebook’s loop: period 1, Ỹ = 1.000 and π = 3.000; period 2, 0.500 and 3.500; period 9, 0.004 and 3.996; period 10, −0.998 and 2.998.

The international transmission of monetary policy: tighter money abroad weakens the dollar, raises U.S. net exports, and works like any positive demand shock. Caveat from the model overview: a recession abroad lowers ā_nx and pushes the other way, so in practice foreign tightening can have either sign.

Mirror image (the book’s “Your turn”): an ECB cut is a negative demand shock in the U.S., a slump first, then a boom when it ends.

Which shock moves which curve

Change	Effect on \(\bar{a}\) or \(\bar{b}\)	What moves	\(\widetilde{Y}\!_1\)
U.S. real rate \(R_t\) up 1 point	none	along the IS curve	\(-\bar{b} = -2\)
foreign rate \(\bar{R}^w\) up 2 points	\(\bar{a}\) up by \(\bar{b}_{nx}\times 2\)	IS and AD shift right	+1
the same, with \(\bar{b}_{nx} = 0\)	none	nothing: no trade channel	0
recession abroad: \(\bar{a}_{nx}\) down 1	\(\bar{a}\) down by 1	IS and AD shift left	−0.5

With a flatter AD curve, output reacts less to a demand shock on impact: the central bank’s response works through trade too.

The comparative statics of the Model overview (the IS table and the AS/AD table combined). A fifth row, left to the takeaway: trade more sensitive, b_nx up, raises b̄, and the IS and AD curves get flatter. The last column is first-period output in the AS/AD model with the baseline parameters; the first row is the IS-curve move with the rate held 1 point higher.

The takeaway uses the AD slope, −1/(b̄ m̄): a larger b̄ makes AD flatter, and the denominator 1 + b̄ m̄ ν̄ larger, so a given ā moves output less on impact.

Long run: the foreign real rate must equal the world marginal product of capital, R^w = r̄, so the foreign term in ā vanishes and ā = 0, as before.

A credible peg ties the interest rate to the anchor country

\[E_t = E^e_{t+1} = E_{fix} \quad\Longrightarrow\quad i_t = i^w\]

\[1 + i_t = (1+i^w)\,\frac{E_{fix}}{E^e_{t+1}} \qquad i_t \approx i^w + \text{expected depreciation}\]

With \(i^w = 5\%\) and an expected devaluation of 10%: \(i_t = 1.05/0.90 - 1 \approx\) 16.7% to keep investors holding the currency.

Fix the exchange rate and keep money free to cross borders, and you give up your own interest rate.

The fixed-exchange-rate block. Assumptions: the central bank promises to trade at E_fix and holds foreign reserves to keep the promise; investors believe it; flows are free, so UIP holds.

Step 1: put E_t = E^e into UIP; the ratio is one, so i = i^w. When the Fed raises its rate, the central bank of a pegging country must follow, whatever its own economy needs.

Step 2: in the long run, PPP with a fixed E makes inflation equal to inflation abroad, and money growth must match. That is why a peg can import a trusted central bank, and why it breaks when a government prints money to pay its bills (Chapter 8).

Step 3, the gold standard: each country fixed the price of an ounce of gold, so the exchange rate was the ratio of the two gold prices. Act III computes it.

Step 4, a peg that markets doubt: the notebook’s check, 16.67 percent (the rough rule says 5 + 10 = 15). If the central bank will not raise its rate that far, investors swap the currency for reserves, and when reserves run out the peg breaks: Mexico 1994, Asia 1997, Argentina 2002.

The policy trilemma

The Fed raises its rate by 1 point, and your economy needs no change. What happens to your interest rate and your exchange rate if you float? If you run a currency board? If you use capital controls?

Let students fill a three-row table: regime, interest rate change, exchange rate change, and what is given up. They have UIP and the peg equation.

Hint: under a float, the interest gap moves the exchange rate by 5 percent per point (the jump of the UIP slide).

Each regime keeps two goals and gives up the third

Model figure: with free flows, \(\%\Delta E = \eta\,(\Delta i - \Delta i^w)\) with an illustrative \(\eta = 5\); \(\Delta i^w = +1\), \(\Delta i^{need} = 0\). Adapted from Fig. 20.7 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the textbook’s triangle, live. The corners are the three goals: exchange-rate stability, an independent interest rate, an open capital account. Each side is a regime that achieves the two goals at its ends and gives up the one opposite. The slider picks the regime; the highlighted side and the box at the top right show the outcome: the change in your interest rate, the change in your exchange rate, and whether capital flows are free.

