The explosion of government budget deficits and debt levels since 2008 has created a lively debate about the effectiveness of fiscal policies in stimulating the economy. The issue is important. It determines whether we are in favour of continuing expansionary fiscal policies or instead believe that it should be ended as soon as possible.
It will not surprise the outsiders that economists do not agree on this issue. This is dramatically illustrated by Figure 1, reproduced from a recent paper surveying the fiscal policy multipliers obtained from different macroeconomic models of the US economy (see Cogan et al. 2009). It shows the effects of a 1% permanent increase in US government spending on US GDP. These effects (the multipliers) are obtained using two different models of the US economy.
After one period, in the Romer/Bernstein model the increase in government spending has a strong multiplier effect. In addition, this increase in GDP is permanent. In the Smets/Wouters model the “multiplier” declines to 0.4 after 4 years, and tends to go to zero.
Figure 1. Impact of permanent increase in US government spending of 1% of US GDP
Source: Cogan, et al. (2009)
In order to better understand the fundamental difference between the two models I present the same results relative to the baseline path of the US GDP (i.e. without the fiscal stimulus, see Figure 2). The baseline GDP path is one that assumes no fiscal stimulus used here as a benchmark.
Figure 2. GDP path – three scenarios
Source: Calculated using data from Cogan, et al.(2009)
Fundamentally different models
In the Romer/Bernstein model an expansionary fiscal policy succeeds in moving the economy permanently onto a higher path of GDP. In the Smets/Wouters model despite the fiscal policy stimulus the economy tends to move back to the pre-policy output path. Thus the two models embody a fundamentally different view about how the economy functions.
The first model has a structure with different equilibria, allowing the government to guide the economy to another path. In the second model, there is only one equilibrium to which the economy will return after a fiscal policy shock. Needless to say that these two views will lead to different prescriptions about how quickly governments should retrench from fiscal stimulus policies.
Three types of macroeconomic model
Figure 1 allows us to categorise three types of macroeconomic model with different predictions for the effectiveness of fiscal policies.
The first type of model is the Keynesian one, which is well represented by the Romer/Berstein model. This model exhibits the typical Keynesian result whereby a fiscal stimulus permanently raises the level of output. It is also a model in which many equilibria are possible, some of which imply less than full employment.
The second type of model is the real business cycle model (Ricardian model). This is a model that assumes Ricardian equivalence. In this model rational and forward looking agents take into account that the increased budget deficit today will lead to more taxation in the future. Thus these agents compute the present value of these future taxes and set that amount aside so as to be able to fund these tax commitments. As a result, the fiscal stimulus is exactly offset by additional savings by private agents so that the net effect is zero. The fiscal policy multiplier is zero, as in the baseline scenario without fiscal stimulus.
The third model is well represented by the Smets/Wouters GDP path. It embodies the New Keynesian result that only allows for a temporary output effect of a fiscal stimulus. At impact, the fiscal stimulus has a very similar effect on output as the Keynesian model. However, this effect tends to disappear and output converges towards the baseline path, which is also the path obtained in the Ricardian model, because rational and forward looking agents do the same Ricardian computation of future tax liabilities. This leads them to reduce private consumption and investment. This process takes more time in the New Keynesian model than in the real business cycle model because of rigidities in wages and prices. Most of today’s “state of the art” macroeconomic models are of this New Keynesian variety and share the same characteristics as the Smets/Wouters model (Cwik and Wieland 2009).
From the previous analysis we can conclude that the difference between the Keynesian and the New Keynesian model is more fundamental than the difference between the Real Business Cycle (Ricardian) model and the New Keynesian model.
In the Keynesian model there is no automatic return to the long run output equilibrium. As a result, policy can have a permanent effect on output. The New Keynesian model, like the Ricardian model, contains a very different view of the economy. In this model fiscal policy shocks lead to adjustments in interest rate, prices and wages that tend to crowd out private investment and consumption. As a result, output is brought back to its initial level. In the Ricardian model this occurs very rapidly; in the New Keynesian models this adjustment takes time because of rigidities in wages and prices. But fundamentally, the structure of these two models is the same.
