Why do voters support unfunded social security systems? Browning (1975) suggested that, since voters consider past contributions as a sunk cost, unfunded systems may outperform alternative assets. To quantify the relevance of Browning's argument I calculate the return from “investing” in social security for the median voter at US Presidential elections from 1964 to 1996. The continuation internal rate of return ranges from 5.7% in 1984 – 43 year old median voter – to 9.8% in 1964 – 46 year old. In ex-post returns social security outperforms the Dow Jones Industrial Average for the 1964 and 1968 median voters, is equivalent for the 1972, and slightly underperforms it for the 1976.