In June 1981 the Fed’s interest rate reached 19.1%

19.1% The federal funds rate, June 1981

An overnight rate that banks charge one another. The Fed pushed it that high on purpose, and a deep recession followed. Why would a central bank do that, and how can one overnight rate move the whole economy?

Open with the number. It is the peak of the effective federal funds rate in the notebook’s Fig. 2: 19.10 percent in June 1981. In 2009–2015 the same rate sat between 0.07 and 0.24 percent, and in September 2026 it was 3.75 percent.

The puzzle the textbook opens with, in our words: the Fed’s main tool is a very short-term nominal rate, used only between banks. Yet a change in it moves stock prices, mortgage rates, investment plans and, in the end, GDP and inflation.

The 1981 answer: inflation had passed 13 percent, and the Fed under Paul Volcker decided to bring it down even at the cost of a recession. Act III of this deck shows why the model says a recession was unavoidable.

Ask the room: who has a loan or a credit card whose rate follows the Fed? (Most variable-rate loans do.)

Today: three questions

• How does the central bank set its interest rate, and why does it matter? The fed funds rate, the money market, the real rate and the yield curve.
• How does that rate move output and inflation? The IS/MP diagram, the Phillips curve, demand shocks and price shocks.
• What does it take to bring inflation down? The Volcker disinflation, the Great Inflation of the 1970s, the pandemic surge, and a policy rule.

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

The headline result: because inflation is sticky, the central bank’s nominal rate sets the real rate, the real rate sets short-run output through the IS curve, and output moves inflation through the Phillips curve. So lowering inflation takes a period of output below potential, and its total cost depends only on the slope of the Phillips curve.

This chapter completes the short-run model: Chapter 9 gave the Phillips curve in a simple form, Chapter 11 the IS curve. Chapter 13 adds a policy rule and expectations (the AS/AD framework).

The central bank’s lever

Act I

The federal funds rate

How much has the Fed’s policy rate moved since 1960, and what does it do in a recession?

Ask for a range: the lowest and highest values students expect. Many will remember the near-zero rates after 2008 and the hikes of 2022; few will guess how high the rate went around 1980.

Ask also: when a recession starts, does the Fed raise or lower its rate?

The Fed’s rate has swung from 19% to zero, and it falls in recessions

Data: Federal Reserve Board, effective federal funds rate (FEDFUNDS), via FRED; BIS central bank policy rates (WS_CBPOL); NBER recessions (USREC). Chart inspired by Fig. 12.2 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the monthly effective federal funds rate in percent since 1960, with the NBER recessions shaded. The menu picks a period; the buttons above the chart list a group of countries in the legend, and a click on a name adds that central bank’s policy rate (BIS data).

Do live:

1. “1960–2026, all the data”: the peak of 19.1 percent in June 1981; near zero in 2009–2015 and again in 2020–2021.
2. Choose “The Volcker years, 1977–1987”: 17.6 percent in April 1980, 18.9 percent in December 1980, above 19 percent in 1981.
3. Choose “Covid-19 and after, 2019–2026”, press G7, and add the euro area, the United Kingdom and Japan. Before the pandemic the ECB was at zero and the Bank of Japan at −0.1 percent. The Bank of England moved first (December 2021), then the Fed (March 2022), then the ECB (July 2022); the Bank of Japan waited until March 2024.

Data caveats: the euro-area series switches in September 2024 from the main refinancing rate to the deposit rate; eight euro countries stop in 1998 (Greece in 2000) because the ECB took over; in hyperinflations some policy rates reached hundreds of percent (Brazil around 1990), and the title warns when a line leaves the chart.

The policy rate is the central bank’s one lever, and it moves a lot

• Range since 1960: about 0.05% (2020–21) to 19.1% (June 1981).
• Volcker’s tight money: 17.6% in April 1980, 18.9% in December 1980, above 19% in 1981.
• Near zero in 2009–15 and 2020–21; fast hikes in 2022–23 against the return of inflation; cuts from September 2024.
• In and after nearly every shaded recession, the rate falls: the Fed leans against the slump.

A rate on overnight loans between banks is the instrument. The rest of the deck explains how it reaches output and prices.

All numbers are the notebook’s Fig. 2 key numbers. The latest value in the data is 3.75 percent (September 2026).

How the Fed makes the rate stick: it stands ready to lend to banks, or to pay them, at its chosen rate. No bank will pay a higher rate to borrow overnight, since it could borrow from the Fed. And no bank can keep lending below the rate, since others would borrow from it cheaply and park the money at the Fed for a riskless profit, an arbitrage.

Bridge: the next slides show the same idea through the money market.

The interest rate is the price of holding money

\[M^d = m_0 - i \qquad M^s = \bar{M}^s\]

\[i^* = m_0 - \bar{M}^s\]

• Cash and checking accounts pay nothing; a savings account pays \(i\). So \(i\) is the opportunity cost of holding money, and money demand slopes down.
• The central bank chooses the supply: a vertical line.
• Baseline: \(m_0 = 10\), \(\bar{M}^s = 5\), so \(i^* = 5\%\).

In the short run prices are sticky, so the money market clears through the interest rate, not through the price level.

The money demand line is illustrative (straight, in made-up units); the book chapter’s Microfoundations section has the derivation. Higher prices or output shift m0 to the right; new ways of paying (cards, phones) shift it to the left.

Above i, people want to hold less money than there is: they buy bonds or move money to savings accounts, and the rate falls. Below i, they want more money than there is, and the rate rises.

Contrast with Chapter 8: in the quantity theory, the price level adjusts to make money demand equal supply. Here inflation is sticky over a short horizon, so the interest rate does the adjusting (and with it the velocity of money).

Experiment: the central bank cuts the money supply

Money demand is \(M^d = 10 - i\) and the supply is 5, so \(i^* = 5\%\). The central bank cuts the supply to 4. What happens to the interest rate, and by how much?

Collect guesses. Students who think of money as “cheaper when there is less of it” may say the rate falls. Ask: if there is less money and people still want the same amount at 5 percent, who has to offer a better deal?

Less money means a higher interest rate

Model figure: money demand \(M^d = 10 - i\), money supply cut from 5 to \(M^{s\prime}\) (illustrative units). Adapted from Fig. 12.19 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the quantity of money across, the nominal interest rate up. Money demand slopes down; the supply is vertical. The old supply stays behind as a dashed line. The title gives the old and the new rate.

Do live:

1. At a new supply of 4: the rate rises from 5 to 6 percent.
2. Move the slider to 3.5: the rate rises to 6.5 percent.
3. Move it to the right of 5 (“loose” money): the rate falls below 5 percent.

The story behind the arrow: at the old 5 percent, people want 5 units but only 4 exist. Households ask their banks for cash; banks raise the rates they offer on savings until people are content to hold just 4 units.

Tight money is a higher rate, carried out by selling bonds

• One unit less of money raises \(i^*\) by one point: 5% → 6%.
• The Fed changes the supply by open-market operations: it sells bonds to take money in, buys bonds to put money out.
• A bond’s price and its interest rate move in opposite directions.

\[P_B = \frac{100}{1+i}\]

Price of a $100 bond	Interest rate
$95	5.26%
$96	4.17%
$98	2.04%
$99	1.01%

When the Fed sells bonds, their price falls and the interest rate rises: tight money.

Numbers: the notebook’s money-market check (supply 5 gives 5 percent, supply 4 gives 6 percent) and its bond-price key numbers. A bond that pays 100 dollars in one year and costs P_B today earns 100 − P_B on P_B, so i = (100 − P_B)/P_B; at 96 dollars that is 4/96 = 4.17 percent.

Why selling bonds lowers their price: the Fed adds to the supply of bonds that the public must absorb, and buyers take more only at a lower price, which is a higher yield.

Ask: the Fed buys bonds. What happens to their price, the money supply and the interest rate? (Price up, money up, rate down.)

Targeting the interest rate

Money demand rises from \(10 - i\) to \(11 - i\): people want more money at every rate. The central bank wants to keep the rate at exactly 5%. What must it do with the money supply?

A good answer: supply one more unit, 6 instead of 5. Ask what would happen if the bank did nothing: the rate would rise to 6 percent, although nothing real changed in the economy.

Pegging the rate makes the supply follow demand

Model figure: money demand \(M^d = m_0 - i\) with the interest rate pegged at 5% (illustrative units). Adapted from Fig. 12.20 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the same axes as before. The central bank now supplies whatever money is demanded at its target, so the supply curve is horizontal at 5 percent. The slider shifts money demand (m0 is demand at a zero rate).

