In September, Congressional Budget Office chief Doug Elmendorf testified to Congress about the effect of the Bush tax cuts on economic growth.* He brought graphs.
The first graph on the economic effect of the tax cuts in 2011 and 2012 found that a permanent tax cut would have the biggest bang for the buck in the next few years because companies like the security of knowing their tax rates won't be changing soon...
... but the second graph, which forecasts the economic impact of these tax cut options in 2020, finds that any permanent extension of the Bush tax cuts would have significant crowding out effect in ten years, reducing GDP by up to 1.2 percent. On the other hand, it found that a partial extension and repeal of the Bush tax cuts would have only a fraction of the drag on economic growth.
*Update: In his testimony, Elmendorf concluded: "Compared with the options examined here for extending the expiring tax cuts, various other options for temporarily reducing taxes or increasing government spending would provide a bigger boost to the economy per dollar of cost to the federal government." I think that's right. Extending the entire Bush tax cuts temporarily is not necessarily the best use of hundreds of billions of Treasury dollars, given a universe of options for tax and spending policy. Rather, dealing with the Bush tax cuts qua Bush tax cuts under the parameters of what's possible in Washington, a short term extension of the law combined with guarantees for jobless benefits is an acceptable deal.