Tax cuts have an advantage over increases in government spending since they can be implemented relatively quickly. Government spending programs take more time to put into place.
But there is also a disadvantage. Unlike government spending, which has an initial impact on demand that is certain and one to one, tax cuts give households and incentive to spend more, but there's no way to ensure that the tax cuts won't be saved instead of spent. And if they are saved instead of spent, then they don't produce the desired increase in aggregate demand. However, if the tax cuts are well targeted, then the amount that is spent rather than saved will go up, but even then the impact on demand is still less certain than for government spending.
The experience from the first attempt at using tax changes to stimulate the economy, the Bush tax rebate of Spring of 2008, suggests that temporary tax cuts of this variety are largely saved. Here's an estimate of the effect of the Bush tax rebates from the Congressional Budget Office. It shows that the rebates drove income up quite a bit, but consumption hardly budged:
The lack of impact on aggregate demand (consumption) is due to how the tax rebates were distributed. The tax rebates went to a lot of people who really didn't need them, and they chose to save the extra money.
But better targeting can fix this. Thus, I thought that the best way to do the second stimulus package, the one Obama put into place a little over a year ago, would be to have well-targeted tax cuts go into effect as fast as possible to give the economy an initial jolt. This would be followed by government spending, which takes a bit longer to put into place. As this spending came online it would sustain or even strengthen the impact on aggregate demand provided initially by the tax cuts, and, importantly, the spending would be sustained until the economy was clearly on its way to recovery.
The stimulus package that was put into place by the Obama administration to fight the recession was constructed along these lines. Though many people don't realize this, just under 40% of the total stimulus package was devoted to tax cuts.
However, though the targeting was better than for the Bush tax rebates, the tax cuts weren't as well targeted as they could have been if the goal is to maximize the impact on aggregate demand (i.e. if the goal is for the tax cuts to go to consumption, not saving), and not all of the tax cuts went into effect immediately when they were most needed.
Initially I was critical of how the tax cuts were targeted since so much ended up going to saving rather than consumption. This is the part I am rethinking.
There are different types of recessions, and this one can be termed "a balance sheet" recession. It had a big impact not just on bank balance sheets, but on household (and, for that matter firm) balance sheets as well. Households were particularly hard hit due to declines in stock prices and declines in the value of housing. These losses were large, they upset plans for things such as retirement, and households needed to refill the holes in their balance sheets that had been created (this includes paying off debt).
How do they refill their balance sheets? By saving more and consuming less (paying off debt is a form of saving). Thus, as the recession took hold, we saw a large increase in the saving rate and a corresponding fall in consumption. The tax cuts were an attempt to reverse the decline in consumption, but instead they mostly raised the amount that went into saving.
But that has a benefit. Households are not going to start consuming normally again until their balance sheets are repaired. The faster the holes in their balance sheets are refilled, and tax cuts can help with this, the faster the households can return to their normal rates of consumption -- a prerequisite for the economy to return to normal.
So the targeting of the tax cuts that was OK after all. You don't see the effects of balance sheet rebuilding in the data initially because the tax cuts are being used to fill up balance sheets, there's no immediate effect on consumption, output, employment, etc., to observe in the data. But since balance sheets are refilled faster, we will emerge from the recession sooner, and that's an important benefit of tax cuts that's often overlooked.
There's one more important aspect of this. While the balance sheet rebuilding is going on, there's very little impact on aggregate demand, and since stimulating aggregate demand is a key to recovery, that must come from somewhere else. If we move the tax cuts from, say, higher income households to lower income households that will increase aggregate demand, but it will also slow the rate of balance sheet rebuilding and hence cause the recession to last longer.
Instead, we should do both. Give tax cuts to some households knowing full well they will use them to fix balance sheet problems, and also give tax cuts to households who will spend rather than save the money (because they need the money to help to pay bills, etc.), and do so in sufficient quantity to provide the necessary stimulus to aggregate demand. Then, as outlined above, follow this up with aggressive fiscal policy action and job creation programs that sustain the impact on aggregate demand until the economy is ready to stand on its own once again.
In balance sheet recessions, tax cuts that are saved will help to end the recession sooner and hence should be part of the recovery package. The most important concern is still aggregate demand, and policies must be devoted to this problem first and foremost, but tax cuts have a role to play in the recovery process even when they are saved.