Do live:

1. At m0 = 11 (the baseline after the shift): the money supplied is 6 units and the rate stays at 5 percent.
2. Move the slider from 8 to 12: the quantity of money changes one for one, the rate never moves.

Compare with the previous figure, where the supply was fixed: there, the same rise in demand would have lifted the rate to 6 percent. The notebook’s interactive version has a toggle for the two policies.

Central banks set a rate, not a quantity of money

• With the rate pegged, demand up by 1 unit is met by 1 more unit: 6 instead of 5, and \(i\) stays at 5%.
• With the supply fixed, the same shift would raise \(i\) to 6%: with sticky inflation, a higher real rate and lower output, for no real reason.
• Money demand shifts all the time: prices, output, new ways to pay.
• IS/LM: hold the supply fixed and the rate rises with output (an upward “LM” curve). Pegging the rate flattens it into our MP line.

Targeting the rate shields output from noise in money demand. That is why the MP curve is a flat line.

Numbers: the notebook’s money-market check: demand 11 − i with the supply fixed gives 6 percent; with the rate held at 5 percent the supply needed is 6.

The IS/LM model is the older textbook version of the short-run model; it assumes a fixed money supply. Because real central banks adjust the supply to hit a rate, the two approaches give the same answers once the LM curve is replaced by the MP curve.

Ask: why did some central banks target the money supply in the past? (Money growth and inflation go together in the long run, Chapter 8. The Fed under Volcker described its policy as targeting money growth, which meant very high interest rates.)

Because inflation is sticky, setting \(i\) means setting \(R\)

\[i_t = R_t + \pi_t \quad\Longrightarrow\quad R_t = i_t - \pi_t \qquad\text{and, with } \Delta\pi_t = 0,\quad \Delta R_t = \Delta i_t\]

• Sticky inflation: over about six months, inflation does not react to a policy move. Firms reprice only now and then, information and calculation are costly, and many wages and rents are fixed in nominal contracts.
• So each point on \(i\) is a point on \(R\). To hit a real rate \(R^T\), set \(i = R^T + \pi\).
• 2% inflation: \(i = 7\%\) gives \(R = 5\% = \bar{r}\). 10% inflation, as in 1979: it takes 15%, and 16% gives \(R = 6\%\).

The MP curve: a horizontal line at the real rate the central bank picks.

The Fisher equation (Chapter 8) says a nominal rate pays for the real return plus the loss of purchasing power. Subtract inflation from both sides and compare the moments just before and after a policy move: if inflation does not move, the real rate moves one for one with the nominal rate. The notebook’s Check 1 confirms: i from 7 to 8 with inflation stuck at 2 raises R by 1.0 point; if inflation also rises to 3, R does not move.

If inflation were not sticky (the classical dichotomy holding even in the short run), a higher nominal rate could simply come with lower inflation and the real rate, and output, would be untouched. Stickiness is what gives monetary policy real effects.

Ex ante and ex post: a borrower decides with expected inflation, i − π^e; the realized rate is i − π. Inflation above what was expected lowers the real rate paid: with i = 7, expected inflation 2 and actual 4, the ex ante rate is 5 percent and the ex post rate 3. The model ignores the difference.

The reasons for stickiness come from the textbook’s microfoundations (a pizza parlor that cannot re-price every day), in our words.

A long rate is roughly the average of the short rates expected over its life

\[(1 + i^{(n)})^n = (1 + i_1)(1 + i^e_2)\cdots(1 + i^e_n) \quad\Longrightarrow\quad i^{(n)} \approx \frac{i_1 + i^e_2 + \dots + i^e_n}{n}\]

• One 5-year bond, or five 1-year bonds in a row: if one paid more, everyone would switch.
• Expected 1-year rates of 2, 3, 4, 5 and 6% give a 5-year rate of about 4% (3.990% exactly).
• A short-rate hike lifts long rates more when markets expect it to last.

The yield curve plots rates against maturity. When it is inverted, markets expect short rates to fall.

Take logs of both sides: the log of a power is the exponent times the log, the log of a product is the sum of the logs, and ln(1 + x) ≈ x for small rates. Divide by n. This is the term structure of interest rates (expectations version).

The notebook’s Check 2 computes the exact five-year rate that matches five one-year bonds in a row: 3.990 percent, against a simple average of 4.000 percent.

This is how an overnight rate reaches the rates firms and homebuyers pay: by changing what markets expect the overnight rate to be over the next years. In practice long rates also include a risk premium, which shows up in the country data two slides ahead.

The yield curve in 2022–2023

In 2022 the Fed raised its rate fast, month after month. What do you expect happened to the yield curve: did short rates or long rates rise more?

Let students reason from the term-structure equation: a hike that markets expect to be reversed later raises short rates more than long ones.

As the Fed raised rates, short yields overtook long yields

Data: U.S. Department of the Treasury, Daily Treasury Par Yield Curve Rates, first trading day of each month. Chart inspired by Fig. 12.6 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the yield of U.S. Treasury securities (percent) against maturity, from 1 month to 30 years. As in the textbook, maturities are equally spaced on the axis, not drawn to scale. Press play, or drag the slider through the months of 2022, 2023 and 2026.

Do live:

1. January 2022: short yields near zero (0.05 percent at 1 month), rising with maturity to 1.63 percent at 10 years: the usual shape.
2. Play through 2022: the left end climbs above 4 percent faster than the right end.
3. Stop in April 2023: 4.70 percent at 1 month, 3.43 at 10 years. Inverted.
4. Jump to 2026: upward-sloping again, at a higher level than in 2022 (on October 15, 2026: 4.17 percent at 1 month, 5.10 at 10 years).

Data caveat: the file has no curves for 2024 and 2025, so the animation jumps from December 2023 to January 2026.

Markets expected the Fed to cut later, so the curve inverted

Spread, 10-year minus 3-month
first inverted month	Nov 2022
most inverted (June 2023)	−1.89 points
September 2026	+1.05 points

• April 3, 2023: the 5- and 10-year yields were 1.18 and 1.27 points below the 1-month yield.
• Short yields rose with the Fed’s rate; long yields rose less, because markets expected rate cuts in the years ahead.
• By 2026 the curve slopes up again.

The yield curve is a market forecast of future policy, read off today’s prices.

Numbers: the notebook’s inversion key numbers (10-year minus 3-month yield, monthly) and the textbook’s two days (January 3, 2022 and April 3, 2023). The September 2026 spread is from the OECD comparison in the next figure.

The term-structure equation in reverse: if the 10-year rate is below the 3-month rate, the average of expected short rates over the next ten years must be below today’s short rate. Markets were betting that the Fed would win against inflation and then cut.

Ask: why should anyone hold a 10-year bond paying less than a 3-month bill? (Because they expect to roll the bill over at lower rates later, and they lock in today’s long rate.)

The slope of the yield curve around the world

Here is the term spread, the 10-year yield minus a 3-month rate, for 47 economies since 2000. Before which U.S. recessions did it turn negative?

Students may have heard that an inverted yield curve “predicts recessions”. Ask them to name the recessions since 2000 (2001, 2007–09, 2020) and to watch where the U.S. line crosses zero.

Inverted curves came before the last three U.S. recessions

Data: OECD, Financial market statistics (DSD_STES@DF_FINMARK), 3-month interbank rate (IR3TIB) and long-term government bond yield (IRLT), on the OECD Data Explorer. Chart inspired by Fig. 12.6 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the term spread in percentage points, monthly since 2000; below zero the yield curve is inverted. The buttons above the chart list a group of countries in the legend; click a name to add it.

Do live:

1. The United States: below zero in 2000, in 2006–2007 and in 2019, each time before a recession; and again from November 2022 to April 2025, this time with no recession so far.
2. Add the United Kingdom, Germany and Canada: their curves inverted too, for about two years from late 2022, as central banks everywhere raised short rates fast.
3. Add Japan: its curve never inverted. The Bank of Japan kept its short rate near zero.
4. Add Greece: +28 points in February 2012, when investors feared default on its 10-year bonds. Russia: −18 points in January 2009, when short rates soared in a banking and currency crisis.

Data caveat: the U.S. 3-month rate here is an interbank rate; it differs from the Treasury bill rate of the previous figure by 0.18 points on average.

