VoxEU Column Global crisis Macroeconomic policy

After the stimulus, the big retrenchment

The staggering growth in public debt as a result of the financial crisis has led many to call for significant fiscal retrenchment. This column argues that such looming expenditure cuts will actually enhance the effectiveness of today’s fiscal stimulus. But if monetary policy is constrained by the zero lower bound on policy rates, the spending cuts should not come too early.

The global crisis has put public debt on a sharply steeper trajectory. Governments have provided large-scale support to the financial system, and implemented discretionary fiscal stimulus, while at the same time experiencing a dramatic fall in tax revenue. With the economic recovery gradually taking hold, the focus is now shifting to fiscal “exit strategies''.

Judging from official announcements, the current period of fiscal stimulus will be followed by significant fiscal retrenchment over the medium term. In the US, for example, the Obama administration's 2010 budget pledges to "cut the deficit in half by the end of [its] first term, and [to] bring non-defence discretionary spending to its lowest level as a share of GDP since 1962'' (Office of Management and Budget 2009). Similarly, the UK government's December 2009 Pre-Budget Report foresees large medium-term deficit cuts, with two-thirds of the fiscal effort on the expenditure side.

Figure 1 plots the outlook for discretionary spending in these two countries as interpreted by the respective national fiscal watchdogs. In both cases, discretionary spending as a share of GDP is projected to fall below the pre-crisis trend, more than offsetting the initial rise in spending related to fiscal stimulus measures.

Figure 1. Discretionary government expenditure (% of potential GDP)

Sources: Institute for Fiscal Studies, HMT 2009 Pre-budget report; Congresssional Budget Office; OECD;and IMF staff calculations.

Of course, the credibility of these official plans cannot be taken for granted. But “spending reversals” are not only a feature of official budget projections. Private forecasters, like the Economist Intelligence Unit, also project that government spending on goods and services will be curtailed over the medium term.

A basic consideration supports this idea: growing market concerns about sovereign debt will induce consolidation one way or another; and voters' resistance to large tax increases may leave policymakers with little other choice but to put some of the burden of adjustment on discretionary expenditure cuts.

Anticipated spending cuts enhance short-run fiscal stimulus

Public spending cuts, when they occur, reduce aggregate demand. The question we address here is whether anticipated future cuts (which would serve to offset at least part of the costs of today’s fiscal expansion) also affect the impact of the upfront stimulus packages.

In a previous column and an earlier paper, we have discussed the macroeconomic transmission of fiscal stimulus followed by a spending reversal (Corsetti and Müller 2008, and Corsetti, Meier, and Müller 2009). The conclusion is that the anticipation of a future spending reversal generally raises the expansionary impact of today’s fiscal stimulus through its effect on long-term real interest rates.

  • Prospective spending cuts reduce expected inflationary pressures, allowing the central bank to set lower nominal (and, effectively, real) short-term interest rates in the future.
  • Anticipated policy rate cuts, in turn, immediately reduce long-term real interest rates today, as these capture market expectations for the entire path of future short-term rates.

Consequently, the future spending cuts boost current demand.

Figure 2 illustrates the macroeconomic dynamics of spending reversals, based on simulation results for a standard new Keynesian business cycle model.

Figure 2. Effect of government spending shock with pure tax finance (dashed lines) vs. spending reversal (solid lines)

Note: Horizontal axes measure quarters, vertical axes deviations form steady state.

The figure contrasts the effects of an exogenous increase in government spending by 1% of GDP across two different medium-term consolidation scenarios.

  • In the first scenario, the temporary upfront stimulus is entirely financed by higher taxes (the dashed lines).
  • In the second scenario, the initial rise in government spending is matched by future spending cuts (solid lines), implying a complete spending reversal.

Demand for private consumption is always higher in the case of the spending reversal than in the pure tax-finance case. The stronger private consumption profile in turn raises aggregate (public plus private) demand for several quarters. Correspondingly, the equilibrium paths of inflation and the policy rate are also higher in the early periods.

