Comparing Federal and California State Retirement Exposures
Californians may be accustomed to living with the specter of a public pension crisis. But the federal government’s problem with its retirement system – including Social Security – is far worse, and yet none of the three remaining major-party candidates for president has a plan to do anything about it.
The California Policy Center generally focuses on state and local issues. But with just days left before California’s June 7, we offer this comparison of California and federal exposure to pension liability.
State Retirement Expenditures
According to the governor’s May Budget Revision, the state will make a total of $8.1 billion in pension contributions during the 2016-2017 fiscal year. This amount represents a sharp increase from the current fiscal year level of $7.1 billion and fiscal 2014-2015 contributions of $6.3 billion. (These numbers exclude Other Post Employment Benefit payments.)
The rapid increase is attributable to lackluster stock market performance, more conservative actuarial assumptions implemented by CalPERS and a teacher’s pension reform that increased the state’s responsibility for CalSTRS. However, even the newly increased contribution levels are unlikely to resolve chronic underfunding in both CalPERS and CalSTRS because these two systems assume a 7.5% annual rate of return, which seems unrealistic in today’s slow growth, low interest rate economy.
Figures 1 and 2 provide a longer term perspective on the growth of state pension costs. These graphs go back to the 1999-2000 budget year when the governor signed SB 400, a bill that provided a large, retroactive increase in pension benefits. In that year, pension contributions were only $1.2 billion.
Because of inflation and the growth of the state economy, it may be more helpful to look at state pension contributions in relation to some broader economic indicator. In previous CPC studies, we have shown pension costs as a percentage of overall government revenue – identifying a number of California cities and counties that devote over 10% of their income to retirement plan contributions.
The state’s position is much better than that of the most burdened counties and cities. In 2014-2015 (the last year for which audited financial statements are available), $6.3 billion of pension contributions represented 2.29% of total state revenues – including general fund revenue, other governmental fund revenue and business type activity revenue – which totaled $276 billion. We project that this ratio will rise to about 2.76% in 2016-2017.
For those interested in general fund statistics only, pension contributions accounted for 5.57% of general fund revenue (on a budgetary basis) in 2014-2015 and are projected to rise to 6.49% in 2016-2017. These ratios overstate California’s pension burden, because many employees are compensated with resources outside the general fund.
On the other hand, some California state spending effectively subsidizes pension costs incurred by city, county, school districts and special districts. For example, most of the state’s $87.6 billion education budget for 2016-2017 will be distributed to local educational authorities, which will use some of these funds to make employer contributions to public employee pension systems.
As Ed Ring reported in a recent CPC study, total California government employer pension contributions in 2013-2014 were $21.2 billion. While only one quarter of this total was directly paid by the state government, some portion of the local government share would not have been made in the absence of state aid payments.
Ring’s report also offers some insight into how much state pension contributions would have to rise if more realistic return assumptions were used. For example, if pension funds used a 5.5% return assumption, pension fund contributions would have to triple from current levels.
The vast majority of federal retirement expenditures take the form of Social Security benefits. Because most American workers are eligibility for Social Security, the program is quite large. In the current federal fiscal year, Social Security expenditures are projected to be $911 billion or just over 27% of federal revenues. About 83% of these costs take the form of retiree and survivor benefits, 16% goes to disabled workers and under 1% covers administrative expenses.
Each year, the Social Security Board of Trustees publishes an actuarial report. The report includes short- and long-term projections, with an emphasis on the status of the Social Security trust fund. The latest report shows that the trust fund contained about $2.8 trillion in assets at the end of calendar year 2014. The report also projects that the trust fund will be exhausted in 2034 based on a set of intermediate cost assumptions. The report also includes projections based on two alternative scenarios: one reflecting higher-cost assumptions (such as greater longevity) and lower cost assumptions. Under the high-cost scenario, the trust fund would be exhausted by 2030, while under the low-cost scenario the trust fund maintains a positive balance throughout the report’s 75-year projection horizon.
