The Costs of Taxation
Miles Kimball, Karl Smith, and Scott Sumner have been discussing the impact of taxation lately, and I wanted to comment a bit. Taxation is a highly complex issue, since the tax code is so huge, and the topic is rife with money illusion problems and other misconceptions.
One approach to thinking about the burden of taxation is to think about the difference between what you put into the web of trade and what you can take out. One of the first things economists learn when studying taxes is that who pays the tax nominally is not the same necessarily as whose consumption is reduced by it. Imagine a firm selling a good with highly inelastic demand. If the government decides to levy a tax on the good, nominally paid for by the firm, the firm can simply raise its prices and pass on the tax to the consumer. Ultimately, consumers of the good have to pay the tax in the form of reduced purchasing power. Similarly, capital gains taxes are paid by more than just investors, and corporate taxes are paid by a combination of a firms owners, workers, and customers.
Incentives change behavior. An even tax rate won’t affect relative prices because what maximizes x also maximizes x/2. The long run labor supply curve is, as near as we can tell, vertical. That means that no matter what the average wage is, people tend to supply the same amount of labor. Taxing particular goods changes the relative price of those goods, and thus can cause inefficiencies. Karl Smith’s point in this post is that benefits reduce work effort, not taxes. If you can get money without working, why work? The larger the benefits, the larger the effect. It is precisely because taxes are typically spent on people, and that causes those people to work less, that welfare causes lower GDP. One of his points is that if welfare is given to low productivity people, their dropping out of the labor force won’t lower GDP by much. Somewhat oddly, giving money to the rich is likely to hurt GDP by more than giving it to the poor, and not just for “marginal utility of the dollar” reasons (money is worth more to poor people).
Adding another layer of complexity, there is a difference between nominal taxes collected and the total real tax burden. Imagine that the government collected 10% of everyone’s income per year, and never spent any of it. No real resources were used up. If prices were flexible, this would result in no change in anyone’s real consumption, investment, employment or any real variable whatsoever. The only consequence would be that prices would fall by 10% per year. The paradigm highlights that government inefficiency occurs if the resources that the government uses to provide goods and services could have been better put to use by the private sector. In real terms, government spending must equal taxation. When the real resources are used up, whoever would have gotten them needs to consume less to make up for it.
In conclusion, the two reasons why taxes reduce private sector consumption and investment are relative price shifts and government spending uses up real resources that otherwise could have been used by the private sector. Taxes are a highly emotional topic for many people, but it is when dealing with highly emotional topics that people most need to step back and try to use reason. High tax rates have some major downsides, so I think it helps to be able to figure out exactly when and how they cause inefficiencies and when they aren’t so bad.