What makes fiscal policy (more) effective?
Sizeable resort to fiscal policy seems justified because of the main features of this recession, namely, its expected severity, and the emergence of widespread credit constraints, which makes the transmission of monetary policy weaker and more uncertain. Simultaneously, the global nature of the crisis and the fact that most economies have become quite open to trade strongly argue for coordinated fiscal expansion.
According to most observers/analysts/economists, fiscal packages need to be large, and implementation rapid. For instance, the simple back-of-the-envelope calculation proposed by Krugman (2008) suggests that, for the US, the order of magnitude of fiscal stimulus should be around 4 percent of GDP, about US$ 600 billion. An intervention of this size may turn out to be too large ex post. But this is not a problem as far as macroeconomic conditions are concerned — Krugman stresses —, because monetary policy can always tighten and correct the mistake. What monetary policy cannot be expected to do effectively under the current circumstances, however, is to compensate for fiscal policy if it is too contractionary: “Fiscal policy should take risks in the direction of boldness.”
While the focus on size is crucial, it is equally important not to lose sight of other necessary conditions for fiscal policy to be most effective — in fact, the current debate emphasizes the requirement that fiscal interventions be ‘timely, targeted and temporary.’ In what follows, we spell two specific points that, in our view, have not received sufficient attention.
When are spending expansions truly temporary?
The effect of a fiscal expansion arguably depends on how the expansion is financed. This applies not only to the short-term debt-tax mix used to finance a current increase in government expenditure, but also—and perhaps even more importantly—to the long-term financing source, i.e., taxes versus spending cuts in the future. The impact of an increase in current expenditure is stronger in the latter case, that is, if complemented with a credible plan ensuring that its cost will be at least partly financed by a reduction in future expenditure. This is so for two reasons. The first relates to an important lesson from both Keynesian and neoclassical models: future spending cuts tend to reduce the long-term interest rate, with positive effects on current consumption and investment decisions. The second reason is that a commitment to lower future spending reduces the overall tax burden implied by the expansion. For both reasons, future spending cuts will help to contain possible negative effects of a fiscal expansion on private expenditure, especially on the spending plans by firms and households.
A commitment to post-crisis expenditure reductions is of particular importance in economies plagued by strained public finances. Indeed, while the cost of a fiscal rescue is highly uncertain, it is likely to raise public liabilities considerably. This may trigger adverse confidence effects that can undermine the expansionary impact of the fiscal intervention. Put differently, it is unlikely that the impact of fiscal expansions will be the same across countries with very different initial budgetary conditions, say, Italy versus France. Empirical work supports this view. Based on a sample of OECD countries, for instance, Corsetti, Meier and Müller (2008b) find that the response of consumption to higher government spending is weaker and sometimes even negative in economies with high debt and deficits, relative to economies with better initial conditions — a point that has motivated the IMF insistence on fiscal consolidation, despite the current downturn, in countries like Hungary, Iceland and Turkey. Yet even in countries where current conditions allow for expansionary fiscal measures, spending interventions must be designed to include a credible commitment to future spending cuts in order to maximize its effectiveness.
Accompanying fiscal expansion with accommodating monetary policy
A second important point is essentially the flip side of the argument by Krugman: fiscal policy is more effective if monetary policy is accommodative. In other words, for fiscal stimulus to work, monetary policy should also be geared towards stimulating growth, consistent with their price stability mandate. A criticism often raised against the Bank of Japan is that in the ‘lost decade’ it was too narrow-minded in pursuing price stability. This risk is small today, as there is widespread awareness about the severity of the crisis and agreement that inflation risks have abated.
Nonetheless, one could envision a situation in which, even with policy interest rates reduced to zero, the overall monetary stance of the economy might still remain too tight. In this situation, the lower bound of zero for nominal interest rates—while providing a rationale for a fiscal expansion—may at the same time limit the effectiveness of fiscal policy.
