If Scott Brown's Massachusetts upset sends health care reform into a free fall, Democrats will start looking for something that feels more solidly populist -- like bank regulation, or job creation. If the unemployment rate is the next object of the Democrats' reform agenda, here's a graph they should probably pay attention to. It's the CBO's jobs-per-dollar analysis of 11 employment-boosting ideas.

Sorry if the graph is a little difficult to read. You can open up the full report here.

 
Here's the good news: Increasing aid to the unemployed is the big winner because the jobless are strapped for cash and more likely than almost anybody else to spend what their checks quickly, sending the money back into the economy. To Congress's credit, they just extended unemployment benefits in November.

Here's the bad news: The CBO doesn't seem to put much stock* in reducing income taxes or extending the AMT exemption, even though Obama's early budget called for reducing income taxes for everybody except the top 5 percent of tax payers. 

Paul Krugman points to this report to hail the job creation tax credit but it's noteworthy that the CBO admits they don't really know how well this policy works in practice. After the 1973-5 recession, the authors note, the New Jobs Tax Credit gave firms a tax break if they increased total employment by at least two percent. The policy was too complex for many firms to apply, and later studies struggled to agree that the tax credit boosted jobs by a significant number. A Department of Labor report ultimately concluded that it was impossible to observe what hiring would have been done without the credit.

It's an important reminder that the CBO is a group of very, very smart individuals, but their conclusions are only as good as their models.