Fiscal Policy

We define fiscal policy as the use of government spending ($G$) and taxation ($T$) to influence The level of aggregate demand, economic activity, and the allocation of resources in the economy.

$\mathrm{Fiscal policy tools: } G, T, \mathrm{ and } (G - T)$

Fiscal policy is conducted by the government (Chancellor of the Exchequer in the UK, Secretary of The Treasury in the US) alongside the central bank”s monetary policy.

The government budget records planned revenue and expenditure:

$\mathrm{Budget balance} = T - G$

• Budget surplus: $T > G$ (government revenue exceeds spending)
• Budget deficit: $G > T$ (government spending exceeds revenue)
• Balanced budget: $T = G$

## 2. Government Spending

Current expenditure (recurring, day-to-day):

• Public sector wages (NHS, education, civil service)
• Transfer payments (pensions, unemployment benefits, child benefit)
• Debt interest payments

Capital expenditure (investment in infrastructure):

• Roads, railways, hospitals, schools
• Research and development
• IT infrastructure

$G = G_{current} + G_{capital}$

Only $G$ in the AD equation ($AD = C + I + G + (X - M)$) represents spending on goods and services. Transfer payments are not directly part of $G$ — they affect $AD$ indirectly through their effect on Disposable income and hence consumption ($C$).

### 2.2 Government Spending as a Share of GDP

$\mathrm{Government spending ratio} = \frac{G}{Y} \times 100\%$

In the UK, this ratio was approximately 45% of GDP in 2023–24, reflecting the expansionary fiscal Response to COVID-19 and the energy crisis. The long-run average is closer to 40%.

Direct taxes are levied on income, wealth, or profits. They are progressive and Difficult to evade.

Tax Type	Description	Examples
Income tax	Tax on earned income	PAYE, self-assessment
Corporation tax	Tax on company profits	19–25% (UK, 2023–24)
Capital gains tax	Tax on profit from sale of assets	10–28% (UK)
Inheritance tax	Tax on transfer of wealth at death	40% above £325,000 threshold
Council tax	Tax on property value (local)	Bands A–H

Indirect taxes are levied on goods and services. They are regressive and easier to Pass on to consumers.

Tax Type	Description	Examples
VAT	Value added tax on most goods and services	20% standard rate (UK)
Excise duties	Tax on specific goods (alcohol, tobacco, fuel)	Per-unit or ad valorem
Customs duties	Tax on imports	Varies by product and trade agreements
Stamp duty	Tax on property and share transactions	Percentage of transaction value

We define these in terms of the average tax rate (ATR) as income changes:

$ATR = \frac{T(Y)}{Y}$

Where $T(Y)$ is the total tax paid by someone earning income $Y$.

Progressive tax: ATR rises as income rises.

$\frac{d(ATR)}{dY} > 0$

Proportional tax (flat tax): ATR is constant regardless of income.

$\frac{d(ATR)}{dY} = 0$

Regressive tax: ATR falls as income rises.

$\frac{d(ATR)}{dY} < 0$

Proof with examples.

Consider three individuals earning £10,000, £30,000, and £100,000.

Progressive tax (UK income tax, simplified):

• £10,000: pays £0 (below personal allowance) → ATR = 0%
• £30,000: pays £3,486 (20% on £12,570–£30,000 after £12,570 allowance) → ATR = 11.6%
• £100,000: pays £27,432 → ATR = 27.4%

ATR rises with income. ✓

Proportional tax (10% flat tax on all income):

• £10,000: pays £1,000 → ATR = 10%
• £30,000: pays £3,000 → ATR = 10%
• £100,000: pays £10,000 → ATR = 10%

ATR constant. ✓

Regressive tax (VAT at 20% on all consumption, assuming lower earners spend a larger share of Income):

• £10,000 earner spends £9,500, pays £1,900 in VAT → ATR = 19.0%
• £30,000 earner spends £22,000, pays £4,400 in VAT → ATR = 14.7%
• £100,000 earner spends £55,000, pays £11,000 in VAT → ATR = 11.0%

ATR falls with income because higher earners save a larger proportion of income. ✓ $\blacksquare$

### 3.3 Laffer Curve

The Laffer curve illustrates the theoretical relationship between the tax rate and tax revenue:

$R = t \cdot Y(t)$

Where $t$ is the tax rate and $Y(t)$ is the tax base (income), which depends on $t$ because higher Tax rates discourage work, investment, and encourage tax evasion.

