The Multiplier

The entire debate about the economic stimulus packages debated by Congress during this recession ought to come down to one key number: the Multiplier. This particular multiplier (there are others) represents an ad hoc estimate of the effectiveness of government spending. What we really want to know is whether the multiplier is significantly larger than 1, significantly less than 1, or approximately equal to 1.

A note of caution before proceeding: the Multiplier has no basis in economic theory. It can be backed out of a number of models, but it would almost never be constant. Theoretically based macroeconomic models - both Keynesian and Neoclassical - will deliver a multiplier less than 1 unless they are specifically built to generate large multipliers. The models used by government economists are ad hoc models, which treat macroeconomic movements like Newtonian physics, rather than more complex general equilibrium models which are currently used for most research.

The Multiplier represents the sum of billions of decisions and interactions that result from government fiscal policy. When the government spends money, it borrows or taxes from some people and gives the money to others, either freely or in exchange for goods and services. This may increase total economic activity (a multiplier greater than 1), decrease it (a multiplier less than 1), or leave it essentially unchanged. Each person and firm is going to react differently to the government's borrowing-and-spending. One firm might find loans for investment unavailable because the government has sucked up all the available funds; another firm might be able to keep its factories open because the government places a huge order for widgets.

We cannot count or measure those billions of decisions, but we can estimate, roughly, the size of the Multiplier. Team Obama is acting under the assumption the assumption that the Multiplier on spending is 1.57 and on taxes 0.99 (source).

This is not without support: Valerie Ramey estimates a multiplier of 1.4 in a recent paper (source).

But Hall and Woodward use wartime data to estimate a multiplier on spending near 1, and the Romers find a tax-cut multiplier of 3! In the most exhaustive study, Barro estimates that the multiplier on spending is 0.8, and near zero for non-wartime spending.

Thus, the Obama team is overshooting the most sanguine estimate of the multiplier on spending, and unaccountably ignoring its own CEA Chair's high estimate of the tax-cut multiplier.

If they are correct, then the spending is a good thing for most Americans. However, the plan still harms some groups: most significantly the young, who will pay the interest on this debt for the rest of our lives. There is no economic theory that suggests government stimulation is helpful for long-run growth - usually, the opposite is true - so we, the young, will be earning less during our lifetimes, and paying higher taxes.

If Team Obama is incorrect, and the multiplier is about 1, then it is spinning its wheels and redistributing income away from the young and toward the old and favored. Perhaps that is good politics, since the young don't vote, and won't see these tax increases until Obama is in the history books.

However, if the administration is drastically wrong, and the multiplier is 0.8, or 0.5, or even lower, as Barro's research suggests, then the government is heavily complicit in worsening the recession. A multiplier as high as 0.8 would still mean that of $1 trillion spent, $200 billion disappears. That would represent a loss of 1.5% of GDP, equal to the entire aggregate loss of GDP from the fourth quarter of 2008. The cure is worse than the disease!

We can - and should - argue about the size of the multiplier. We can cite Nobel prize-winning economists on both sides. But until we have real intelligence about the size of the multiplier and the severity of the punishment on future taxpayers, isn't it irresponsible to be pushing a huge spending package? Remember the last time a president pushed a major spending and policy action on cherry-picked intelligence? How well did that work out?