CalSTRS is 79% Funded—Well, Maybe

CalSTRS earned 13.9% on investments for the year ended June 30, 2026, up from 8.5%, and called it a significant step toward full funding. The National Association of State Retirement Administrators said this is “a marathon, not a sprint.”[1]

Thirty years in local government finance taught me to distrust a figure that measures the wrong thing. A return measures the portfolio, not the obligation. In a defined benefit plan the employer promises a lifetime payment stream and owes whatever the investments and plan contributions fail to produce. With a public employer, that means the taxpayer is at risk for the life of the plan. A 13.9% investment return year means only that the risk did not show up this year. A deeper look reveals that CalSTRS is not performing well in this tax-funded marathon.

Contribution Rates Doubled and the Hole Got Bigger.

Three parties pay into CalSTRS: school districts, teachers, and the state. District and teacher contribution rates had been frozen for more than a generation, at 8.25% of district payroll since 1986 and 8% of teacher pay since 1972. The state’s rate, 3.041%, was lower than it had been in 1998, cut by earlier legislation. The 2014 rescue legislation, the CalSTRS Funding Plan,[2] raised all three: districts to 19.1%, teachers to 10.25%, the state to 8.328%.[3] More than half of the district rate now goes to obligations for teaching already done, not to the cost of teaching today.[4]

The invested funds performed well: the five-, ten-, twenty- and thirty-year returns all met or beat the 7% target. In fact, ten-year returns reached 9.4%. Yet the gap between what CalSTRS has and its obligations was $73.7 billion in 2013, and it is now $82.0 billion.[5] How does a plan double the money going in, beat its investment target for a decade, and end up further behind?

Two reasons, and CalSTRS discloses both. The plan was built to fall behind at the start: for eight years the payments came in below what CalSTRS’ own actuary said was needed just to keep the gap from growing, adding about $29 billion. And the board changed how the gap is measured, adding about $17.6 billion.[6]

The Board Controls the Yardstick

The 7% is not only a forecast of earnings. It is also used to quantify the pension obligations: actuaries measure a pension payable decades from now by asking how much must be set aside today to cover it, growing at 7% a year. If the plan assumes more growth, then less must be accumulated, so the reported funding gap shrinks without a dollar changing hands.[7]

The plan funded ratio is simply plan assets divided by the current cost of pension obligations. In April 2017 CalSTRS reported the plan 63.7% funded as of June 30, 2016, with a $96.7 billion gap, using a 7.25% investment assumption. The same announcement disclosed that at 7.00%, today’s rate, the identical measurement would have shown a $105.1 billion gap and 61.8% funded. Same day, same assets, same obligations. A quarter of a percentage point in the assumption moved the reported obligation by $8.4 billion.[8]

Back in 2013, CalSTRS reported itself 66.5% funded, at 7.5%, and the Legislature sized the rescue to that number. Four years later, with plan contribution rates rising every year, the measurement came in at 62.6% and the gap at $107.3 billion, worse than when the rescue began. The market had not fallen. The board had changed its assumptions.[9]

So when CalSTRS announces an eighth straight year of funding improvement, from 62.6% to 79.3% since 2017, we should consider where the count starts. It starts at the bottom that the board created through its own policy changes. Had the plan truly been 66.5% funded in 2013, four straight years of rising contributions would have moved the reported number up. Instead, it went the other way. The teachers and the pension promises did not change; the yardstick did. Measured by the standard the board itself adopted four years later, the plan was not 66.5% funded in 2013, it was closer to 62%.[10]

The Board Overrules Its Own Actuary.

Milliman, CalSTRS’ outside actuarial firm, recommends the plan’s assumptions. The Teachers’ Retirement Board decides whether and when to adopt them, a power it holds exclusively under the state constitution.[11] At the December 2010 CalSTRS board meeting a staff recommendation of 7.5% had been pending since spring. Milliman’s actuary told the board that 7.5% “would still only result in about a 50 percent chance of achieving the earning target.” The California Teachers Association’s retirement committee chair said the reduction should stop at 7.75%, a number the actuary had not studied. The board adopted 7.75%.[12] A higher target is harder to hit, so the board settled on worse-than-even odds, at a number its own actuary had never evaluated.