Proposition: There exists a tax rate $t^*$ that maximises revenue, and beyond this rate, further Increases reduce revenue.

Proof sketch. At $t = 0$Revenue $R = 0$. At $t = 100\%$No one works (all income is taxed Away), so $R = 0$. Since $R$ is continuous, by the Intermediate Value Theorem there exists at least One maximum $t^*$ where $\frac{dR}{dt} = 0$. At this point, the marginal gain from a higher rate (more tax per unit of income) exactly equals the marginal loss from a smaller tax base. $\blacksquare$

$\frac{dR}{dt} = Y(t) + t \cdot \frac{dY}{dt} = 0 \quad \mathrm{at } t^*$

The key debate is where $t^*$ lies. Supply-side economists (e.g., Arthur Laffer) argue that many Economies are already to the right of $t^*$So cutting rates would increase revenue. Empirical Evidence is mixed.

Budget deficit: the amount by which government spending exceeds revenue in a given year.

$\mathrm{Deficit}_t = G_t - T_t$

National debt: the accumulated total of all past budget deficits minus surpluses.

$D_t = D_{t-1} + (G_t - T_t)$

Debt-to-GDP ratio:

$\frac{D}{Y} = \frac{D_t}{Y_t}$

This is the preferred measure of debt sustainability, since a large absolute debt is manageable if GDP is also large.

### 4.2 Debt Dynamics

The evolution of the debt-to-GDP ratio is given by:

$\frac{D_{t}}{Y_{t}} = \frac{(1 + r) D_{t-1} + (G_t - T_t)}{(1 + g) Y_{t-1}}$

Where $r$ is the average interest rate on government debt and $g$ is the GDP growth rate.

Key insight: If $g > r$The debt-to-GDP ratio can fall even with a primary deficit (before Interest payments), because GDP is growing faster than the debt stock. If $r > g$Debt dynamics are Unstable — the debt ratio grows unless offset by primary surpluses.

A government’s debt is sustainable if it can continue to service it without default. Indicators of Concern:

• Debt-to-GDP ratio above 90% (Reinhart & Rogoff threshold, though contested)
• Interest payments exceeding 10% of tax revenue
• Rising $r - g$ differential
• Loss of market confidence (rising bond yields)

Used to increase aggregate demand during a recession:

Tool	Action	Effect on AD
Increase $G$	Build infrastructure, hire public workers	AD shifts right directly
Decrease $T$	Cut income tax, VAT, corporation tax	AD shifts right via higher $C$ and $I$
Increase transfer payments	Raise benefits, pensions	AD shifts right via higher $C$

The total effect on output:

$\Delta Y = k \cdot \Delta G \quad \mathrm{or} \quad \Delta Y = k \cdot MPC \cdot \Delta T$

Where $k$ is the complex multiplier.

Used to reduce aggregate demand during an overheating economy (to combat inflation):

Tool	Action	Effect on AD
Decrease $G$	Cut public spending programmes	AD shifts left directly
Increase $T$	Raise income tax, VAT	AD shifts left via lower $C$
Decrease transfer payments	Reduce benefits	AD shifts left via lower $C$

Proposition: Equal increases in $G$ and $T$ increase output by exactly the amount of the Increase.

Proof. The increase in $G$ directly adds $\Delta G$ to AD. The increase in $T$ reduces disposable Income by $\Delta T = \Delta G$Reducing consumption by $MPC \times \Delta G$. The net injection Is:

$\Delta A = \Delta G - MPC \cdot \Delta G = (1 - MPC) \cdot \Delta G = MPS \cdot \Delta G$

The total change in output:

$\Delta Y = k \cdot \Delta A = \frac{1}{MPS} \cdot MPS \cdot \Delta G = \Delta G$

So $\Delta Y / \Delta G = 1$ when $\Delta G = \Delta T$. The balanced budget multiplier equals 1. $\blacksquare$

Crowding out refers to the reduction in private sector spending (particularly investment) that Results from an increase in government spending.

When the government runs a deficit, it must borrow by selling bonds:

$G > T \Rightarrow \mathrm{government issues bonds} \Rightarrow \mathrm{demand for loanable funds rises}$

This increases the demand for loanable funds, pushing up the real interest rate:

$r \uparrow \Rightarrow I \downarrow$

The rise in $r$ reduces private investment, partially (or fully) offsetting the expansionary effect Of $\Delta G$ on AD.

Financial crowding out: the mechanism described above — government borrowing raises interest Rates, reducing private investment.

