Keynesian Versus New Keynesian Multipliers

A new paper from Cogan, Cwik, Taylor and Wieland suggests that the estimated multipliers are inflated:

Renewed interest in fiscal policy has increased the use of quantitative models to evaluate policy. Because of modelling uncertainty, it is essential that policy evaluations be robust to alternative assumptions. We find that models currently being used in practice to evaluate fiscal policy stimulus proposals are not robust. Government spending multipliers in an alternative empirically-estimated and widely-cited new Keynesian model are much smaller than in these old Keynesian models; the estimated stimulus is extremely small with GDP and employment effects only one-sixth as large and with private sector employment impacts likely to be even smaller.

Here is the non-gated link.

Estimating the Multiplier

As many of you know, I have expressed skepticism that the stimulus package will work. At the heart of the debate is whether the multiplier associated with government spending is greater than one. Some recent literature should allow us to do some back-of-the-envelope calculations of the multiplier.

A recent paper by Matthew Shapiro and Joel Slemrod (non-gated link) uses survey data to determine the effects of the 2008 stimulus package in which individuals received tax rebates. They find that 20% of those interviewed suggested that they would “mostly spend” the rebate. 32% would mostly save the check and the remaining 48% would mostly pay down their debt. At first glance, this would seem to suggest that the fraction of the population likely to increase their consumption is 20%. The authors, however, note that there is a difference between “save” and “mostly save”. Thus, they estimate the marginal propensity to consume (MPC) to be roughly .3.

In Mankiw-Campbell land, this roughly translates to the idea that the fraction of consumers that are rule-of-thumb consumers (those that live paycheck-to-paycheck) is about .3. We can then use this measure of the number of rule-of-thumb consumers to estimate the multiplier based on a recent study by Jordi Gali, et. al (HT: Ambrosini).

The paper by Gali, et. al provides two ways to estimate the multiplier based on the fraction of rule-of-thumb consumers. First, with competitive labor markets, the fraction of rule-of-thumb consumers would have to be roughly .65 to generate a multiplier of 1. Clearly, this is well above the estimate provided above. Second, with non-competitive labor markets, the fraction of rule-of-thumb consumers would have to be .25 to generate a multiplier of 1. Further, a rule-of-thumb fraction of .3 implies a multiplier greater than 1, but less than the administration’s estimate of 1.5.

Naturally, this raises the question as to whether the labor market in the United States is competitive. Gali et. al define a non-competitive labor market as that in which unions set the wage and employment is determined by the labor demand curve. The idea that the wage is determined by labor demand is certainly a reasonable assumption, especially given the current circumstances (Thanks Will). Thus, based on the work presented, the multiplier may be slightly above 1.

(A BRIEF NOTE: The paper by Shapiro and Slemrod examines the impact of a temporary government stimulus. The current stimulus package contains both temporary and permanent aspects. Nevertheless, I think that their estimates provide a meaningful guide. The current stimulus provides permanent tax cuts through the “making work pay” rebate. It also includes temporary spending provisions such as infrastructure spending etc. In addition, some of the spending side is likely to be permanent. Thus, the increase in consumption is likely to be greater than that estimated by Shapiro and Slemrod because the tax cut is permanent. However, this effect might be offset by the permanent increases in government spending in the package, which will crowd out private investment. Also, the package has been sold as a temporary stimulus to the American people and thus treating the entire thing as temporary when evaluating the behavior of consumers is likely a good first-approximation. Finally, these are merely back-of-the-envelope calculations and should be taken with a grain of salt.)

The Stimulus Will Fail

Recently, I have been receiving a great deal of email that has resulted in responses in which I find myself either defending the stimulus package or defending the proponents of the stimulus package. These defenses, however, have much more to do with the flawed reasoning of fellow stimulus skeptics than with any sort of favorable outlook toward the stimulus package itself. With this in mind, it is perhaps time to clarify my position.

