Crowding Out, Ricardian Equivalence, and the Evidence

Two theoretical arguments against active fiscal policy -- crowding out and Ricardian equivalence -- have shaped policy debates for four decades, particularly in justifying the austerity programs implemented in the United States and Europe after 2010. Both have genuine intellectual foundations; both fail under conditions that frequently apply in real economies.

Crowding out holds that government borrowing reduces private investment by competing for available savings, raising interest rates, and thereby offsetting the demand stimulus the deficit was intended to provide. In a closed economy at full employment with a fixed savings rate, this argument has genuine force: additional government demand for loanable funds will raise interest rates and crowd out some private investment. But the argument breaks down in several common circumstances: when the economy is operating below full employment (when there are unemployed resources the government can deploy without displacing private use), when the central bank accommodates government borrowing by purchasing securities (as in QE), when foreign capital flows in to finance the deficit without raising domestic rates, or when government investment raises the productivity of private capital (as infrastructure investment typically does). The evidence from 2009-2015 -- when the United States ran large deficits with low and declining interest rates -- is inconsistent with the crowding-out mechanism in that context (Blanchard and Leigh, 2013).

Ricardian equivalence, associated with Robert Barro (1974), is more subtle: it holds that deficit-financed fiscal expansion is offset by increased private saving, as rational households anticipate the future tax increases needed to service the debt. If true, the demand stimulus of a tax cut is exactly offset by the increase in household saving to meet future tax obligations. The empirical evidence rejects full Ricardian equivalence: the 2001 and 2003 tax cuts produced increases in consumption, not the saving increases the theory predicts; the 2008-2009 rebate payments generated measurable consumption responses; and the COVID transfer payments in 2020-2021 produced the rapid consumption and recovery that full Ricardian offset would have precluded (Mankiw, 2000).

The theoretical case against fiscal policy assumed full employment, closed economies, and rational infinite-horizon consumers. Real economies are not fully employed in recessions, are not closed, and are populated by households that face liquidity constraints and do not behave as immortal optimization machines. The theory does not apply.Adapted from N. Gregory Mankiw, review of fiscal multiplier literature

The fiscal multiplier -- the ratio of additional GDP to the additional government spending that produces it -- has been the empirical battleground for these theoretical disputes. Pre-crisis estimates of the multiplier varied widely: estimates below 1.0 (consistent with substantial crowding out or Ricardian offset) and above 1.0 (consistent with significant demand-creating effects) appeared in the literature simultaneously. The 2008 crisis provided a large natural experiment: the American Recovery and Reinvestment Act of 2009 allowed economists to examine spending multipliers in an economy at the zero lower bound, where crowding-out through interest rates was impossible. The accumulation of post-crisis evidence has generally supported multipliers above 1.0 in depressed economies with constrained monetary policy -- precisely the conditions of 2009-2015 (Christiano, Eichenbaum and Rebelo, 2011).

The IMF's 2013 acknowledgment that it had systematically underestimated fiscal multipliers in designing austerity programs for European countries -- an acknowledgment made possible only when the severity of the consequent output contractions could not be explained by the models the IMF had used -- was a significant concession. The theoretical framework that had guided a decade of austerity policy had been empirically wrong in the conditions to which it was applied. The practical consequence was prolonged depression in several European economies, with unemployment rates exceeding 25% in Greece and Spain.