Do live:

1. “floating exchange rate”: interest rate +0.0, exchange rate −5.0 percent, flows free. Lost: a stable exchange rate. The United States today.
2. “currency board”: interest rate +1.0 point, exchange rate unchanged, flows free. Lost: monetary autonomy. Argentina 1991–2001; the euro.
3. “capital controls”: interest rate +0.0, exchange rate unchanged, flows blocked. Lost: free financial flows. China 1996–2005.

The trilemma bites only when you need a different policy from the anchor

• The prediction: a float keeps the rate and lets the currency fall 5%; a currency board raises the rate 1 point; controls keep both by blocking flows.
• UIP links three things: the exchange rate, the domestic rate, and free flows. Fix two and the third is set.
• If your economy needs the same change as the anchor’s, a peg costs nothing.

An open economy can have at most two of a stable exchange rate, its own monetary policy and free capital flows.

Reveal the prediction first; the slider’s readouts are the answer.

The model overview’s trilemma table: floating keeps autonomy and free flows and gives up a stable rate (E moves with i − i^w); a currency board or monetary union keeps a stable rate and free flows and gives up autonomy (i = i^w); capital controls keep a stable rate and autonomy and give up free flows (i ≠ i^w with E fixed breaks UIP, so trade in the currency must be rationed).

The book’s “Your turn”: the foreign rate rises 2 points and your economy needs a cut of 1 point. Under a currency board you get +2 instead of −1, a 3-point miss; under a float you get −1 and the currency falls by 15 percent (5 × 3).

Act III follows these choices through history and through the crises when they broke.

Regimes, crises and the euro

Act III

Two centuries of the pound

How did the price of a dollar in pounds behave under the gold standard, under Bretton Woods, and since 1973?

Ask for a sketch on paper: three periods, three shapes. Under gold, each country fixed the price of an ounce of gold in its money: from 1834, 20.67 dollars and 4.25 pounds. Under Bretton Woods (1944 to the early 1970s), other currencies were pegged to the dollar, and the dollar to gold at 35 dollars an ounce. Since 1973 the major currencies float.

One parity under gold, a few steps under Bretton Woods, then constant motion

Data: Lawrence H. Officer, MeasuringWorth, Dollar-Pound Exchange Rate From 1791, annual, inverted to pounds per dollar. Chart inspired by Fig. 20.6 in Jones (2024), Macroeconomics, 6th ed.

How to read it: pounds per dollar each year since 1791. The dashed line is the gold parity, 4.25/20.67 = 0.206 pounds per dollar; the shaded labels mark the gold era, Bretton Woods (1946–1971) and the float since 1973.

Do live:

1. Hover on 1864: the U.S. left gold in the Civil War, and the dollar fell to 0.100 pounds, half its value. It came back to the parity after the war.
2. Bretton Woods: long flat steps, 0.357 pounds (2.80 dollars per pound) in the 1950s and 1960s, 0.417 (2.40 dollars) after the pound’s devaluation in 1967.
3. After 1973 the line moves every year.

Over two centuries the dollar appreciated, from 0.22 pounds in 1791 to 0.76 in 2025, as the U.S. economy overtook the British one.

Floating rates move about 30 times as much a year as rates did under gold

Regime	Years	Average, pounds per dollar	Typical yearly change
Gold standard	1880–1913	0.206	0.26%
Bretton Woods	1951–1966	0.357	0.42%
Floating rates	1974–2025	0.629	8.51%

Each regime is a choice of two goals: gold and Bretton Woods chose stability; the float chose autonomy.

Numbers: the notebook’s “how stable is each regime?” key numbers, the standard deviation of yearly percent changes.

Read the trilemma into history. The gold standard combined fixed rates with free capital flows, and gave up monetary autonomy. Bretton Woods kept fixed rates and some autonomy with widespread capital controls. Since the 1970s, the big economies chose free flows and autonomy, and the exchange rate floats.

Beyond the U.K. (the website version in the book chapter, the Jordà-Schularick-Taylor Macrohistory Database for 17 countries): under gold, the U.K., Germany, France and the Netherlands moved by less than half a percent a year; Spain, Portugal and Japan (on silver until 1897) by about 6 percent. Germany’s index rises a trillion-fold in the 1923 hyperinflation. After 1973 all 17 float, with typical yearly changes of 8 to 13 percent.

Data lesson: for 1946–1955 that database records the pound at 0.33–0.42 per dollar, well above the official parity that MeasuringWorth reports. Two careful sources can measure “the exchange rate” differently.

When a peg breaks

A currency that has been fixed for years comes under attack and the peg breaks. How much of its dollar value can it lose, and how fast?

Ask for a percentage and a time span. Remind the room of the peg slide: when markets doubt a peg, the central bank must raise rates sharply or spend reserves; when reserves run out, it has no choice.

