Thus an income tax that taxes interest can be viewed as a differential commodity tax in which future consumption is taxed more heavily than current consumption. The question of whether it is desirable to tax interest income is then equivalent to the question of whether it is desirable to tax future consumption at higher rates than current consumption.
Just as, with a well-designed income tax, there may be little to be gained by adding differential commodity taxation, so too there is little to be gained from taxing consumption at different dates at different rates. This means, in effect, that interest income should be exempt from taxation. An income tax that exempts interest income is, of course, equivalent to a wage tax, and we showed in Chapter 17 that a wage tax was equivalent to a consumption tax (in the absence of bequests. This suggests that it may be optimal to have a consumption tax.
That’s very tight reasoning on Stiglitz’s part.
Even though I hate taxes, I love analyzing the economics of taxation.
Co-blogger Scott Sumner is, I believe, an advocate of consumption taxes. If you’ve read me closely over the years, you’ll know that I’m not. I’ve been very critical of a VAT, for example, which comes close to a consumption tax.
So why am I not a fan? Because the analysis of a VAT that finds a VAT to be superior, on efficiency grounds, to other taxes, holds constant something that empirical evidence says should not be held constant: the amount of revenue raised. Because VATs tend to be less visible, they lead to less political resistance and result in a much larger amount of government tax revenue and government spending. Because so much of government spending is inefficient–how much do you value, relative to cost, the occupation of Iraq and Afghanistan or defending wealthy Germany or the middle-speed train in California’s central valley or agricultural subsidies?–the inefficiencies from higher government spending probably outweigh the inefficiency of a less-efficient form of collecting government revenue.