When the government spends more, aggregate demand rises immediately by the same dollar amount, because government purchases are one of its four components. What happens to aggregate demand when government spending rises after that first-stage bump depends on how much of the money keeps circulating: taxes, imports, and saving pull dollars out at every round; borrowing costs can push private investment down; and the state of the economy decides whether the extra demand produces real output or just higher prices. The Congressional Budget Office puts the eventual multiplier for federal purchases somewhere between 0.5 and 2.5, well below the neat figures that appear in textbooks.1Congressional Budget Office. The Fiscal Multiplier and Economic Policy Analysis in the United States
The Immediate Shift
Aggregate demand is total spending on all final goods and services in an economy, split into household consumption, business investment, government purchases, and net exports. Because government purchases are one of those four terms, an increase in them mechanically pushes aggregate demand higher by the same dollar amount at the first stage.
If the federal government awards 10 billion dollars in new highway construction contracts, the aggregate demand curve shifts to the right by 10 billion dollars before anything else happens. That is new money entering the economy’s circular flow. Federal spending totaled roughly 7 trillion dollars in fiscal year 2025, about 23 percent of GDP, so even small percentage changes involve very large sums.2U.S. Treasury Fiscal Data. Federal Spending
How the Multiplier Amplifies That First Dollar
The initial shift is only the starting point. When the government pays a construction firm, that payment becomes income for the firm’s workers and suppliers. Those workers spend part of the income at grocery stores, restaurants, and other businesses, creating income for a second set of people. The second group spends part of what it receives, and the chain continues in diminishing rounds.
The fraction of each new dollar of disposable income that a household spends rather than saves is called the marginal propensity to consume, or MPC. If the MPC is 0.75, the simple textbook multiplier is 1 divided by (1 minus 0.75), which equals 4. In that setup, each dollar of new government spending would eventually generate four dollars of activity. Cut the MPC to 0.5 and the multiplier falls to 2. The formula is useful for seeing the mechanism, but it assumes away several forces that reduce the multiplier sharply in practice.
Why Real-World Multipliers Are Smaller
The CBO’s estimated range of 0.5 to 2.5 for federal purchases sits well below the textbook figure. A multiplier of 1.0 means each dollar of government spending produces exactly one dollar of additional output. Below 1.0, the spending partially displaces private activity rather than adding to it.
The gap between theory and reality comes from leakages, meaning money that drops out of the domestic spending chain at every round:
- Taxation. When a worker earns new income, federal and state income taxes take a share before any spending happens. Higher marginal rates shrink disposable income at each round and reduce the effective multiplier.
- Imports. Income spent on foreign-made goods leaves the domestic economy entirely. A high propensity to import weakens the domestic multiplier because each round sends more dollars overseas.
- Saving. Money set aside in savings accounts, retirement funds, or debt repayment exits the spending chain. Empirical U.S. MPC estimates vary widely by population studied, from about 0.05 to 0.9, which is part of why the multiplier range is so wide.
These leakages compound. By the third or fourth round, the amount being re-spent domestically can be a small fraction of the original injection.
The Type of Spending Matters
Not all government spending enters the economy the same way. Direct purchases, like building a bridge or buying military equipment, put the full dollar into the market for goods and services right away because the government is buying something real.
Transfer payments work differently. Unemployment benefits or Social Security checks hand money to individuals, who then decide how much to spend and how much to save. Because recipients save some portion, not all of the initial dollar enters the spending stream. The CBO estimates the multiplier for transfer payments to individuals at 0.4 to 2.1, slightly below the range for direct purchases.1Congressional Budget Office. The Fiscal Multiplier and Economic Policy Analysis in the United States
The gap widens further for other categories. One-time payments to retirees carry a multiplier of just 0.2 to 1.0, and tax cuts aimed at higher-income individuals range from 0.1 to 0.6. Higher-income households tend to save a larger share of windfalls, so less money recirculates. A dollar spent hiring workers to repair infrastructure generates more aggregate demand than a dollar sent as a tax rebate that ends up in a savings account.
