- Effects of inflation
- Monetary policy
- Supply shocks and stagflation · coming soon
Two stories about the same inflation
In October 2022, consumer price inflation in the United Kingdom reached 11.1%, its highest rate in 41 years. People argued about the reason.
One story said: households had saved during the pandemic lockdowns, the government had supported incomes, and interest rates had been close to zero. Too much spending was chasing too few goods. Another story said: gas and electricity prices had exploded, ships and computer chips were in short supply, and firms were simply passing on their costs.
The two stories lead to different cures. If spending is the problem, higher interest rates will cool it. If costs are the problem, higher interest rates cannot make gas any cheaper. So before anyone can fight inflation, they must ask where it comes from. This lecture gives the three main answers, and shows how to tell them apart.
A price rise is not yet inflation
The last lecture defined inflation as a sustained rise in the general price level. The word "sustained" matters here. Suppose the price of oil jumps once and then stays at its new, higher level. Prices across the economy rise for a while, as the extra cost works through. Then they stop rising. That is a one-off rise in the price level, not a lasting inflation.
So every cause of inflation must answer two questions. What pushes the price level up in the first place? And what keeps it rising, year after year? The first two causes in this lecture answer the first question. The third answers the second.
We will use the AD-AS diagram from the lecture on economic growth. The vertical axis is the price level, an index. The horizontal axis is real GDP. Aggregate demand (AD) slopes down. Short-run aggregate supply (SRAS) slopes up: gently when the economy has spare capacity, and steeply as it gets close to its capacity. Long-run aggregate supply (LRAS) is a vertical line at potential output.
Cause 1: demand-pull inflation
Demand-pull inflation happens when aggregate demand grows faster than the economy's ability to produce. An old saying puts it as "too much money chasing too few goods". On the diagram, AD shifts to the right.
Anything that raises one part of can do this. Low interest rates make borrowing cheap, so households and firms spend more. Tax cuts and extra government spending add income and orders. A boom abroad raises exports. Rising house and share prices make owners feel richer, so they spend more (economists call this the wealth effect).
But a rise in demand does not always mean much inflation. It depends on how much spare capacity the economy has. When many workers are unemployed and many machines are idle, firms can sell more without paying more for anything. So most of the extra demand becomes extra output. Near capacity, firms must pay overtime, compete for scarce workers, and run old machines. Their costs rise with every extra unit, so most of the extra demand becomes higher prices.
That is why the SRAS curve bends. Try it below: push demand up in equal steps, and watch how each step splits between output and prices.
The economy starts with spare capacity. Press "+ 5" on demand five times, one push at a time. After each push, read how it split between output and prices. Then raise costs.
Presets:
An invented economy starts with spare capacity: real GDP 84, potential output 100, price level 100. Demand is pushed up in equal steps of 5. Then costs rise by 10.
| Shift in AD | Rise in costs | Real GDP | Price level | This push: output, prices |
|---|---|---|---|---|
| 0 | 0 | 84.0 | 100.0 | start |
| +5 | 0 | 88.2 | 102.5 | +4.2, +2.5 |
| +10 | 0 | 91.9 | 106.3 | +3.7, +3.8 |
| +15 | 0 | 95.1 | 111.6 | +3.2, +5.3 |
| +20 | 0 | 97.9 | 118.4 | +2.7, +6.8 |
| +25 | 0 | 100.2 | 126.5 | +2.3, +8.1 |
| 0 | +10 | 81.0 | 108.9 | costs alone: output −3.0, prices +8.9 |
Each equal push of demand gives less extra output and more extra inflation as the economy nears capacity. A rise in costs raises prices and lowers output together.
What should you have seen? The first push of demand raised output by about 4.2 and the price level by about 2.5. The fifth push, close to capacity, raised output by only about 2.3 but the price level by about 8.1. The same push gave far more inflation once the spare capacity was used up. Then a rise in costs did something demand never does: it raised prices and lowered output at the same time.
Cause 2: cost-push inflation
Cost-push inflation starts with the firms, not the buyers. When the cost of making each unit rises, firms must charge more to stay in business. They will supply the same output only at a higher price. So the SRAS curve shifts up, which is the same as shifting it to the left.
Look at what happens. The price level rises. But real GDP falls, because at the higher prices buyers can afford less. Rising prices together with falling output, and usually rising unemployment, is called stagflation (stagnation plus inflation).
The main sources of rising costs are these.
