Underfunded pension systems are one of the biggest challenges facing state and local budgets, resulting in an estimated $1 trillion to $5 trillion in total U.S. pension debt, depending on the practice of measurement used. In an effort to examine how we arrived in this predicament and what steps can be taken going forward to ease the overwhelming financial burden on governments, the Retirement Security Initiative (RSI) will talk with leading fiscal and pension experts to get their opinion and analysis on the funding crises. The below Q&A with Donald Boyd, director of Fiscal Studies at the Rockefeller Institute of Government, is the first in our series.


RSI: In the January 2014, Rockefeller Institute report, Strengthening the Security of Public Sector Defined Benefit Plans, you attribute inaccurate financial reporting, investment risk incentives and lax rules as allowing public sector pension underfunding to develop. In your opinion, how have these risks grown and why are potential consequences of these funding flaws greater now than before?

DB: The risks have grown enormously – risks to taxpayers; to people who benefit from government-financed education, health care, police, courts, roads or water systems; and to people who work for or have retired from government.

As measured by the Federal Reserve Board, public pension plan unfunded liabilities are at a near-record $1.9 trillion (Figure 1). That is conservative; alternative estimates, discounted using risk-free rates, exceed $5 trillion (1). This debt will be paid mostly by current and future taxpayers or by cuts in services.

Figure 1. Unfunded state and local government pension liability, relative to taxes, is near a record.

In an effort to reduce these liabilities, public pension funds are exposing taxpayers and others to much greater risks than before. The risks have gone up for two reasons.

First, public pension funds are larger now relative to the economy: In 1975, public pension fund investible assets were about six percent of gross domestic product; today, they are approximately 20 percent. This asset growth reflects the aging public workforce and maturation of pension funds.

Second, as nominal investment returns on safe assets fell over the last 20 years, it became more difficult to achieve assumed investment returns of 7 to 8 percent. Public plans moved into riskier assets rather than reduce earnings assumptions (2). If this risk-taking is successful, taxpayers and other stakeholders will benefit; if not, they will be worse off.

Rules and institutional arrangements encourage public pension funds to take investment risk. U.S. public funds, unlike other funds, calculate liabilities and actuarially determined contributions using discount rates based on what they assume their portfolios will earn. High assumed rates keep reported liabilities and requested contributions low. This encourages investing in risky assets that can justify high expected returns. There are few constraints on this assumed rate except for what prudence will allow, and prudence is quite generous. A recent statistical analysis concluded that as Treasury yields fell, U.S. public plans increased risk-taking in ways that U.S. private plans and plans in other countries did not, and that “gradually, U.S. public funds have become the biggest risk-takers among pension funds internationally.” (3)

This heightened risk-taking with a larger asset base means that taxpayers and other stakeholders bear far more risk now than in the past. According to a forthcoming Rockefeller Institute report, a single-year investment shortfall of one standard deviation (a common measure of risk) would now be about three times as large relative to state and local government tax revenue as in 1995 and 10 times as large as a decade earlier (4). A one standard deviation shortfall has about a one in six chance in any given year (5).

Such a shortfall would be about $427 billion. If amortized (spread out with interest) in a common fashion, it would require increased contributions from governments of about $23 billion now, rising 3 percent annually for 30 years, after which it would be paid off (6). This is equivalent to about a 24 percent cut in highway capital spending for 30 years – the consequence of a single year of moderately bad investment returns (7).

Whether taxpayers really want public pension funds taking risks of this magnitude on their behalf – risks that affect not only them but their children – is an open question.


RSI: Your recent policy brief with Yimeng Yin, Public Pension Funding Practices, draws on simulations to show that there is a one-in-six chance that a typical pension fund will fall below 40 percent funding sometime in the next 30 years. What, if anything, can we do to change this scenario?

DB: A typical 75 percent funded pension fund runs about a one in six chance of falling below 40 percent in the next 30 years, assuming an average portfolio and a common method of funding. Portfolios are risky enough that even with full payment of contributions, bad returns could drive funding this low. This is a crisis level: only a few plans are below 40 percent and they and their governments are in deep crisis (e.g, Illinois and Kentucky). We used a simulation model to estimate this, and I think we were conservative. A one in six chance of a crisis seems like a big risk to me.

Using risky assets to fund guaranteed benefits creates additional concerns. People who pay taxes and benefit from government services and infrastructure must be willing to put up with ups and downs in taxes, services and infrastructure quality in response to ups and downs in investment markets. No amount of funding smoothing can prevent this and also keep funds solvent. If we’re all ok with that, it’s not a problem.

The way to reduce this risk is for pension funds to lower earnings assumptions and move into investments similar to their bond-like liabilities. That means they’ll expect lower returns and be more confident that they’ll get them. Contributions will be much more stable, but at a big cost: It requires the very unpopular action of requesting much higher contributions from governments now, which could damage public and political support for defined benefit plans. Benefits that weather any reaction would be more secure.

While some de-risking is going on, I don’t expect a big shift. Incentives in the system encourage risk taking, and there are no rules to counter those incentives - it’s the Wild West out there.


