A Failure to Communicate, Not a Failure to Stimulate

It’s a little difficult to reply to Prof. Mishra’s latest op-ed because it doesn’t really have a point. It goes all over the place. As far as I can tell, the only actual argument he makes against President Obama’s American Jobs Act is:

…the first stimulus bill in 2008, a $700 billion package geared toward government spending to stimulate the economy, and financed with borrowed money, has obviously failed to create new jobs.

He never offers any evidence to support this claim.

I’ve disproven this hypothesis before, but I’ll do so again — first by repeating what I said last time, then with even more evidence. If you’ve already read the first part, you might want to skip to the new stuff, though it can’t hurt to refresh your memory…  

President Obama signed the American Recovery and Reinvestment Act (ARRA) on February 17, 2009. Almost immediately, GDP growth turned around:

Private sector job growth also turned the corner immediately following the ARRA:

A few weeks after Congress passed the ARRA, the stock market unexpectedly began a strong, steady rise:

It’s true that GDP and job growth has been slower during this recovery than previous post-WWII recoveries, but recoveries following financial crises are almost always weaker than recoveries that follow traditional recessions. Compared to other debt-burdened recoveries, this recovery has been better than usual.

Another reason for the slow GDP and job growth is the lack of stimulus. No, that’s not a typo. The ARRA may have injected $787 billion into the economy, but state and local governmentscut their budgets by roughly the same amount. Overall, government spending hardly budged from its long-term trend:

A lot of economists have tried to measure the historical “government multiplier,” or “bang for your buck,” but only a few have studied periods that are comparable to today: deep recessions where interest rates can’t go any lower. Such responsible studies estimate that the multiplier ranges from 1.5 to 2, which means that $1 in government spending increases GDP by $1.50 to $2.

Stimulus critics predicted that interest rates would rise because the government borrowed too much, competing with private investors for scarce lending. Instead, the opposite happened:

In fact, interest rates are the lowest they’ve been in many decades:

Some stimulus critics also predicted that inflation would soar because the government would print money to pay for the stimulus. Actually, inflation is lower than it was before the recession, lower than it was in the 1980s and most of the 1990s, and out of dangerous deflation territory, thanks to the ARRA:

But if you want to measure the effects of the stimulus, says conservative economist Robert Barro,  you need “‘experiments,’ in which the government changes transfer in an unusual way — while other factors stay the same.” Fortunately, we have four such “econometric” research papers.* (We won’t discuss papers that use “models” because, as Barro says, they tend to assume what they’re trying to prove.) Here are their findings:

  1. James Feyrer and Bruce Sacerdote, both from Dartmouth, compared jobs growth in each state and county to the amount of stimulus funds spent in that state or county. They used the seniority of each state’s/county’s legislators to determine whether they received the funding because of economic need or political patronage. That allowed them to correct for any possible reverse causation, whereby the jobs growth determined how much stimulus funds were received (rather than the other way around, which is what we’re trying to measure). They found a multiplier of 1.96 to 2.31 for low-income spending, 1.85 for infrastructure spending, and 0.47 to 1.06 for the stimulus overall.
  2. Gabriel Chodorow-Reich (Berkeley), Laura Feiveson (MIT), Zachary Liscow (Berkeley), and William Gui Woolston (Stanford) compared jobs growth in each state to the amount of federal Medicaid matching funds spent in that state. They only looked at aid that states received based on pre-recession Medicaid spending, not the aid distributed based on economic need during the recession, to avoid the reverse causation problem mentioned earlier. They found a multiplier of 2.
  3. Daniel J. Wilson, of the Federal Reserve Bank of San Francisco, compared jobs growth in each state to the amount of stimulus funds spent in that state. He used pre-recession Medicaid spending, school-age population (a proxy for education need), and highway act (unrelated to unemployment) to determine whether they received the funding because they were hit harder during the recession or because of these non-cyclical needs. That allowed him to correct for any possible reverse causation. He found that the ARRA “created or saved about 2 million jobs in its first year and over 3 million by March 2011.”
  4. John B. Taylor, of Stanford, measured the change in consumption at the time when the tax cuts boosted personal disposable income the most. He found no significant increase in consumption due to the tax cuts. He also measured the change in state and local government spending at the time when federal aid gave them the biggest boost. He found that the increase in federal spending was too small to significantly outweigh the decrease in state and local spending.

There you have it: All the evidence points to a stimulus that worked exactly as it was supposed to. If anything, it was too small and had too many tax cuts and not enough spending.

On the tax cuts, it’s worth mentioning, however, that this is not a regular recession. It’s a “balance-sheet recession,” where consumers won’t start spending again until they pay off the debt hanging over them. The tax cuts may not boost consumption immediately, but they help consumers pay off debt and thus speed up the recovery. Of course, this only works if the tax cuts are directed to low- and middle-income Americans with debt to pay off, not to the rich. This is why President Obama is advocating payroll tax cuts, not tax cuts for the rich that Prof. Mishra and Congressional Republicans are so fond of.

==========

* A hat tip to Ezra Klein for compiling these papers (and several more) in one handy blog post.