A spread mixes expected policy with risk

• An inversion says markets expect short rates to fall, often because they expect a slump that will make the central bank cut.
• That forecast was right in 2000, 2006–07 and 2019. In 2022–25 it was not, at least so far.
• A long yield also carries a risk premium: Greece’s +28 points in 2012 was fear of default, not a forecast of Greek monetary policy.

Read the yield curve as expectations plus risk, not as a recession alarm.

The predictive record of the term spread is a well-known empirical result: Estrella and Hardouvelis (1991, Journal of Finance) and Estrella and Mishkin (1998, Review of Economics and Statistics) find that the spread forecasts U.S. recessions a year or more ahead better than most indicators. The 2022–25 inversion without a recession is a warning against reading it mechanically: the inversion may have reflected an expected fall in inflation, and with it in nominal rates, rather than a slump.

The notebook’s U.S. dates: below zero before the recessions of 2001, 2007–09 and 2020; again from November 2022 to April 2025.

Bridge to Act II: the overnight rate now reaches the real rate and longer rates. Next, how the real rate moves output and inflation.

IS, MP and the Phillips curve

Act II

Experiment: the whole chain at once

Inflation is stuck at 2%, the real rate that keeps output at potential is \(\bar{r} = 5\%\), and the central bank sets \(i = 7\%\). It raises \(i\) to 8%. What happens to the real rate \(R\), to short-run output \(\widetilde{Y}\) and to the change in inflation \(\Delta\pi\)?

Ask for three numbers, or at least three signs. The slope parameters are b̄ = 2 (IS curve) and ν̄ = 0.5 (Phillips curve); give them if students want numbers.

Expected answer: R from 5 to 6 percent, output −2 percent, inflation falling by 1 point a year.

Each link of the chain turns one number into the next

Model figure: \(R = i - \pi\), \(\widetilde{Y} = \bar{a} - \bar{b}(R - \bar{r})\), \(\Delta\pi = \bar{\nu}\widetilde{Y}\), with \(\pi = 2\%\), \(\bar{r} = 5\%\), \(\bar{b} = 2\), \(\bar{\nu} = 0.5\). Adapted from Fig. 12.1 in Jones (2024), Macroeconomics, 6th ed.

How to read it: four circles, from left to right: the nominal rate the central bank sets, the real rate (MP curve), short-run output (IS curve) and the change in inflation (Phillips curve). The title repeats the chain in numbers. The slider sets the nominal rate; inflation is held at 2 percent and there is no demand shock.

Do live:

1. At i = 7: R = 5 = r̄, output at potential, inflation steady.
2. Move to 8: R = 6, Ỹ = −2 percent, Δπ = −1 point.
3. Move to 5: R = 3, Ỹ = +4 percent, Δπ = +2 points: a boom that speeds up inflation.

The notebook’s interactive version also has sliders for inflation and the demand shock: raise inflation and the nominal rate by the same amount, and nothing real changes.

Monetary policy in a nutshell: \(\uparrow i \;\Rightarrow\; \uparrow R \;\Rightarrow\; \downarrow \widetilde{Y} \;\Rightarrow\; \downarrow \Delta\pi\)

Building block	What it does	At \(i = 8\%\), \(\pi = 2\%\)
MP curve	the nominal rate sets the real rate, \(R = i - \pi\)	\(R = 6\%\)
IS curve	a real rate above \(\bar{r}\) lowers spending and output	\(\widetilde{Y} = -2\%\)
Phillips curve	output below potential slows inflation	\(\Delta\pi =\) −1 point

What matters is \(i - \pi\). With inflation at 10%, the same chain needs \(i = 16\%\).

The central bank moves inflation only indirectly: through the real rate, then output.

Numbers: the live figure at i = 8 and the notebook’s “model in a nutshell” table, which runs the chain with inflation stuck at 10 percent: i = 15 gives R = 5, Ỹ = 0, Δπ = 0; i = 16 gives R = 6, Ỹ = −2, Δπ = −1; i = 17 gives R = 7, Ỹ = −4, Δπ = −2.

A negative demand shock lowers output at any interest rate; the central bank can offset it by cutting i. That is the next sequence.

Ask: inflation is stuck at 4 percent and there is no demand shock. What nominal rate keeps output at potential? (9 percent: R must equal r̄ = 5.)

The IS/MP diagram: the central bank picks the height, the IS curve picks the output

\[\text{IS: } \widetilde{Y}\!_t = \bar{a} - \bar{b}\,(R_t - \bar{r}) \qquad\qquad \text{MP: } R_t \text{ set by the central bank}\]

\[\widetilde{Y}\!_t = 0 \quad\Longleftrightarrow\quad R_t = \bar{r} + \frac{\bar{a}}{\bar{b}}\]

• \(\bar{a}\) is the demand shock (zero in normal times); \(\bar{b}\) is how strongly spending reacts to the real rate.
• When borrowing costs more than capital earns (\(R > \bar{r}\)), output falls; a fall in demand is offset by a cut of \(1/\bar{b}\) points per point of \(\bar{a}\).

The IS curve comes from Chapter 11. When borrowing costs more than new capital earns, firms and households postpone investment. The baseline of every diagram: ā = 0, b̄ = 2, r̄ = 5 percent.

The second equation solves the IS curve for the rate that delivers output at potential (equation 3 of the book chapter). The notebook’s Check 1 confirms it: ā = −2 gives R = 4.00 percent, ā = −1 gives 4.50, ā = +1.5 gives 5.75, and at each the IS curve returns Ỹ = 0.

In terms of the nominal rate: i = r̄ + ā/b̄ + π.

Experiment: the central bank raises the real rate

The economy is at potential with \(R = \bar{r} = 5\%\) and \(\bar{b} = 2\). The central bank raises \(R\) to 6%. Where does output go? Which curve moves?

Two answers to collect: the number (−2 percent) and the geometry. Many students say “the IS curve shifts”. Ask them to wait for the figure.

A higher real rate moves the economy along the IS curve into a recession

Model figure: \(\widetilde{Y} = \bar{a} - \bar{b}(R - \bar{r})\) with \(\bar{a} = 0\), \(\bar{b} = 2\), \(\bar{r} = 5\%\); MP line raised to \(R'\). Adapted from Fig. 12.4 in Jones (2024), Macroeconomics, 6th ed.

How to read it: short-run output Ỹ across (percent), the real interest rate R up. The downward line is the IS curve; the horizontal line is the MP curve. A is the starting point at R = r̄ = 5 percent; the old MP line stays behind as a dashed gray line, and B is the new point.

Do live:

1. At R′ = 6: B at Ỹ = −2 percent. Only the MP line moved.
2. Move the slider to 7: Ỹ = −4 percent.
3. Back to 5: the economy returns to A.

The starting point A is the textbook’s first IS/MP figure (12.3), which the book chapter draws separately.

Each point on the real rate costs \(\bar{b}\) points of output

• \(R\) from 5% to 6%: \(\widetilde{Y}\) from 0 to −2%. The prediction: −2, and the MP line moves, not the IS curve.
• A change in \(R\) is a movement along the IS curve; a demand shock shifts it.
• A flatter IS curve (larger \(\bar{b}\)) makes the same hike more painful: with \(\bar{b} = 2.5\), output falls to −2.5%.

The central bank can cause a recession on purpose, simply by holding the real rate above \(\bar{r}\).

Numbers: the notebook’s model checks (R = 6 gives Ỹ = −2.0 percent; the slope of the IS curve is −2.0) and the Fig. 4 “What to notice”.

Why would it want to? To bring inflation down, as in 1979–1982 (Act III), or to cool a boom: with ā = 1.5, the rate that brings output back to potential is 5.75 percent.

Experiment: a housing bubble bursts

Households lose wealth and confidence, and demand falls: \(\bar{a}\) drops from 0 to \(-2\). With \(\bar{b} = 2\), where does output go if the central bank does nothing? What real rate brings it back to potential?

Collect both numbers. The first is −2 percent (output moves one for one with ā at an unchanged rate). The second needs the formula R = r̄ + ā/b̄ = 5 − 1 = 4 percent.

Cutting the real rate can offset a fall in demand

Model figure: the IS curve shifts left by the demand shock \(\bar{a}'\); the MP line is cut to \(R' = \bar{r} + \bar{a}'/\bar{b}\). Adapted from Fig. 12.5 in Jones (2024), Macroeconomics, 6th ed.

