Fiscal policy and the national debt Lesson 9 of 10
Economics · Fiscal policy and government
Crowding out: do government deficits raise interest rates?
How government borrowing can raise interest rates and displace private investment, what Laubach and CBO estimate, why it is weak in slumps, and what US data show.
On this page
Crowding out is the idea that when the government borrows more, it leaves less for everyone else. To sell more bonds, the Treasury has to offer investors a better deal, which means higher interest rates. Higher rates make some private projects, such as a factory, a house or a business loan, no longer worth doing. The government's spending has then displaced, or crowded out, private spending, and the economy's capital stock grows more slowly than it otherwise would.
It is one of the oldest arguments about deficits and one of the hardest to measure. A glance at US history seems to refute it: deficits were largest in 2009 and 2020, when interest rates were lowest. This lesson explains the mechanism with a worked example, shows why that raw comparison misleads, summarizes the best-known estimate (Laubach's: about 25 basis points per point of deficit) and CBO's long-run working assumption, and explains when the effect is likely to be weak.
The mechanism
Think of a market for saving. Households, firms and foreigners supply saving; businesses, home buyers and the government demand it. The interest rate balances the two.
When the government runs a larger deficit, its demand for saving rises at every interest rate. Two things then happen together: the interest rate rises, which draws out somewhat more saving, and other borrowers, facing a higher rate, borrow less. The second part is crowding out.
The open-economy channel
The US borrows from the whole world. If higher US rates attract foreign saving, the rise in rates is smaller, but the inflow tends to strengthen the dollar, which makes US exports dearer and imports cheaper. Net exports fall instead of investment. Economists sometimes call the pair of deficits this creates the "twin deficits".
What the data show
The raw comparison misleads
Across fiscal 1977 to 2026, the correlation between the deficit and the 10-year yield is about −0.3: bigger deficits went with lower yields. Fiscal 2020 had the largest deficit in the sample, 14.7% of GDP, and an average 10-year yield of 1.13%. Fiscal 1982 had a deficit of 3.9% and a yield of 13.86%. Fiscal 2000, a surplus year (2.3% of GDP), averaged 6.17%.
This does not show that deficits lower interest rates. It shows that something else drives both. In a recession, tax receipts fall and benefits rise (Automatic stabilizers: taxes and benefits that move with the cycle), so the deficit widens; at the same time the Fed cuts rates and investors buy safe bonds, so yields fall. And over decades, yields followed inflation, which was high around 1980 and low after 2008. Any test of crowding out has to strip these out.
Laubach's estimate: projections, not current deficits
Thomas Laubach of the Federal Reserve Board handled the problem by using expected deficits and expected future rates. He took CBO's and OMB's five-year-ahead projections of the deficit and debt, which are little affected by today's recession, and compared them with the 10-year rate expected five years ahead, read from the Treasury yield curve. That removes most of the business cycle and the Fed's current stance.
He found that a 1-percentage-point increase in the projected deficit-to-GDP ratio raised that forward 10-year rate by 20 to 40 basis points, with a typical estimate of about 25. A 1-point increase in projected debt-to-GDP raised it by about 4 to 5 basis points. His sample covered projections from 1976 to 2003.
CBO's long-run assumption
CBO's March 2009 letter on the Recovery Act stated its basic assumption plainly: in the long run, each dollar of additional federal debt crowds out about a third of a dollar of private domestic capital, with the rest offset by more private saving and by inflows of foreign capital. That is why CBO estimated the act would slightly lower GDP after 2015, by between zero and 0.2%, even though it raised GDP in 2009–2011.
When crowding out is weak
- In a slump. The same CBO letter noted that crowding out was unlikely in the short run under early-2009 conditions, because firms were cutting investment in response to weak demand anyway. Extra government spending can even crowd in investment by raising sales.
- At the zero lower bound. When the central bank holds its policy rate near zero and buys bonds, the rate does not rise with the deficit. Ramey's 2019 survey finds some evidence that fiscal multipliers can be larger in such periods, the mirror image of weak crowding out, though the estimates vary.
- With ample foreign saving. If the rest of the world is eager to hold US Treasuries, the domestic rate rises less, though the dollar and net exports may adjust instead.
And when is it strong? At full employment, with a central bank raising rates to contain inflation, and with debt already large, crowding out works through both the market and the Fed's response.
Pitfalls and limitations
- Don't use the raw scatter. Deficit and yields move together for reasons other than crowding out; the correlation has the "wrong" sign.
- Small per year, large over time. 25 basis points sounds minor, but applied to every new loan and every refinanced Treasury for years, it adds up, including to Interest on the national debt: net interest, rates and r versus g.
- Forward rates are not today's rates. Laubach's effect is on the rate expected five years ahead; today's 10-year yield also reflects current Fed policy and the term premium.
- What the money buys matters. Borrowing that pays for productive public investment can raise future output, partly offsetting lost private capital. CBO's 2009 range assumed between zero and the same return as private investment.
Crowding out versus related ideas
| Idea | What is displaced | Through |
|---|---|---|
| Crowding out | Private investment | Higher interest rates |
| Open-economy crowding out | Net exports | Stronger currency |
| Monetary offset | Overall demand | Central bank raising rates |
| Ricardian equivalence | Private consumption | Households saving for future taxes |
| Crowding in | Nothing; investment rises | Higher sales in a slump |
How Finabase does it
Finabase does not estimate crowding out; the scatter and diagram on this page are ours, from public Treasury, BEA and Fed data and invented numbers. The app shows the market where any effect would appear:
- The 10-year Treasury yield has its own page and chart.
