President Obama proposes to increase tax rates on high-income households while making the existing tax rates permanent for taxpayers below the top tax brackets. While the increase would hit only a relatively small fraction of all households, that group represents a large share of total taxes and of private spending. Raising their tax rates would be a substantial blow to overall spending and therefore to GDP growth.
While it is true that high-income households represent a large share of consumption, the issue is what is the marginal propensity to consume out of a change in disposable income created by a two-year increase in current tax obligations? For households who do not face borrower constraints and would therefore most likely behave in a fashion predicted by life cycle models of consumption, the impact on consumption demand would most likely be very modest at best. The Barro-Ricardian Equivalence proposition would go so far as to suggest that delaying a tax increase needed to restore long-run fiscal sustainability would have no effect on aggregate demand.
Fortunately Mark Thoma offers us this more reasoned assessment of the role of fiscal policy and aggregate demand:
Shifting the tax cuts to people who are more likely to spend the extra money rather than put it into savings would provide an even larger boost to the economy. It's also worth noting that if the worry is about the effect on the economy and the deficit, it would also be possible to allow the tax cuts to expire and then replace the missing demand with additional temporary government spending