The tax burden faced by consumers obviously differs in the two scenarios depicted in Figure 2. Taxes increase in the first scenario, but not in the second. Yet, this “wealth effect" is of little consequence for the dynamics of private consumption, as households’ permanent income falls very little in the case of temporary fiscal stimulus. Instead, the difference across the two scenarios is chiefly driven by the distinct behaviour of interest rates. Demand is higher with spending reversals because long-term real interest rates are lower, triggering strong intertemporal substitution effects.

What if monetary policy is constrained by the zero lower bound?

Monetary policy plays an essential part in the transmission mechanism outlined above. The implicit assumption – as embodied in a standard Taylor rule – is that the central bank effectively controls short-term real rates, by varying nominal policy rates in response to the inflation outlook.

Today’s circumstances pose a formidable challenge in this regard. In the US, the UK, and a number of other economies, policy rates are effectively at the zero lower bound and thus cannot be cut any further even if new deflationary shocks hit the economy.1

This fact also complicates fiscal exit strategies relying on a spending reversal. Indeed, anticipated spending cuts enhance the short-term expansionary impact of fiscal stimulus only insofar as their future deflationary effect, all else equal, leads to lower expected real rates. With short-term nominal rates at the zero lower bound, lower inflation (caused by the public spending cut) instead raises real rates. As a result, the prospect of fiscal adjustment via spending cuts might actually weaken current demand and prolong the recessionary dynamics. This raises the question to what extent our earlier findings about the beneficial impact of spending reversals must be qualified at the lower bound.

We take up this issue in a second recent paper (co-authored with Keith Kuester). Borrowing from Christiano, Eichenbaum, and Rebelo (2009), and Erceg and Lindé (2010), we consider an economy hit by a strong recessionary shock; a sudden marked rise in desired household saving depresses demand and creates deflationary pressure. Once the central bank has brought its policy rate to zero, any further deflationary shock raises, rather than lowers, real interest rates. Higher rates, in turn, further discourage demand, compounding the initial deflationary shock. A discretionary increase in government spending can halt this negative spiral by counteracting the deflationary impact of falling private demand. Indeed, fiscal multipliers can be unusually large under such circumstances.

Timing is crucial

We study the interaction of government spending reversals with the zero lower bound in the new Keynesian model introduced above. In response to the recessionary shock, the government is assumed to increase public spending by one per cent of GDP for eight quarters. The stimulus is financed either fully, by higher taxes, or by a combination of somewhat higher taxes and a partial spending reversal accounting for 50% of the upfront stimulus cost. In addition, we allow for the spending reversal to be phased in at different dates. This aspect of our simulations is crucial given that it interacts directly with the constraint on monetary policy imposed by the lower bound.
Policymakers’ focus in the midst of a deep recession is arguably on the expansionary effect of the fiscal stimulus. We define the relevant fiscal multipliers as the increase in output or, respectively, private consumption relative to the government spending impulse in the initial period of the stimulus.

Figure 3 summarises the results from our simulations for different fiscal exit strategies. In particular, the figure shows how fiscal multipliers vary with and without reversals ("no" marks the case of no reversal), as well as with the timing of the spending reversal. The x-axis reports the number of quarters between the withdrawal of the stimulus and the beginning of spending cuts below trend levels, i.e., the actual reversal. The two panels of Figure 3 refer to reversals spread over four and twelve quarters, respectively, allowing us to compare fiscal exit strategies along one further dimension, i.e., the degree of gradualism in the spending reversal.

Figure 3. Fiscal multipliers under alternative reversal patterns.

Note: Horizontal axes measure quarters between the end of the stimulus and the beginning of the reversal; “no” refers to the case of no reversal (full tax finance).

The results show that, in spite of the zero lower bound constraint, the anticipation of spending reversals can still enhance the expansionary effect of fiscal stimulus in the short run: fiscal multipliers are higher in most spending-reversal scenarios than in the pure tax-finance case. But, for a positive effect to be obtained, spending cuts need to be implemented with suitable delay and/or at least with some gradualism.

