Should we privatize Social Security?

Brad DeLong argues that we should privatize social security. His ingenious analysis suggests that privatization could help overcome market failures of insufficient savings and insufficient investment in equities (the “equity premium paradox”). Turning the usual debates on their head, he argues that market fans should have no reason to support privatization, but interventionists might. Note that Brad was replying to earlier posts by Matt Yglesias and Atrios, both worth reading in their own right on the issue.

I agree with most of Brad’s analysis (though the equity premium matter is complex and poorly understood), but I am less sanguine about the privatization idea. I wish to privatize many things, but forced savings is not one of them.

Most of all, I am worried about the fiscal implications of this privatization. Current plans need not, in the long run, cost us any money, as Arnold Kling reminds us. But they do require a big tax increase in the short to medium run. In essence the proposed reforms stop collecting “pay as you go” contributions and move everyone possible toward private accounts. But during the transition an influx of money is still needed to pay off the current elderly, therefore the tax increase. And we are talking trillions here, not just small change.

Of course these taxes must come sooner or later, given that privatization is simply shifting future claimants out of the public sector expenditure nexus. So if we raised taxes today, to finance a transition, and cut them twenty years from now, social security privatization would be revenue-neutral in this regard (of course it could influence revenues along other dimensions).

Now you can see my fear. If we move to privatize social security, I predict that higher taxes today will mean still higher taxes tomorrow, not an eventual tax cut. People would grow to tolerate the higher taxes, and our ever-diligent Congress would find new ways to spend the money (don’t trust any promises to the contrary — remember when they told us that social security numbers would never be used for purposes of personal identification?).

I think of social security as having two parts: a welfare system for old people, plus a regime of forced savings for the young. The Bush plan cuts back on the welfare angle, but also would put the forced savings in the private sector.

If we were starting from scratch, I could imagine that a fully vested system of private accounts could make sense. But given where we are, I would like to see social security evolve into a system of welfare for the elderly, and junk the forced savings aspect. Keep in mind those “private” accounts will be regulated rather than truly private in the libertarian sense. They will channel benefits to government-approved providers, thus leading to bureaucracy, regulation, and costly commissions. And if anyone’s account goes bust, do you really think there will be no secondary safety net to bail them out? What if the whole market went bust for about ten years’ time?

One of the original virtues of social security was its minimum administrative costs and its relatively “clean” fiscal nature. Let’s not lose their properties of the system just to privatize something — forced savings — that should be a matter of voluntary choice in the first place.

So let’s push for means-tested benefits, and hope that social security slowly but surely shrinks and evolves to a welfare system for the needy elderly. It should not be a stranglehold over every mainstream employment relationship.

It may sound like a strange world where Brad DeLong endorses social security privatization and I oppose it. But given the other views we hold, this is where we each ought to end up.