How to read it: two panels with the same axes as before. Panel (a): the bust shifts the IS curve left; at the unchanged rate the economy moves from A to B. Panel (b): the central bank cuts the rate to R′; the economy moves down the new IS curve to C, back at potential. The slider sets the size of the shock.

Do live:

1. At ā′ = −2: B at −2 percent; R′ = 4 percent brings C back to 0.
2. Move to −3: the cut must go to 3.5 percent.
3. Move to −1: a cut to 4.5 percent is enough.

In the model, policy can undo any demand shock; in practice it is late

• Bust of \(\bar{a}' = -2\): output −2% at \(R = 5\%\); cut to \(R' = 4\%\) and output is back at potential.
• The larger the shock, the larger the cut. The flatter the IS curve, the smaller the cut needed.
• But the size of a shock is seen only with a delay, and rates take roughly 6 to 18 months to move spending: Friedman’s “long and variable lags”.

The rule \(R = \bar{r} + \bar{a}/\bar{b}\) is easy to write and hard to apply in real time.

Numbers: the notebook’s model checks and the Fig. 5 “What to notice”.

In 2008 the housing bust turned out to be much more than an ordinary demand shock: a financial crisis that pushed the policy rate to zero (Chapter 14).

Ask: the Fed waits and cuts only to 4.5 percent after a bust of −2. Where does output end up? (On the new IS curve: −2 − 2 × (4.5 − 5) = −1 percent.)

The Phillips curve: output moves the change in inflation

\[\pi_t = \pi^e_t + \bar{\nu}\widetilde{Y}\!_t \qquad \pi^e_t = \pi_{t-1} \qquad\Longrightarrow\qquad \Delta\pi_t = \bar{\nu}\,\widetilde{Y}\!_t + \bar{o}\]

• Price setting: each firm raises prices by the inflation it expects elsewhere, plus more when demand is strong and less when it is weak.
• Adaptive expectations: firms expect this year’s inflation to equal last year’s.
• Price shocks \(\bar{o}\): oil, steel, a big union settlement (\(\bar{o} > 0\)); a wave of immigration that holds wages down (\(\bar{o} < 0\)).

Output sets the change in inflation. At potential and with no shock, inflation stays wherever it was.

Inflation here is the rise of prices over the coming year, π_t = (P_{t+1} − P_t)/P_t, set by firms during year t when they choose next year’s prices. That timing is why a recession this year already lowers this year’s inflation.

The slope ν̄ is illustrative: 0.5 in all the figures. Chapter 9 estimated about 1/3 for the United States.

Two names for the two terms: ν̄Ỹ is demand-pull inflation (strong demand pulls prices up); ō is cost-push inflation (higher costs push them up).

Experiment: a boom

Inflation has been 2% for years, and \(\bar{\nu} = 0.5\). This year output is 2% above potential. What happens to inflation this year? And next year, if output is still 2% above potential?

Ask for both numbers. A common wrong answer for the second part: “inflation stays at 3 percent”. Ask what firms expect next year after a year of 3 percent inflation.

Above potential, inflation rises; below it, inflation falls

Model figure: \(\Delta\pi = \bar{\nu}\widetilde{Y} + \bar{o}\) with \(\bar{\nu} = 0.5\), \(\bar{o} = 0\). Adapted from Fig. 12.7 in Jones (2024), Macroeconomics, 6th ed.

How to read it: short-run output Ỹ across, the change in inflation Δπ up. The line rises through the origin; the dot is the economy at the output on the slider. The title does the arithmetic.

Do live:

1. Ỹ = 0 (point A): Δπ = 0, inflation steady.
2. Ỹ = +2: Δπ = +1 point.
3. Ỹ = −2: Δπ = −1 point. This is the textbook’s recession figure (12.11), which the book chapter draws separately.
4. Ỹ = −4: Δπ = −2 points.

The notebook’s version also has a slider for the slope: with ν̄ = 1, the textbook’s example of firms raising prices by 8 instead of 10 percent in a recession of 2 percent.

A boom raises inflation every year it lasts

• At \(\widetilde{Y} = +2\%\): inflation rises by \(0.5 \times 2 =\) 1 point, from 2% to 3%.
• Next year firms expect 3%, so a boom that continues takes it to 4%, then 5%…
• To keep output above potential for ever, inflation must rise without limit: there is no lasting trade-off.
• History: the 1960s curve linked the level of inflation to activity. Friedman and Phelps argued that firms would build higher inflation into their prices; the 1970s proved them right.

A steeper curve means prices react more to demand; a flat curve means a deep slump is needed to lower inflation.

Numbers: the live figure and the Fig. 7 “What to notice”: output 2 percent above potential raises inflation by 0.5 × 2 = 1 point a year.

The algebra: with Ỹ held at x every year, inflation after t + 1 years is the starting inflation plus ν̄x(t + 1). Inflation is a running total of all past gaps.

Friedman (1968, American Economic Review) and Phelps (1967, Economica) are the classic statements of the natural-rate hypothesis.

Quantity theory or Phillips curve? In Chapter 8, faster growth of real GDP lowers inflation; here a boom raises it. The quantity theory is about the long run, where more output is more supply; here a positive Ỹ is demand above what the economy normally supplies.

Experiment: oil prices jump

The economy is at potential with inflation steady. Oil prices jump: a price shock of \(\bar{o} = 1\) point. What happens to inflation this year? To keep inflation steady instead, where would output have to be?

Students usually get the first part. The second part is the policy dilemma: ask for a number, Ỹ = −ō/ν̄.

A price shock shifts the whole curve up

Model figure: \(\Delta\pi = \bar{\nu}\widetilde{Y} + \bar{o}\) with \(\bar{\nu} = 0.5\), the curve shifted by the price shock \(\bar{o}\). Adapted from Fig. 12.8 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the same axes as the previous figure. The old curve stays behind dashed; the new curve is shifted up by ō for one period. The economy stays at potential, Ỹ = 0.

Do live:

1. At ō = +1: at potential, inflation rises by 1 point.
2. Move to +2: the curve shifts up 2 points.
3. Move to −1 or −2: a favorable shock (cheaper energy, an inflow of workers) shifts it down.

Ask: where on the new curve is Δπ = 0? (At Ỹ = −2 percent.)

Demand-pull and cost-push inflation add up

Case	Demand-pull \(\bar{\nu}\widetilde{Y}\)	Cost-push \(\bar{o}\)	\(\Delta\pi\)
oil shock at potential	0	+1	+1
boom of +2%	+1	0	+1
both	+1	+1	+2
slump of −2% with the shock	−1	+1	0

• Holding inflation steady after the shock takes output of \(-\bar{o}/\bar{\nu} =\) −2%.
• Keeping output at potential lets inflation rise by \(\bar{o}\), and it stays up.

A price shock leaves the central bank a hard choice: a recession now, or higher inflation for good.

The table is the notebook’s demand-pull and cost-push key numbers, with ν̄ = 0.5.

Why it stays up: by the Phillips curve, a one-year shock raises this year’s inflation; next year firms expect the higher rate, and with output at potential nothing pulls it back. The notebook’s Check 4: a one-year shock of 1 point at R = r̄ takes inflation from 2 to 3 percent, and it stays at 3. Undoing it later costs a recession of 2 percent-years in total.

This dilemma returns in Act III as a policy question (oil shocks) and in the pandemic surge.

Which shock moves which curve

A rise of 1 in …	Effect on \(\widetilde{Y}\)	Effect on \(\Delta\pi\)	What moves
the real rate \(R\) (nominal rate, with \(\pi\) sticky)	\(-\bar{b} = -2\)	−1	MP line up, along the IS curve
the demand shock \(\bar{a}\)	+1	+0.5	IS curve right
the price shock \(\bar{o}\)	0	+1	Phillips curve up, one period
\(\pi\) and \(i\) together	0	0	nothing

Policy moves output through \(R\); output and price shocks move inflation.

The table combines the book chapter’s “What the model predicts” tables for the MP curve, the IS/MP diagram and the Phillips curve, at the baseline b̄ = 2 and ν̄ = 0.5.

Read the rows as a checklist for any news headline: is it a change in the policy rate, a change in spending at every rate, or a change in costs? Then trace it through the chain.

The last row is the classical dichotomy: if inflation and the nominal rate rise together, the real rate, output and the change in inflation do not move.

