Appropriate Discount Rate for Public Plans Is Not Simple

The New York Times recent article “A Sour Surprise for Public Pensions: Two Sets of Books” told the tale of the tiny six-employee Citrus Pest Control No.2 pension fund that thought it was overfunded but instead received a large bill when CalPERS calculated its termination liability.  The article has already been criticized as “remarkably fact challenged.”  First, CalPERS does not keep two sets of books.  Rather, it uses a very conservative methodology to calculate withdrawal liability for plans (covering about one third of participating workers) that have essentially contracted out the management of their pension fund to CalPERS and then decide to withdraw.  Second, the termination liability for these participating plans has been disclosed on CalPERS’ website for many years, so the tiny plan was not the victim of a nefarious plot.

More serious than the factual errors, however, is the implication that the “two sets of books” is an admission by CalPERS that it should be using a riskless rate for valuing and funding its plan.  That is a simplistic assessment of a complex issue.

For financial reporting purposes, the argument is compelling that the benefits of public pension plans, most of which are guaranteed under state law, should be discounted by a rate that reflects their relatively riskless nature.  Such a valuation would produce relatively low funded levels.  A more conservative measure of the funded status would deter plans from offering more generous benefits in response to supposed “excess” assets, as happened during the 1990s when seemingly “overfunded” plans increased benefits.  A lower assumed return would also reduce the incentive for plans to invest aggressively.  And using a near riskless rate of return would inspire confidence in the reports of state and local plans by reporting liabilities in accordance with the latest finance principles. 

The argument for using a riskless discount rate, however, pertains to reporting – investing and funding are different issues.  Discounting the stream of future payments by the riskless rate does not mean that plans should hold only riskless assets.  A number of considerations suggest that they should continue to invest in equities.  For example, if the future resembles the past, the cost of funding pension liabilities will be lower than with an all-bond portfolio.

Determining funding contributions is a trickier issue.  Academic models suggest the calculation should use a riskless rate.  But contributing based on the riskless rate while investing part of the portfolio in equities produces ever growing funded levels.  That outcome may sound great, but one of two things will happen.  Politicians will raise benefits, increasing the commitments of public plans.  Or they will reduce the contributions for successive generations, creating serious equity issues.  Calculating contributions based on the long-run expected rate of return is probably the least bad option and does not conflict with using the riskless rate for reporting purposes.  

That said, the long-run returns assumed by state and local plans – currently the average is 7.6 percent – are too high.  These rates should be reduced to 6 percent.  In 2015, CalPERS’ board took steps that could reduce the rate from 7.5 percent to 6.5 percent, but the transition could take more than 20 years.  

Public plans could reduce a lot of the criticism of not using a riskless rate for funding if they adopted more realistic return assumptions and explicitly recognized the risk that they hold in their investments by establishing some risk-sharing provisions if things turn out badly.  Articles like “Sour Surprise,” however, do not help the process one iota.