The Unseen Effects of the Stimulus

Did the American Recovery and Reinvestment Act destroy jobs or create them?


According to an old joke, if you ask any two economists a question you will get back at least three opinions. As humor goes it's only mildly funny. But as a gloss on the discussion of the stimulus, it's terrifically on point.

Conservatives have slammed the American Recovery and Reinvestment Act of 2009 (ARRA) as a grotesque waste of money, claiming it has not created any jobs. Liberals counter that, to the contrary, it has created or saved millions. For instance, the White House Council of Economic Advisers claims the stimulus deserves credit for perhaps as many as 3.6 million jobs created or saved. Writing Tuesday in The Wall Street Journal, economist Alan Blinder says, regarding the claim that federal spending of more than $600 billion had zero effect, "that would be quite a trick."

He's right—at least in a very superficial sense. If you throw a rock that big into a pond it is bound to make some ripples. And it is categorically undeniable that the stimulus has funded positions which otherwise might not exist. State governments alone have increased payrolls by more than 40,000 since the recession began, generally using stimulus dollars to cut the paychecks. Just the other day a Republican lawmaker who had blasted Obama's "failed stimulus" attended a ceremony at a jobs placement center in Rome, New York. Stimulus money paid local youth to refurbish the building. Would they have been jobless otherwise? Impossible to say.

Still, more than 7 million jobs have disappeared from the economy since Barack Obama took office. He will be only the second president since Herbert Hoover to face re-election with fewer people working than when he started. (George W. Bush was the other.) So it seems fair to ask whether stimulus projects have increased the net number of jobs in the United States—or whether they simply have moved a diminishing number of jobs around.

Analogy time. Consider a robber who steals a purse containing $500, who then uses the money to buy himself a new TV. It is categorically undeniable that the theft has created a sale for the TV store. Conservatives who pretend the stimulus has not created any jobs whatsoever stand in the position of an observer trying to deny the TV has been sold.

Yet the liberal analysis lacks any recognition that the purse owner now has $500 less to spend on the laptop computer she was going to buy. The theft has generated one sale only by destroying another.

The first effect is easily seen. The second is not. But only the economically illiterate would conclude that just the first effect occurred, and that therefore the way to increase consumption is to encourage more purse-stealing. So in addition to looking at the number of jobs created or saved by the stimulus, shouldn't we also consider the number of jobs destroyed or forestalled?

That all might sound rather speculative. But it is no more speculative than the methods used by the administration's supporters. If you think they audited every single one of the jobs the stimulus ostensibly created to make sure each position can be traced back to the ARRA, then you have another think coming. Note how the various pro-stimulus studies estimate the program created somewhere between 800,000 and 4.2 million jobs. That's a huge variation. What accounts for it? Huge variations in the assumptions built into the studies.

For instance, the White House Council on Economic Advisers supports its claims by comparing real changes in GDP to projections—i.e., against educated guesses. Then it declares the stimulus deserves the credit for any difference. This is known as assuming the conclusion.

True, other sources—such as the Congressional Budget Office—also say the stimulus has created or saved jobs. Yet still more sources say the opposite. Last month economists Timothy Conley of the University of Ontario and Bill Dupor of Ohio State University found that while the ARRA created or saved about 443,000 (mostly government) jobs, it destroyed or forestalled more than 1 million private-sector jobs—for a net loss of 595,000.

Liberals say that's bunk. "I don't find that very compelling," says a director of the liberal Center for Economic and Policy Research. (So there!) Conservatives retort that the pro-stimulus studies are programmed to produce a positive result, so they're no good either.

That's one of the fun things about economics: You can always find a study to support your position. And the human tendency toward confirmation bias means you'll probably believe the studies that support your view, and dismiss the ones that don't.

Given how economists still vigorously debate the economic effects of the New Deal, it's not surprising that they would differ over the 2009 stimulus as well. So remember the old joke. You might ask for a definitive answer about the stimulus. The best you'll get back is conflicting opinions.

A. Barton Hinkle is a columnist at the Richmond Times-Dispatch. This column originally appeared at the Richmond Times-Dispatch.