The whole model is one equation for inflation

\[\pi_t = \pi_{t-1} + \bar{\nu}\left[\bar{a} - \bar{b}\,(R_t - \bar{r})\right] + \bar{o}\]

• Given last year’s inflation, this year’s policy \(R_t\) and the shocks, it gives this year’s inflation: a difference equation.
• Within a year the chain runs one way: \(R_t \Rightarrow \widetilde{Y}\!_t \Rightarrow \pi_t\).
• Once policy returns to \(R = \bar{r} + \bar{a}/\bar{b}\), output is at potential and inflation stays where it has got to. Any inflation rate can be a resting point.

Nothing inside the model pulls inflation back to 2%. Only policy can move it.

Substitute the IS curve for Ỹ in the Phillips curve in levels. The book chapter’s simulate function applies this equation year after year; the notebook checks it against the closed form for inflation as a running total of past gaps and shocks.

Act III runs this equation for the Volcker disinflation and for the 1970s.

Disinflation and the Great Inflation

Act III

U.S. inflation and the real interest rate

U.S. inflation since 1960, with the fed funds rate and the real rate \(R = i - \pi\). In 1979–80, with inflation above 11%, was monetary policy tight or loose?

Most students guess “tight”, because the nominal rate was very high (13.4 percent on average in 1980). Ask them to compute the real rate before the reveal: subtract inflation.

Inflation fell only after the real rate jumped in 1981

Data: BLS CPI (CPIAUCSL) and Federal Reserve Board fed funds rate (FEDFUNDS), via FRED; World Bank WDI inflation (FP.CPI.TOTL.ZG); BIS policy rates; NBER recessions. Chart inspired by Fig. 12.9 in Jones (2024), Macroeconomics, 6th ed.

How to read it: annual CPI inflation as a filled area, with U.S. recessions shaded. Click the legend to add the policy rate and the real interest rate (the ex post rate, policy rate minus actual inflation). The menu picks one of the 49 countries with a policy rate.

Do live:

1. Inflation: 1.1 to 1.5 percent in the early 1960s, 11.0 percent in 1974, 13.5 percent in 1980, down to 3.2 percent by 1983.
2. Turn on the real rate: −0.06 percent in 1979 and −0.15 in 1980; then 6.00 in 1981, 6.10 in 1982, 5.93 in 1983.
3. 2021–2022: inflation back (8.0 percent in 2022), with a deeply negative real rate.
4. Choose Türkiye: 72 percent inflation in 2022 with a policy rate averaging 12.6 percent.

Data caveats: the October 2025 CPI was never collected (government shutdown), so 2025 averages the other 11 months. The real rate is ex post: investors decide with the inflation they expect, which the chart cannot show.

High nominal rates were not tight money until the real rate rose

Year	Inflation	Fed funds rate	Real rate \(R = i - \pi\)
1979	11.25%	11.19%	−0.06%
1980	13.50%	13.36%	−0.15%
1981	10.38%	16.38%	6.00%
1982	6.16%	12.26%	6.10%
1983	3.16%	9.09%	5.93%

With a real rate near 6% for four years, inflation fell from 13.5% to 3.2%, and the recession of 1981–82 was the deepest in decades.

The table is the notebook’s Fig. 9 key numbers (annual averages). 1984: inflation 4.37, fed funds 10.22, real rate 5.86.

The model’s reading: before 1981 the nominal rate rose roughly with inflation, so the real rate stayed near zero and output was not pushed below potential. In 1981 the Fed raised the nominal rate far above inflation; because inflation is sticky, the real rate jumped, and through the IS and Phillips curves inflation came down.

Inflation usually falls after recessions (the shaded bands), as the Phillips curve predicts.

A high interest rate is not the same as tight money

Interest rates are very high, so monetary policy must be very tight.

Tightness is the real rate. In 2022 Türkiye’s policy rate averaged 12.6% while inflation reached 72%: a real rate of about −60%. Inflation began to fall only after the rate was raised to about 49% in 2024.

Numbers: the notebook’s Fig. 9 “What to notice” for Türkiye: inflation 72 percent in 2022 with a policy rate averaging 12.6 percent; after the central bank raised its rate to about 49 percent in 2024, inflation started to fall (35 percent in 2025).

The U.S. case is the same lesson: a fed funds rate of 13.36 percent in 1980 sounds very tight, but with 13.5 percent inflation the real rate was −0.15 percent.

The reverse mistake also happens: near-zero nominal rates in a deflation can be tight money, because the real rate is positive. Ask: with inflation at −1 percent and a nominal rate of 0, what is the real rate? (+1 percent.)

Experiment: the first step of the Volcker disinflation

It is 1979. Inflation is stuck at 10%, and a nominal rate of 15% keeps the real rate at \(\bar{r} = 5\%\). The Fed raises the nominal rate to 16%. What happens to the real rate and to output?

The point is that the numbers look scary but the logic is the same as with 2 percent inflation: only i − π matters. Collect answers: R = 6 percent, Ỹ = −2 percent.

With inflation stuck at 10%, a 16% rate is a real rate of 6%

Model figure: \(R' = i' - \pi\) with sticky \(\pi = 10\%\); IS curve \(\widetilde{Y} = -\bar{b}(R - \bar{r})\) with \(\bar{b} = 2\), \(\bar{r} = 5\%\). Adapted from Fig. 12.10 in Jones (2024), Macroeconomics, 6th ed.

How to read it: the IS/MP diagram again. The slider sets the new nominal rate; the title subtracts the sticky 10 percent inflation to get the real rate, and reads output off the IS curve.

Do live:

1. At i′ = 15: R′ = 5, output at potential.
2. At 16: R′ = 6, Ỹ = −2 percent.
3. At 17: R′ = 7, Ỹ = −4 percent.

Only the real rate matters, and stickiness is what makes it move

• \(i' = 16\%\) with \(\pi = 10\%\) does what \(i' = 7\%\) does with \(\pi = 1\%\): \(R' = 6\%\), \(\widetilde{Y} = -2\%\).
• If inflation fell at once with the nominal hike (the classical dichotomy in the short run), the real rate and output would not move.
• In 1981 the fed funds rate averaged 16.4% and the real rate was about 6%.

Step one of a disinflation: a real rate above \(\bar{r}\), which means a recession.

Numbers: the live figure, the notebook’s “nutshell” table with inflation at 10 percent, and the Fig. 9 key numbers for 1981 (16.38 and 6.00).

Step two is the Phillips curve: with Ỹ = −2 and ν̄ = 0.5, inflation falls by 1 point, from 10 to 9 percent in the first year. The textbook’s own example (firms raise prices by 8 instead of 10 percent) corresponds to a steeper curve, ν̄ = 1.

Experiment: the disinflation year by year

Inflation starts at 10% and \(\bar{\nu} = 0.5\). The Fed holds output at −2% for five years, then returns the real rate to \(\bar{r}\). Sketch the path of inflation. Where does it end up after the Fed lets go?

Ask students to draw the path on paper. Two common mistakes: inflation falling in one jump and then flat, or inflation drifting back up to 10 percent after the Fed lets go.

Five years at −2% bring inflation from 10% to 5%, and it stays there

Model figure: the IS, MP and Phillips curves simulated year by year, \(\pi_t = \pi_{t-1} + \bar{\nu}\widetilde{Y}\!_t\), with \(\pi_{-1} = 10\%\), \(\bar{\nu} = 0.5\), \(\bar{b} = 2\). Adapted from Fig. 12.12 in Jones (2024), Macroeconomics, 6th ed.

How to read it: three panels over time: the real interest rate, short-run output and inflation. The Fed raises R in year 0 and keeps it high until year t*, set by the slider; then it returns R to r̄. The title gives the final inflation and the output lost.

Do live:

1. t* = 5: inflation 9, 8, 7, 6, 5 percent, then flat at 5. Output lost: 10 percent-years.
2. t* = 2: inflation stops at 8 percent and stays there.
3. t* = 8: inflation falls to 2 percent; the cost grows to 16 percent-years.

Inflation already falls in year 0, because π_0 is the price rise from year 0 to year 1: firms set next year’s prices when they see this year’s recession.

The notebook’s interactive version also lets you choose the depth of the recession and the slope.

