PGL and others have addressed many of the important points of contention in the debate between Mankiw and Baker-DeLong-Krugman (henceforth BDK, or simply The Trio) on the returns that privatized Social Security accounts could be expected to enjoy, so I have relatively little to add.
But for those who are new to the whole discussion, let me just offer a bit of context with the following graph, which shows corporate profits as a share of national income over the past 75 years. (Yes, I’ll admit it, basically this is just an excuse to put up a graph with a really long time series, which I always think is cool.)
The central point that this picture makes is this: net profits as a share of national income is historically very stable. Very few years over the past 75 have they been outside of the 5%-8% range. This means that profits have never demonstrated an ability to systematically grow faster than national income, which is the starting point for The Trio‘s entire argument.
If profits don’t grow faster than national income, it’s hard to see how stock returns will grow faster than national income. (Yes, there are a few possible ways, some of which Mankiw and The Trio discuss, but all seem rather tortured to me.) Those who have argued that privatized social security returns can indeed be higher than the growth in national income have still not been able to offer any substantial response to this logic.
As a side note, it is striking that net profits in 2004 were the largest as a share of national income of any year in the past 75. Some may take this as a suggestion for where they will head from here…