Seventy-nine Percent Funded is Not Seventy-nine Percent Sure

A funded ratio is not a confidence level. It says that under the assumptions the board has chosen, assets cover about 79% of the calculated liability.

Actuaries can assess how likely an assumption is to hold. In 2024 CalSTRS hired Segal, a second firm, to review the experience analysis behind Milliman’s work. Segal put the confidence level of the 7% investment assumption at 53.2%: the odds that earnings match or beat 7% over fifteen years, roughly the duration of CalSTRS’ obligations. Segal said 7% “does not include much margin for future adverse deviation” and recommended 6.75%. The board kept 7%, and has kept it for two years since.[13]

So the 7% return is the assumption, the 53.2% the odds it holds, and the 79.3% a measurement that assumes it does hold. Nobody is claiming pensions will go unpaid. The question is who covers the difference when the assumption misses, and CalSTRS supplies the answer: run the plan on more conservative assumptions and “contributions would have to increase for the state, for employers and possibly for CalSTRS 2% at 62 members.”[14] If the current assumptions hold, the plan will reach 100% funded, but will have no cushion.[15]

A Defined Contribution Plan Has Nothing to Guess

Even a board beyond reproach could not fix this. A defined benefit plan cannot be run without estimates: decades of investment returns, life expectancy, payroll growth. Honest assumptions describe the problem better. They do not remove it. Adopting Segal’s 6.75% would not have reduced the promised benefit by a dollar. It would have revealed more of the bill, addressed to the same payer. That is the design: the party that sets the estimate is not the party that pays when the estimate is wrong.

Some will object that a defined contribution plan shifts investment risk to the worker. But the teachers already carry risk. Their rate rose more than two points with no better benefit, and CalSTRS’ own report puts its newest members next in line. A defined contribution participant owns the balance and can track its value in real time. A defined benefit member is told 79% on an assumption rated 53.2%, and finds out decades later.

In a defined contribution retirement plan, the employer’s obligation is discharged the moment the contribution is deposited. No vote on an assumed return. No estimate to revise later. No recommendation overruled for political convenience. No leftover risk for a future taxpayer to inherit.

The only way for taxpayers to win at a public employee defined benefit plan such as CalSTRS is not to play.

Mark Moses is a senior fellow with California Policy Center. He has thirty years of experience in local government administration and finance. His book, The Municipal Financial Crisis – A Framework for Understanding and Fixing Government Budgeting, was published by Palgrave Macmillan.

