Do I disagree with Brad DeLong and Michael Froomkin about pre-funding social security too ?


This is getting ridiculous.

On social security, Brad does advocate eliminating the $ 90,00 ceiling and more boldly advocates taxing all income. Still he doesn't explain what he wants to do with all the money. He says benefits are bsically OK so it must be that he wants to cut the Social Security tax rate.

I agree that this would be good, but I think pre-funding would be better. I have two reasons. First, the US national saving rate is too low. Other things equal pre funding would help that. Pre funding has made Singapore one of the richest countries in the world.

Now other things are not equal. The strength of this argument for pre-funding depends on how one thinks pre-fuding would affect the general fund surplus/deficit. I am confident that, the indirect effects of a social security administration surplus on the general fund deficit are less than one for one. Believing in one for one (or more) offset would lead one to imagine that any tax increase is useless, because it will just lead to tax cuts elsewhere.

Second, in steady state, prefunding offers a better deal to workers than pay as you go if the rate of growth of economy is less than the rate of return on capital. This is clearly true in the us In as noted by Abel Mankiw Summers and Zeckhauser some time ago, since capital is a source of funds not a sink. The only way to make the social security administration earn a rate of return less than the growth rate of the wage bill (so prefunding is costly to beneficiaries) is to force it to invest badly. This is the current approach, but it doesn't have to be that way. Like Ronald Reagan and Bill Clinton, I think the SSA should buy stock and corporate bonds as well as t-bills.

One might argue that this would give the SSA too much power over firms. However, there is no need for the SSA to be allowed to vote its shares (actually it should be required to vote them proportionally to how other shareholders vote theirs). There would have to be a rule for which shares the SSA would buy. That means I have to address the concerns of the renegade Froomkin.

Froomkin wrote

"Start with the buying end. Buying will either be mechanistic, delegated, or discretionary. If it is mechanistic, we have the wrangle over the formula, which then distorts markets unless the formula is to buy a basket consisting of the entire stock market (even this, arguably, has secondary effects on the bond market, but let’s not go there).

In fact, this is the only buy/sell formula that doesn’t create huge problems right off: have the feds buy a basket that represents all the shares in a multiplicity of exchanges (not just the big ones). But this isn’t easy to do, especially for small-cap stocks, without distorting markets."

The rest of the post argues that any rule other than buy a basket consisting of the entire stock market would create distortions. I consider the rest of the post to consist of setting up straw men and knocking them down. I argue that the SSA should buy the market. Froomkin's argument against is that "this isn’t easy to do, especially for small-cap stocks, without distorting markets." Now the SSA is not agile and congress is anything but agile. There will not be a problem with a sudden huge order for stock. Rather purchases will be highly predictable. In theory this should not affect relative prices of different shares (note that for me theory is a pejorative term). In practice, it will, as suggested by Froomkin drive up the price of low cap stocks compared to high cap stocks. Curiously there is strong evidence that this price is currently too low. Since I don't think the market is efficient, I don't think that a public intervention which changes stock prices necessarily moves the market away from efficiency. Indeed the guess that stock is still underpriced compared to its value to the US Treasury (which has extra special risk bearing capacity) is the reason I think the SSA should buy stock.

Now I think there are risks (Froomkin hints at them). If SSA bureaucrats are required to buy stock they are required to ignore their suspicions about market manipulation. There could be two forms. One is driving up the price of a stock by other than arms length purchases. The other is cornering the market for a stock. Both are illegal, but both are not always easy to detect. In both cases, the markets ability to survive such tricks might depend on private agents deciding not to buy stock even if it can't be proven beyond reasonable doubt that its price is manipulated. A legal requirement to buy unless manipulation is proven will make such frauds more profitable. I don't know enough to estimate so I just guess that this is a tolerable cost.