$\Delta I = -\frac{1}{MPS} \cdot \Delta G \quad \mathrm{(full crowding out in extreme case)}$

Resource crowding out: if the economy is at full employment, government spending uses resources That would otherwise be employed by the private sector. The increase in $G$ bids up wages and Prices, reducing private sector profitability.

Full vs partial crowding out:

• Full crowding out: $\Delta I = -\Delta G$ in the simple model. The increase in $G$ is exactly offset by the decrease in $I$. Output does not change (LRAS is vertical).
• Partial crowding out: $\Delta I > -\Delta G$. Some private investment is displaced, but the net effect on AD is still positive.

Factor	More crowding out	Less crowding out
State of the economy	At full employment	In deep recession (idle resources)
Central bank response	Keeps money supply constant	Accommodates (increases money supply)
Elasticity of investment demand	Investment is interest-elastic	Investment is interest-inelastic
Source of financing	Domestic borrowing	Foreign borrowing / monetary financing
Type of spending	Current spending (consumption)	Capital spending (productivity-enhancing)

## 7. Automatic Stabilisers

Automatic stabilisers are features of the tax and benefit system that automatically dampen Fluctuations in economic activity, without any deliberate policy action.

$\mathrm{Automatic stabilisers: } T(Y) \mathrm{ and } B(Y) \mathrm{ where } \frac{dT}{dY} > 0 \mathrm{ and } \frac{dB}{dY} < 0$

During a boom ($Y \uparrow$):

• More people are employed and earn higher incomes $\Rightarrow$ tax revenue rises automatically (progressive income tax)
• Fewer people claim unemployment benefits $\Rightarrow$ transfer payments fall
• Net effect: disposable income rises by less than GDP $\Rightarrow$ consumption grows more slowly $\Rightarrow$ AD is dampened

During a recession ($Y \downarrow$):

• People lose jobs or earn less $\Rightarrow$ tax revenue falls automatically
• More people claim benefits $\Rightarrow$ transfer payments rise
• Net effect: disposable income falls by less than GDP $\Rightarrow$ consumption falls more slowly $\Rightarrow$ AD is supported

The budget balance as a function of output:

$BB(Y) = T(Y) - G - B(Y)$

Where $T'(Y) > 0$ and $B'(Y) < 0$.

The cyclical component of the budget balance:

$BB_{cyclical} = BB(Y) - BB(Y^*)$

Where $Y^*$ is potential output. During a recession ($Y < Y^*$), $BB_{cyclical} < 0$ (the deficit Widens automatically). During a boom ($Y > Y^*$), $BB_{cyclical} > 0$ (the deficit narrows or a Surplus emerges).

Feature	Automatic stabilisers	Discretionary fiscal policy
Timing	Immediate (no lag)	Subject to recognition, decision, and implementation lags
Reversibility	Automatic (reverse when cycle turns)	Politically difficult to reverse (e.g., spending cuts)
Scope	Limited to built-in tax/benefit structures	Can target specific sectors or problems
Political	Non-controversial (no active decision)	Subject to political debate and lobbying

Fiscal policy is subject to three significant lags:

1. Recognition lag: time taken to identify that the economy is entering a recession or overheating. Data is published with a delay and is often revised. ( 3–6 months)

2. Decision lag: time between recognition and the political decision to act. Parliamentary processes, coalition negotiations, and political disagreements can delay action. ( 3–12 months)

3. Implementation lag: time between the decision and the actual impact on the economy. Infrastructure projects take years; tax changes are faster. ( 6–18 months)

Total lag: potentially 12–36 months, by which time the economic cycle may have turned.

1. Electoral cycle: governments may pursue expansionary policy before elections and contractionary policy after, creating a political business cycle.

2. Voter myopia: voters reward short-term gains (tax cuts, spending increases) and punish short-term pain (tax rises, spending cuts), biasing policy toward expansion.

3. Interest group pressure: powerful groups (pensioners, public sector unions, business lobbies) resist changes that affect them.

4. Institutional constraints: EU Stability and Growth Pact (deficit < 3% of GDP, debt < 60% of GDP), UK fiscal rules.

The effectiveness of fiscal policy depends on the size of the multiplier:

$k = \frac{1}{MPS + MPT + MPM}$

• Large multiplier: economy in deep recession, MPC high, economy relatively closed (low MPM), interest rates at the zero lower bound (no crowding out)
• Small multiplier: economy near full employment, open economy (high MPM), high interest rates (significant crowding out)

IMF (2012) estimated multipliers of 0.9–1.7 for advanced economies post-2008, larger than previously Assumed.