- Energy and raw materials. Oil and gas are used to make and move almost everything. A rise in their price raises costs across the whole economy.
- Imported inputs. If a country's currency falls, everything it buys from abroad costs more in its own money. This is called imported inflation.
- Wages rising faster than productivity. This needs care, and Worked Example 2 shows why.
- Supply disruptions and indirect taxes. Shortages of parts or shipping, poor harvests, and rises in taxes on spending all raise what firms must pay or charge.
A bakery's cost per loaf is $1.00: wages $0.40, flour $0.30, energy $0.10, and other costs $0.20. In one year, energy prices rise by 150% and flour by 20%. Other costs do not change. By how much must the price of a loaf rise to keep the same profit per loaf?
Energy: extra. Flour: extra. Total: $0.21 extra per loaf.
So the cost per loaf rises from $1.00 to $1.21, which is 21%. A loaf that sold for $1.50 must now sell for $1.71 to keep its profit of $0.50.
Check: , so the shop price rises by 14%. That is less than 21%, because the profit part of the price did not rise. Notice that energy is only 10% of the cost, yet it caused most of the increase. A huge rise in a small cost can still matter. This is how an energy shock spreads into goods that seem to have nothing to do with energy.
In a factory, wages rise by 6% in a year. In case A, output per worker rises by 2%. In case B, output per worker rises by 6%. What happens to the labour cost of each unit made?
The labour cost of one unit is the wage divided by the output per worker. Economists call this the unit labour cost.
Case A: . Unit labour cost rises by about 3.9%. This is a cost push.
Case B: . Unit labour cost does not change at all. Each worker is paid 6% more, but each one also makes 6% more.
So a pay rise is not inflationary in itself. What matters is pay growing faster than productivity. Check case A with numbers: a worker paid $600 a week who makes 300 units has a labour cost of $2.00 per unit. Next year: $636 for 306 units, which is about $2.078 per unit, about 3.9% more. ✓
A country's currency falls by 20% against the dollar: one unit used to buy $1.00, and now buys only $0.80. A machine part costs $100 abroad. By how much does its price rise in the country's own money?
Before: 100 units of the currency. After: units. The price rises by 25%, not 20%.
A fall of 20% in the currency is a rise of 25% in import prices, because 0.80 must be multiplied by 1.25 to get back to 1. This is the same lopsidedness of percentages that you met in the formula effect in the last lecture.
Suppose gas prices double in one year, and then stay at the new, high level for the next year. In the second year, what does the jump in gas prices add to the 12-month inflation rate?
Show the answer
Nothing. The inflation rate compares prices with the same month one year earlier. In the second year, gas costs the same as twelve months before, so its contribution to the rate is zero. Prices are still high, but they are no longer rising. When a jump drops out of the 12-month comparison like this, statisticians call it a base effect. This is why a one-off cost shock gives a burst of inflation that fades by itself, unless something else keeps prices rising. The third cause is that something else.
Cause 3: built-in inflation and the wage-price spiral
Workers notice when prices rise. If prices rose by 8% last year and their pay rose by 2%, they can buy less than before. So at the next pay round they ask for more. Firms that pay more face higher costs, so they raise prices again. Then workers ask for more again.
This is the hard part of the lecture. Spend time here. The spiral is not caused by greedy workers or greedy firms. Each step is a reasonable response to the one before. Workers want to protect what their pay can buy. Firms want to cover their costs. The fuel for the loop is expectations: what people believe inflation will be next year.
If everyone believes that inflation will soon return to 2%, a one-off shock fades. Workers accept modest pay rises, and the base effect brings inflation back down. Economists say that expectations are anchored. If everyone instead expects next year to look like last year, each year's inflation is built into the next year's pay and prices. Then a one-off shock becomes a lasting inflation. Try it in the model below.
An invented economy has 2% inflation. In year 1, an energy shock adds a few percentage points. After that, every year, people set pay and prices to match the inflation they expect. Choose how much they expect next year to look like last year, and watch the next eight years.
Presets:
In an invented economy with 2% inflation, an energy shock adds 6 percentage points in year 1. Each later year, inflation = 2% × (1 − w) + w × last year's inflation, where w is how much people look back at last year.
| w | Year 1 | Year 2 | Year 3 | Year 4 | Year 8 |
|---|---|---|---|---|---|
| 0 (anchored) | 8.0% | 2.0% | 2.0% | 2.0% | 2.0% |
| 0.5 (partly) | 8.0% | 5.0% | 3.5% | 2.8% | 2.0% |
| 1 (fully) | 8.0% | 8.0% | 8.0% | 8.0% | 8.0% |
| 1.2 (catch-up) | 8.0% | 9.2% | 10.6% | 12.4% | 23.5% |
When expectations are anchored, the shock fades at once. When people look back partly, it fades slowly. When they look back fully, it stays for ever. When they also try to catch up, it grows into a spiral.
