Government spending crowds out private investment

The claim

When the government runs a deficit, it must borrow from the same pool of savings that would otherwise flow to private investment. This increased demand for loanable funds pushes up interest rates. Higher interest rates raise the cost of capital for private businesses, leading them to cancel or defer investment projects. The result is a net loss: the economic activity generated by government spending is offset by the private investment it displaces. In the strong form of the argument, the offset is complete — government spending produces no net increase in output. In moderate versions, the offset is substantial enough to make deficit spending a poor policy tool compared to tax cuts or spending restraint.

This claim is a staple of classical and neoclassical macroeconomics and has been central to fiscal austerity arguments from Ricardian equivalence (Barro, 1974) through to the expansionary austerity literature of the early 2010s (Alesina & Ardagna, 2010). Proponents point to the period of high deficits and high interest rates in the early 1980s as empirical confirmation.

The mechanism

The crowding-out mechanism is straightforward in the classical loanable funds model: savings are a fixed pool; government borrowing competes with private borrowers for that pool; competition raises the price of borrowing (the interest rate); higher rates reduce private investment. The model is internally consistent and predicts a clear observable signature: when government deficits expand, interest rates rise.

The mechanism breaks down, or inverts, under at least three well-documented conditions.

The zero lower bound. When the central bank has pushed short-term interest rates to zero (or near zero), the economy is operating below potential, and idle savings and unemployed resources exist, government borrowing does not displace private investment — it draws idle resources into productive use. The interest rate channel is inoperative because the central bank is already setting rates at the floor. Under these conditions, the fiscal multiplier can exceed 1.0: each dollar of government spending produces more than one dollar of GDP, which increases private sector income and can stimulate additional private investment. This is the condition Summers (2014) described as secular stagnation: a persistent tendency toward excess savings and inadequate private demand that requires fiscal intervention to maintain full employment.

The complementarity channel. Government investment in infrastructure, basic research, and human capital can raise the private return to investment rather than displacing it. A highway reduces private logistics costs. A publicly funded semiconductor research program creates spillovers that private chip manufacturers capture. A public health investment raises labor productivity. In these cases, public and private investment are complements, not substitutes. Mazzucato (2013) documents this systematically for US federal R&D: the technologies that made Apple’s iPhone commercially viable — GPS (Defense Advanced Research Projects Agency), the internet (ARPA), lithium-ion batteries (Department of Energy), SIRI (DARPA), touchscreen interfaces (National Science Foundation-funded university research) — were each funded by government before any private firm identified them as commercially viable. Crowding out assumes private investment would have occurred absent government spending; the complementarity evidence shows it often would not have.

Developmental state investment. In late-industrializing economies, private capital markets frequently underprovide investment in capital-intensive sectors because individual projects are too large, returns are too distant, and risk cannot be priced without incumbents who do not yet exist. State investment creates those incumbents, after which private capital follows. South Korea’s POSCO, established with state capital in 1968 when no private Korean investor would fund a steel plant, became the nucleus of an integrated industrial complex that private firms built around. The crowding-out model assumes pre-existing private investors who are then denied access to funds; the developmental state evidence shows that in many cases, those investors did not exist until public investment created the conditions for them.

The evidence

Post-2008 United States: a direct test of the interest rate prediction

The 2008–2012 period provides a near-ideal test of the crowding-out mechanism. Federal debt held by the public doubled, rising from 36% of GDP in 2007 to 72% of GDP in 2012. If crowding out were the dominant mechanism, 10-year Treasury yields should have risen substantially as government competed for the available pool of savings. Instead, the 10-year yield fell from 3.7% in 2007 to a low of 1.8% in 2012. The interest rate prediction failed. The reason is not mysterious: the economy was operating well below potential, the Federal Reserve was holding short rates at zero, and private demand for investment funds had collapsed. In the absence of government borrowing, the loanable funds market would have cleared at lower investment levels, not higher.

