Monthly Archives: May 2012

The Sumner Critique and Taxes

Scott Sumner has been promoting the idea that fiscal policy is irrelevant when monetary policy does its jobs. Matt Yglesias has labeled this idea the Sumner Critique, although for those familiar with the macro literature, this conclusion is widespread (see, for example, Bullard’s “Death of a Theory” paper.) However, Yglesias misunderstands the conclusion. In particular, here is his claim:

Conventional wisdom in DC is that not only would the full expiration of the Bush tax cuts make people grumpy as they find themselves needing to pay more taxes, it would also provide the macroeconomy a job-killing dose of fiscal drag. The chart above from Goldman Sachs illustrates the idea clearly.

I don’t buy it.

The problem is that this chart ignores what I think we’re now going to call the Sumner Critique. In other words, it assumes that the Federal Reserve is somehow going to fail to react to any of this. You can probably construct a scenario in which the Fed is indeed caught unawares, or is paralyzed by conflicting signals, or is confused by errors in the data, or any number of other things. But Ben Bernanke knows all about the scheduled expiration of these tax cuts. He’s discussed it several times in public. Maybe he and his colleagues won’t do anything to offset this drag on demand, but if they don’t as best I can tell that’s on them. This is the very essence of a predictable demand shock, and the policymakers ultimately responsible for stabilizing demand are the ones who work at the Fed.

Yglesias is correct in applying the so-called Sumner Critique to demand, but seriously underestimates any impact on supply and thereby ignores the potential downsides from both the supply side effects as well as the potential for over-expansionary monetary policy that results. Allow me to explain.

A basic representation of the permanent income hypothesis for a household that lives T periods and pays income taxes, \tau, would suggest that consumption is determined by:

C_t = {{1}\over{T}}\big(A_0 + \sum_{t=0}^{T} (1 - \tau_t) Y_t\big)

where C is consumption, Y is income, and A is the initial stock of wealth.

It is straightforward to see that an increase in the tax rate leads to a reduction in consumption. Ceteris paribus, this leads to a reduction in “aggregate demand.” However, if one accepts the Sumner critique, Yglesias is correct that monetary policy would be more expansionary than it would have been in the absence of the tax increase thereby correcting the “aggregate demand” shortfall.

Nonetheless, what Yglesias ignores is that these taxes will have supply side effects. (They will have supply side effects, this is not disputable. Critics could argue that there is a question of magnitude.) The reason that they have supply side effects is because these are taxes on income and investment, which affect the labor supply and capital accumulation and thereby long-run economic growth. This is problematic because we know that taxing capital in and of itself is a bad idea. In addition, if this tax increase were to reduce the trend of real GDP growth by a non-significant amount, a monetary policy that gives weight to the the output gap — as in the Taylor Rule, for example — might create an over-expansionary because of an erroneously belief that the output gap was negative.

Monetizing the Debt?

Amar Bhidé writes in the WSJ:

Governments may pay lip service, but in times of stress they face a strong temptation to force central bankers to cover budget deficits through the printing press. Often the bankers don’t wait to be told (witness the Federal Reserve’s recent rounds of quantitative easing).

Suppose that we are uncertain as to the force that is driving the Federal Reserve’s behavior. What probability should we assign to this chain of events? I suspect it is very low. Is there evidence to support this hypothesis other than simple correlations? There are a lot of reasons to criticize the Fed, I’m not sure this is one.

Evans on Nominal GDP Targeting

Charles Evans has come out in favor of nominal GDP level targeting.