Although discussion of Social Security often revolves around the trust fund, this emphasis is misplaced. Unlike CalPERS or CalSTRS, the Social Security trust fund does not contain real assets. Instead, it holds special-issue U.S. Treasury bonds. Since the trust fund is part of the federal government, its assets are merely IOUs issued by its owner. The situation is analogous to an individual removing money from his piggy bank and replacing it with a note showing the amount he plans eventually to put back. This may be a good commitment device, but any financially knowledgeable third party would not consider the note a meaningful asset.
One might argue that the Treasury bonds in the trust fund represent a claim on federal assets, but as shown in its latest audited financial statements, the federal government has a negative net position. Total federal assets of $3.2 trillion are easily exceeded by $13.2 trillion of federal debt securities held by the public and $8.2 trillion of other liabilities. So the IOUs held by the Social Security trust fund compete with claims held by many external parties for a relatively small pool of federal assets.
While the trust fund assets are not economically meaningful, they do have a legal significance – but even that is less than meets the eye. Under current law, if the trust fund is exhausted, benefit payments must be immediately reduced so that they are equivalent to Social Security revenues, which mostly derive from Federal Insurance Contribution Act (FICA) taxes paid by employees and employers. Under the trustee’s intermediate scenario, benefits would fall to 79% of the then-current level when the trust fund is exhausted in 2034.
However, this sudden, sharp reduction is extremely unlikely. Given the large number of Social Security recipients, the high voting propensity of older voters and the power of AARP, the benefit cut would almost inevitably be reversed, with additional costs borne by the general fund. There is a recent precedent for general fund transfers of this type: when Congress temporarily reduced FICA taxes in 2011 and 2012, the loss of trust fund income was offset by general fund transfers.
Rather than view Social Security through the trust fund prism, its fiscal impact is better understood in terms of its net impact on the consolidated federal budget. In other words, we should look at the difference between Social Security revenues and expenditures. The trustee report includes interest on the Treasury bonds held by the Social Security trust fund, but this notional income should be disregarded: the interest is paid and received by the same entity, the federal government.
Figure 3 shows Social Security’s net cash flow in constant dollars back to 1957. Projected revenues are depicted by three lines, with shaded areas in between. The middle line reflects the trustee’s intermediate assumptions, with the low cost and high cost scenarios shown by the lowest and highest lines respectively. As the chart shows, program revenues and expenditures were roughly equal for the first three decades. Between the late 1980s and the last decade, revenues exceeded expenditures, often by large margins. In the late 1990s, this surplus helped balance the federal budget; later, it offset budget deficits that developed under the George W. Bush Administration.
Increased disability insurance claims associated with the Great Recession and the beginning of baby boomer retirements ushered in a series of negative net balances beginning in 2010. These deficits are expected to continue under all three trustee scenarios, and to become quite large under the intermediate and high cost assumptions. By 2040, the shortfall reaches $371 billion under the intermediate scenario and $610 billion under the high cost scenario – in 2015 constant dollars.
Unprecedented deficits of this magnitude have very serious implications for the federal budget, especially when combined with escalating Medicare and Medicaid costs. Last year, the Congressional Budget Office projected that the ratio of publicly held debt to GDP will increase from 74% currently to 107% by 2040.
Federal Employee Retirement Programs
The federal government also has a large number of employees and retirees eligible for defined pension benefits. According to its latest annual report, the Civil Service Retirement and Disability Fund, paid $81 billion of retirement benefits in fiscal year 2015, or 2.49% of federal revenues. The system reported an Unfunded Actuarial Liability of $804.3 billion and Assets of $858.6 billion, implying a funded ratio of only 51.6%. Further, the fund’s assets are almost entirely invested in U.S. Treasury securities. Similar to the Social Security Trust Fund, the economic meaning of these investments is questionable.