RSI: Throughout your work, the resounding conclusion is that taxpayers—both current and future—and government employees and retirees ultimately bear the burden of investment risk. How do assumptions and gimmicks of pension administrators, as well as the different methods of funding policies and practices to repay shortfalls, vary in impact to taxpayers and employees?

DB: It is not just pension fund administrators; it is executive and legislative branch politicians, and government finance officials, too. And they often have the tacit support of unions, taxpayer groups, advocates for services and others who would rather see taxes used for current services than for services delivered in the past. All share an interest in smoothing out governmental contributions and in pushing off the cost of any shortfalls to the future.

The motive for stable contributions is understandable and appropriate: there are costs to society from instability in taxes and services. Unfortunately, we know from our modeling and from the work of economist Andrew Biggs that imposing stability on contributions when investment portfolios are extremely volatile is dangerous to the long-run health of pension funds and to future taxpayers. The second motive – the desire to push shortfalls to the future – is understandable but comes at the expense of future taxpayers.

Pension plans and governments use two main mechanisms to smooth out the consequences of investment shortfalls and overages. I focus primarily on shortfalls because they can cause distress that gains do not.

First, most plans acknowledge investment gains and losses over multiple years, rather than all at once. A common approach is to recognize one fifth of a gain or loss over each of five years. This pushes the full pain of paying for investment shortfalls until year six and beyond. If tax revenue is likely to bounce back sharply from the depths of a recession – and often it will – then this defers the greatest contribution increases until the revenue recovery, which can make sense. (This bounce-back is not assured: tax revenue recovered very slowly after the last two recessions.) The downside is moral hazard: current politicians and financial managers take risk, but bear relatively little consequence and have little incentive to minimize risk.

The second mechanism is repaying losses, once recognized, over periods that can be 30 or more years. Some methods start with payments so low that they allow unfunded liabilities to grow for years before they begin to be paid down. Other methods pay down liabilities more quickly. The most deeply underfunded plans tend to use methods that push repayments far into the future. Again, this creates a moral hazard: it softens the near-term consequences of risk-taking and thereby encourages risk. And it pushes the potential costs and benefits of uncertain investment returns to future taxpayers who did not benefit from services they may have to pay for.


RSI: You estimate that state and local government contributions to pension funds have increased $0.59 for every $1.00 in additional tax revenues between 2007 and 2015 and by about 40 cents for every dollar of available own-source revenue. Yet many public pension plans fall short of their actuarial determined contributions, with 33 plans underpaid by $100 million or more in 2014, according to your most recent presentation on funding risks. If this trend continues, what is your prediction for the future solvency of public pensions?

DB: Most public plans are nowhere near insolvency – out of assets and unable to pay benefits as they come due. But a plan is in trouble long before it is insolvent. If funding falls low enough, the needed contribution increases will become huge – perhaps so large that the resulting tax increases, service cuts, or infrastructure neglect will drive economic activity out of state. Politicians may understandably balk at paying contributions that large, despite legal and moral commitments to pay promised benefits honestly earned. If politicians do balk, insolvency could become possible and the courts would have to sort out who will lose the most and who the least. Governments and plans must act long before insolvency.

Will governments act? The political and funding risks differ from place to place. Some governments are underpaying actuarially determined contributions by a great deal. Other governments are paying actuarially determined contributions that are too low to keep unfunded liabilities from growing, as we know from work by Moody’s, Pew and the Rockefeller Institute. Some governments and their plan partners are behaving well.

Further increases in actuarial contributions are coming. Most pension plans fell short of their investment assumptions in their 2015 and 2016 fiscal years, totaling about 10 percent across the two years. That suggests increased underfunding of $350-$400 billion nationwide. Even if paid off slowly, that would be roughly the equivalent of, for example, cutting highway capital spending by 22 percent for the next 30 years or raising all sales taxes in the nation by 0.4 cents. That sounds like a lot of political pain.

When these increases come, will governments pay, will they try to cut benefits already earned or will they underpay and let plans drift further downward hoping that investments in risky assets will bail them out? A little of each I suspect. Even though solvency is not a near term risk, there is plenty to worry about.


RSI: Our organization believes that all workers deserve safe and secure retirements and retirement plans should place employees on a path to secure retirement, yet defined benefit plans, used by the majority of public pension plans, have inherent flaws that could lead to dangerous consequences for workers. You are currently working on a two-year project on modeling and disclosing public pension fund risks. Do you have recommendations so far to better guide the funding and disclosure of defined benefit plans?