That is it. That is the whole idea. Demand-pull and cost-push can start an inflation. Expectations decide whether it stays. This is why central banks work so hard to be believed when they promise to bring inflation back to target.
A real case: three causes at once (UK, 1972–75)
In its 1972 Budget, the UK government cut taxes and raised spending, adding an estimated 2% of GDP to demand (Bank of England, 1998). Real GDP grew by more than 7% in 1973, much faster than the economy's capacity could grow. That was demand-pull. Then, from October 1973, oil prices rose from about $3 to nearly $12 a barrel within a few months. That was a huge cost push.
The third cause was built in by design. A 1973 incomes policy included "threshold agreements": when the RPI rose above a set level, pay rose automatically. From April 1974 these were triggered every month. Average hourly earnings rose by about 30% during 1974 (Cambridge Economic Policy Review). A later economic history judged that this close link between pay and prices was disastrous. A system meant to bring inflation down produced an upward spiral instead (Britton, 1994). RPI inflation peaked at 26.9% in August 1975.
A real case: the UK, 2021–24
The situation. In 2021 the UK economy reopened after the lockdowns. Households had built up savings, and the Bank of England's interest rate (Bank Rate) had been 0.1% since March 2020. At the same time, wholesale gas prices rose sharply, and rose again after Russia invaded Ukraine in February 2022. The regulator's cap on a typical household's yearly energy bill rose from £1,277 in October 2021 to £1,971 in April 2022. It was due to reach £3,549 in October 2022, until the government held it at £2,500 (Ofgem; Energy Price Guarantee).
The policy. From December 2021 the Bank of England raised Bank Rate fourteen times in a row, reaching 5.25% in August 2023.
The prediction. If the inflation was mainly demand-pull, higher interest rates should bring it down by slowing spending. If it was mainly cost-push, interest rates could not lower energy prices. Inflation would fall when the energy jump dropped out of the 12-month comparison, provided that expectations stayed anchored and no spiral began.
What happened. CPI inflation peaked at 11.1% in October 2022. A year later, in October 2023, it fell from 6.7% to 4.6% in a single month. The Office for National Statistics explained that last year's steep rise in energy costs had been followed by a small fall in the price cap. Gas was 31% cheaper than a year earlier (ONS, November 2023). That is a base effect. CPI inflation reached 2.0% in May 2024. But the price level did not come back down. Prices were about 16% higher in October 2023 than two years before (Resolution Foundation, 2023).
The verdict. Ben Bernanke and Olivier Blanchard studied this period with ten central banks, including the Bank of England. They concluded that the inflation was due mainly to supply shocks: sharp rises in energy and food prices, and shortages. Unlike in the 1970s, the effects did not persist, partly because people believed the central banks' inflation targets. As the shocks faded, tight labour markets and wage growth became a larger part of what remained (Bernanke and Blanchard, 2024). So the first story in the hook was partly right, and the second was mostly right. The rate rises could not make gas cheaper. Their main job was to stop the burst of cost-push inflation from turning into a lasting spiral.
Three classic mistakes
A rise in costs moves SRAS, not AD. Higher oil prices do not make buyers want more; they make producing each unit dearer. If your diagram shows output rising after a cost shock, you have shifted the wrong curve. Cost-push gives higher prices with lower output.
Wages that rise in line with productivity leave unit labour costs unchanged. Only pay rising faster than output per worker pushes up costs. And in a spiral, wages often rise because prices rose first.
Real inflations almost always mix causes. A strong answer names the main driver, the secondary forces, and what kept prices rising. Then it explains what that means for policy, because the cure depends on the cause.
Work each one on paper first. Open a hint only when you are stuck.
-
For each event, say whether it would start demand-pull or cost-push inflation, and which curve moves. (a) The government cuts income tax. (b) A drought ruins the wheat harvest. (c) The currency falls sharply. (d) A boom abroad raises orders for the country's exports. (e) The tax on spending (such as VAT or GST) rises.
Hint 1
Ask: does this change how much people want to buy, or what it costs to produce and sell?
