of the program involves no distortions. That is unlikely. After all, the rationale for the current Social Security system is that individuals are myopic and would not save an adequate amount on their own. The fact that individuals would not undertake the saving voluntarily means that they undervalue it. As a result, they would probably view at least a portion of their top-tier contribution as an implicit tax.
More generally, if you asked most individuals if they view their FICA payment as a tax, they would undoubtedly say yes -- even though they will inevitably get benefits based on these payments. The same would most likely be true for as mandatory payment to an account to which the taxpayer has no access until retirement. It would be fool- hardy to construct policy on Shoven's assertion that a tax is not a tax.
Shoven recommends a defined-contribution component for Social Security because he fails to consider the advantages of the most realistic alternative -- namely accumulating assets through the existing defined-benefit plan and investing some of those reserves in equities. He also mistakenly argues that a defined-contribution component will make the system efficient.
Shoven is right on one point. Two tiers, or even three tiers, are better than one. Defined-benefit and defined-contribution plans are subject to different types of risks. A system that combines the two approaches will function better than a system that relies on a single model. But the United States has never tried to provide retirement income through a sin-