Proposition (Barro, 1974): Tax-financed and debt-financed government spending have the same effect On AD.

Argument. Rational, forward-looking households anticipate that current government borrowing Implies future tax increases to repay the debt. They increase saving by exactly the amount of the Deficit to pay the expected future taxes:

$\Delta G \mathrm{ (deficit-financed)} \Rightarrow \Delta S_{private} = \Delta G \Rightarrow \Delta C = 0$

Therefore, the multiplier is zero — fiscal policy is completely ineffective.

Critique. Ricardian equivalence requires:

• Perfect capital markets (households can borrow against future income)
• Infinite horizons or intergenerational altruism
• Lump-sum taxes (not distortionary)
• Rational expectations

These assumptions are unrealistic. Empirical evidence suggests partial Ricardian effects at best — Fiscal policy does affect AD, but the multiplier may be smaller than predicted by the simple Keynesian model.

In response to the Global Financial Crisis, the UK government introduced:

• Temporary reduction in VAT from 17.5% to 15% (December 2008)
• Bank recapitalisation (£500bn)
• Increased public spending (infrastructure, education)

The budget deficit rose from 2.7% of GDP (2007–08) to 10.2% (2009–10). The OBR estimated the Multiplier for the VAT cut at approximately 0.3–0.5 (small, due to the open economy and temporary Nature).

The Coalition government (2010) implemented austerity measures:

• Spending cuts (departmental budgets reduced by ~25% on average)
• Welfare reforms (benefit cap, universal credit)
• VAT increase from 17.5% to 20% (January 2011)

This was contractionary fiscal policy aimed at reducing the deficit. Critics argue it prolonged the Recovery; supporters argue it was necessary for market confidence and debt sustainability.

The largest peacetime fiscal expansion in UK history:

• Furlough scheme (£70bn): paid 80% of wages for furloughed workers
• Self-Employment Income Support Scheme (£25bn)
• Business support grants and loans
• Universal Credit uplift of £20/week

Budget deficit peaked at 14.8% of GDP in 2020–21. The multipliers were estimated to be relatively Large (1.0–1.5) because the economy was in a deep recession with the zero lower bound binding.

Problem 1. An economy has MPC = 0.8, MPT = 0.2, MPM = 0.15. The government increases spending by £50 billion. (a) Calculate the multiplier. (b) Calculate the total change in GDP. (c) If the Government had instead cut taxes by £50 billion, what would the change in GDP be? (d) Compare the Two approaches.

Hint

(a) $k = 1/(MPS + MPT + MPM) = 1/(0.2 + 0.2 + 0.15) = 1/0.55 = 1.82$. (b) $\Delta Y = 1.82 \times 50 = £91$Bn. (c) Tax cut: $\Delta C = MPC \times \Delta T_{disposable} = 0.8 \times 50 = £40$Bn initial injection. $\Delta Y = 1.82 \times 40 = £72.7$Bn. (d) Government spending is more effective per pound because it is a direct injection, whereas tax cuts are partially saved. The "spending multiplier" > "tax multiplier."

Problem 2. A country’s income tax schedule is: 0% on the first £12,570, 20% on £12,571–£50,270, 40% on £50,271–£125,140, 45% above £125,140. Calculate the average tax rate for individuals earning (a) £20,000, (b) £60,000, and (c) £150,000. Is the system progressive?

Hint

(a) £20,000: Tax = $0.20 \times (20,000 - 12,570) = 0.20 \times 7,430 = £1,486$. ATR = $1,486/20,000 = 7.4\%$. (b) £60,000: Tax = $0.20 \times 37,700 + 0.40 \times 9,730 = 7,540 + 3,892 = £11,432$. ATR = $11,432/60,000 = 19.1\%$. (c) £150,000: Tax = $0.20 \times 37,700 + 0.40 \times 74,870 + 0.45 \times 24,860 = 7,540 + 29,948 + 11,187 = £48,675$. ATR = $48,675/150,000 = 32.5\%$. Yes, progressive — ATR rises with income.

Problem 3. Explain why VAT is considered regressive despite being charged at a flat rate. Use a Numerical example with two individuals earning £15,000 and £80,000, both spending 90% and 60% of Their income respectively on VAT-able goods at 20%.