Which model is the most useful?
I argue that in “normal recessions” the New Keynesian model is probably the right way to look at the world. In “abnormal recessions” it is the Keynesian model. Equilibrium models are useful to understand “normal recessions”. These are recessions that maintain an equilibrating mechanism, e.g. a change in interest rate or prices that tend to bring the economy back to its potential output level. These models allow us to understand how much fiscal policy can help us in guiding the economy back to equilibrium. For example, suppose that the economy is subject to a negative demand shock. New-Keynesian models tell us that ultimately the economy will find its long term output level, but that this will take some time. A fiscal policy stimulus can speed up this process, and thus knowledge of the multiplier is important for the policymaker. The evidence from these models is that the impact multiplier may be around one, but that it quickly declines. Thus these models teach us that in normal recessions, fiscal policy can be useful but that caution should be exerted to avoid overshooting. These models also teach us that the “exit strategy” should be organised rather quickly.
Is the present downturn a normal recession?
Probably not. In order to show this I distinguish between three deflationary spirals that have played a role in the present recession. These are:
- Keynesian savings paradox: Individuals save as a result of a collective lack of confidence, leading to a self-fulfilling fall in output.
- Fisher’s debt deflation: Individuals try to reduce their debt driven by a collective movement of distrust. They all sell assets at the same time, thereby reducing the value of these assets. This leads to a deterioration of the solvency of everybody else and self-defeating assets sales.
- Bank credit deflation: Banks are gripped by extreme risk aversion and simultaneously reduce lending, increasing the riskiness of their loan portfolios.
The three deflationary spirals have the same structure. The actions by economic agents create a negative externality that makes these actions self-defeating. This spiral is triggered by a collective movement of fear, distrust or risk aversion. Individuals (savers, firms, banks) are unable to internalise these externalities because collective action is costly. There is thus a failure to coordinate individual actions to avoid a bad outcome (see Cooper and John(1988)).
These market failures are triggered by waves of correlated beliefs (“animal spirits” see Akerlof and Shiller 2009). When these beliefs are not correlated the market will work fine in coordinating the different beliefs of individuals. But when animal spirits occur, the market will fail to coordinate individuals’ actions towards a “good equilibrium” (see also Farmer and Guo 1994).
The three deflationary spirals, although similar in structure, are different in one particular dimension. The savings deflation spiral can be called a “flow deflation”. It arises because consumers want to change a flow (savings). The Fisher debt deflation and the bank credit deflation involve the adjustment of stocks (the debt levels and the levels of credit). We call them “stock deflations”. Problems arise when the flow and stock deflations interact with each other.
In “normal” recessions such as the ones we have experienced in the postwar period prior to the present crisis, only the flow deflations were in operation. As a result, households, firms and banks were not trying to adjust their balance sheets. The pessimism of households and firms was related to expected shortfalls in income and profits, and led to increased savings. In such an environment, there were sufficient automatic equilibrating mechanisms that prevented the flow deflation from leading to an unstoppable downward spiral. The most important equilibrating mechanism occurred through the banking system.
The problem the world economy has been facing since 2007 is that flow and stock deflations have been interacting and reinforcing each other. The period prior to the crisis was one of excessive build-ups of private debt that activated the stock deflation processes with full force. The equilibrating mechanism that exists in normal recessions did not function as the lower interest rates engineered by central banks were not transmitted to consumers and firms by the banking sector.
We are now confronted with the interaction of flow and stock deflations. As a result of the excessive debt accumulation of the past, households want to reduce their debt levels and attempt to save more. Since these attempts are self-defeating, households fail to save more and fail to reduce their debt. This leads them to try to save more. The fact that the banks do not pass on the lower deposit rates into lower loan rates makes things worse. There are no incentives for firms to increase their investments and nothing stops the deflationary spiral (see also Minsky 1986 and Fazzari, et al. 2008).