The American Recovery and Reinvestment Act (ARRA) of 2009 deployed approximately $840 billion in spending and tax cuts. CBO analysis estimated it raised GDP by 1.5–3.5% and employment by 1.0–2.9 million jobs at its peak effect (2010 Q1–Q2). Private nonresidential fixed investment, which had collapsed 22% in 2009, began recovering in mid-2010 — concurrent with, not after, the stimulus deployment. The crowding-out prediction would have been that stimulus prolonged the slump in private investment. The data show the opposite temporal pattern.

IMF infrastructure multiplier estimates

The International Monetary Fund’s 2014 World Economic Outlook included a systematic review of public investment multipliers across advanced economies. For spending at or near the zero lower bound, the IMF found a short-run multiplier of approximately 1.5 and a long-run multiplier above 2.0, driven substantially by productivity-enhancing infrastructure investment. An IMF Working Paper (Abiad, Furceri, & Topalova, 2015) specifically examining public investment found that a 1-percentage-point increase in public investment as a share of GDP raised output by about 1.5% over four years and increased private investment by about 2% over the same period — a crowding-in effect of 2:1. The conditions required for this result include economic slack and productive public investment; the result does not hold at full employment with wasteful public spending.

The German KfW model

Germany’s Kreditanstalt für Wiederaufbau (KfW), a state-owned development bank with €540 billion in assets, provides a sustained cross-national test. KfW lends at below-market rates for infrastructure, housing, climate investment, and small business development, funded by government-guaranteed bonds. German business investment rates have historically exceeded the EU average despite — or plausibly because of — KfW’s scale. The co-movement of public development lending and private investment in Germany runs contrary to the crowding-out prediction. KfW’s model has been explicitly cited by the European Investment Bank and the Biden Administration’s infrastructure program designers as evidence that patient public capital catalyzes rather than displaces private capital.

Japan: three decades of deficits and falling yields

Japan expanded its government debt-to-GDP ratio from approximately 60% in 1990 to over 250% by 2020 — the largest peacetime debt accumulation by any major economy. Across this period, 10-year Japanese government bond yields fell from 6.5% to below 0%. The crowding-out mechanism predicts the opposite. The Japanese case is extreme, but it demonstrates clearly that the relationship between deficit spending and interest rates is not a simple positive one: it is mediated by savings behavior, central bank policy, domestic versus external financing, and the output gap. Japan’s experience is consistent with the secular stagnation hypothesis: excess private savings require sustained government borrowing to prevent the interest rate from falling below the level consistent with positive investment.

South Korea and developmental crowding-in

South Korea’s post-1961 industrialization provides the clearest example of public investment creating, rather than displacing, private investment. State-directed credit through the Korea Development Bank financed POSCO (steel, 1968), KEPCO (electricity), KIST (science and technology research), and Hyundai’s early shipbuilding facilities. Private chaebol (conglomerates) built on the infrastructure, trained workforce, and adjacent industries that state investment created. By 1996, when South Korea joined the OECD, private investment as a share of GDP was among the highest in the world — the opposite of the crowded-out outcome. Amsden (1989) and Wade (1990) document this pattern across the East Asian developmental states. The crowding-out model simply does not describe how industrialization occurred in South Korea, Taiwan, or Japan.

Where crowding out does appear

The evidence does not show that crowding out never occurs. In economies at or near full employment, with tight credit conditions and a central bank unwilling to accommodate fiscal expansion, government borrowing can raise interest rates and displace some private investment. The United States in the early 1980s is plausibly such a case: Reagan deficits coincided with Federal Reserve Chairman Volcker’s aggressive rate increases, and real interest rates rose sharply. However, even here the causal attribution is contested — the rate increases were primarily a deliberate Federal Reserve policy response to inflation, not a loanable funds market phenomenon. Standard macroeconomic models (including the IMF’s and CBO’s) estimate some crowding out in full-employment conditions, but quantify it at far below the complete offset implied by Ricardian equivalence.

Who benefits

The crowding-out argument is strategically useful to any political or financial constituency that prefers fiscal contraction. The financial sector — particularly holders of long-duration fixed-income instruments — benefits from lower inflation expectations and fiscal restraint, which protect the real value of existing bond portfolios. Peter Peterson’s deficit-hawk network, including the Peterson Foundation and its funded projects at the Committee for a Responsible Federal Budget, has spent hundreds of millions of dollars promoting fiscal austerity. The argument frames deficit reduction as economically necessary rather than distributionally motivated, obscuring the fact that austerity disproportionately affects public services used by lower-income households.

