Municipal and state governments provide input to the organization that creates their accounting standards. Such input by stakeholders can be helpful, but their influence has produced some accounting rules that diverge from both economic reality and private sector rules. These deviations allow governments to understate budget liabilities, including pensions plans, and put at risk the financial health of states and cities across the United States.
This article is part of a series examining the financial conditions of local governments in the United States and the forces that shape them. We will publish a new addition to the series every Monday for the next few weeks.
If someone asked whether we can accurately calculate the present value, and thus accounting liability, of an annuity benefit that pays $100 per month for someone currently age 65, the answer would be a resounding yes. After all, insurance companies and sponsors of defined benefit pension plans have been providing these types of benefits since the 1930s.
Perhaps surprisingly, however, the reported liability for the payer associated with this annuity benefit currently differs depending on the entity that provides the benefit. For a private-sector employer or an insurance company or similar financial institution, the liability will be roughly the same. However, for a public-sector employer, the reported liability will be substantially less. The reason the same annuity benefit has a lower liability for public-sector employers is not because they are able to provide annuity benefits more cost effectively than insurance companies or private-sector employers. Instead, it is due to non-sensical government accounting practices that have serious implications for government budgeting and explains why some cities today are in financially precarious positions.
Accounting standards, at least in part, are influenced by regulatory capture—famously introduced into the economics literature by George Stigler, who forcefully argued that “regulation is acquired by the industry and is designed and operated primarily for its benefit.” In the governmental setting, what this means is that the entity responsible for developing accounting rules, the Governmental Accounting Standards Board (“GASB”), a private non-government organization, develops rules that are designed, at least in part, to meet the preferences of state and local governments. The preparers of financial statements are clearly important stakeholders and their views should be considered. However, when the rules that are ultimately adopted place too much emphasis on the preferences of preparers, the inevitable outcome includes distortions that understate both costs and cost variability—as is the case with public sector pension benefits. The design of governmental accounting rules presents a case of governments capturing their own rules, however unmalicious the intent may be.
To illustrate this distortion, let’s return to the example of an individual currently age 65 who has earned a lifetime benefit of $100 per month. Private-sector employers are required to report a liability equal to the settlement cost of the promised benefit, which is effectively the cost of purchasing an annuity benefit from a highly rated insurance company. For the typical company in 2020 (before COVID), this entailed discounting the expected payments at approximately 3%. The discount rate captures the risk associated with the promised payments and is based on prevailing AA corporate bond rates, in which insurance companies would typically invest the annuity premiums to back the annuity contracts. For this example, the company would report a liability or present value of around $21,500 for the $100 monthly benefit. We can calculate the present value using a standard industry equation that factors in the monthly benefit, discount rate, and number of payments the annuity promises.
Importantly, if that employee did not participate in an employer-sponsored plan and instead went to an insurance company to purchase an annuity that paid $100 monthly for the rest of their life, the cost would also be around $21,500. The liability associated with the pension benefit is effectively the same whether it is provided by an insurance company or by a private sector employer, an outcome that seems entirely reasonable from an economic perspective as the underlying economics of the payments are also the same.
However, even though the underlying economics of the payments are also the same in the public sector, the reported cost in the public sector is much lower. Unlike private-sector reporting, governmental reporting allows the present value of the payment stream to be determined by discounting the anticipated payments using the expected return on the investments held by the pension trust. These investments are primarily composed of equity securities, and as a result, the typical government was using a discount rate of around 7% in 2020 to match the long-term average return of the stock market. A higher discount rate produces a lower present value and thus liability for the same $100 monthly payment.
However, this higher discount rate does not reflect the risk associated with the promised payments, and therefore it substantially understates the cost of providing those payments. From the participant’s perspective, the promised payments are effectively the same whether they are provided by a private sector employer, an insurance company, or a public sector employer. Therefore, the liability should also be the same.
The liability is not the same because the process used by public sector employers is equivalent to assuming that the pension trust will earn guaranteed returns of 7% per year, something that would far exceed the long-run performance of every asset manager, including the fake performance data provided by Bernie Madoff. This implicit assumption means that even though all insurance companies and private sector employers would agree that a $100 per month life annuity at age 65 has a liability of $21,500, a government would report a liability of only $14,000, a savings of about 35%.
The use of a different discounting methodology is almost entirely due to the preferences of governments, as it ensures that the pension cost is understated relative to economic reality and that the accounting liability is relatively stable over time.
The liability is understated because it doesn’t reflect the guaranteed nature of the promised payments. In some states, such as Illinois, the pension benefits are guaranteed by the state constitution, meaning that the appropriate discount rate is even lower than AA corporate bond rates, and is likely close to the risk-free rate. Using such a rate would increase costs substantially, thus either requiring the government to increases taxes or make cuts to other spending programs. Governments, like most entities, prefer understated costs for long-term obligations because it provides more flexibility to expend resources on other initiatives in the short term and reduces the need to increase taxes, something that most elected officials want to avoid.
The liability is stable because the GASB methodology results in a discount rate of 7% that is relatively unchanged over time. Governments prefer stable costs because it facilitates annual budgeting. The use of market rates expressed in AA corporate bonds results in costs that vary substantially over time because prevailing interest rates change, and as they do they change the actual costs of the benefits being provided. Over the last few years, private sector employers have reported large one-time gains (as interest rates increased) and one-time losses (as interest rates decreased). Governments are immune from these market fluctuations for financial reporting purposes. Since they are not immune from an economic perspective, they can still be downgraded by credit rating agencies who do not rely on the unrealistic pension liability reported by governments. The appearance of financial statement stability is not economic reality.
These two distortions, which clearly align with the preferences of governments, generate substantial fiscal problems.
My research shows that using understated pension costs means that governments are using artificially low values to determine the cost of each employee. For example, in the example above using the $100 monthly annuity benefit, the benefit cost of one employee will appear to be $7,500 less than reality. If the understated pension costs and salary are $75,000 per employee, then a government with a budget of $825,000 will believe it can hire 11 employees. However, if it used the correct pension costs, it would only be able to hire 10 employees. My research shows that understated pension costs are associated with additional hiring, which is indicative of a negative feedback loop where current understated costs exacerbate future public sector fiscal stress. In addition, public sector pension plans tend to hold riskier investment portfolios than comparable private sector plans. While this additional risk exposure can lower current costs, it does so at the risk of substantially higher future costs.
While the GASB has recognized the problems with pension accounting, it has not yet made the necessary changes. GASB Statement 67, the last standard to address problems with the discount rate, had virtually no effect on the selection of the discount rate when it was issued more than 10 years ago. Through substantial lobbying on the part of governments, my research shows that the final rules included provisions that allowed governments to circumvent the use of a lower discount rate with simple actuarial and legislative adjustments.
An annuity benefit has clear measurable economic value, and the cost of providing that benefit should not be substantially lower simply because it is provided by a governmental entity. While updating the accounting for pensions should be a priority, a greater priority should be adjusting the process for producing governmental accounting rules to ensure that they do not continue to be overly influenced by the preferences of governmental entities. The cost to governments of readjusting pension accounting standards may be high in the short term, but the cost will be much steeper in the long term if they do not.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