The cost of a disinflation: \(1/\bar{\nu}\) percent-years per point

\[R = \bar{r} + \frac{\pi_{-1} - \pi^T}{\bar{\nu}\,\bar{b}\,T}\]

\[\text{lost output} = \frac{\pi_{-1} - \pi^T}{\bar{\nu}}\]

• 10% to 5% in 5 years: \(R = 6\%\), \(\widetilde{Y} = -2\%\) a year, path 9, 8, 7, 6, 5.
• Cost: \(5/0.5 =\) 10 percent-years of output.
• Once \(R = \bar{r}\), output returns to potential and inflation stays at 5%.

Disinflation is a temporary recession, not a permanent loss. Its total size is fixed by the slope.

The first equation: hold the rate constant for T years and solve the running-total formula for the output gap, then the IS curve for the rate. The second: add up the gaps; the T cancels.

The notebook’s checks: holding R at 6.00 percent for 5 years gives 9, 8, 7, 6, 5 and an output loss of 10.0 percent-years, equal to equation (9). Another design: from 10 to 4 percent in 4 years needs R = 6.50 percent, with inflation 8.5, 7, 5.5, 4.

In reality U.S. inflation fell from 13.5 percent in 1980 to 3.2 percent in 1983, faster than in this example, and the recession of 1981–82 was the deepest in decades. Chapter 13 explains part of the speed with expectations.

A faster disinflation is deeper, not cheaper

Going slowly spreads the pain thin, so a gradual disinflation costs less output.

In the model the total is the same: 10 percent-years to go from 10% to 5%. Output at −4% gets there in 2.5 years; at −2% in 5 years. Only a steeper Phillips curve makes it cheaper.

Numbers: the Fig. 12 “What to notice” (a deeper recession, Ỹ = −4, takes 2.5 years at about the same total cost). The general result: any path of output whose gaps add up to (π^T − π_{−1})/ν̄ does the job, whatever its shape.

Real-world caveats, which the simple model leaves out: if the central bank is credible, expectations may fall faster than last year’s inflation and the cost may be lower (Chapter 13). Ball (1994) found smaller sacrifice ratios in faster disinflations across OECD countries, a contested finding discussed in the Chapter 9 deck.

Ask: which costs more in human terms, two years at −5 percent or five years at −2 percent? (Same output in total; the answer depends on who loses jobs and for how long.)

Should a central bank follow a rule instead of its judgment?

Instead of choosing each year’s rate by judgment, the central bank could commit to a rule: raise the real rate by \(\bar{m}\) points for each point of inflation above its target, \(R_t - \bar{r} = \bar{m}(\pi_t - \bar{\pi})\).

What the model says

• Each year the inflation gap shrinks to \(\lambda\) of itself: 2/3 at \(\bar{m} = 1/2\).
• From 10%: 7.3, 5.6, 4.4%… toward the 2% target.
• Any \(\bar{m}\) costs the same 16 percent-years.

What is debated

• How strongly to respond.
• Whether to respond to output as well.
• A rule is predictable; judgment can react to surprises.

The model numbers are the book chapter’s preview of Chapter 13 (equations 10 and 11) and the notebook’s Check 5: λ = 1/(1 + ν̄b̄m̄), so with ν̄ = 0.5, b̄ = 2 and m̄ = 1/2, λ = 2/3. Starting from 10 percent with a 2 percent target, inflation runs 7.33, 5.56, 4.37, 3.58 percent and output −5.33, −3.56, −2.37, −1.58 percent. For m̄ = 0.25, 0.5 and 1, output in year 0 is −3.20, −5.33 and −8.00 percent, but the total loss is 16.00 percent-years in every case: a tougher rule front-loads the recession without making it cheaper.

The nominal rate rises by 1 + m̄ points for each point of inflation, more than one for one, so that the real rate rises: the Taylor principle.

The research. Taylor (1993, Carnegie-Rochester Conference Series on Public Policy) proposed the rule that bears his name, responding to inflation and the output gap, and showed that it tracked Fed policy in 1987–1992. Clarida, Galí and Gertler (2000, Quarterly Journal of Economics) estimate that before 1979 the Fed raised nominal rates by less than inflation, violating the Taylor principle, and that it raised them by more after 1979. Orphanides (2003, Journal of Monetary Economics) argues that with the output-gap estimates available in real time, 1970s policy looked closer to a rule than it does with today’s data. How mechanically central banks should follow rules is debated.

Ask the policy students: what would you lose by tying the central bank to a formula? (Room to respond to a financial crisis or a pandemic; a rule also needs an estimate of r̄ and of the output gap.)

A productivity slowdown

After 1973 potential output grew more slowly, but policymakers still believed in the old trend. Actual output fell behind that believed trend. What did they think they saw, what did they do, and what happened to inflation?

This is the textbook’s explanation of part of the Great Inflation. Let students reason: if you think potential is higher than it is, an economy at its true potential looks like a recession. You stimulate. Then?

Policymakers saw a slump, but output was above its true potential

Model figure: believed potential growth 3.5% a year; true growth after 1973 on the slider; output index 1973 = 100 (illustrative). Adapted from Fig. 12.13 in Jones (2024), Macroeconomics, 6th ed.

How to read it: output index (1973 = 100). The believed potential path keeps the old 3.5 percent growth; true potential grows more slowly after 1973, by the rate on the slider; actual output is pushed gradually above true potential by the stimulus. The title gives the perceived and the true gap in 1980 and the rise in inflation.

Do live:

1. At 2.5 percent true growth: in 1980 output is 3.7 percent below the believed path (a “recession”) but 3.3 percent above true potential; inflation is up about 8 points since 1973 with ν̄ = 0.5.
2. Move to 3.5 percent: no slowdown, the two dashed paths coincide, and the stimulus is plainly a boom.
3. Move to 1.5 percent: a larger slowdown, a larger illusion of a slump.

The model is stylized: the three growth rates are illustrative.

A misread slowdown turned stimulus into inflation

• 1980: perceived \(\widetilde{Y} =\) −3.7%, true \(\widetilde{Y} =\) +3.3%.
• The true gap was positive every year, so inflation rose every year: about 8 points by 1980.
• Actual U.S. inflation: 6.3% in 1973, 13.5% in 1980, helped by the oil shocks of 1974 and 1979.
• The warning sign is rising inflation, but oil shocks also raise inflation, so the lesson came slowly.

Potential output is not observed. Mistake a slower trend for a slump, and policy feeds inflation.

Numbers: the live figure’s title and the Fig. 13 “What to notice”.

The research. Orphanides (2002, American Economic Review Papers and Proceedings; 2003, Journal of Monetary Economics) uses the Fed’s real-time data to show that output gaps believed in the 1970s were far more negative than today’s estimates, so policy that followed a sensible rule on those numbers was too loose. Romer and Romer (2002) stress instead policymakers’ beliefs, such as an optimistic natural rate of unemployment. The weight of each is debated.

The Great Inflation in the IS/MP diagram

Policymakers believe demand fell by \(\bar{a}' = -2\) and cut the real rate to \(R' = 4\%\). But demand never fell. Where does output really go, and what happens to inflation each year?

This is the textbook’s Exercise 12, worked in the book chapter. Ask for the true output gap and the change in inflation, with b̄ = 2 and ν̄ = 0.5.

The cut moves the economy down the true IS curve, into a boom

Model figure: believed IS curve shifted by \(\bar{a}'\), true IS curve unchanged, MP cut to \(R' = \bar{r} + \bar{a}'/\bar{b}\), Phillips curve with \(\bar{\nu} = 0.5\). Adapted from Fig. 12.21 in Jones (2024), Macroeconomics, 6th ed., the worked answer to Exercise 12.

How to read it: four panels. (a) What policymakers believe: the IS curve shifted left to IS′, the economy at B below potential. (b) What they do: cut the real rate to R′, which they think brings the economy back to C at potential. (c) The truth: the IS curve never moved, so the lower rate takes the economy down the true IS curve to a point above potential. (d) The Phillips curve: output above potential raises inflation. The slider sets the shock policymakers believe in.

Do live:

1. Believed −2: R′ = 4 percent, true Ỹ = +2 percent, Δπ = +1 point a year.
2. Believed −1.5: R′ = 4.25 percent, true Ỹ = +1.5 percent.
3. Believed −3: a bigger cut (R′ = 3.5 percent), a bigger boom (+3 percent), faster inflation.