Crowding Out From Government Borrowing
When the government finances new spending by borrowing rather than raising taxes, it competes with private borrowers for the available pool of savings. The Treasury issues bonds to cover the gap, and the added demand for funds pushes interest rates higher.3U.S. Department of the Treasury. Financing the Government
Higher rates make borrowing more expensive for businesses weighing new factories, equipment, or hiring, so some private investment projects that would have been profitable at lower rates get shelved. Households also pull back on interest-sensitive purchases like homes and cars. The rise in government spending is partly offset by a fall in private investment and some consumption, an effect economists call crowding out.
How severe it gets depends on how sensitive private investment is to rate changes. If businesses have few funding alternatives and thin margins, even a small rate increase can kill projects. If capital markets are deep and businesses are flush with cash, the effect is milder. Over longer horizons, one modeling exercise found that a 1 trillion dollar increase in government debt could reduce the capital stock by roughly 0.8 percent and hourly wages by about 0.2 percent by 2050 in a partially open economy.
Do Consumers Cancel the Stimulus?
Ricardian equivalence argues that deficit-financed government spending may not boost aggregate demand much at all. Rational households, the theory goes, know that today’s borrowing must eventually be repaid through higher future taxes, so they increase their saving now by the amount needed to cover those future tax bills instead of spending freely.
If the theory held perfectly, every dollar of new deficit-financed spending would be matched by a dollar of reduced private consumption, and fiscal policy would be useless as a stimulus tool. In practice, most people do not calculate their lifetime tax liabilities and adjust their saving to match. Credit constraints, short planning horizons, and uncertainty about future tax policy all weaken the mechanism. Some households do save more when deficits grow, which partially offsets the stimulus, but the full offset that pure Ricardian equivalence describes does not show up in the data.
Timing Delays the Boost
Even when a spending increase will eventually raise aggregate demand, the gap between the decision and the economic impact can be long. Data arrives late, so recognizing that stimulus is needed takes months. Legislation to authorize it routinely takes six months to a year to move through Congress. And once money is authorized, it still has to actually flow out to reach the economy.
The Infrastructure Investment and Jobs Act illustrates that last delay. As of early 2026, the Department of Transportation had obligated about 73 percent of the law’s funding under binding agreements with recipients, but only about 43 percent had actually been paid out.4US Department of Transportation. Infrastructure Investment and Jobs Act (IIJA) Funding Status The law was signed in November 2021, so more than four years later over half the money had still not reached the economy. Large infrastructure projects move through design, permitting, and phased construction, with payments trickling out over years.
Fiscal stimulus often arrives after the recession it was meant to fight has already ended. That mistiming can push demand into an economy that no longer needs it, feeding inflation rather than recovery.
The Economy’s Starting Point Changes the Answer
The same dollar of government spending produces very different results depending on whether the economy is slumping or running near capacity. During a deep recession, unemployment is high, factories sit idle, and businesses have room to expand production without bidding up prices. New government spending puts idle resources to work and generates real output gains with little inflationary pressure. The multiplier tends toward the higher end of its range.
The International Monetary Fund has found that fiscal stimulus using public spending is “particularly potent when there is economic slack” and interest rates are low.5International Monetary Fund. Countering Future Recessions in Advanced Economies When the central bank is already holding rates near zero, the crowding-out problem largely disappears because government borrowing does not push rates meaningfully higher from an already-low baseline.
The picture reverses near full employment. With most workers already employed and factories running near capacity, new government demand competes with private demand for the same limited resources, and the result is higher prices rather than higher output. The multiplier shrinks, and in extreme cases can fall below 1.0, meaning the spending displaces more private activity than it creates.
Monetary policy sits on top of all this. If the Federal Reserve accommodates fiscal expansion by keeping rates low, the multiplier is larger because crowding out stays muted. If the Fed responds to inflationary pressure by raising rates, it works against the fiscal expansion and shrinks the multiplier further. The interaction between the two is often more important than either policy in isolation, which is where most simple multiplier stories go wrong.