States usually report unfunded liabilities on their pension plans of roughly $500 billion. However, many economists, joined by a growing number of actuaries, have been very vocal about the fact that this is a gross underestimation of the problem. For instance, economists such as AEI’s Andrew Biggs, the Mercatus Center’s Eileen Norcross, the Kellogg School of Management’s Joshua Rauh, and the University of Rochester’s Robert Novy-Marx have noted that unfunded liabilities could be as high as $3 trillion. This difference is the product of the difference between market valuation of pension liabilities (the one economists favor) and government accounting valuation of pension liabilities. Much of the debate between the two schools boils down to a disagreement over the value of the discount rate. To understand what the discount rate is and why it is so important, I will quote once again the work of economist Douglas Elliott of the Brookings Institution. He explains:
A dollar today is worth more than the promise of a dollar a year from now, even if you are sure the promise will be kept. A dollar today could be invested and would therefore be worth more in a year. Alternatively, a dollar today could allow you to avoid borrowing a dollar from someone else, on which you would have to pay interest. If the promise is less than certain of being kept, then receiving the money up-front becomes even more valuable in comparison to the promise. Economists and other experts dealing with long-term promises use a “present value” approach to reckoning the value in today’s dollars of future payments. This involves estimating the future payments and then reducing the payments in each year by a discount factor based on: (1) the number of years from now until the payment and (2) the chosen interest rate, known as a “discount rate” . . . The discount rate is absolutely crucial to measuring the cost in today’s dollars, since the pension payments are spread over so many years, with the average payment typically occurring decades into the future. The compound effect of a discount rate being applied over so many years means that, for example, one dollar received 20 years from now would be worth 46 cents if discounted at a 4% rate or 21 cents at an 8% rate.
Economists argue that the discount rate should be 4 percent, yet, in most cases, states use an 8 percent rate.
The great news is that the Congressional Budget Office has endorsed the market valuation of pension liabilities.
By accounting for the different risks associated with investment returns and benefit payments, the fair-value approach provides a more complete and transparent measure of the costs of pension obligations [than GASB accounting.] . . .
[E]ven over very long periods, higher returns on risky investments are not a sure thing. Although a strategy of investing in relatively risky assets produces higher returns on average, compared with a strategy of investing in relatively safe assets, there is no certainty of obtaining higher returns for a particular portfolio or over any given period of time. Those potential higher returns would be obtained by imposing added risk on taxpayers — particularly the risk of experiencing severe losses during protracted economic downturns — which represents an equal and offsetting cost. Such long-run risk, which investors view as costly, is reflected in the premium that investors require to bear it. Fair-value measures account for that cost.
Over at Public Sector Inc., Josh Barro writes a good post about what the implications would be if states adopted market valuation of pension liabilities:
CBO notes, correctly, that there would be certain challenges associated with a shift to fair-value accounting. One is that such a shift might lead pension funds to require significantly greater pension contributions from participating governments in the near term. That’s true, but it would really be a reflection of the fact that governments have long been understating the cost of public employee pensions (and got lucky with an outstanding 20-year run of equity returns from 1988 to 2008.)
They also note that a fair-value standard could make required pension contributions more volatile. That’s also true, but (as CBO also notes) pension funds do not have to use the same figures for reporting funding ratios and determining required contributions. As I discuss in my paper on pension transparency, pension funds should report their funding status based on the market value of assets, but should continue to use a smoothed asset value to determine contributions.
States’ fiscal outlooks would be darker for sure, and it would make it obvious that hard choices need to me made. Continuing to hide the size of the unfunded liabilities does not make them disappear, as the Illinois example shows.