Pensions and So Forth
There's been a fair amount of news about pensions recently. For a start, the Department of Labor wants legislation to shore up the PBGC (Pension Benefits Guarantee Corporation — to traditional, defined-benefit pension plans what the FDIC is to banks), which has recently taken over an unusually large number of underfunded pension plans. One component of the proposal is to increase the premium paid by a healthy plan from $18 per employee per year to $30, and to start indexing it to wage increases.

It seems odd that Congress specifies the premium, rather than instructing the PBGC to charge a premium high enough to keep itself solvent and not substantially higher. It also struck me as odd that the base rate is expressed "per employee" and not, say, "per million dollars liability".

One reason for that might be that pension plans have a fair amount of flexibility in setting their discount rates. I'm a little bit unclear on the specifics of this, but I believe pension plans actually use two discount rates, one for the assets and a different one for the liabilities. If I'm not mistaken, this establishes the pension-accounting community as an opportune target for the war on drugs. Anyway, as I understand things, Labor also wants to set at least one of these discount rates itself based on an index of investment-grade corporate bonds. While I usually oppose centralized decision-making, this would make pension plans easier for investors to compare plans and would make it harder to make an insolvent plan look solvent, and I can't see how the lack of flexibility would harm anyone honest.

One of my professors, Zvi Bodie, who is known for his risk-aversion, has written an article for next month's Milkin Institute Review in which he argues that pension funds should be required to invest entirely in investment-grade bonds. It's an interesting idea, which would clear up the key question of what to do about downside risk in the plan assets. Currently, companies that sponsor plans keep most of this risk, with the PBGC as insurance. Plans could be required to buy put options, or, equivalently, required to put a certain portion of investment assets in safe bonds and allowed to buy calls with the rest. The PBGC insurance acts like a put option, but the premiums aren't thought out, or even related to the level of risk inherent in a plan's assets.

In other news, the Governor of California has suggested that the defined-benefit plan run by CalPERS in behalf of state employees be phased out in favor of 401(k)s. Which is the way the world is going. The employee gets all of the downside risk, but gets to decide how much to take (by investing in stock funds or bond funds or some of each).

The President and the Congress are, of course, going to spend a lot of the next year on a key plank in Bush's "Ownership Society", the diversion of some Social Security payments into private accounts. I generally support this, but I support some restrictions, again because of downside risk. If people are allowed to squander their social security accounts, I think the government is likely to bail them out, and I don't think the taxpayer ought to bear a participant's downside risk. So I'd like to see mutual fund companies offer mututal funds with floors, with unusually large management fees to include the cost of the put option. I don't know precisely how these would be structured. I don't envision guaranteeing a minimum return on a daily basis, for example, but on a yearly or multi-year basis; I don't want the fund companies to go under if the market takes a crash; and there are moral hazard issues with respect to the performance of the fund managers. But I think something like this could be worked out.
- posted by Steven @ 9:21 PM