Four textbook cases: Mexico 1994, South Korea and Indonesia in the Asian crisis of 1997, Argentina 2002 (the end of the currency board).

When pegs broke, currencies lost half or more of their dollar value

Data: International Monetary Fund, Exchange Rates dataset (ER), national currency per U.S. dollar, end of period (XDC_USD.EOP_RT), annual. Chart inspired by Fig. 20.8 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the dollar value of each currency as an index, 1991 = 100. A fall to 40 means the currency buys only 40 percent of the dollars it bought in 1991. The buttons switch between the textbook’s window (1991–2004), the full window to 2025, and the full window on a ratio scale.

Do live:

1. “1991–2004 (textbook)”: Mexico’s peso falls to 40 in 1995; Korea’s won to 54 in 1997; Indonesia’s rupiah to 25 in 1998; Argentina’s peso, fixed at exactly 1 per dollar for ten years, to 30 in 2002.
2. “1991–2025”: the Argentine line sits on zero.
3. “1991–2025, ratio scale”: equal distances are equal percent changes. Argentina lost most of its value again and again; by 2025 its index is 0.07.

A defended peg holds until the reserves run out, then falls at once

• Mexico 1995: 40.2. Korea 1997: 53.8. Indonesia 1998: 24.8. Argentina 2002: 30.1 (1991 = 100).
• The mechanism: an expected devaluation demands a much higher interest rate. When the central bank stops paying it, investors take the reserves.
• Many firms and banks had borrowed in dollars, so the fall in the currency raised their debts and deepened the recessions.

A peg is only as strong as the reserves and the willingness to raise rates behind it.

Numbers: the notebook’s “size of each crisis”: Mexico 1995, a loss of 60 percent of the 1991 dollar value; South Korea 1997, 46 percent; Indonesia 1998, 75 percent; Argentina 2002, 70 percent. Korea’s 53.8 differs from the textbook’s “about 45”: the won was extremely volatile in late December 1997, and the IMF’s current data record a less extreme end-of-year rate than the vintage the textbook used.

Other crises in the book chapter’s IMF data (base: the year before): Russia 1998 (29), Türkiye 2001 (46) and 2018 (72, and 8.8 by 2025), Iceland 2008 (51), Ukraine 2014 (51), Egypt 2016 (43). In 1992 even the pound (81) and the lira (78) fell when they left the European exchange-rate mechanism.

The dollar debts are the textbook’s point about Asia: a depreciation that should help exporters also bankrupts firms whose debts are in dollars.

Does a floating exchange rate buy monetary independence?

Small open economies float to keep their own interest rate. But when the Fed tightens, capital flows out of them anyway. How much freedom does a float really give?

What the model predicts

• A float lets you set \(i\) for your own economy.
• The exchange rate absorbs the foreign shock.
• A hard peg must copy the anchor’s rate.

What is debated

• A global financial cycle may move credit even under floats.
• Do partial capital controls help, or leak?
• Currency falls hurt dollar borrowers.

The trilemma says floating restores autonomy. The history supports it: Obstfeld, Shambaugh and Taylor (2005, Review of Economics and Statistics) show, across the gold standard, the interwar years, Bretton Woods and the float, that pegs with open capital markets followed the base country’s interest rate much more closely than floats did.

The challenge: Rey (2013, Jackson Hole symposium), “dilemma not trilemma”, argues that a global financial cycle driven by U.S. monetary policy moves capital flows, credit and asset prices even in countries with floating rates, so true independence may need capital-flow management, whatever the exchange rate regime.

The middle ground: Klein and Shambaugh (2015, American Economic Journal: Macroeconomics) find that floats and wide capital controls give some autonomy, while limited, partial controls give little. The IMF’s “institutional view” (2012) accepts temporary capital-flow measures in some circumstances.

Name it as contested: the size of the global cycle, and how much a float insulates, are active research questions.

Ask the policy students: if you advised a central bank in a small emerging economy with large dollar debts, would you let the currency fall when the Fed tightens?

Bond yields in the euro area

Before the euro, Italy and Spain paid 4 to 6 points more than Germany to borrow for ten years. What happened to those gaps after 1999? And after 2008?

Ask for two predictions: one for the first decade of the euro, one for the years after the global financial crisis.

Hint: before the euro, part of the gap was expected inflation (Fisher: i = R + π^e). With one currency and one central bank, what happens to that part?

One currency closed the gaps in bond yields; fear of default reopened them

Data: OECD, Main Economic Indicators, 10-year government bond yields, via FRED (IRLTLT01DEM156N, IRLTLT01ESM156N, IRLTLT01ITM156N and 31 other countries); Eurostat (irt_lt_mcby_m). Chart inspired by Fig. 20.9 in Jones (2024), Macroeconomics, 6th ed.