Corporations in sectors where government investment competes with private provision (broadband, energy, healthcare) also benefit from crowding-out framing, since it reframes public investment programs as harmful to growth rather than competitive threats to their market position. The American Legislative Exchange Council (ALEC), funded substantially by the Koch network and telecom incumbents, has consistently opposed public broadband investment using crowding-out arguments.

Conservative think tanks — the Cato Institute, the Heritage Foundation, the American Enterprise Institute — have produced a sustained literature applying crowding-out logic to argue against infrastructure spending, the Inflation Reduction Act’s clean energy investments, and federal R&D programs. These organizations receive major funding from fossil fuel interests and financial sector donors with direct stakes in the policy outcomes these arguments support.

The counter

The classical crowding-out model deserves more credit than its post-2008 performance implies. The zero lower bound episode was unusual: a once-in-several-generations coincidence of financial crisis, debt deflation, and monetary policy exhaustion. In normal cyclical conditions — when the economy is at or near full employment and the central bank is not constrained — government borrowing does raise interest rates and does compete with private investment. The mechanism is real. The error is not in the model but in its advocates’ insistence on applying it regardless of the cyclical position of the economy.

Ricardo’s equivalence theorem (Barro, 1974) offers a sophisticated version of the argument: rational households, anticipating future tax increases to service current debt, increase their savings now to offset government dissaving, neutralizing any fiscal stimulus. This is a stronger prediction than simple interest rate crowding out, and it is not fully refuted by post-2008 evidence — some Ricardian offset is likely, and the net fiscal multiplier is probably lower than Keynesian models without rational expectations predict. The dispute is about magnitude: complete offset is implausible; zero offset is also implausible; the contested space is how large the net effect is under different conditions.

The secular stagnation critique (Summers, 2014; Rachel & Summers, 2019) itself implies a limit: if fiscal expansion successfully raises inflation expectations and returns the economy to full employment, then continued deficit expansion after that point would indeed produce crowding out. The argument for large fiscal multipliers is not an argument for unlimited deficit spending — it is a conditional argument that depends on slack in the economy. Crowding-out skeptics are on firmest ground during recessions and weakest ground during booms.

References

Abiad, A., Furceri, D., & Topalova, P. (2015). The macroeconomic effects of public investment: Evidence from advanced economies (IMF Working Paper WP/15/95). International Monetary Fund. https://doi.org/10.5089/9781513512907.001

Alesina, A., & Ardagna, S. (2010). Large changes in fiscal policy: Taxes versus spending. Tax Policy and the Economy, 24(1), 35–68. https://doi.org/10.1086/649828

Amsden, A. H. (1989). Asia’s next giant: South Korea and late industrialization. Oxford University Press.

Barro, R. J. (1974). Are government bonds net wealth? Journal of Political Economy, 82(6), 1095–1117. https://doi.org/10.1086/260266

Blanchard, O., & Leigh, D. (2013). Growth forecast errors and fiscal multipliers. American Economic Review, 103(3), 117–120. https://doi.org/10.1257/aer.103.3.117

International Monetary Fund. (2014). World economic outlook: Legacies, clouds, uncertainties (Chapter 3: Is it time for an infrastructure push?). IMF. https://www.imf.org/en/Publications/WEO/Issues/2016/12/31/Legacies-Clouds-Uncertainties

Mazzucato, M. (2013). The entrepreneurial state: Debunking public vs. private sector myths. Anthem Press.

Rachel, L., & Summers, L. H. (2019). On secular stagnation in the industrialized world. Brookings Papers on Economic Activity, 2019(1), 1–76. https://doi.org/10.1353/eca.2019.0000

Summers, L. H. (2014). US economic prospects: Secular stagnation, hysteresis, and the zero lower bound. Business Economics, 49(2), 65–73. https://doi.org/10.1057/be.2014.13

Wade, R. (1990). Governing the market: Economic theory and the role of government in East Asian industrialization. Princeton University Press.