NRO's Andrew Stiles flags this exchange from Congressional Budget Office director Douglas Elmendorf's testimony before the Senate Budget Committee ealier this week:

The quote that matters starts around 1:25, when Elmendorf says that, according to CBO's estimates, with the stimulus legislation in place, "the level of GDP would be a little lower at the end. That is, a net negative effect on the growth of GDP over 10 years." Elmendorf then confirms that CBO estimates that the economic drag will continue in the following decade: 

SESSIONS: And in the next 10 years, since you’re carrying that debt and paying interest on it and the stimulus value is long since gone, it would be a continual negative of some effect? 

ELMENDORF: Yes, it would represent a drag on the level of GDP beyond that, if no other actions were taken. 

As with all CBO's projections, you have to take these with a grain of salt, especially when you start looking out two decades into the future. I've been critical of the way folks have used the CBO's stimulus job-creation numbers in the past, and this sort of long-term economic forecasting is not the most precise tool either, to say the least. Counterfactuals—like asking what economic growth would have looked like the absence of the stimulus—probably tell us more about our current economic assumptions than about alternate economic timelines.

Still, it's worth noting, if only because even the mildly Keynesian congressional scorekeeper agrees that borrowing $800 billion dollars ultimately creates a drag on the economy and a net loss in economic performance relative to what otherwise might have been. And yet the administration went ahead with the legislation anyway, arguing that it would be more or less a free lunch in the long run.