Ezra Klein posted this chart, which Henry Blodget of Business Insider calls “the chart that will get Obama fired.”
This chart is an updated version of the one the Obama administration released back in 2009 to help sell the stimulus bill. The original chart came from a report by economists Christina Romer and Jared Bernstein on the impact of the stimulus, showing projected unemployment with and without the proposed spending bill. As you can see, the prediction was that if the stimulus plan was passed, the unemployment rate would top off in mid-2009 at 8 percent and drop to 7 percent by the third quarter of 2010. Obviously, that didn’t happen.
There are many running theories about why the stimulus failed (and Klein does a good job of listing many of them): No one knew how bad the recession was, the stimulus was poorly designed, the stimulus was too small, stimulus spending can’t work, politics gets in the way, and so on.
My position on the stimulus, as you know, is that we shouldn’t ignore the lack of consensus among economists about whether government spending is capable of jump-starting the economy. Certainly we shouldn’t claim that economists agree. (On that point, I would suggest reading this very good post by Matt Mitchell over at Neighborhood Effects.) Also, those who believe that stimulus could work shouldn’t ignore that spending bills are mostly designed, because of politics or other factors, in ways that prevent them from having anywhere near promised impacts. Ever.
This is the reality we live in. We can’t wish it away. And yet, over and over again, the same policies that failed in the past find advocates for their implementation once again. The belief that the government can get it right this time is a powerful force — even among those who are willing to admit that government policies have failed in the past. This belief exists on both sides of the political aisles.
However, I think the main lesson from this chart is that we know much less than we claim we know. The pretense of knowledge is a powerful thing.