Endnotes

1. CalSTRS news release, “CalSTRS earns 13.9% net return, exceeds benchmark in fiscal year 2025–26”, August 4, 2026. Total fund $415.4 billion at June 30, 2026. CalSTRS states the five-, 10-, 20- and 30-year returns “all surpass the actuarial assumption of 7.0%”; the five-year figure is 7.0%, at the assumption rather than above it. Keith Brainard’s “marathon, not a sprint” comment appeared in the Sacramento Bee, August 4, 2026. The 79.3% funded ratio reported alongside the return is from the June 30, 2025 valuation; the ratio reflecting the 2025–26 return will not be published until May 2027.
2. AB 1469 (Bonta), Chapter 47, Statutes of 2014, the CalSTRS Funding Plan. Before it, every contribution rate was fixed in statute and the board had no authority to raise the employer rate. At June 30, 2013 the Defined Benefit Program was 66.5% funded with a $73.7 billion unfunded actuarial obligation, and the fund was projected to exhaust its assets around 2046. Increases phase in over three to seven years, estimated at $237.7 billion in added contributions over 32 years, with the pre-July-2014 obligation to be eliminated by June 30, 2046. Source: AB 1469 as amended June 12, 2014.
3. CalSTRS Funding Plan fact sheet. Employers: 8.25% since 1986, phased to 8.88%, 10.73%, 12.58%, 14.43% and 16.28%, with temporary relief in 2019–20 through 2021–22 funded by a $2.2 billion state supplemental payment, reaching 19.10% in 2022–23. Members: 8% since 1972, now 10.250% for CalSTRS 2% at 60 and 10.205% for 2% at 62. State: 3.041% of payroll in 2013–14, a 2.017% base plus increments tied to the 1990 benefit structure, itself the product of earlier legislation that had reduced the state’s rate, now 8.328% to the Defined Benefit Program, plus about 2.5% of payroll for the inflation-protection account, a 2026–27 total of 10.828%.
4. The 19.100% employer rate is 8.000% base, 0.250% for unused sick leave, and a 10.850% supplemental rate that Education Code §22950.5 dedicates to amortizing the unfunded obligation attributable to service before July 1, 2014. The board may move the supplemental rate up or down one point a year; the total is capped at 20.25%. The June 30, 2025 valuation calculated an 18.100% employer rate for 2026–27. The board adopted 19.100%, the fifth consecutive year it has declined to reduce the rate.
5. The $73.7 billion figure was measured at a 7.5% assumption; the $82.0 billion at 7.0%. The June 30, 2025 valuation reports an actuarial obligation of $395.5 billion, an actuarial value of assets of $313.5 billion, an unfunded actuarial obligation of $81.992 billion, and a funded ratio of 79.3%. Both dollar figures are nominal, not inflation-adjusted.
6. CalSTRS’ Review of Funding Levels and Risks decomposes the roughly $15 billion net increase in the unfunded obligation from 2013 to 2024: negative amortization added about $29 billion and assumption changes about $17.6 billion, offset by roughly $31.4 billion from investment performance, growth in active membership, and state supplemental payments. CalSTRS states the negative amortization rule itself: “When a pension plan is less than 100% funded, contributions toward the unfunded actuarial obligation must exceed the interest on the unfunded actuarial obligation to prevent it from increasing over time… payments toward the unfunded actuarial obligation must be more than 7% of the unfunded actuarial obligation.” This is why the funded position matters and not only the direction of travel. Seven percent of today’s $82.0 billion gap is about $5.7 billion a year, owed before the debt shrinks by a dollar; at 95% funded against the same $395.5 billion obligation, the annual toll would be roughly $1.4 billion. In 2023–24, employer and state contributions exceeded that interest by roughly $1.6 billion against interest of roughly $6 billion, so four of every five dollars aimed at the debt bought standing still. Contributions first exceeded the interest in fiscal year 2022–23, the plan’s ninth year. AB 1469’s own analysis noted that contributions “have been less than the amount necessary to fully amortize the UAO in the DB Program since 2002.” Two of the three offsetting factors are also taxpayer money: on membership growth, the 2024 edition explains that an 8% payroll increase produced “approximately $550 million more in contributions from employers than anticipated” in 2023–24, and the supplemental payments were General Fund checks, including $2.246 billion paid in 2019–20 on employers’ behalf.
7. CalSTRS’ 2025 Review of Funding Levels and Risks, prepared annually for the Teachers’ Retirement Board, calls investment risk “the most significant risk facing the CalSTRS Funding Plan.” The Legislative Analyst puts it the same way: “The most important assumption concerns future investment returns,” in part because that assumption is also used to discount future benefit payments to present value. The same report tests the funding plan against investment return assumptions of 6.75% and 6.5%, and payroll growth of 3.0% and 2.75%. It tests nothing above 7.0%.
8. CalSTRS news release on the June 30, 2016 actuarial valuation, April 6, 2017: the plan “would have had a $105.1 billion funding gap as of the June 30, 2016, valuation and would have been 61.8 percent funded” at a 7.0% assumption, against the 63.7% reported at 7.25%. Both figures flow from a single February 1, 2017 board action that phased the assumption from 7.5% to 7.0% across two valuations. The board went past Milliman’s recommended 7.25% to 7.0%, and CalSTRS said the changes “reflect the less than 50-percent probability that current return assumptions will be met over the long term.”
