March 29, 2005,
On Social Security, our elected officials appear implicitly divided into two camps. One group favors restoring balance with lower promised benefits. This group also hopes to maintain or increase benefits with the higher returns associated with private accounts. The second group opposes private accounts and hopes to finance current promised benefits with higher taxes.
With tax hikes and benefit cuts forming vivid lines in the sand, it is useful to pause for a moment and ask a simple question: To what extent are Social Security taxes really “taxes” in the traditional sense? Do the lines in the sand make sense?
The question is not as odd or revolutionary as it may sound, and the answer has a significant impact on all sorts of debates. For example, opponents of President Bush’s first-term tax cuts often included payroll taxes in their incidence calculations. Since the reforms did not change payroll taxes, which are capped, the share of federal taxes raised by regressive payroll taxes increased after the tax cuts. This fact lead to several analyses that made the tax cuts appear far more regressive than they would have if only the income-tax reduction were analyzed.
At the time of this debate, Larry Lindsey, former chairman of President Bush’s National Economic Council, gave a well-reasoned speech at the American Enterprise Institute where he mused on the proper definition of the word “tax.” Since Social Security also includes benefits that are calculated in a progressive manner, was it really right to do an incidence analysis that excluded the benefits?
Now that Social Security is being addressed head on, careful thinking about this question is again sorely needed.
The economics of the question is (sadly) far from simple. Suppose, to start, that we live in a world of absolute certainty and rational individuals. In this world, everyone knows what their income will be until the day they die. In this world, if an individual pays $10 in Social Security tax today, but gets back $10 in present value when he retires, then his net benefit is zero. A rational individual in this case would not think of the $10 as a tax, or as anything at all. It’s the net benefit that matters. If he pays in $14 and gets out $16, then the system increases his lifetime income by $2. The same is true if he pays in $2 and gets out $4.
Restoring balance in this example requires that the net benefit be reduced. Politicians might obsess over the labels, but rational individuals would not. Money is money. Since the net benefit is the true tax, a benefit reduction is as much of a tax hike to a rational individual as an explicit tax hike.
If individuals are irrational and save too little for retirement, the size of the benefit and tax will matter, but since individuals are irrational, it is hard to say how these things will matter.
In a world of uncertainty, things get a little trickier. Now, a worker will have to value the many different characteristics of Social Security that may or may not affect his lifetime welfare. For example, Social Security only provides a benefit if you live to retirement. Social Security requires you to save for retirement when you are young, something that may delay your first purchase of a house. These two examples are features that are unattractive, and the cost of them would have to be factored in when estimating the true cost associated with Social Security taxes. On the other hand, Social Security may offer some benefits as well. It offers insurance against having a lower income, for example.
A full accounting of the system would have to impute a value for all of the relevant features of Social Security. One attempt to take a step in this direction is a recent National Bureau of Economic Research paper by economists Erik Hurst ad Paul Willen. They found that the total costs of the constraints of the current system are very large, perhaps as much as 7.5 percent of a typical individual’s lifetime consumption.
With regard to the question at hand, the Hurst and Willen result, if true, is enormously important. It means that even if Social Security paid individuals back an amount of money that looked comparable to the amount of money paid in, the true costs of the system could be high. Even for those individuals who get $1 out (in present value) for each $1 put in, some goodly portion of that $1 should be thought of as a tax. This is because the transfer has all sorts of strings attached, and those strings are unattractive and costly.
This does not reverse Lindsey’s main point. There may certainly be some individuals who receive a large enough redistribution from Social Security to offset these costs. But even if individuals are rational, raising taxes and cutting benefits are not easy substitutes for one another. Higher taxes would, even if matched with higher benefits, maintain a high gross benefit level that would do little to reduce the costs identified by Hurst and Willen. Thus, if a larger Social Security system is adopted to protect irrational individuals who do not save enough for retirement, it will impose a significant cost on rational individuals. This cost will be significant even if the rational individuals have access to the higher benefits themselves.
Kevin A. Hassett is director of economic-policy studies at the American Enterprise Institute.