How to read it: monthly 10-year government bond yields in percent. Germany, Spain and Italy are shown; a click on a name in the legend adds any of 42 countries. The markers show the start of the euro and Draghi’s speech of July 2012. The vertical axis starts at the textbook’s range; press “fit the lines shown” after adding Greece.

Do live:

1. 1995: Italy and Spain 4 to 6 points above Germany. By the end of 1998 all three are near 4 percent, and they move together until 2008.
2. 2011–2012: Italy above 7 percent, Spain 6.8 percent, Germany falling.
3. After July 2012 the gaps close again.
4. Add Greece, Portugal and Ireland, then press “fit the lines shown”: Greece 29.2 percent in February 2012, Portugal 13.9 in January 2012, Ireland 12.4 in July 2011.

German yields were negative in mid-2016 and from April 2019 to January 2022: investors paid the German government to hold their money.

The spread over Germany measures the fear of default

• The prediction: the Spanish spread fell from 5.00 points (April 1995) to 0.21 (December 1998); the Italian, from 6.32 to 0.13.
• After 2008 it reopened: Spain 4.32 and Italy 5.18 in November 2011; Spain 5.55 in July 2012. By August 2026: 0.45 and 0.80.
• One currency removed the inflation gap. What was left was default risk, \(i \approx i^{safe} + p\).

Inside a currency union, the spread is the price of the fear that a government will not repay.

Reveal the prediction first. Numbers: the notebook’s spreads over Germany.

Equation (14) of the book chapter: a lender who cares only about the expected return must earn the same on a risky bond as on the safe German one, (1 − p)(1 + i) = 1 + i^safe, so i ≈ i^safe + p. The notebook’s check: a safe rate of 2 percent and a default probability of 3 percent give a risky yield of 5.15 percent, a spread of 3.15 points.

Before the euro, spreads also held the expected inflation gap; one currency removed that part, so yields converged. After 2008, fear of default opened the spreads again.

Ask: why did the gap close after July 2012, when nothing had changed in Italy’s or Spain’s budgets? (Next slides.)

High debt can make a debt crisis self-fulfilling

\[\frac{\text{interest}}{Y} = i \times \frac{D}{Y} \qquad\qquad i \approx i^{safe} + p\]

Debt in 2012	Yield, November 2011	Interest bill
Italy: 137% of GDP	7.06%	9.7% of GDP a year
Germany: 88% of GDP	1.87%	1.7% of GDP a year

Fear raises \(p\), \(p\) raises \(i\), \(i\) raises the bill, and the bill makes default more likely.

The euro-crisis arithmetic of the Model overview. With debt at 100 percent of GDP, every point on the interest rate costs one point of GDP a year: at 2 percent, 2 percent of GDP; at 7 percent, 7 percent.

The table is the notebook’s: Italy’s and Germany’s debt in 2012 at their bond yields of November 2011, the peak of the Italian yield. It is an upper bound, since only new borrowing pays the new rate.

The spiral: investors’ beliefs can confirm themselves. A government that would be fine at 2 percent can be pushed toward default at 7 percent. A central bank that promises to buy the bonds lowers p and can stop the spiral: the first policy slide of this act.

The book’s “Your turn”: Greece’s debt was 171 percent of GDP in 2012. What rate makes its interest bill 10 percent of GDP? About 5.8 percent (10/171).

Money flows inside the euro

In the first years of the euro, banks in southern Europe lent much more at home. Where did the money come from?

Hint: a country that spends more than it produces borrows from the rest of the world. Its current account is in deficit. Ask which countries were in surplus.

Define the measure before the reveal: the current account balance is the trade balance plus the interest and dividends a country earns on its past lending abroad, minus those it pays. A deficit means the country borrows from the rest of the world.

Northern Europe lent to the south: deficits there, surpluses in Germany

Data: International Monetary Fund, World Economic Outlook (October 2026), current account balance, % of GDP (BCA_NGDPD); World Bank country groups. Chart inspired by Fig. 20.11 in Jones (2024), Macroeconomics, 6th ed.

How to read it: each bar is a country’s current account balance averaged over a period, in percent of GDP: 1993–1997, 2003–2007, 2008–2011 and the latest years, 2021–2025. Below zero is borrowing from abroad. The menu picks a group; a click on a period in the legend hides or shows it.

Do live:

1. Greece: from −2.1 percent of GDP in 1993–97 to −9.0 in 2003–07 and −11.5 in 2008–11.
2. Spain from −1.4 to −6.9; Portugal from −2.5 to −8.8.
3. Germany: from a small deficit (−1.1) to a surplus of 4.4, then 5.7.
4. Latest years: Portugal, Spain and Italy have small surpluses; Greece still borrows 7.5 percent of GDP.