9. CalSTRS’ Report to the Legislature puts the June 30, 2013 funded ratio at 66.5%; AB 1469’s bill analysis said 66.9%. The June 30, 2017 valuation reported 62.6% funded and a $107.3 billion unfunded obligation. Employer rates over those four years ran 8.88%, 10.73%, 12.58% and 14.43% of payroll, and CalSTRS’ investment returns ran 18.7%, 4.8%, 1.4% and 13.4%. In April 2016, before the assumption change took effect, CalSTRS reported the plan 68.5% funded with a $76.2 billion gap and said the gap was $8.9 billion smaller than the funding plan had projected. The Legislative Analyst estimated that the first phase of the 2017 assumption changes alone added roughly $7 billion to unfunded liabilities.
10. CalSTRS, “Funded status continues to rise; contribution rates remain same”: “This is the eighth consecutive year the CalSTRS funded status has increased. The funded status has grown by nearly 17 percentage points (from 62.6% to 79.3%) since 2017.” CalSTRS has not published a restatement of the June 30, 2013 valuation on the assumptions in force today. Applying the sensitivity CalSTRS did disclose — a quarter-point reduction raised the June 30, 2016 obligation by about 3.3% — twice to the 2013 figures (a $220 billion obligation, $146 billion of assets, 66.5% funded) yields an obligation of roughly $235 billion, a gap of roughly $88 billion, and a funded ratio of roughly 62%. The estimate isolates the investment return assumption; the February 2017 package also lengthened assumed life expectancy, which raised the obligation further, and lowered the inflation and wage growth assumptions, which reduced it. The 2014 funding plan had projected 70.2% for June 30, 2025. The May 2026 board item notes that about $665 million of the past year’s $6.7 billion reduction in the unfunded obligation “was the result of the previous decision made by the board to not reduce the contribution rates for both the employers and the state.”
11. California Constitution, article XVI, section 17, added by Proposition 162 in 1992, gives the board “sole and exclusive power to provide for actuarial services in order to assure the competency of the assets” of the system, and provides that its duty to participants and beneficiaries “shall take precedence over any other duty.” The twelve-member board seats three educators elected by the membership; a retiree, three public representatives and a school board representative, all appointed by the Governor; and four ex officio members, the Director of Finance, State Controller, State Treasurer and Superintendent of Public Instruction. The board’s own governance manual lists “determining actuarial assumptions and contribution rates” among its responsibilities.
12. Ed Mendel, “CalSTRS funding gap grows, new earning forecast,” CalPensions, December 3, 2010. The assumption had stood at 8% since 1995. Milliman’s actuary told the board that 7.5% “would still only result in about a 50 percent chance of achieving the earning target,” and had no comparable calculation for 7.75%. The chair of the California Teachers Association’s retirement committee told the board the reduction should stop at 7.75%. The board adopted 7.75% on an eight-to-three vote; the dissenters included the Department of Finance. The board did not adopt 7.5% until February 2012, nearly two years after its actuaries first recommended it.
13. Segal, a second actuarial firm retained by the board, reviewed Milliman’s 2024 experience analysis and put the confidence level of the 7.0% assumption at 53.2%, the likelihood that investment earnings equal or exceed 7.0% over a fifteen-year period, which Segal takes as an approximation of the duration of CalSTRS’ liabilities. At 6.75% the confidence level would be 57.1%. Segal wrote that 7.0% “does not include much margin for future adverse deviation” and recommended 6.75%. The board retained 7.0%.
14. CalSTRS, Review of Funding Levels and Risks, finds that the plan still reaches full funding on more conservative economic assumptions, but “contributions would have to increase for the state, for employers and possibly for CalSTRS 2% at 62 members.” Of the $82.0 billion shortfall, employers owe $78.4 billion and the state $4.0 billion, down from $10.2 billion a year earlier. The state’s share is projected to be eliminated next year, dropping its supplemental rate to zero and returning it to a 2.017% base; employers pay to 2043. The employer rate stands at 19.1% of payroll against a statutory ceiling of 20.25%. The payroll the rate is collected on is itself at risk: K–12 enrollment is down about 367,000 students, or 6%, since 2019, and the state’s November 2025 projection anticipates a further decline of about 16% over twenty years.
15. Funding standards elsewhere add a margin above the best estimate rather than funding to the estimate itself. Office of the Superintendent of Financial Institutions (Canada), guidance on the preparation of actuarial reports for defined benefit pension plans: a provision for adverse deviations “determined at the 80% confidence level ensures a high probability that the pension promise will be met.” Canadian actuarial standards define a best-estimate assumption as unbiased, neither conservative nor non-conservative, which is why a funding standard normally adds a margin on top of the estimate rather than funding to the estimate itself. At 6.75% the confidence level Segal calculated would be 57.1%.