Believing in a shock that never happened creates a boom of the same size

• The cut to \(R' = \bar{r} + \bar{a}'/\bar{b}\) gives true output \(\widetilde{Y} = -\bar{b}(R' - \bar{r}) = -\bar{a}'\): +2% for a believed −2.
• Inflation rises by 1 point a year for as long as the mistake lasts: \(\downarrow R \Rightarrow \uparrow\widetilde{Y} \Rightarrow \uparrow\pi\).
• Add oil shocks and the belief in a lasting trade-off, and you have the Great Inflation.

The same tool that cures a demand shock causes inflation when the shock is imaginary.

Numbers: the notebook’s Check 4 (believed −2: R cut to 4.00 percent, true Ỹ = +2.0 percent; believed −1: 4.50 percent, +1.0 percent).

The book chapter’s practice question: once the Fed discovers the mistake, to what rate should it return? (To r̄ = 5 percent, which stops the rise in inflation; to bring inflation back down by 1 point a year it must go higher, to 6 percent.)

Should a central bank fight an oil shock or let it pass?

Oil prices jump and inflation rises at once, with output at potential. Should the central bank raise rates and cause a recession, or accommodate the shock and accept higher inflation?

What the model says

• Accommodate: inflation up by \(\bar{o}\), and it stays up.
• Fight: output at \(-\bar{o}/\bar{\nu}\); undoing a 1-point shock costs 2 percent-years.
• No option avoids both costs.

What it leaves out

• Expectations: if people trust the target, a one-off shock may not stick.
• Oil can lower potential output too.
• Part of the 1970s damage came from the policy response.

Model numbers: the notebook’s Check 4 (a one-year shock of 1 point raises inflation from 2 to 3 percent for good; undoing it costs 2.0 percent-years) and the Fig. 8 key numbers.

The research. Bernanke, Gertler and Watson (1997, Brookings Papers on Economic Activity) argue that much of the output loss after the oil shocks of the 1970s came from the Fed’s tightening response rather than from oil itself, a conclusion others dispute. Blanchard and Galí (2007, NBER Working Paper 13368; published 2010 in Galí and Gertler, eds., International Dimensions of Monetary Policy) find that the oil shocks of the 2000s did far less damage than those of the 1970s, and attribute it partly to more credible monetary policy, more flexible labor markets and a smaller oil share. In the model of this chapter, with adaptive expectations, credibility has no role; Chapter 13 adds it.

Today’s version: in 2022 energy prices jumped after Russia’s invasion of Ukraine, and central banks had to decide how much of the inflation to fight (Fig. 15, next).

Ask the policy students: how would you explain to the public that you are causing a recession to fight an inflation that came from abroad?

Inflation during the pandemic

In the first half of 2020 U.S. real GDP fell 9.1%. What does the Phillips curve predict for inflation in 2020? And what actually happened in 2021–22?

Let students apply the curve: output far below potential should lower inflation. Then ask whether they remember what happened two years later.

Inflation dipped in 2020, then rose to a 40-year high

Data: BLS CPI, all items (CPIAUCSL) and less food and energy (CPILFESL), via FRED; BIS consumer prices (WS_LONG_CPI); OECD and Eurostat core CPI; NBER recessions. Chart inspired by Fig. 12.15 in Jones (2024), Macroeconomics, 6th ed.

How to read it: 12-month CPI inflation, all items and core (without food and energy, the most volatile prices), monthly, with the 2020 recession shaded. The menu picks another of the 64 economies.

Do live:

1. Early 2020: from 2.6 percent in January to 0.2 percent in May, as the Phillips curve predicts for a slump.
2. June 2022: 9.0 percent, the highest since November 1981 (9.6 percent). Core peaked later, 6.6 percent in September 2022.
3. Mid-2024 through 2025: between 2.3 and 3.0 percent, with no recession.
4. 2026: headline back above 3 percent from March (4.2 percent in May) while core stayed near 2.5 percent: the sign of a new price shock.
5. Choose the euro area or the United Kingdom: peaks of 10.6 and 11.1 percent in October 2022, pushed by energy prices; core peaked only in 2023 (5.7 percent in March and 6.5 percent in May).

Data caveats: October 2025 has no CPI (government shutdown), so the line breaks there. Core data are missing or stop early for some countries.

Supply shocks and strong demand brought inflation back

• The dip of 2020 fits the curve: a slump slows inflation (2.6% → 0.2%).
• The surge to 9.0% (June 2022) had three forces: broken supply chains and energy prices (\(\bar{o} > 0\)), and demand above a reduced potential.
• Headline above core means price shocks: in 2022, and again in 2026.
• Central banks, not expecting it, were slow to raise rates.

The pandemic surge was a 1970s mix in miniature: price shocks plus demand pressure plus a late response.

Numbers: the Fig. 15 “What to notice”.

The textbook’s three forces, in our words: supply-chain disruptions and energy shocks are temporary increases in ō; demand pressure came because the pandemic also lowered potential output, so a recovering economy could end up above potential; government transfers (next figure) and low interest rates added to demand.

The Fed’s rate stayed near zero until March 2022 (Fig. 2), while inflation was already high: the real rate was deeply negative in 2021–22 (Fig. 9).

Income and consumption in the pandemic

In April 2020 U.S. unemployment reached 14.8%. What do you expect happened to households’ disposable income, income after taxes and government transfers?

Most students expect income to collapse with jobs. Ask them to commit before the reveal.

Government payments raised household income in a deep recession

Data: BEA real disposable personal income (DSPIC96) and real personal consumption expenditures (PCEC96), via FRED; NBER recessions. Chart inspired by Fig. 12.17 in Jones (2024), Macroeconomics, 6th ed.

How to read it: real disposable personal income and real consumption, trillions of chained 2017 dollars at annual rates, monthly. The three markers are the government payment months.

Do live:

1. Consumption: down 16.8 percent from February to April 2020, back above its February level by early 2021.
2. Income: three spikes, April 2020, January 2021 and March 2021, the months of the payments of 1,200, 600 and 1,400 dollars per person. The three largest monthly jumps in the data are exactly those months.
3. After March 2020 income never fell below its February 2020 level.

Data check: the textbook’s caption says the payments totaled 3,000 dollars per person; 1,200 + 600 + 1,400 = 3,200.

Background, from Fig. 16 of the book chapter: federal outlays rose from 20.7 percent of GDP in FY2019 to 30.5 percent in FY2020 and 28.8 percent in FY2021.

Transfers kept income up, and the savings fed the recovery

• Income rose while unemployment hit 14.8%; consumption fell 16.8% in two months.
• Households saved much of the money: in the model, transfers raised \(\bar{a}_c\) and so \(\bar{a}\).
• Per person, U.S. real income rose 17.6% from 2019Q4 to 2021Q1; consumption only 1.1%.
• Germany and the United Kingdom paid firms to keep workers instead: their income barely moved.

Fiscal support turned a collapse in demand into a fast recovery, and later into demand pressure on prices.

Numbers: the Fig. 17 “What to notice” and its cross-country key numbers (OECD household dashboard, real income and consumption per person). Canada and Japan also show income jumps in 2020Q2: emergency benefits in Canada and 100,000 yen to every resident in Japan.

The Covid-19 shock in the model, from the textbook: partly a fall in demand (people avoided shops and restaurants), which lowered inflation at first; partly a fall in potential output (people stopped going to work), which by itself leaves Ỹ unchanged and does not push inflation down. That mix helps explain why inflation fell so little in 2020.

Real GDP, Fig. 14 of the book chapter: down 9.1 percent from 2019Q4 to 2020Q2, back within about 1 percent of its pre-Covid trend from 2021Q2. Chapter 10 told that story in detail.

What caused the 2021–22 inflation, and should the Fed have moved sooner?

Was the surge a temporary supply shock that a central bank should look through, or excess demand from stimulus that it should have fought earlier?

Supply shocks

• Energy, food and supply chains: a rise in \(\bar{o}\).
• Headline peaked before core, and core stayed lower.
• Fighting it means a recession.

Demand

• Transfers, savings and near-zero rates.
• Potential output was reduced, so demand ran above it.
• A late response let inflation spread.

The two columns are the textbook’s forces, sorted. Numbers on the slides: Fig. 15 (headline peak 9.0 percent in June 2022, core 6.6 percent in September).

The research. Bernanke and Blanchard (2023, Brookings Hutchins Center Working Paper 86, “What Caused the U.S. Pandemic-Era Inflation?”) estimate that most of the surge came from price shocks in energy, food and supply chains, with tight labor markets playing a growing and more persistent role later. Shapiro (2022, Federal Reserve Bank of San Francisco Working Paper 2022-18) splits PCE inflation into supply- and demand-driven parts and finds that supply factors explained about half of the rise, demand about a third. The share each side deserves, and whether the Fed waited too long to raise rates, is still argued.

