CBO has estimated the law’s impact on employment and economic output using evidence about how previous similar policies have affected the economy and various mathematical models that represent the workings of the economy. On that basis, CBO estimates that in the third quarter of calendar year 2009, an additional 600,000 to 1.6 million people were employed in the United States, and real (inflation-adjusted) gross domestic product (GDP) was 1.2 percent to 3.2 percent higher, than would have been the case in the absence of ARRA.
This was trumpeted as conclusive evidence that the stimulus is working and used to smack opponents upside the head. Other than overconfidence in the CBO (they estimated at the time of stimulus debates that ARRA would increase employment and GDP, and this report, unsurprisingly, states that the CBO was right on the money with that prediction), it's just not that big of a bombshell.
As Brian Riedl of the Heritage Foundation noted,
If this seems completely detached from reality, it’s because CBO did not actually analyze the performance of the economy — including the rapid increase in the unemployment rate since the stimulus was enacted. They simply took their economic model that predicted last January the stimulus would work, re-plugged in the bill’s provisions, and (surprise!) got the same result. CBO is effectively saying the evidence the stimulus worked is that they predicted it would work.And Arnold Kling backed him up.
Back in March, the CBO ran a simulation model of the economy, with and without the stimulus. The difference between the two simulations gives you the predicted increase in employment and GDP.This CBO report says nothing meaningful about the efficacy of the stimulus. We'll likely not know just how bad things have been and the effects the stimulus have had until very far down the road. In the meantime, the stimulus has only paid out a little more than $150 billion of the total $787 billion package.
Recently, the CBO repeated the exercise. Lo and behold, the differences were the same. This says nothing about what happened in the real world. It tells you that the simulation model that they used did not change.