According to the Weekly Standard, "two quarters ago, the "stimulus" had added or saved just under 2.7 million jobs -- or 288,000 more than it has now. In other words, over the past six months, the economy would have added or saved more jobs without the 'stimulus' than it has with it. In comparison to how things would otherwise have been, the 'stimulus' has been working in reverse over the past six months, causing the economy to shed jobs."
But this is a fundamental misreading of the data. Over the past six months, and even before then, the stimulus has been winding down, and the fact that it's now responsible for the existence of fewer jobs than it was six months ago is a function of that phase-out, and the (slowly) recovering economy, not of its inherent ineffectiveness.
Here's Moody's Chief economist, Mark Zandi, in an email to me, explaining the data
"It's not that ARRA [the stimulus] is now costing the economy jobs, it is that the economy is now creating jobs without ARRA's help," Zandi says. "This is exactly the objective of fiscal stimulus, namely to end recession and jump-start economic recovery. The Great Recession ended in June 2009, the same month that ARRA was providing its maximum benefit to the economy. Stimulus was never intended to be a source of long-term economic growth, it was intended to stop the free-fall in the private sector. It did that."
In its report, CEA isn't revising its claim about ARRA's effectiveness downward -- it's accounting for the fact that the stimulus' impact on current employment is less than it was when the stimulus was pumping out more money, and will continue to decline. Years from, when the stimulus is responsible for few if any existing jobs, it won't be evidence of the stimulus' failure either.
"CEA's estimate that payroll employment was 2.4 million higher to as of [the first quarter of 2011] than it would have been without ARRA is very similar to my estimate," Zandi says. "I consider this to be a significant benefit to the economy's performance over the past more than two years. This benefit is fading, but this is by design."
Republicans are also citing the Weekly Standard article to revisit one of their oldest claims about the stimulus: that every job the Recovery Act saved or created cost taxpayers about $278,000. That's the number you get if you clumsily divide the cost of the stimulus by the number of jobs it's responsible for. The White House is aggressively pushing back on this argument by pointing out that the stimulus paid for more than jobs -- it paid for new infrastructure, materials and so on.
Zandi adds another layer to this critique: "[I]t is inappropriate to say that ARRA cost taxpayers $278k per job. The cost should be calculated based on the number of full-time equivalents," he said. Specifically, the stimulus' efficiency shouldn't be calculated based only on how many full-time jobs it was responsible for, but by how many part-time or seasonal jobs it created as well. "I'm not arguing that ARRA was as efficient as it could have been, although given the political constraints and the crisis environment it is a significant achievement."