Menzie Chinn looks at the numbers for consumption and durable consumption, which are in free fall (follow the link to see the graphs), especially in comparison to the 2001 recession. Chinn notes that

the behavior of consumption has some implications for the debate over the proper course of fiscal policy. If consumption is falling because the marginal propensity to consume out of disposable income (MPC) is declining (the b parameter in this post), then -- in terms of maximizing the impact on aggregate demand arising from a dollar's worth of budget deficit -- it makes sense to favor government spending on goods and services, as opposed to tax rate cuts. In the extreme, if MPC were zero, then the multiplier is 1 for government spending, and 0 for tax cuts (assuming accommodative monetary policy). Hence, those who argue that the MPC is small and in favor of general tax cuts as a means of stimuluating the economy are not providing an internally consistent argument.

Further, if the decline in the MPC is concentrated on the non-liquidity constrained households, then whatever tax cuts (or transfers increases) occur should be aimed at the most liquidity constrained households.

I.e., food stamps and unemployment compensation are much more effective than tax cuts. Exactly what conservative economist Mark Zandi of Moodys was forced to conclude last month when he looked at the effectiveness of various types of stimulus.

In fact, Zandi was directly investigating the Heritage Foundation's idea of the bonus depreciation, and wrote,

The economic bang-for-the-buck of bonus depreciation is very modest (see table).[7] Indeed, of all the tax and spending policies considered, it provides the least amount of stimulus.