Ask the policy students: if most of the surge was supply, was the Fed’s fast hiking in 2022 a mistake? (The model’s answer: it depends on whether the shock would have stuck in expectations. That is Chapter 13’s subject.)

Inflation does not return to target on its own

Once the recession is over and output is back at potential, inflation drifts back to 2% by itself.

In the model inflation stays wherever it has got to: after a 1-point oil shock it stays at 3%; after five years of tight policy it stays at 5%, not 2%. Only policy moves it.

Numbers: the notebook’s Check 4 (one-year price shock: inflation 3 percent in every later year) and Fig. 12 (inflation stays at 5 percent once R returns to r̄).

Why: with adaptive expectations, firms expect last year’s inflation; at potential they have no reason to raise prices faster or slower than that. Any inflation rate can be a resting point.

What changes this: a policy rule that responds to inflation (the policy slide on rules), and expectations anchored by a credible target (Chapter 13). Both are ways to put a “pull” toward 2 percent into the model.

“Inflation fell without a recession, so the model is wrong”

A skeptic says: “U.S. inflation fell from 9.0% in June 2022 to between 2.3% and 3.0% from mid-2024 through 2025, and there was no recession. The model says each point of disinflation costs \(1/\bar{\nu}\) percent-years of lost output. It cannot be right.”

Reply: the cost formula is for inflation that has become expected. A price shock that reverses (\(\bar{o} < 0\)) lowers inflation with output at potential, at no cost. Much of 2022–24 was shocks unwinding, with expectations that stayed close to target, which the next chapter adds to the model.

Take the objection seriously: it is the main empirical challenge to this chapter’s model, and economists argued about it in real time. Blanchard, Domash and Summers (2022, Peterson Institute Policy Brief 22-7) argued from the Beveridge curve that bringing inflation down would likely require a sizable rise in unemployment; it did not happen.

The reply in the model’s own terms: in Δπ_t = ν̄Ỹ_t + ō, inflation can fall because Ỹ < 0 (costly) or because ō < 0 (free). When energy prices and supply chains normalized, ō turned negative. The key caveat of the model is adaptive expectations: if people expected inflation to return to 2 percent, expected inflation did not follow last year’s 8 percent, and the Phillips curve did not have to grind it down.

The research. Benigno and Eggertsson (2023, NBER Working Paper 31197) argue that the Phillips curve is steeper when labor markets are very tight, so that inflation rose fast in 2021–22 and could fall back quickly as the labor market cooled, without large job losses. Bernanke and Blanchard (2023) find that the price shocks that caused much of the surge faded on their own. Neither view is settled.

The book’s numbers: Fig. 15 (9.0 percent in June 2022; 2.3 to 3.0 percent from mid-2024 through 2025).

Ask: what would the 2022–24 episode have looked like if expectations had followed the adaptive rule? (Inflation would have stayed high until a recession brought it down, as in 1981–82.)

What to remember from this chapter

• The central bank sets an overnight nominal rate by supplying whatever money is demanded there. Because inflation is sticky, that sets the real rate: \(R = i - \pi\) (the MP curve).
• Through the IS curve the real rate sets output; through the term structure it reaches long rates.
• The Phillips curve, \(\Delta\pi = \bar{\nu}\widetilde{Y} + \bar{o}\): booms and price shocks raise inflation, and it stays up.
• Lowering inflation takes a temporary recession of \((\pi_{-1} - \pi^T)/\bar{\nu}\) percent-years, however fast it is done: Volcker, 13.5% to 3.2%.
• Oil shocks, a misread slowdown and a late response fed the Great Inflation, and in miniature the 2021–22 surge.

Ask each student to write one sentence answering the opening question: how does the central bank’s rate move output and inflation, and why does bringing inflation down take a recession? A good answer: with sticky inflation, a higher nominal rate is a higher real rate, which lowers spending and output; output below potential slows inflation; so inflation built into expectations can be lowered only by a period of output below potential.

For the policy students, three habits: judge policy by the real rate, not the nominal one; ask whether a rise in inflation is demand or a price shock; and ask how sure anyone is about potential output.

Walk the three branches of the mind map in about a minute each: the central bank’s lever (cyan), IS, MP and the Phillips curve (lavender), disinflation and the Great Inflation (green). Click the image to zoom in.

Sticky inflation lets the central bank steer output, so bringing inflation down takes a recession

Next: Chapter 13, Stabilization Policy and the AS/AD Framework

This chapter completed the short-run model: the MP curve, the IS curve and the Phillips curve. Chapter 13 replaces the central bank’s year-by-year judgment with a policy rule, adds expectations, and turns the model into the AS/AD framework.

Sources: data and models

• Fig. 2, the federal funds rate. Federal Reserve Board, effective federal funds rate (FEDFUNDS), via FRED; BIS, central bank policy rates (WS_CBPOL); NBER recessions (USREC; NBER dating page). Inspired by Fig. 12.2 in Jones (2024).
• Fig. 6, the yield curve. U.S. Treasury, Daily Treasury Par Yield Curve Rates; OECD, Financial market statistics (DSD_STES@DF_FINMARK: IR3TIB, IRLT), OECD Data Explorer. Inspired by Fig. 12.6 in Jones (2024).
• Fig. 9, inflation and the real rate. BLS CPI (CPIAUCSL) and the fed funds rate, via FRED; World Bank WDI (FP.CPI.TOTL.ZG); BIS policy rates. Inspired by Fig. 12.9 in Jones (2024).
• Fig. 15, pandemic inflation. BLS CPI (CPIAUCSL, CPILFESL), via FRED; BIS, consumer prices (WS_LONG_CPI); OECD core CPI (DSD_PRICES@DF_PRICES_ALL); Eurostat HICP (prc_hicp_minr). Inspired by Fig. 12.15 in Jones (2024).
• Fig. 17, income and consumption. BEA, real disposable personal income (DSPIC96) and real consumption (PCEC96), via FRED; OECD household dashboard (DSD_HHDASH@DF_HHDASH_INDIC). Inspired by Fig. 12.17 in Jones (2024).
• Models, Figs. 1, 4, 5, 7, 8, 10, 12, 13, 19, 20 and 21. The MP curve, the IS/MP diagram, the Phillips curve, the money market and the Great Inflation exercise follow Sections 12.2 to 12.7 of Jones (2024) and are adapted from Figs. 12.1, 12.4, 12.5, 12.7, 12.8, 12.10, 12.12, 12.13, 12.19, 12.20 and 12.21; the numerical baseline, the disinflation-cost and rule derivations are the book chapter’s own.
• 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 12.

Sources: research cited

• The yield curve and recessions. Estrella, A. and G. Hardouvelis (1991), Journal of Finance 46(2); Estrella, A. and F. Mishkin (1998), Review of Economics and Statistics 80(1).
• No lasting trade-off. Phelps, E. (1967), Economica 34(135); Friedman, M. (1968), American Economic Review 58(1).
• Disinflation. Ball, L. (1994), “What Determines the Sacrifice Ratio?”, in N. G. Mankiw (ed.), Monetary Policy, University of Chicago Press.
• Rules and the 1970s. Taylor, J. (1993), Carnegie-Rochester Conference Series on Public Policy 39; Clarida, R., J. Galí and M. Gertler (2000), Quarterly Journal of Economics 115(1); Orphanides, A. (2002), American Economic Review 92(2); Orphanides, A. (2003), Journal of Monetary Economics 50(3); Romer, C. and D. Romer (2002), Federal Reserve Bank of Kansas City, Rethinking Stabilization Policy.
• Oil shocks. Bernanke, B., M. Gertler and M. Watson (1997), Brookings Papers on Economic Activity 1997(1); Blanchard, O. and J. Galí (2007), NBER Working Paper 13368.
• The pandemic inflation. Bernanke, B. and O. Blanchard (2023), Hutchins Center Working Paper 86, Brookings; Shapiro, A. H. (2022), Federal Reserve Bank of San Francisco Working Paper 2022-18; Benigno, P. and G. Eggertsson (2023), NBER Working Paper 31197; Blanchard, O., A. Domash and L. Summers (2022), Peterson Institute for International Economics Policy Brief 22-7.