Hint 2
One event works through both curves. Think about what a currency fall does to import prices and to export sales.
Full answer
(a) Demand-pull: more disposable income, so consumption rises and AD shifts right. (b) Cost-push: flour costs more, so SRAS shifts up. (d) Demand-pull: exports are part of AD, so AD shifts right. (e) Cost-push: a tax on spending raises the price firms must charge, so SRAS shifts up. (c) Both. Imported inputs cost more, so SRAS shifts up (cost-push). But the country's exports also become cheaper for foreign buyers, so AD shifts right (demand-pull). Both push the price level up.
-
Wages in a country rise by 5% in a year. Find the change in unit labour cost if output per worker rises by (a) 1%, (b) 5%, (c) 7%.
Hint 1
Unit labour cost = wage ÷ output per worker.
Hint 2
Divide the growth factors: for (a), 1.05 ÷ 1.01.
Full answer
(a) , so a rise of about 4.0%: a cost push. (b) , no change. (c) , so a fall of about 1.9%: firms could pay 5% more and still cut prices. Check (a): $500 for 100 units is $5.00 per unit; $525 for 101 units is about $5.198 per unit, about 4.0% more. ✓
-
A firm imports 30% of its costs. The country's currency falls by 25% (one unit now buys $0.75 instead of $1.00). Other costs do not change. By how much do the firm's total costs rise?
Hint 1
First find how much import prices rise in the country's own money. It is not 25%.
Hint 2
Import prices are multiplied by . Then weight the rise by 30%.
Full answer
Import prices rise by a factor of , so by about 33.3%. Imports are 30% of costs, so total costs rise by . Check with $1000 of costs: imports $300 become $400, the rest stays $700, total $1100, which is 10% more. ✓
-
Two invented countries both have inflation of 7%. In Arvelia, real GDP grew by 5% last year, unemployment is at a record low, and firms report that they cannot find staff. In Belmora, real GDP fell by 1%, unemployment is rising, and the price of imported fuel has doubled. (a) Name the main cause of inflation in each country, and sketch the AD-AS shift. (b) The central bank in each country raises interest rates. In which country does this work better, and what does it cost? (c) Suppose Belmora's fuel price stays at its new level. Explain what will happen to its inflation rate next year, and what could stop that from happening.
Hint 1
Look at what output is doing. In which country is it rising with prices, and in which is it falling?
Hint 2
Interest rates work mainly through AD. For (c), think about base effects and about expectations.
Full answer
(a) Arvelia: demand-pull. AD has shifted right close to capacity, so output and prices are both rising and labour is scarce. Belmora: cost-push. SRAS has shifted up because fuel costs more, so prices rise while output falls: stagflation. (b) Higher interest rates reduce spending, so they shift AD left. In Arvelia this attacks the cause, at the cost of slower growth. In Belmora it cannot make fuel cheaper. It lowers inflation only by cutting output further, which deepens the recession and raises unemployment. This is the policy dilemma of stagflation. (c) If fuel stays at its new price, its contribution to the 12-month rate falls to zero next year, a base effect, so inflation should fall back. That can fail if workers expect 7% to continue and win 7% pay rises: unit labour costs rise and a wage-price spiral keeps inflation high. A central bank whose target is believed makes that less likely. That is the main reason it may still raise rates a little in Belmora.
The cause decides the cure. Demand-pull inflation can be treated by cooling demand, at the cost of slower growth. Cost-push inflation cannot be cured by any tool that works on demand without making the recession worse. Built-in inflation is the most dangerous of the three, because it no longer needs a cause from outside: it runs on what people expect. That is why the most powerful tool a central bank has may not be the interest rate at all, but being believed. The next lecture asks who gains and who loses while prices rise.
- Demand-pull inflation happens when AD grows faster than capacity, and the closer the economy is to capacity, the more of each rise in demand becomes higher prices.
- Cost-push inflation happens when the cost of producing each unit rises, so SRAS shifts up, giving higher prices with lower output, which is called stagflation.
- Pay rises push up costs only when wages grow faster than output per worker, because unit labour cost is the wage divided by output per worker.
- A one-off jump in costs gives a one-off rise in the price level, and the inflation rate falls back when the jump drops out of the 12-month comparison (a base effect).
- Built-in inflation, the wage-price spiral, keeps inflation going when people expect last year's inflation to continue, so anchored expectations are what stop a shock from becoming a lasting inflation.