The collective action solution and the multiplier
The intensity of the present economic downturn is the result of a coordination failure as the market fails to coordinate private actions towards an attractive collective outcome.
This market failure can in principle be solved by collective action organised by the government. The key to economic recovery is the stabilisation of the banking sector. This appears to have been achieved by now, although the evidence suggests that we are still in a liquidity trap. As for the collective action failure implicit in the Keynesian savings paradox and the Fisher debt deflation mechanism, governments solved it by dissaving (higher budget deficits) and debt accumulation by the public sector. Without this, the private sector would not have been able to increase its saving and to reduce its debt. As a result, the different interacting spirals have been stopped.
If we accept the previous analysis as representing the underlying dynamics of the recession that started in 2007, the estimates of fiscal policy multipliers obtained from models assuming stable equilibria are pretty much useless (see for example Wieland 2009, Cogan et al. 2009, Fatás and Mihov 2009, Hassett 2009). By allowing government deficits and debts to increase, governments solved a private sector coordination problem and made it possible for private agents to realise their desire of saving more and deleveraging without making the economy unstable. The “multiplier” effect of these government actions is potentially very large, but also difficult to estimate.
The issue that arises and that has been gripping financial markets is whether the substitution of private debt by government debt will lead to unsustainable government debt levels. If that is so, exit strategies should be started as soon as possible. This issue, however, cannot be seen separately from the sustainability of the private debt.
If we believe that private debt is still too high and that the private sector must continue to deleverage, then it is not clear exit strategies should be started as soon as possible. An early exit strategy would expose the excessive and unsustainable private debt levels, and would set in motion a new deflationary process.
The issue then boils down to the question of whether private sector debt is now sustainable so that the deflationary spirals will not be set in motion when governments start to reduce their deficit and debt levels. Unfortunately, it is difficult to answer this question today because it is tricky to decide when a debt level is sustainable. As the Fisher debt deflation paradox has highlighted, the sustainability of the debt level of one agent depends on the actions of others. These externalities explain why it will remain difficult to identify sustainable debt levels.
Akerlof, George, and Robert Shiller (2009), Animal Spirits: How Human Psychology Drives the Economy and Why It Matters for Global Capitalism, Princeton University Press, 264.
Cogan, John, Tobias Cwik, John B Taylor and Volker Wieland (2009), “New Keynesian versus Old Keynesian Government Spending Multipliers”, CEPR Discussion Paper 7236, March.
Cwik, Tobias, and Volker Wieland (2009), “Keynesian Government Spending Multipliers and Spillovers in the Euro Area”, CEPR Discussion Paper 7389.
Cooper, Russell W and John, Andrew (1988), “Coordinating coordination failures in Keynesian models”, Quarterly Journal of Economics, 103:441-463
Farmer, Roger and Jang-Ting Guo (1994), “Real Business Cycles and the Animal Spirits Hypothesis”, Journal of Economic Theory, 63, 42-73.
Fatás, Antonio and Illian Mihov (2009), “Why Fiscal Policy is Likely to Work”, International Finance 12:1, Spring.
Fazzari, Stevan, Pierro Ferri and Edward Greenberg (2008), “Cash flow, investment, and Keynes–Minsky cycles”, Journal of Economic Behavior and Organization, 65:555–572.
Fisher, Irvine (1933), “The Debt-Deflation Theory of Great Depressions”, Econometrica, 1:337-57, October.
Leijonhuvud, Axel (1973), “Effective demand failures”, Swedish Journal of Economics, 75:27-48.
Minsky, Hyman (1986), Stabilizing an Unstable Economy, McGraw Hill, 395pp.
Reinhart, Carmen and Kenneth Rogoff (2009), “The Aftermath of Financial Crises”, NBER Working Paper 14656.
Smets, Frank and Raf Wouters (2007), “Shocks and Frictions in US Business Cycles: A Bayesian DSGE Approach”, American Economic Review 97, 3: 506-606.
Wieland, Volker (2009), “The fiscal stimulus debate: “Bone-headed” and “Neanderthal”?”, VoxEU.org, 31 March