Fiscal policy — government spending and taxation

Paying people not to work

In March 2020, the United Kingdom government made an offer that would have sounded absurd a month earlier. If a firm kept a worker on its books instead of firing them, the government would pay 80% of that worker's wage, up to £2 500 a month. The worker did not have to do any work. By the time the scheme closed in September 2021, it had supported 11.7 million jobs at 1.3 million employers, and claims totalled about £70 billion (HM Treasury and HMRC, final evaluation, 2023).

That year the government borrowed more than at any time since the Second World War. Was this reckless? Or was it exactly what economics says a government should do when spending collapses?

Economists still argue about parts of the answer. But the tools for judging it are clear. They are the tools of fiscal policy: how a government uses its spending and its taxes to steer the whole economy.

Two levers: spending and taxes

The national income lecture split total spending on an economy's output into four parts: $C+I+G+(X-M)$. The growth lecture called this total aggregate demand (AD). A government can move AD in two ways.

• Government spending (G). When the government buys more goods and services, such as teaching, hospitals or roads, AD rises directly, because G is one of its parts.
• Taxes (T). A cut in income tax leaves households more to spend, so C and AD rise, but indirectly. A cut in taxes on profits can raise investment I in the same way.

Transfers, such as pensions, benefits and the furlough payments, are not part of G, because the government buys nothing with them. They put money into households' hands, so they work like a tax cut.

Expansionary fiscal policy raises AD: more spending, lower taxes, or both. Contractionary fiscal policy lowers AD: less spending, higher taxes, or both. The first suits a slump. The second suits an economy whose spending is running ahead of capacity.

Deficits and debt

Compare what the government collects with what it pays out in a year. If tax revenue is less than its spending, including transfers, it has a budget deficit. It covers the gap by borrowing, usually by selling bonds: promises to repay with interest. If revenue is greater than spending, it has a budget surplus.

The national debt is different. It is the total the government owes, built up from all its past borrowing. A useful picture is a bath. The deficit is the water flowing in this year. The debt is the water already in the bath. A smaller deficit means the bath fills more slowly, but it still fills. Only a surplus lets some water out.

Both are usually given as a percentage of GDP, because a debt that is large for a small economy is easy for a large one. In the financial year to March 2021, UK borrowing was about £311 billion, about 14.7% of GDP. That was the highest share since 1945–46 (ONS, as summarised by the House of Commons Library, 2026).

Why one dollar can become more than one dollar

Suppose the government spends an extra $10 billion building schools. That $10 billion is income for builders, suppliers and their workers. They spend part of it, and that spending is income for other people, who spend part again. Each round is smaller, because some money leaks out at every step.

There are three leaks. Some extra income is saved. Some is paid in taxes. And some is spent on imports, so it becomes income abroad, not at home. Call the share of each extra dollar that is spent again on the country's own output $c$. Then the rounds are $10$, then $10c$, then $10c^2$, and so on.

Adding all the rounds gives the first injection times the multiplier.

Many textbooks write $\frac{1}{1-\text{MPC}}$, using the marginal propensity to consume (MPC), the share of extra income households spend. That version leaves out taxes and imports, so its numbers are too big. (Pre-University economics calls the total leak the marginal propensity to withdraw.) Watch below how much the leak matters.

What should you have seen? The size of the leak decides almost everything. With 80% spent again, $10 billion becomes $50 billion. With 25%, it becomes about $13 billion. And a tax cut always does less than the same amount of spending in this model, because the first round is smaller: households save part of the tax cut and spend part on imports, so that part never becomes income at home.

That is it. That is the whole idea. Spending is someone else's income, so a change in spending travels round the economy several times before it fades.

The automatic part

The lecture on the costs of unemployment met automatic stabilisers. In a recession, benefit payments rise and tax revenue falls without any new decision, because both are built into the rules. This puts money into households' hands just when their incomes are falling, so spending falls less than it would have done. In a boom the opposite happens. Tax revenue rises and benefit payments fall, which slows spending down.

Economists call a deliberate change, such as a new building programme or a new tax rate, discretionary fiscal policy. The furlough scheme was discretionary. The rise in other benefit claims in 2020 was automatic. Sort the examples below to see the difference.

What should you have learned? Automatic stabilisers always push in the right direction, and they start at once. A decision takes time.

When interest rates cannot fall

The deflation lecture ended with a promise. It showed that a central bank cannot cut its interest rate far below zero, and said that this lecture would show what a government can do instead. Here is the answer.

When spending collapses, the usual first step is a cut in interest rates. But in March 2020 the Bank of England had already cut Bank Rate to a record low of 0.1% (Bank of England). Fiscal policy does not need interest rates to fall. A government can put money straight into workers' wages, as the furlough scheme did. So when interest rates are close to the zero lower bound, fiscal policy becomes the main tool. Some studies also find that multipliers are larger at these times, because interest rates do not rise to offset the extra spending. That finding is still debated.

A real case: the United Kingdom, 2020

The Bank of England did not lend to the government directly. It bought bonds from investors, which helped keep the government's interest costs low. Whether that counts as "paying for" the deficit is debated, and the monetary policy lecture returns to it.

Numbers to work with

Why fiscal policy is not a perfect tool

Crowding out. Heavy government borrowing competes with firms for lenders' money. That can raise interest rates and reduce private investment, so part of the extra G is cancelled by a fall in I. Crowding out is weak in a deep recession, when firms borrow little, and stronger near capacity.

Time lags. The government must notice the problem (recognition lag), agree what to do (decision lag), and start the spending, and roads take years to plan (implementation lag). Then the multiplier takes time to work (effect lag). Help that arrives after the recovery has begun can make the economy overheat.

Politics. Tax cuts and new spending are popular, and tax rises and spending cuts are not. So governments may spend more before elections, whatever the economy needs. Economists call this the political business cycle.

Debt. A government can borrow only while lenders believe it will repay. Greece shows what happens when that belief goes. Its deficit in 2009 was about 15% of GDP and its debt about 127% of GDP (Eurostat). Greece used the euro, so it had no central bank of its own. In 2010 lenders demanded interest it could not afford, and it took emergency loans from other euro countries and the International Monetary Fund, with deep spending cuts and tax rises attached. Real GDP fell by about a quarter between 2008 and 2013, and unemployment peaked at about 27% (IMF).