Data warning: Ireland’s surplus of 10.8 percent in 2021–25 does not mean the Irish lend a lot abroad. The profits of multinationals registered in Ireland inflate both its GDP and its current account.

Once exchange-rate risk was gone, savings poured from north to south

• Greece borrowed 9% of its GDP a year from abroad in 2003–07; Spain about 7%. Germany lent about 4% of its GDP.
• The money went through banks: lending rose from 32% to 84% of GDP in Greece and from 83% to 168% in Spain, 1998–2007.
• The boom was wider than the south: Latvia, Estonia and Bulgaria, pegged to the euro, borrowed even more.

A capital inflow is a loan, and a loan must be repaid: the boom built the debts of the bust.

Numbers: the notebook’s period means for Fig. 11, and the bank-lending table of Fig. 10 in the book chapter (not shown in the deck): Greece 32.15 → 84.40 and Spain 82.60 → 167.95 percent of GDP from 1998 to 2007, while in Germany lending fell (113.50 → 97.13). Iceland’s banks lent 247 percent of GDP in 2007 just before they collapsed.

Latvia, Estonia and Bulgaria borrowed 15.4, 12.7 and 11.3 percent of GDP a year in 2003–07; on the lending side the Netherlands ran a surplus of 6.5 percent, more than Germany.

The link to Chapter 19: S = I + NX. Southern investment (much of it housing) exceeded southern saving, and northern saving filled the gap. With one currency, lenders saw no exchange-rate risk.

Data lesson from Fig. 10: the textbook’s figure uses a 2010 edition of the World Bank’s database; today’s edition revises Ireland’s 2007 lending down by 26 points. The story survives the revision.

Government debt after the crisis

What happened to government debt in Europe after the financial crisis of 2008, and has it come back down?

Ask students to name the countries they expect at the top in 2012 and 2017. Many will say Greece; few will guess Ireland’s jump.

The crisis pushed public debt above 100% of GDP in the south

Data: OECD, Economic Outlook No. 119, general government gross financial liabilities (GGFLQ), OECD Data Explorer; IMF, World Economic Outlook (October 2026), gross debt (GGXWDG_NGDP). Chart inspired by Fig. 20.12 in Jones (2024), Macroeconomics, 6th ed.

How to read it: government debt in percent of GDP for 2007, 2012, 2017 and 2025. The menu picks a group of countries and a measure: the OECD’s gross financial liabilities (the textbook’s) or the IMF’s gross debt. A click on a year in the legend hides or shows it.

Do live:

1. Ireland: 28 percent of GDP in 2007, 131 in 2012, when the government took over the debts of its banks.
2. By 2017 Italy (152), Greece (195) and Spain (118) are above 100.
3. 2025: debt is below its 2017 level in all six; Germany’s (63) is even below its 2007 level (66).
4. Switch to “Textbook six, IMF”: France in 2017 is 125 percent in the OECD measure, 99 in the IMF’s. Pick “G7, IMF”: Japan at 198 percent in 2025 and the U.S. at 126, still rising.

Debt jumped where banks failed and where output fell

• From 2007 to 2012 Ireland’s debt rose by 103 points of GDP: the state took over its banks’ losses.
• Greece’s debt reached 195% of GDP in 2017 (239% in the Covid year 2020, which the bars skip).
• Two measures, two numbers: the OECD also counts trade credits and values bonds at market prices.

A banking crisis becomes a debt crisis when the state stands behind the banks.

Numbers: the notebook’s debt table and the figure’s title (largest rise 2007–2012: Ireland, +103 points).

The doom loop: banks hold their own government’s bonds, and the government stands behind the banks. When bond prices fall, banks weaken; when banks need rescue, the debt rises. This is why the euro area later created a banking union, a policy the textbook mentions only briefly.

Always check which definition a debt number uses before comparing countries: France in 2017, 125 percent (OECD) or 99 (IMF).

Should the central bank stand behind government debt?

In July 2012, the ECB’s president promised to do “whatever it takes” to preserve the euro. Spreads fell before any bond was bought. Was that the right promise?

What the model predicts

• A credible buyer lowers \(p\), so \(i\) and the bill fall.
• It can stop a self-fulfilling run.
• A country with its own currency has this backstop.

What is debated

• Moral hazard: less pressure to fix budgets.
• Who bears the losses if a state defaults?
• Is it monetary policy or fiscal aid?

The facts: Mario Draghi’s London speech of 26 July 2012; the ECB then announced Outright Monetary Transactions (OMT), conditional bond purchases, in September 2012. OMT was never used, yet spreads fell (Fig. 9).