The Defense Department also provides retirement benefits. The latest available actuarial report shows $54.8 billion of benefits paid in fiscal year 2013 and a 35% funded ratio. Last year, President Obama signed a Defense Authorization Bill containing a military pension reform. Instead of a straight defined-benefit plan, new recruits joining the armed forces after January 1, 2018 will be placed in a hybrid plan containing a 401(k)-style component with an employer match. The defined benefit component will remain, but will be reduced by 20%. This reform should improve the program’s funded ratio, but won’t reduce military pension costs by very much – if at all. Under the current system, service members must remain in the military for 20 years to become eligible for pension benefits. Vesting in federal matching payments under the new defined contribution plan will begin after two years.
Comparing the Federal and State Governments
Overall, the federal government has much greater exposure to pension costs that does the state of California. Civilian and military pension benefits consume a proportionately larger amount of the federal revenue than the share of total state revenue absorbed by CalPERS and CalSTRS contributions. Further, the federal government is responsible for providing most American workers pension benefits through Social Security, which absorbs more than a quarter of federal revenue and has an inadequate level of pre-funding, even if one considers Treasury securities to be an acceptable investment vehicle for a federal retirement system.
That said, it is worth considering some advantages the federal government has relative to the state in dealing with pension costs. First, the U.S. constitution does not provide a right to accrued benefits. In an emergency, Congress and the president could cut or terminate benefits to Social Security recipients, federal civilian retirees or veterans. This is not the case for the state of California.
As Alexander Volokh points out: “In California, when a public employee begins work, he not only acquires a right to the pension accumulated so far — presumably zero on the first day, and increasing as he works longer — but also the right to continue to earn a pension on terms that are at least as generous as the ones then in effect, for as long as he works. And if pension rules become more generous in the future, then those more generous terms are the ones that are protected.”
As I discussed earlier, I do not expect Social Security benefits to be reduced when the trust fund runs out, so the fact the Social Security recipients do not have access to the courts may be a distinction without a difference. But it is still true that the federal government has a tool for reducing benefit costs – especially during a fiscal emergency – that is not available to the state.
Further, there is a widespread belief that the federal government is less vulnerable to a fiscal emergency than California because it has access to the printing press. In other words, if the federal government cannot obtain enough tax revenue to pay retirement benefits, it could do so with newly created money.
While this is a fair distinction, it comes with a couple of caveats. First, at the national level, money creation has become the role of the Federal Reserve, which has some degree of political independence. Strictly speaking, the president cannot order the Fed Chair to create money. Second, U.S. state and local governments have been able to create circulating IOUs in the past. During the Depression, numerous cities issued scrip, while, in 2009 the state issued IOUs to vendors amidst a budget crisis. These IOUs were eventually traded on a secondary market.
These caveats notwithstanding, it is true that a central government controlling an international reserve currency does have more fiscal flexibility than a state which is legally obligated to balance its budget each year. So the federal government’s ability to absorb pension obligations is greater than California’s. This is fortunate, because the federal governments exposure is so much greater.
We have seen that both California and the federal government face high and rising pension costs, and that each has not fully accounted for these obligations. The drivers of these problems are similar, and are duplicated throughout much of the developed world: retirement of the large baby-boom generation, increased longevity and a failure of political institutions to deal effectively with long-term problems.
While the specific policies to improve pension sustainability differ across jurisdictions, the basic ideas are similar. These include:
- Paring back benefit levels, especially for the most highly paid, most affluent beneficiaries.
- Increasing retirement ages and then indexing them to longevity.
- Increasing employee contributions.
- Replacing deceptive accounting techniques and rosy actuarial assumptions, with conservative, fact-based financial reporting.
Finally, libertarians and fiscal conservatives working on these issues should re-evaluate their tactics. In 2005, George W. Bush’s strategy of using the impending Social Security crisis to justify a partial switch to personal accounts was roundly rejected by Democrats and Republicans alike. While many of us in the public-sector pension reform community like the idea of 401ks, we need to understand that employees – especially those who are risk-averse or financially unsophisticated – prefer defined benefits. Rather than attacking defined-benefit plans, we should try to fix these plans so that they don’t bankrupt the governments that offer them.