Hint

£15,000 earner: spends $0.9 \times 15,000 = £13,500$. VAT paid = $0.20 \times 13,500 = £2,700$. ATR = $2,700/15,000 = 18.0\%$. £80,000 earner: spends $0.6 \times 80,000 = £48,000$. VAT paid = $0.20 \times 48,000 = £9,600$. ATR = $9,600/80,000 = 12.0\%$. The lower earner pays a higher proportion of income as VAT because they spend a larger fraction of income (lower savings rate). Therefore VAT is regressive. Note: essentials are zero-rated (food, children's clothes), which partially mitigates regressivity.

Problem 4. A government increases spending by £100 billion, financed entirely by borrowing. If The economy is at full employment, the MPC is 0.75, and investment is relatively interest-elastic, Analyse the likely effects on (a) output in the short run and long run, (b) interest rates, (c) Private investment, and (d) inflation.

Hint

(a) SR: AD shifts right, but at full employment SRAS is vertical or very steep, so most of the effect is on prices not output. LR: no change in output (LRAS vertical). (b) Government borrowing increases demand for loanable funds → interest rates rise. (c) Private investment falls due to higher interest rates (financial crowding out). Since investment is interest-elastic, the fall is significant — potentially full crowding out. (d) Inflation rises due to demand-pull pressure. The policy is ineffective at raising output and harmful for investment and inflation. This illustrates the Classical critique of fiscal policy.

Problem 5. “Automatic stabilisers are superior to discretionary fiscal policy.” Evaluate this Statement.

Hint

Arguments for: (1) No time lags — operate immediately as the cycle turns. (2) No political bias — not subject to electoral manipulation. (3) Automatically reverse, avoiding pro-cyclical policy. Arguments against: (1) Limited scope — can only operate through existing tax/benefit structures. (2) Cannot target specific problems (e.g., regional unemployment, structural issues). (3) May not be sufficient for severe recessions (the 2008 crisis required discretionary stimulus). (4) The strength of automatic stabilisers varies across countries (stronger in Nordic countries with generous welfare states). Best answer: automatic stabilisers are the first line of defence; discretionary policy is needed for exceptional circumstances. *Revision: see [Aggregate Demand and Aggregate Supply](02-aggregate-demand-and-supply).*

Problem 6. The national debt is £2.7 trillion and GDP is £3.0 trillion. Interest payments on the Debt are £110 billion. GDP growth is 2.5% per year. (a) Calculate the debt-to-GDP ratio. (b) If the Primary deficit is £50 billion, will the debt-to-GDP ratio rise or fall next year? (c) Explain your Reasoning.

Hint

(a) Debt-to-GDP = $2,700/3,000 = 90\%$. (b) Approximate change: $\Delta(D/Y) \approx (r - g) \times (D/Y) + \mathrm{primary deficit}/Y$. $r = 110/2,700 = 4.1\%$. $r - g = 4.1\% - 2.5\% = 1.6\%$. $(r-g) \times D/Y = 1.6\% \times 90\% = 1.44\%$. Primary deficit ratio = $50/3,000 = 1.67\%$. Total change $\approx 1.44\% + 1.67\% = 3.1\%$. The debt ratio rises. (c) Because the interest rate on debt exceeds the growth rate ($r > g$), and the primary deficit adds to borrowing, the debt burden grows faster than the economy.

Problem 7. Using the concept of the Laffer curve, explain why a government might increase tax Revenue by cutting tax rates. Under what conditions is this most likely to be true?

Hint

If the economy is to the right of $t^*$ (tax rates so high that they discourage work and investment), cutting rates increases the tax base by more than the rate reduction, raising total revenue. Most likely when: (1) marginal tax rates are very high (e.g., 80%+ top rate in 1970s UK). (2) The tax base is elastic (people can relocate, retire early, or work less). (3) There is significant tax evasion that would decline at lower rates. (4) The economy is small and open (mobile capital). Less likely at moderate tax rates (most evidence suggests the UK and US are to the left of $t^*$ for most taxes). Thatcher's cut of the top rate from 83% to 40% (1980s) is a historical example where revenue may have increased.

Problem 8. Explain the concept of Ricardian equivalence. Why do most economists believe it does Not fully hold in practice?