The case for: De Grauwe (2012, Australian Economic Review) argues that members of a monetary union issue debt in a currency they do not control, so they are prone to self-fulfilling liquidity crises that stand-alone countries with their own central bank avoid. De Grauwe and Ji (2013, Journal of International Money and Finance) find that a large part of the rise in southern spreads in 2010–2011 was not explained by fundamentals such as debt ratios, consistent with panic, and that spreads fell after the ECB’s announcement.

The case against: moral hazard (governments relax once the backstop exists) and the legal question of whether a central bank may finance states. OMT was challenged in the German Constitutional Court; the EU Court of Justice upheld it in 2015 (Gauweiler).

Name it as contested: how much of the 2010–2012 spreads was panic and how much fundamentals is still debated.

Ask: should a central bank that promises to buy bonds attach conditions? Who should set them?

Wage costs in a currency union

Inside the euro, Spain and Greece cannot devalue against Germany. How did their wage costs per unit of output move compared with Germany’s before 2008, and after?

Define the measure: the unit labor cost is the nominal wage divided by real output per worker, the wage cost of producing one unit of output. If firms price as a markup over it, its growth rate is inflation, and inside a union a gap in unit-labor-cost growth is a change in the real exchange rate.

Southern unit labor costs ran far ahead of Germany’s before 2008

Data: OECD, Productivity Database, unit labour costs, total economy (ULCE), OECD Data Explorer. Chart inspired by Fig. 20.13 in Jones (2024), Macroeconomics, 6th ed.

How to read it: unit labor costs as an index. The slider changes the base year of every line (1995, 1999, 2008, 2015): the index is 100 in that year for every country. A click on a name in the legend adds any of 43 economies.

Do live:

1. Base 1995: by 2007 Greece 170, Spain 142, Germany 98.
2. After the crisis: Greece’s costs fell from 194 in 2011 to 167 in 2019, Spain’s from 151 in 2009 to 142 in 2014. Wages adjust only slowly, even with very high unemployment.
3. Base 2008: since the crisis Germany’s costs rose the most and Greece’s the least. The gap is closing, slowly.
4. Add Portugal (144 by 2007), Ireland (149) and Latvia (235, on a fixed rate to the euro).

Without an exchange rate, only wages and productivity can restore competitiveness

\[g_{ULC} = g_W - g_{Y/L} \qquad\qquad g_{RER} = g^{Spain}_{ULC} - g^{Germany}_{ULC}\]

• By 2007 unit labor costs had risen 70% in Greece and 42% in Spain; in Germany, not at all.
• Wages 4% and productivity 1% a year in the south, 1.5% and 1.5% in Germany: \(1.03^{10} \approx 1.34\), a third dearer in a decade.
• With \(g_E = 0\), the way back is an internal devaluation: wages growing slower than productivity, for years.

A country with its own currency can devalue in a day; inside a union, the same correction takes years of high unemployment.

Equation (15) of the book chapter, the growth rule for a ratio applied to ULC = W/(Y/L). Inside the euro the nominal exchange rate between Spain and Germany is fixed at one, so their real exchange rate is P^Spain/PGermany; with prices a fixed markup over unit labor costs, the real exchange rate grows at the gap in unit-labor-cost growth. The notebook’s check: the south’s costs times 1.340 after 10 years, Germany’s times 1.000.

The takeaway’s devaluation “in a day” is what the crises of Fig. 8 show, with their own costs (dollar debts, imported inflation): a devaluation is fast, not free.

Southern Europe’s high unemployment in the 2010s is the slow version of the same adjustment.

Productivity behind the costs

Is the euro area’s problem only that wages ran ahead, or did productivity also fall behind? How has productivity grown since 2000 in Italy, Spain, France, Germany and the United States?

Total factor productivity is the Ā of the production model of Chapter 4: how much output an economy gets from its capital and labor. Ask for a ranking of the five countries’ growth since 2000.

Productivity stalled or fell in Italy, Spain and France

Data: Groningen Growth and Development Centre, Penn World Table 11.0 (Feenstra, Inklaar and Timmer 2015), TFP at constant national prices (rtfpna); World Bank country groups. Chart inspired by Fig. 20.14 in Jones (2024), Macroeconomics, 6th ed.

How to read it: total factor productivity as an index with 2000 = 100 for every country, so the lines show growth, not levels. The menu picks a group; a click on a name in the legend hides or shows it.

Do live:

1. “Textbook: U.S. and Europe”: in 2023 the U.S. is at 113 and Italy at 82.
2. France, Spain and Italy are lower in 2019 than in 2000, and lower still in 2023. In Italy and Spain the decline started in the 1990s; in France after 2006.
3. Germany kept up with the U.S. until 2019. The Covid year 2020 is a dip everywhere except the U.S.
4. Pick “Large emerging economies”: China almost doubled (198); Korea and India rose by about 20 percent.