In our experiments, the highest fiscal multiplier arises when the spending reversal begins, some eight quarters after the stimulus is withdrawn. Although this exact numerical finding is sensitive to the parameters of our model, the basic insight is fairly robust: for the spending reversal to enhance short-term fiscal stimulus, it should occur when the economy is already emerging from recession, so that policy rates are above zero again. In this case, the prospective reversal allows the central bank to worry less about inflation, and thus maintain an accommodative monetary stance.

Very early and rapid reversals can instead be counterproductive. This possibility is illustrated by the graph in the top panel of Figure 3, where the reversal is spread over four periods: fiscal multipliers for very early spending reversals reduce the fiscal multiplier relative to the pure tax-finance case. This effect occurs because at the time of the reversal the economy is still struggling to recover from the initial recessionary shock that ties monetary policy to the lower bound. The additional deflationary shock lengthens the zero lower bound episode and worsens the recession. Sufficient gradualism, meanwhile, restores the beneficial impact – even for early spending reversals. As the bottom panel of Figure 3 shows, a spending reversal always raises fiscal multipliers if it is spread over 12 quarters.


The effectiveness of short-run fiscal stimulus depends not only on the specific fiscal measures taken today, but also the on medium-term fiscal outlook, notably the government’s expected strategy for fiscal consolidation.

We find that anticipated spending cuts generally enhance the expansionary effect of current fiscal stimulus. This result still holds when monetary policy is constrained by the zero lower bound on policy rates. In this situation, however, the spending reversal must not come too early on the recovery path, or at least must be suitably gradual.

It is worth noting that our analysis abstracts from the additional complication of default risk premia on sovereign debt. Modelling default risk is challenging, but we think that it would essentially reinforce our main conclusion about the macroeconomic benefits of spending reversals. In fact, to the extent that credible consolidation plans attenuate concerns about default risk, current long-term interest rates would ease above and beyond the effects captured in our analysis, further boosting current demand. That said, very acute concerns about fiscal sustainability may in some cases argue for immediate consolidation even where this imparts a procyclical impulse to the recessionary economy.

Our results suggest that current stimulus and credible plans for medium-term expenditure restraint are complementary elements of a strategy to move the economy out of a deep recession. While care should be taken not to program very early spending cuts when monetary policy is constrained by the zero lower bound, this does not call into question the general case for identifying upfront credible consolidation measures. Policymakers undoubtedly face a difficult task in this regard, but a combination of smart stimulus design – for example, a strong emphasis on bringing forward future investment spending, which quasi-automatically implies subsequent spending cuts – and robust fiscal rules strikes us as a promising approach to pursue.

Disclaimer: The views expressed herein are those of the authors and do not necessarily represent those of the IMF.


1 This is not to deny the possibility that central banks can affect economic conditions even when the short-term nominal interest rate is at the lower bound, as indeed several central banks have attempted through various unconventional operations since 2008. However, the significant uncertainty about the effectiveness and risks of such operations may itself be regarded as a policy constraint.


Corsetti, Giancarlo, Keith Kuester, André Meier, and Gernot J. Müller (2010), “Debt Consolidation and Fiscal Stabilization of Deep Recessions”, CEPR Discussion Paper 7649.

Corsetti, Giancarlo, André Meier, and Gernot J. Müller (2009), “Fiscal Stimulus with Spending Reversals”, IMF Working Paper 09/106.

Corsetti, Giancarlo and Gernot J. Müller (2008), “The Effectiveness of Fiscal Policy Depends on the Financing and Monetary Policy Mix”, VoxEU.org, 12 November.

Christiano, Lawrence J, Martin Eichenbaum, and Sergio Rebelo (2009), “When is the Government Expenditure Multiplier Large?” NBER Working Paper 15394.

Erceg, Christopher J and Jesper Lindé (2010), “Is There a Fiscal Free Lunch in a Liquidity Trap?”, CEPR Discussion Paper 7624.

Office of Management and Budget (2009), “Presidential Transmittal Letter: Budget FY 2010”.

105 Reads