Hint

Ricardian equivalence (Barro, 1974): households anticipate future tax liabilities from current government borrowing, so they save the full amount of a deficit-financed tax cut → no change in consumption → multiplier is zero. Why it fails in practice: (1) **Liquidity constraints** — poor households cannot borrow against future income, so a tax cut raises current consumption. (2) **Myopia** — households do not fully anticipate future taxes. (3) **Finite lives** — if taxpayers do not care about future generations, they will not save for taxes they won't pay. (4) **Distortionary taxes** — future taxes create deadweight loss, so the equivalence is not exact. (5) **Uncertainty** — households may not know when or how future taxes will be raised. Empirical evidence: consumption responds to tax cuts, suggesting partial but not full Ricardian equivalence. *Revision: see [Aggregate Demand and Aggregate Supply](02-aggregate-demand-and-supply).*

Problem 9. The government is considering two options to stimulate the economy: (A) increase Spending on infrastructure by £80 billion, or (B) cut income tax by £80 billion. The economy has MPC = 0.7, MPT = 0.15, MPM = 0.1. (a) Which option has a larger impact on GDP? (b) Which option might Have greater long-run benefits? (c) Evaluate the trade-offs.

Hint

$k = 1/(0.3 + 0.15 + 0.1) = 1/0.55 = 1.82$. (a) Option A: $\Delta Y = 1.82 \times 80 = £145.5$Bn. Option B: initial consumption boost $= 0.7 \times 80 = £56$Bn. $\Delta Y = 1.82 \times 56 = £101.9$Bn. Option A has a larger impact (£145.5bn vs £101.9bn). (b) Option B might be better long-term if it incentivises work and investment. But Option A (infrastructure) also has long-run supply-side benefits — better transport raises productivity, shifting LRAS right. (c) Trade-offs: Option A has higher multiplier but adds to debt and may suffer from implementation lags. Option B is faster to implement and may improve work incentives but has a smaller multiplier. The best choice depends on the economic context.

Problem 10. “Fiscal policy was ineffective during the 2008 financial crisis because of crowding Out.” Evaluate this statement.

Hint

Partially false. Crowding out was limited during 2008–09 because: (1) The economy was in deep recession with large output gap → idle resources available. (2) Interest rates were cut to near zero (the zero lower bound) → central bank accommodated fiscal expansion by keeping $r$ low. (3) Private investment was already depressed ( pessimism) → crowding out was minimal. However: (1) The UK's high MPM (~0.3) reduced the multiplier. (2) Some financial crowding out occurred as government borrowing increased. (3) Confidence effects may have amplified or dampened the policy. The effectiveness of fiscal policy was moderate — sufficient to prevent a deeper recession but not enough to deliver a rapid recovery. *Revision: see [Macroeconomic Performance](01-macroeconomic-performance) for data on UK GDP and unemployment post-2008.*

• Macroeconomic Policy: Fiscal policy is part of macroeconomic management
• Demand and Supply: Fiscal policy affects aggregate demand
• Theory of the Firm: Taxation affects firm behaviour

Fiscal policy is the government’s most direct lever on the economy — it’s the ability to spend money or collect less of it, right now, to change how much people and businesses are buying. Think of it like a thermostat for the economy: when things are too cold (recession), the government turns up the heat by spending more or cutting taxes; when things are overheating (inflation), it turns the heat down by spending less or raising taxes. The key insight is that government spending doesn’t just add one pound to the economy — it circulates. When the government builds a road, the construction workers get paid, they spend that money at shops, the shops hire more staff, and so on. This ripple effect is the multiplier, and it’s why fiscal policy can have a much larger impact than its initial size suggests.

But the thermostat isn’t perfect. It takes time to recognise when the economy needs heating or cooling, time to agree on what to do, and time for the policy to actually work. By the time you’ve fitted a new radiator, the weather may have changed. This is the lag problem, and it means fiscal policy can sometimes end up making things worse — stimulating an economy that’s already recovering, or tightening when a recession is just beginning. There’s also the crowding out problem: when the government borrows to fund its spending, it competes with private firms for loans, pushing up interest rates and potentially cancelling out some of the stimulus. The size of this problem depends on how much spare capacity the economy has — in a deep recession with idle workers and factories, crowding out is minimal, but at full employment it can be severe.

The deepest question in fiscal policy is about debt. Borrowing today means someone has to pay tomorrow. The key relationship is between the interest rate on government debt and the economy’s growth rate. If the economy grows faster than the interest on the debt, the debt burden shrinks over time even without running a surplus — like a company whose profits grow faster than its loan interest. But if interest rates exceed growth, the debt snowball can become unsustainable. This is why the composition of spending matters enormously: borrowing to build roads and schools that boost future growth is very different from borrowing to fund day-to-day consumption. The first makes future repayment easier; the second just pushes the problem forward.