- The Treasury page lists the whole curve from 1 month to 30 years, so you can see whether long yields move more than short ones around budget news.
- Coverage of budget and borrowing debates collects under the Economy news topic.
Frequently asked questions
What is crowding out in economics?
It is the reduction in private investment or other private borrowing caused by higher interest rates when the government borrows more. The government's demand for saving pushes up its price, and some private projects no longer pay.
Do budget deficits raise interest rates?
Studies that control for the business cycle find a modest effect: Laubach estimated about 25 basis points on the expected 10-year rate per percentage point of projected deficit-to-GDP. Raw data show the opposite sign because recessions raise deficits and lower rates at once.
Is crowding out a problem in a recession?
Usually much less. When firms are cutting investment and the central bank is holding rates down, government borrowing does not push rates up much, and extra demand may even raise private investment.
What is crowding in?
It is the opposite case, where government spending raises private investment, for example by lifting sales in a slump or by building infrastructure that makes private projects more profitable.
Summary
Crowding out is the displacement of private borrowing and investment by government borrowing, working through higher interest rates or, in an open economy, a stronger currency. Raw US data show the opposite of what the theory predicts, because recessions raise deficits and lower yields together. Careful estimates find a real but modest effect: about 25 basis points on the expected 10-year rate per point of projected deficit, 4 to 5 per point of projected debt (Laubach), and a long-run loss of about a third of a dollar of private capital per dollar of debt (CBO's 2009 assumption). The effect is weakest in a slump at the zero lower bound and strongest at full employment.
Check your understanding
4 questions on this lesson. Each answer comes with the reason, and nothing is scored.
A chart shows that in years with the largest US deficits, 10-year yields were among the lowest. Does this disprove crowding out?
- Yes: deficits clearly lower interest rates
- No: recessions widen deficits and lower rates at the same time, so the raw comparison mixes up cause and effect
- Yes, but only after 2008
Show the answer
B. No: recessions widen deficits and lower rates at the same time, so the raw comparison mixes up cause and effect Automatic stabilizers widen deficits in a recession while the Fed cuts rates and investors buy bonds. Laubach used projected deficits and forward rates to strip those effects out and found a positive effect.
Using Laubach's typical estimate, roughly how much would a rise in projected deficits from 4% to 6% of GDP raise the 10-year rate expected five years ahead?
- About 5 basis points
- About 50 basis points
- About 2 percentage points
Show the answer
B. About 50 basis points 2 points × about 25 basis points per point ≈ 50 basis points. It is the paper's average effect for 1976–2003, not a forecast.
In early 2009 firms were cutting investment and the Fed's policy rate was near zero. What did CBO say about crowding out from the Recovery Act in the short run?
- It would be unusually large because debt was rising fast
- It was unlikely, because firms were reducing investment anyway and stimulus could partly offset that
- It would be exactly one third of each dollar spent
Show the answer
B. It was unlikely, because firms were reducing investment anyway and stimulus could partly offset that CBO's March 2009 letter said crowding out was unlikely in the short run under those conditions; its one-third-of-a-dollar assumption applied to the long run.
Foreign investors buy most of a new wave of Treasury debt, so US interest rates barely rise. Is there any crowding out?
- No, because rates did not rise
- Possibly, through a stronger dollar that lowers net exports instead of investment
- Yes, investment falls by the full amount of the deficit
Show the answer
B. Possibly, through a stronger dollar that lowers net exports instead of investment In an open economy the adjustment can come through the currency: capital inflows raise the dollar, making exports dearer and imports cheaper.
Exercise
One task to do in Finabase. Nothing is graded or saved.
Open the 10-year Treasury yield on its longest range. Find the level in 1982, in 2000 and in 2020. The lesson gives the deficits: 3.9% of GDP in fiscal 1982, a 2.3% surplus in fiscal 2000 and 14.7% in fiscal 2020. Rank the three years by deficit and by yield. Does the raw comparison support crowding out?
Open the 10-year yield chartWhat you should find
The ranking runs the wrong way: the biggest deficit (2020) came with the lowest yield and a modest deficit (1982) with the highest. Recessions widen deficits while the Fed cuts rates, and over decades yields followed inflation. That is why tests of crowding out use projected deficits and forward rates, as Laubach did, rather than this raw picture.
Sources and further reading
- Laubach, T. — New Evidence on the Interest Rate Effects of Budget Deficits and Debt, Federal Reserve Board FEDS 2003-12 (Journal of the European Economic Association 7(4), 2009), retrieved 2026-11-18
- CBO — Letter to Sen. Charles E. Grassley on the macroeconomic effects of ARRA, March 2, 2009 (UNT web archive), retrieved 2026-11-18
- Ramey, V. — Ten Years after the Financial Crisis, NBER Working Paper 25531 (2019), retrieved 2026-11-18
- US Treasury / OMB — Federal surplus or deficit, fiscal year (FRED FYFSD), downloaded 2026-10-30
- Federal Reserve Board H.15 — 10-year Treasury constant maturity, monthly (FRED GS10), downloaded 2026-10-30
- BEA — Gross domestic product, nominal (FRED GDP), downloaded 2026-10-30