A look at the mechanism
The argument in favour future spending cuts and accommodative monetary policy can be illustrated through simulation experiments carried out within a standard new-Keynesian model. Specifically, we track the macroeconomic consequences of an unexpected increase in government spending in an economy which is otherwise undisturbed (Corsetti, Meier and Müller 2008a). Of course, government spending can mean a lot of different things (infrastructure investment, public employment…), and governments are actually discussing policy interventions in response to a strong deterioration of overall macroeconomic conditions, rather than an exogenous (“surprise”) policy change. Yet this experiment illustrates how the transmission of fiscal policy depends both on the financing mix and monetary stance. The mechanism is illustrated for the case of a small open economy. The exercise also assumes away credit-constrained agents, whose presence would increase further the consumption multiplier above what is reported. The results from our exercise are shown in the graphs below.
The graphs show the evolution of government consumption, private consumption, output, the government budget balance and debt, the real exchange rate, inflation and interest rates, over 40 quarters in response to an increase in government spending by one percent of (quarterly) GDP. All variables are expressed relative to their trend values (note that a negative value for the nominal interest rate means a fall relative to the initial value). Quantity variables are expressed in percent of quarterly GDP; the real exchange rate is measured in percentage deviations relative to its pre-intervention value; and interest rates and inflation are measured in annualized percentage points. Each graph includes three lines. The dashed black line refers to a spending shock which is entirely financed by taxes (such taxes may be levied today or in the future—this is irrelevant in the exercise given that taxes are non-distortionary and that households and firms are not credit-constrained). The solid blue and the dash-dotted red lines refer to a spending shock which is partly financed by future spending cuts: in the upper left panel this becomes apparent from government spending falling below trend about 3-4 years after the initial measures were taken. The dash-dotted lines refer to the case of no monetary accommodation as the central bank pursues complete price stability; the solid blue line to the case of accommodative monetary policy — in the sense that central banks adopt a Taylor rule with a relatively low coefficient on inflation (1.2 in this example, with prices remaining fixed for 5 quarters on average).
The message from our Figure above is unequivocal: the response of consumption is positive if both monetary policy is accommodative and financing is partially through future spending cuts: consumption falls if either spending is entirely financed through higher taxes (the dashed lines), or when the monetary reaction is non-accommodative (the dash-dotted lines). In fact, comparing the difference in the response of consumption and output across monetary stances (accommodative, non-accommodative), one can observe a gap of about half a percentage point of GDP throughout several quarters.
Monetary accommodation is measured by the difference in the response of real interest rates, depicted by the solid and the dash-dotted lines in the lower right panel (ex ante real rates): under the accommodative stance (solid lines) real rates are lower by about a quarter of a percentage point (annualized) relative to the tight monetary stance (dash-dotted line). Importantly, under the ‘right’ policy mix the path of real short term interest rates implies a fall in the long-term real interest rate, because future short rates fall below their long-term average value.
A new item for international policy coordination
Looking at the results of the exercise above, one could express a concern regarding real exchange rate depreciation, which, in the exercise, accompanies the fiscal expansion cum monetary accommodation. Depreciation may be seen as an unwelcome beggar-thy-neighbour effect of domestic policies — domestic economic activity being sustained by ‘stealing’ foreign demand.
However, exchange rate depreciation is not crucial for the size of fiscal multipliers as our results also hold in a closed economy. More important still, depreciation will be contained, or eliminated altogether, when fiscal expansion is coordinated across borders. In fact, the main conclusions from the analysis apply at a global level, as well: for the world as a whole, fiscal policy is most effective when implemented with the right financing and monetary policy mix.
Coordinated fiscal policy need not mean identical fiscal policy everywhere, however. As stressed by many observers, current-account surplus countries, especially those with strong public finances, should do more than others (e.g. Willem Buiter or Martin Wolf in the Financial Times). Nevertheless our arguments highlight two factors which a coordinated fiscal initiative needs to include in order to be most effective: a commitment to match current expenditure increases at least in part by lower spending in the future; and the need for accommodative monetary conditions.
Corsetti Giancarlo, André Meier and Gernot Müller (2008a) “The transmission of fiscal policy: the role of financing and policy mix”, mimeo European University Institute.
Corsetti Giancarlo, André Meier and Gernot Müller (2008b) “The transmission of fiscal policy in open economy”, mimeo European University Institute. Corsetti Giancarlo, and Gernot Müller (2008) “The effectiveness of fiscal policy depends on the financing and monetary policy mix”, www.voxeu.org, November
Krugman, Paul (2008) Stimulus math (wonkish), The conscience of a liberal, November 10, 2008