Working Papers logo 130

State and local public sector employee pensions are widely known to be underfunded, but pension financial reports do not reveal the true extent of funding shortfalls. Pension accounting methods assume that plan investments can earn high returns without taking any account of the market risk involved. This gives a false sense of the financial strength of public sector pensions and understates risks to taxpayers.

This paper first uses a Monte Carlo simulation of current pension assets and projected market returns to calculate the probability that public sector pension assets will be sufficient to fund accrued benefits. The typical public sector pension has only a 16 percent probability of paying full accrued benefits with assets on hand. A larger number of public pension plans have zero probability of paying accrued benefits than have a probability in excess of 50 percent.

But since accrued pension benefits are legally and constitutionally protected, any pension funding shortfalls must be met by taxpayers. This benefit guarantee amounts to an effective put option on plan investments, the cost of which is not disclosed under current actuarial accounting.

This paper uses an options pricing method to calculate the market value of taxpayer guarantees underlying public sector pensions. The average funding ratio declines from 83 percent under actuarial accounting to 45 percent under this options pricing approach. The typical state has unfunded public pension liabilities three times larger than its explicit government debt. Public pension shortfalls equal an average of 27 percent of state gross domestic product, posing a significant fiscal challenge in coming years.

Accurate measures of public pension liabilities are important for policymakers, taxpayers, investors considering the economic environment in which to start or locate a business, and bond purchasers considering the risk premia appropriate to municipal government bonds that are in practice subordinate to public pension liabilities.

Click here to view the full text of this working paper as an Adobe Acrobat PDF.

Andrew G. Biggs is a resident scholar at AEI.