As I have previously noted, there are two questions that seem to be at the center of the debate:

1. Can a stimulus package be designed to give a boost to the economy in the short-run?

2. Can the stimulus package get us out of the recession?

Under the current circumstances, with unemployment approaching 8%, the answer to (1) is undoubtedly “yes.” HOWEVER, this does not imply that (2) can similarly be answered in the affirmative as many stimulus advocates would have you believe. Unfortunately, the debate regarding stimulus has diverged to the point in which those on each side are answering different questions and ultimately giving the impression that there is widespread disagreement as to the effectiveness of fiscal stimulus. Thus, this post is my attempt to explain why the answer to (1) is “yes” and the answer to (2) is “no.”

The traditional Keynesian multiplier analysis suggests that if the government spends money on a particular project it not only increases output by the amount of the increase in government spending, but also by the increase in consumption of those employed to work on the project. Given the idea of the multiplier, government spending can clearly be seen as providing an affirmative answer to both (1) and (2) since $1 of government spending producers more than $1 of additional output.

Unfortunately, this analysis is entirely too simplistic. If fails to take into account the fact that the government must compete for resources (and credit) with private enterprise. The traditional Keynesian response to the criticism is that during times of less than full employment, there are idle resources (e.g. unemployed workers) that can be put to use without competing with private firms. The failure, however, is that to rule out any sort of crowding out effect, all resources utilized by the government must be idle. This is not to say that government spending cannot promote a temporary boost to the economy. Even if there is crowding out, so long as the multiplier is greater than zero, output will increase in response to an increase in government spending. The question as to the value of the multiplier is thus an empirical question.

John Taylor recently presented a paper (non-gated link here) at the AEA meetings on the revival of fiscal policy. In his introduction, Taylor writes:

A decade ago in a paper, “Reassessing Discretionary Fiscal Policy,” published in the Journal of Economic Perspectives, I concluded that “in the current context of the U.S. economy, it seems best to let fiscal policy have its main countercyclical impact through the automatic stabilizers….It would be appropriate in the current circumstances for discretionary fiscal policy to be saved explicitly for longer term issues, requiring less frequent changes.” This was not an unusual conclusion at the time. As Martin Eichenbaum (1997) put it, “there is now widespread agreement that countercyclical discretionary fiscal policy is neither desirable nor politically feasible,” or, according to Martin Feldstein (2002), “There is now widespread agreement in the economics profession that deliberate ‘countercyclical’ discretionary policy has not contributed to economic stability and may have actually been destabilizing in the past.”

Taylor also discusses some of his own empirical work from his 1992 book Macroeconomic Policy in a World Economy on the actual value of the multiplier:

. . . simulations of my (1992) empirically estimated multi-country dynamic model with rational expectations indicates that multiyear changes in government spending phased in at realistic rates have a maximum government spending multiplier less than one because of offsetting reductions in the other components of GDP.

Further, he provides empirical evidence of the utter failure of the tax rebate checks issued in 2008.

Taylor also echoes the point made by Casey Mulligan that there are some types of government spending that can be justified by their intrinsic value, but are unlikely to cause economic stimulus:

To be sure, it may be appropriate to increase government purchases in some areas including for infrastructure as in the 1950s when the interstate highway system was built. But such multiyear programs did not help end, mitigate, or prevent the recessions of the 1950s. In sum, there is little reliable empirical evidence that government spending is a way to end a recession or accelerate a recovery that rationalizes a revival of discretionary countercyclical fiscal policy.

Another issue at hand here is the idea of Ricardian equivalence, which essentially posits the claim that there is no fundamental difference between spending financed by deficits or taxation. Suppose that the government increases spending. It follows that they must either tax current economic agents or issue government bonds to pay for the increased spending. The intuition behind Ricardian equivalence is that when the government increases spending individuals recognize that, even if the spending is financed through government borrowing, the spending creates an increase in their future tax liability.