Data warning: Ireland’s TFP rises by 15 percent from 2019 to 2023, an accounting effect of multinational profits, the same one as in its current account.

The deeper problem is slow productivity, not only high wages

• Average TFP growth, 2000–2023: U.S. 0.53%, Germany 0.29%, France −0.28%, Spain −0.31%, Italy −0.87% a year.
• U.S. growth itself slowed: from 0.88% a year in 1990–2003 to 0.38% in 2003–2019.
• Falling TFP raises unit labor costs too: by \(g_{ULC} = g_W - g_{Y/L}\), slow productivity has the same effect as fast wages.

A currency union can remove exchange-rate crises, but it cannot make productivity grow.

Numbers: the notebook’s average TFP growth table. 1990–2003: U.S. 0.88, Germany 0.96, Japan −0.19, France 0.70, Spain −0.42, Italy −0.38 percent a year. 2003–2019: U.S. 0.38, Germany 0.42, Japan 0.11, France −0.15, Spain −0.28, Italy −0.82. 2000–2023: as on the slide, plus Japan 0.07.

A margin note of the textbook claims that Italian and Spanish TFP is now below its 1990 level. With the latest Penn data, Italy’s index is 1.2131 in 1990 and 0.9869 in 2023: still true.

Why TFP fell is debated: misallocation of capital to less productive firms after the credit boom, slow adoption of information technology, rigid product and labor markets. Chapter 6 (ideas) and Chapter 4 (TFP) give the tools.

Can a currency union work without exchange rates between its members?

Greece and Germany share one interest rate and one currency. When one booms and the other slumps, what adjusts instead of the exchange rate?

What theory asks of a union

• Workers who move to where the jobs are.
• Wages and prices that can fall.
• Shared budgets that send money to slumping regions.

What the euro had

• Little labor mobility across borders.
• Slow wage adjustment (Fig. 13).
• No large shared budget until 2020.

The optimum currency area: Mundell (1961, American Economic Review) asked when a region should share a currency. His answer: when shocks hit its parts alike, or when factors (above all labor) move freely among them. Later work added wage flexibility and fiscal transfers.

Krugman (2013, NBER Macroeconomics Annual, “Revenge of the Optimum Currency Area”) argues that the euro crisis played out as the theory predicted: large asymmetric shocks (the credit boom and bust of Figs. 9 to 12), little labor mobility, no fiscal union, so adjustment had to come through slow internal devaluation and high unemployment. The U.S. states share a dollar but also a federal budget and a mobile labor force.

Other side: the euro also brought benefits (no currency crises inside it, lower transaction costs, credibility for high-inflation members), and the union has added tools since: the ECB’s backstop, a banking union, and the joint recovery fund agreed in 2020 (NextGenerationEU). The right-hand column’s last bullet refers to that.

Ask: Latvia kept a fixed rate to the euro and went through a sharp internal devaluation after 2008, then joined the euro. Would you advise a new member to join?

“Exchange-rate models can’t even beat a coin toss”

A skeptic says: “Exchange rates swing every day for no visible reason. Famous studies found that no model forecasts them better than a random walk, and high-interest currencies don’t depreciate as interest parity says. Why teach theories that fail?”

Reply: the theories are not day-trading tools. They explain what forecasts cannot: over decades low-inflation currencies gained (Fig. 2), pegs that copy the wrong interest rate break (Fig. 8), and spreads track default fear (Fig. 9). The model says which forces move the rate, not where it will be next month.

Take the objection seriously: it is right on the facts. Meese and Rogoff (1983, Journal of International Economics) found that the standard exchange-rate models of the 1970s forecast no better than a random walk at horizons up to a year, even when given the actual future values of their fundamentals. Fama (1984, Journal of Monetary Economics) found that currencies with high interest rates tend, if anything, to appreciate, the opposite of UIP: the “forward premium puzzle” behind carry trades.

The reply uses the deck. PPP fails in the short run but its direction holds over decades: Act I. Rogoff (1996, Journal of Economic Literature), “The Purchasing Power Parity Puzzle”, documents that deviations from PPP do die out, but only slowly, over years. The trilemma and the peg arithmetic explain the crises of Act III, which are the events policymakers care most about. And the open-economy IS curve gives the right sign of a monetary policy shock.

The honest summary: in the short run exchange rates are hard to forecast because they are asset prices driven by expectations, like stock prices; the models tell us how they respond to policy and why they return toward fundamentals in the long run.

