Will the Stimulus Work? Reasons to Doubt the Romer-Bernstein Spending Multiplier Estimates

US economic advisers called for aggressive fiscal stimulus, and some support further measures. But many macroeconomists are not so sure. This column analyses fiscal stimulus using a New Keynesian model that exemplifies contemporary academic thinking on the subject. It says that the spending multiplier is much lower than the Obama administration’s estimates – government spending may quickly crowd out private consumption and investment.

Not long ago, Paul Krugman warned European governments that:

“We’re rapidly heading toward a world in which monetary policy has little or no traction, … Fiscal policy is all that’s left… if Germany prevents an effective European response, this adds significantly to the severity of the global downturn. … in short, there’s a huge multiplier effect at work; unfortunately, what it’s doing is multiplying the impact of the current German government’s boneheadedness.”

Financial Times columnist Martin Wolf, taking a look at the US and Japan (17 Feb 2009), asserted similarly:

“The bad news is that the debate over fiscal policy in the US seems even more Neanderthal than in Japan: it cannot be stressed too strongly that in a balance-sheet deflation, with zero official interest rates, fiscal policy is all we have.”

This urgent, almost desperate call for aggressive fiscal stimulus was reinforced by the economic analysis of President Obama’s advisers Christina Romer and Jared Bernstein, which underscored the power of discretionary fiscal policy. In a paper circulated in January 2009, Romer and Bernstein provide numerical estimates of the impact of an increase in government spending on US GDP and employment. Such estimates are a crucial input for the policymaking process. They help determine the appropriate size and timing of countercyclical fiscal policy packages, and they inform parliaments and their constituents about whether a vote for a policy is appropriate.

Romer and Bernstein make use of two macroeconomic models – one from the staff of the Federal Reserve Board and the other from an unnamed private forecasting firm. Averaging the impacts obtained from these two models, they estimate that increasing government spending permanently by an amount equal to 1% of GDP would induce an increase in GDP of 1.6% above what it would have been otherwise. They conclude that a package similar in size to the American Recovery and Re-investment Act passed in February 2009 would raise GDP by 3% to 4% and create 3 to 4 million additional jobs by the end of 2010.

Stimulus doubts

Nevertheless, many macroeconomists still admit to substantial uncertainty about the quantitative effects of fiscal policy. This uncertainty derives not only from the empirical estimates of model parameters and shocks but also from different views on the appropriate theoretical framework and empirical method. In light of such model uncertainty, it is crucial to evaluate the robustness of particular policy proposals in different models with different assumptions. Cogan, Cwik, Taylor, and Wieland (2009) conduct such a robustness analysis with New Keynesian macroeconomic models. Nowadays such models are used by many central banks and international institutions. We report findings from two models, Taylor (1993) and Smets and Wouters (2007), but focus more on the latter model, which has been described as representative of the current New Keynesian macroeconomic thinking (see Woodford 2009).

Unfortunately, we find substantially smaller government spending multipliers than those used by Romer and Bernstein. For example, the multiplier associated with a permanent increase in government spending by the end of 2010 lies between 0.5 and 0.6. In other words, government spending does not induce additional private spending but instead quickly crowds out private consumption and investment.

We also provide an assessment of the impact of the American Recovery and Re-investment Act. This legislation implies measures amounting to $787 billion and spread over 2009 to 2013 but peaking in 2010. Our estimate of the total impact is closer to 1/6 of the effect estimated by Romer and Bernstein. By 2010 we project output to be about 0.65% higher. Using the same rule-of-thumb as Romer and Bernstein, this increase in GDP would translate to about 600,000 additional jobs rather than three to four million.

Why is our assessment of government spending multipliers so different? Well, first of all Romer and Bernstein constrain the Fed to keep interests rates constant at zero forever.

Such an interest rate peg would lead to explosive behaviour and instability in New Keynesian models. Instead we allow the Fed to raise rates eventually, starting in 2011, or more realistically, in 2010.

Committing to 1 or 2 years of zero interest rates still implies much additional monetary stimulus.

Furthermore, people out there worry about the future. Thus, the models we use take into account that forward-looking households and firms will modify their expectations and change their behaviour in response to the new fiscal policy measures.

Finally, at least some people out there realise that higher government spending and debt today ultimately require raising more taxes in the future. Such households will consume less today. This negative wealth effect is particularly strong in the Smets and Wouters analysis. The model by Taylor implicitly allows for the presence of some consumers who consume all of their current income.

Fiscal policy focus

In light of these findings, European policy makers are well-advised to question the usefulness of further stimulus packages. They ought to carefully monitor the impact of decisions already taken on the burden imposed on future taxpayers. The available funds and remaining borrowing capacity should be utilised where it is still most needed – to prevent a collapse of the financial system and finance the necessary re-capitalisations and toxic asset removals. If governments exhaust their fiscal space in measures that have little aggregate effect, they will instead stimulate scepticism of their capability to back up the financial system.

Thus, it remains crucial to focus fiscal efforts on the financial front.

What else can be done? Monetary policy is still an option. Sure, nominal interest rates cannot decline below zero. This is a serious constraint on conventional interest rate policy. However, monetary expansion remains feasible, and increasing the relative supply of base money to other assets will lower its value. In other words, the central bank can stimulate inflation and reduce real interest rates by means of quantitative easing if necessary (see Orphanides and Wieland 2000).

References

Cogan, John, Tobias, Cwik, John B. Taylor and Volker Wieland (2009), “New Keynesian versus Old Keynesian Government Spending Multipliers”, CEPR Discussion Paper 7236, March 2009.

Orphanides and Wieland, (2000), Efficient Monetary Policy Design Near Price Stability, Journal of the Japanese and International Economies, 14, 327-365.

Romer, Christina and Jared Bernstein (2009), “The Job Impact of the American Recovery and Reinvestment Plan”, January 8, 2009.

Smets, Frank and Raf Wouters (2007), “Shocks and Frictions in US Business Cycles: A Bayesian DSGE Approach”, American Economic Review 97, 3: 506-606.

Taylor, John B. (1993), Macroeconomic Policy in a World Economy: From Econometric Design to Practical Operation, WW Norton, New York.

Woodford, Michael (2009), “Convergence in Macroeconomics: Elements of the New Synthesis”, American Economic Journal: Macroeconomics, Vol. 1, No. 1, 267-279.

• Become A Funded Futures Trader