DB: Let me offer observations rather than recommendations. I said earlier that it’s the Wild West because there are virtually no rules - no police – for public pensions:
• Public plans are not subject to the Financial Accounting Standards Board (FASB) standards that govern calculation of financial statement liabilities and expenses for private plans. Instead, Governmental Accounting Standards Board (GASB) standards generally allow them to calculate liabilities and expenses based upon plan-chosen investment return assumptions.
• GASB, unlike FASB, is not overseen by the Securities and Exchange Commission - no external authority constrains GASB standards (8).
• Public plans are not subject to the private sector rules in the Employee Retirement Income Security Act (ERISA) and later acts that restrict the discount rate used for funding purposes (9). While those rules have flaws, public pension plans don’t face any rules, except for a few in state-specific laws.
• Many governments have abused the rules-free environment, paying less than actuarially determined contributions, which in turn are too low. The most troubled plans invariably have had governments that underpay substantially, often when times are tough but sometimes habitually.
• The Actuarial Standards Board provides guidance to actuaries in developing interest rate assumptions, but that has not prevented actuaries and public pension boards from choosing assumptions that have contributed to this risky situation (10).

Even if we find money to fix current problems, unless we arrange police for public pensions we likely will wind up in this situation again.

Can the Wild West be tamed? I do not believe that disclosure of liabilities, costs and risks could be enough to change the behavior of pension plans and governments. Rules and other institutions will be needed:
• Liabilities and annual costs in government and pension plan financial statements should be measured properly, with discount rates that reflect the characteristics of the liability. This would not directly affect plan funding, but would provide information that could affect behavior. Understating the annual cost of pension benefits can lead governments to negotiate or legislate higher pension benefits than they otherwise might, and sunlight might discourage that.
• Discount rates used to determine contributions should be based on market interest rates rather than a plan’s portfolio. This could reduce the incentive to invest in risky assets to keep requested contributions low.
• Plans and governments should disclose this information clearly, along with measures of risk, in comprehensive annual financial reports and in municipal bond disclosures.
• Constraints should be imposed upon the earnings assumption used for funding to further counter incentives to assume high returns. This would make it more practical for plans to invest in assets that are similar in character to their liabilities.
• Governments should be required to pay actuarially determined contributions (11).

These changes would cause huge increases in contributions by state and local governments. That is not the fault of the changes, but of the situation we are in. Resulting tax increases probably would weaken public and political support for defined benefit pensions. Almost no one wants rules, particularly these rules, imposed on public pension plans – not the plans, not governments, and likely not unions or current taxpayers. Thus, they are extremely unlikely, absent a trigger event such as nonpayment of benefits. No one wants that.

One thing plans can do on their own, at political risk to boards, is gradually reduce their risk taking at every opportunity.

None of these changes would fix unfunded “legacy” liabilities that governments currently face. Most of those liabilities have strong legal protections and properly must be paid. But elected officials and others need to think about how to fix the system, so that the problem does not recur.


  1. The Bureau of Economic Analysis, which prepared the estimates, discounted liabilities of recent years with a 5 percent rate. Risk-free rates, which are more appropriate, are much lower and currently lead to estimates over $5 trillion. Updated estimates prepared by Josh Rauh of Stanford University were reported in Nicole Bullock, “The Crumbling Assumptions of US Public Pension Plans,” Financial Times, August 26, 2016, https://www.ft.com/content/456172b4-6b11-11e6-ae5b-a7cc5dd5a28c. I confirmed via email with the article author that the estimates were prepared by Professor Rauh.
  2. There have been minor downward moves in earnings assumptions, but they are small compared to the movements in market interest rates.
  3. Aleksandar Andonov, Rob Bauer, and Martijn Cremers, “Pension Fund Asset Allocation and Liability Discount Rates,” Available at SSRN 2070054, March 2016, http://papers.ssrn.com/sol3/Papers.cfm?abstract_id=2070054.
  4. I use taxes as a rough proxy for state and local governments’ ability to pay. They have other sources of income they can use to pay pension contributions, but these are not easy to measure consistently over time and are much smaller than taxes. I estimate that taxes are used to pay for roughly 80 percent of pension contributions. Over longer periods state and local governments can raise or lower taxes, and so measures of economic capacity are important. If the risk were measured relative to gross domestic product, the conclusions would be essentially the same.
  5. The 1 in 6 statement assumes normally distributed investment returns.
  6. Based on 30-year amortization as a level percentage of pay. I ignore asset smoothing for purposes of the example. Assumes a shortfall of $426.5 billion, a 7.5 percent interest rate, and 3 percent annual growth in payments.
  7. Authors’ analysis of U.S. Census Bureau, 2013 Annual Surveys of State and Local Government Finances.
  8. For a good discussion, see James P. Naughton and Holger Spamann, “Fixing Public Sector Finances: The Accounting and Reporting Lever,” Harvard University John M. Olin Center for Law, Economics, and Business Discussion Paper, no. 814 (2015), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2574308.
  9. Private-sector rules have been manipulated by Congress for politically expedient purposes.
  10. See “Actuarial Standard of Practice No. 27: Selection of Economic Assumptions for Measuring Pension Obligations - Adopted September 2013 (ASOP No. 27)” (Actuarial Standards Board, September 2013), http://www.actuarialstandardsboard.org/pdf/asops/asop027_172.pdf.
  11. For one analysis of this issue, see Natalya Shnitser, “Funding Discipline for U.S. Public Pension Plans: An Empirical Analysis of Institutional Design,” Iowa Law Review 100 (2015): 663–714.