What to remember from this chapter

• The nominal rate \(E\) is the price of a currency; the real rate \(EP/P^w\) is the price of goods. In the long run \(\bar{E} = \bar{P}^w/\bar{P}\) and \(g_E \approx \pi^w - \pi\).
• The law of one price holds only loosely: Japan’s real rate rose 2.5-fold to 1995, and rich countries look expensive (Balassa-Samuelson).
• In the short run UIP moves \(E\) with interest rates. Net exports flatten the IS curve, and foreign rates are a demand shock at home.
• The trilemma: at most two of a stable rate, monetary autonomy and free flows. Gold, Bretton Woods, pegs and the euro are different choices of two.
• When a choice breaks, currencies collapse or spreads explode; inside a union, the cure is a slow fall in unit labor costs.

Ask each student to write one sentence answering the opening question: what sets the price of a currency, and what does a country give up when it fixes it? A good answer: in the long run relative price levels, set by money; in the short run interest rates and expectations. A country that fixes its rate with open capital markets gives up its own monetary policy, and must defend the peg with interest rates and reserves.

For the policy students, three habits: write down the units of an exchange rate before you interpret a change; ask which two corners of the triangle a country has chosen; and when a peg or a union is under stress, ask what will adjust instead of the exchange rate.

Walk the three branches of the mind map in about a minute each: the long run, prices set the rate (cyan); the short run, interest rates move the rate (lavender); regimes, crises and the euro (green). Click the image to zoom in.

Money and prices set the exchange rate in the long run; interest rates move it in the short run, and a fixed rate costs a country its own monetary policy

Next: Chapter 21, Parting Thoughts

This chapter added the price of a currency to the book’s models: the quantity theory and the law of one price for the long run, and uncovered interest parity inside the short-run model. The last chapter, Parting Thoughts, looks back at the whole book and at the open questions of macroeconomics.

Sources: data and figures

• Figs. 1 and 3, exchange rates and the real rate. Federal Reserve Board, H.10, via FRED (EXJPUS, EXUSEU); Groningen Growth and Development Centre, Penn World Table 11.0 (pl_gdpo; Feenstra, Inklaar and Timmer 2015, American Economic Review 105(10)). Inspired by Figs. 20.1 and 20.3 in Jones (2024).
• Fig. 2, inflation. World Bank, WDI (FP.CPI.TOTL.ZG), via FRED for the U.S. and Japan. Inspired by Fig. 20.2 in Jones (2024).
• The Big Mac scatter. The Economist, Big Mac index open data (July 2026); Penn World Table 11.0 (GDP per person, 2023). Inspired by Table 20.1 in Jones (2024).
• Figs. 4, 5 and 7, the models. The book chapter’s open-economy IS curve, AS/AD and trilemma models, adapted from Figs. 20.4, 20.5 and 20.7 in Jones (2024).
• Figs. 6 and 8, regimes and crises. Lawrence H. Officer, MeasuringWorth, dollar-pound rate from 1791; IMF, Exchange Rates dataset (XDC_USD.EOP_RT). Inspired by Figs. 20.6 and 20.8 in Jones (2024).

Sources: the euro area and research cited

• Figs. 9, 11 to 14, the euro area. OECD 10-year yields via FRED (IRLTLT01{CC}M156N) and Eurostat (irt_lt_mcby_m); IMF WEO, October 2026 (BCA_NGDPD, GGXWDG_NGDP); OECD Economic Outlook No. 119 (GGFLQ) and Productivity Database (ULCE); Penn World Table 11.0 (rtfpna); bank lending (in the notes): World Bank, Global Financial Development Database (GFDD.DI.01). Inspired by Figs. 20.9 to 20.14 in Jones (2024).
• The trilemma. Obstfeld, M., J. Shambaugh and A. Taylor (2005), Review of Economics and Statistics 87(3); Rey, H. (2013), “Dilemma not Trilemma”, Jackson Hole Economic Policy Symposium; Klein, M. and J. Shambaugh (2015), American Economic Journal: Macroeconomics 7(4); IMF (2012), The Liberalization and Management of Capital Flows: An Institutional View.
• The euro crisis. Draghi, M. (2012), speech at the Global Investment Conference, London, 26 July; De Grauwe, P. (2012), Australian Economic Review 45(3); De Grauwe, P. and Y. Ji (2013), Journal of International Money and Finance 34; Mundell, R. (1961), American Economic Review 51(4); Krugman, P. (2013), NBER Macroeconomics Annual 27.
• Do the models work? Meese, R. and K. Rogoff (1983), Journal of International Economics 14(1–2); Fama, E. (1984), Journal of Monetary Economics 14(3); Rogoff, K. (1996), Journal of Economic Literature 34(2).
• Textbook. Jones, C. I. (2024). Macroeconomics, 6th ed. W. W. Norton. Book chapter, with the data, code and every figure: Data Science and Macroeconomic Models, Chapter 20.