The Perils of the PBGC

Since the financial plight of the Pension Benefit Guaranty Corporation is in the news these days, I suppose that, as a PBGC alumnus (director of policy and regulations, 1984-87), I ought to weigh in. A good starting point is Dartmouth economist Andrew Samwick’s diagnosis and proposed cure:

Pension insurance – not the idea but its implementation, and certainly not the dedicated people who work at the PBGC – is a complete joke. There are three problems with the PBGC’s setup:

1) The premium amounts are too low. On average, companies do not pay enough to cover the risk to which they expose the PBGC.

2) The premium formula is inadequately linked to underfunding. Pension sponsors whose plans are underfunded do pay slightly more in premiums than pension sponsors whose plans are fully funded, but the amount of additional premiums does not adequately compensate the PBGC for the added risk of a claim.

3) The premium formula is unrelated to the PBGC's risk exposure – the portfolio allocation between stocks and bonds and the bankruptcy risk of the company. . . .

With PBGC insurance, the company has an incentive to invest in a portfolio heavily weighted toward stocks. If the stocks do well, the company can cut back on future contributions. If the stocks do poorly, then in some cases, the company can terminate the plan and leave the liability with the PBGC. Classic moral hazard. When the economy goes through a period of weak stock market returns (so the pension fund’s assets fall in value) and low interest rates (so the present value of the future liabilities rise in value), we get tremendous underfunding. And the laws governing minimum pension contributions don't require pension sponsors to make up the difference quickly enough.

What to do? Impose a levy on each DB pension plan sponsor that is proportional to the current value of all past PBGC premiums paid for current participants. Impose the levy based on 2004 data, so there is no rush to the exit. The levy should be enough to put the PBGC at a zero balance position. Then retire the PBGC and allow companies to obtain pension insurance privately if they so desire. For current sponsors, pass a law that moves pension participants’ claims in bankruptcy ahead of all unsecured creditors. If this means that fewer firms offer DB pensions, then so be it. Unhealthy companies – like United – ought not to be making promises to pay beneficiaries decades into the future.

Professor Samwick’s view of the roots of the PBGC’s problems is the same as the PBGC staff’s back when the insurance program was reformed in 1987. At its inception in 1975, the PBGC charged a flat premium that bore no relationship to exposure or risk of loss. As the initial rate was only a dollar per participant per year, trying to apply insurance rate-setting principles would have been fiddly. Nor was that low number pulled out of thin air. It was based on a Department of Labor study of participant losses in plan terminations in the 1960’s and early 1970’s. One reason why Congress enacted pension insurance, brushing aside “moral hazard” objections, was that it looked like an inexpensive fix for a small-bore problem.

Now it is thirty years later, and moral hazard has proven to be more than a theoretical concept. Had Congress anticipated that premium revenue would rise from $30 million to $1.46 billion a year, without any significant increase in the number of workers covered by insured plans, and that the program would nonetheless find itself nearly $30 billion in the red, the whole idea would have been aborted.

The original program carried even more moral hazard than the present one. What the PBGC provided was not insurance but a put option against the federal government: A company could rid itself of its pension liabilities by paying the PBGC the present value of guaranteed benefits. Since not all benefits were guaranteed, that was prima facie an attractive proposition: Paying x dollars relieved x+y dollars of liability.

That this system was not catastrophically abused was due primarily to three factors: (i) good luck (economic circumstances were such that pension overfunding was a much bigger issue than underfunding throughout the 1970’s and 1980’s); (ii) the assumptions used to value liabilities, which were more conservative than most companies would use for internal projections (so that the cost of exercising the put option was artificially high); and (iii) the PBGC’s inhospitable attitude toward companies that terminated plans in conditions short of extreme financial distress. Still, it was obvious by the early 1980’s that the existing structure was unsound. The reforms that the PBGC wanted then sound like Professor Samwick’s in several respects: Among them were higher premiums to eliminate the existing deficit, a change from a flat-rate premium to one related to risk, and an eventual shift to private pension insurance.

The package that Congress was willing to enact was less radical. Premiums were increased and based partly on underfunding (9/10ths of one percent of unfunded vested benefits), though without regard to the likelihood that claims would arise (i. e., the employer’s financial condition), and companies were effectively forbidden to terminate underfunded plans (and thus impose costs on the PBGC) unless they were in bankruptcy. The put option was thus replaced by a program of bankruptcy insurance.

The obvious flaw in a company’s buying insurance against its own insolvency is that paying the premiums deprives it of capital and increases the probability that the undesired event will come to pass. That dilemma is compounded if premiums escalate with financial weakness. Naturally, as Professor Samwick observes, premiums that compensated the PBGC for the full risk that it bore were politically impossible. It was also impossible to charge healthy plan sponsors the amount needed to cover claims likely to arise from weaker companies’ plans. If that had been possible, it in any case wouldn’t have been a wonderful idea: Making profitable firms pay the debts of those that fail is a classic perverse incentive.

So I’m not at all sure that I agree with Professor Samwick that the weakness of pension insurance is “not the idea but its implementation”. In anything like its present form, implementation is unfeasible, which suggests that the idea isn’t a sound one. Certainly, there is no real prospect of private alternatives. The PBGC’s staff and its outside advisory committee devoted quite a bit of effort to trying to devise a privatization scheme in 1985 and 1986. We could never surmount the hurdle of formulating a product that companies would voluntarily buy or insurers willingly underwrite.

Professor Samwick’s proposal is dramatic: Zero out the PBGC by imposing a one-time levy on current plan sponsors to fund the current deficit; accept no new claims from plans that terminate in the future; give pension claims priority over other unsecured creditors in bankruptcy; and leave pension insurance to the private market.

As a realistic political scenario, that is about what one expects from an academic economist. The deficit-closure levy would be nearly $1,000 a participant, an intolerable imposition, though it could be reduced a bit by valuing the PBGC’s liabilities on a less conservative basis. United Airlines’ competitors would denounce the unfairness of letting one company escape liabilities that the rest of the industry must continue to bear. Critics would rightly point out that secured claims take almost all of the assets available in most business bankruptcies, leaving only pennies on the dollar for unsecured debts. Private pension insurance is no more attractive to carriers or plan sponsors in 2005 than it was in 1987.

The Bush Administration’s proposal, which I have described elsewhere, would raise premiums sufficiently to eliminate the PBGC’s deficit over a period of years, but its principal feature is severe tightening of minimum funding standards. It would come very near to mandating that plans be fully funded for accrued benefits at all times, which would make contributions extremely volatile. All commentators expect a regime like that to discourage companies from maintaining defined benefit pension plans, and the Administration doesn’t seem to disagree. For all practical purposes it has given up on this form of retirement plan and is looking only for a PBCG exit strategy.

Now, it may be that traditional pension plans have no niche in the current marketplace environment. On the other hand, it may not be coincidental that their decline follows a significant expansion of government regulation, beginning in a modest way with ERISA in 1974, then accelerating with the Retirement Equity Act of 1984, the Tax Reform Act of 1986 and the deluge of legislation in the 1990’s. Regulation of individual account plans has increased, too, but beginning from a lower base and to a far lesser extent. If, as I have suggested before, pension plans were subject to less federal micromanagement, they might spring back to life. If they didn’t, we would at least know that genuine market forces rather than government intrusions were responsible for their demise.

A pension plan revival would give the PBGC a larger base of premium payers and ease the amortization of its deficit. I don’t deny that the moral hazard posed by pension insurance would remain worrisome. Fundamental reform or disastrous collapse are the PBGC’s only possible long-run fates. While we wait to see which will occur, I suggest that we do what we can to make the short run less bleak.