Consider the incredibly over-simplified example. Suppose the government decides to issue a check for $500 for each taxpayer. They pay for these checks by issuing bonds that will be paid off next period. Under Ricardian equivalence, the individuals recognize that while they will receive $500 this period, but will have their taxes increased to pay for the bonds next period. Thus individuals would simply buy the bonds (save the $500) in order to meet their increased tax liability in the next period.

Critics of Ricardian equivalence have long argued that this is incorrect to the point of being trivial. For example, if individuals have finite lives and the government lives forever, this would seem to refute the idea of Ricardian equivalence — unless, of course, the finite-lived agents view their offspring as mere extensions of their lifetime. This exception, of course, leads to criticism based on the idea of myopic economic agents or the failure of operative bequests to future generations. The list of criticisms far exceed the very small sample above, but they do seem to suggest that Ricardian equivalence fails in simple theoretical applications. Again, however, we must turn to the empirical investigations.

In regards to the validity of Ricardian equivalence, John Seater at N.C. State wrote an excellent survey that appeared in the Journal of Economic Perspectives in 1993 (non-gated link here). Regarding the theoretical issues above, Seater explains:

Finite horizons, nonaltruistic or inoperative bequest motives, childless couples, liquidity constraints, and uncertainty all can lead to failure of Ricardian equivalence, and it seems virtually certain that some of these sources of nonequivalence are operative. It appears likely that the world is not Ricardian. (p. 155 – 156)

Upon examination of the empirical evidence, however, Seater concludes:

Nevertheless, equivalence appears to be a good approximation. Although some of the early empirical literature sent conflicting signals, recent work generally supports Ricardian equivalence. It is true that existing data cannot distinguish the Ricardian model based on altruism, from one of approximate equivalence, based on pure selfishness, but there seems little practical significance to that fact.

[...]

Empirical success and analytical simplicity make Ricardian equivalence an attractive model of government debt’s effects on the economic activity. (p. 184)

Thus, while Ricardian equivalence is seems literally untrue, the idea “holds as a close approximation” (Seater, p. 143) based on empirical evidence.

In short, the fact that Ricardian equivalence serves as a close approximation to actual results and the widespread failure of discretionary fiscal stimulus described by Taylor above, suggest that the stimulus package will be a failure. Further, given the pork-laden nature of the stimulus package, it is unlikely that the spending can even be justified on its intrinsic value, let alone on stimulation.

The stimulus will fail.

An Interview With Robert Barro

Some of you may recall that Paul Krugman referred to Robert Barro’s analysis of the multiplier associated with World War II spending as “boneheaded.” Admittedly, I did agree with Krugman to the extent that the period in question is not ideal for such measurement given the variety of other simultaneous changes (e.g. price controls, rationing, etc.). Nonetheless, I did acknowledge that “Robert Barro essentially wrote the book on government from a macro perspective.” Further, Barro’s analysis was not “boneheaded”, but merely less than ideal.

Now over at The Atlantic, there is an excellent interview with Robert Barro where he discusses the stimulus package, his analysis, and Krugman. First, in response to Krugman:

He said elsewhere that it was good and that it was what got us out of the depression. He just says whatever is convenient for his political argument. He doesn’t behave like an economist. And the guy has never done any work in Keynesian macroeconomics, which I actually did. He has never even done any work on that. His work is in trade stuff. He did excellent work, but it has nothing to do with what he’s writing about.

On stimulus:

This is probably the worst bill that has been put forward since the 1930s. I don’t know what to say. I mean it’s wasting a tremendous amount of money. It has some simplistic theory that I don’t think will work, so I don’t think the expenditure stuff is going to have the intended effect. I don’t think it will expand the economy. And the tax cutting isn’t really geared toward incentives. It’s not really geared to lowering tax rates; it’s more along the lines of throwing money at people. On both sides I think it’s garbage. So in terms of balance between the two it doesn’t really matter that much.

Read the whole thing.