Figure 6 tracks annual investment returns for public pension plans from 2001 through 2024, showing the incredible volatility of these assets.

This volatility creates problems for pension systems. The plans need stable, predictable returns to meet long-term obligations. When pension boards assume they'll earn 7% or higher annually, they're betting on averaging out these wild swings over time. However, just a few bad years can wipe out gains from several good ones, leaving pension systems further behind than their assumptions predicted.

The chart also reveals why pension systems remain vulnerable even after years of economic recovery. While strong performance years help repair damage from market downturns, they don't eliminate the underlying risk that another major downturn could erase years of economic appreciation.

Figure 7 compares what pension plans assumed they would earn against the average of what they actually earned from 2001 to 2024. It shows that over the past 24 years, public pension plans consistently assumed they would earn more than their portfolio delivered. When return assumptions are higher returns than actual returns, plans underestimate both their future liabilities and the contributions needed to fund those promises.

However, the narrowing gap in recent years shows that pension systems have been adopting more realistic assumptions. In the fiscal year 2001, the average assumed rate of return was 8.02%, while as of 2024, it is 6.87%. Yet this adjustment comes after decades of overestimating returns, which is the source of much of the unfunded liability burden that taxpayers now face.

Even with more conservative assumptions, pension systems still project returns that exceed what they've actually achieved over the last 24 years. The average assumed rate of return is 6.87%, while the average 24-year rate of return is 6.62%.

Until assumptions align more closely with demonstrated performance, pension systems risk continuing to underestimate obligations, generating unfunded liabilities that taxpayers must eventually cover through higher contributions or benefit adjustments.

Figure 8 compares the distribution of what pension plans actually earned over the last 24 years against what they assumed they would earn, both historically and today. It shows three distributions: actual 24-year average returns for each plan (orange), their historical average assumed returns over that period (red), and their current assumed returns (blue).

Average historical assumptions (orange) sit well to the right of average actual performance (orange), confirming that most pension plans spent two decades assuming they would earn more than markets delivered.

The current assumptions (blue) show that pension systems have begun adjusting to market realities. There's been a clear shift leftward from historical assumptions, with many plans now converging around a 7.0% assumption. This represents a welcome move toward more realistic expectations. But the adjustment may not be enough. Even with this more conservative approach, the median of current assumptions still sits to the right of where actual long-term performance peaked. Many pension systems are still betting they can achieve returns that exceed what they historically delivered over the last 24 years.

This persistent optimism, even after the recent adjustment, suggests that pension systems haven't fully reckoned with the current market environment. Plans that assume 7.0% returns when their long-term average is closer to 6.5% are still setting themselves up to accumulate unfunded liabilities over time. Given that both contributions and debt amortization payments are based on assumed rather than achieved returns, each percentage point of overoptimism translates to billions in underestimated costs that taxpayers will eventually be obligated to pay.

Figure 9 shows "excess returns" for individual pension plans, calculated as each plan's 24-year average return minus its current assumed return rate. Plans that assume higher returns than they have achieved over the past 24 years appear in orange to the left of zero, while the few that assume lower returns than their previous performance show up in blue to the right.

As of the latest recorded fiscal year, 83% of public pension funds assume higher rates of return than their 24-year average returns.

Most public pension plans haven't achieved the returns they're currently assuming they can deliver going forward. The distribution is heavily weighted toward the left side, with most plans showing long-term performance shortfalls between 0.5% and 1.5% relative to their current assumptions.

In our sample, 7 public pension plans assume rates at least 200bps (2 percentage points) higher than their historical investment performance, while only 1 assumes returns 200bps lower than its historical average.

Figure 10 shows average annualized excess returns for pension plans relative to the S&P 500 across 5-, 10-, 15-, and 20-year periods. The dots represent individual fund performances and the black dashed line indicates the S&P 500 benchmark (zero excess return).

One hundred percent of public pension funds have underperformed the S&P 500 over the past 20 years. While the S&P 500 delivered an average return of 10.4% over this period, public pension funds averaged 7.5%.

It's worth noting that this comparison isn't entirely fair to pension funds. These systems maintain diversified portfolios across multiple asset classes to manage risk, not to replicate a 100% stock index. Pension funds hold bonds, real estate, and alternative investments specifically to reduce volatility, even if it means accepting lower long-term returns. However, as seen in Figure 6 (annual investment returns), it is not clear whether pension systems have been successful in reducing the volatility of their returns.

Figure 11 compares pension fund performance against a simple 60/40 portfolio (60% stocks, 40% bonds) over 5-, 10-, 15-, and 20-year periods. Each dot tracks an individual pension fund's excess returns relative to this basic benchmark, with the black dashed line representing the 60/40 index performance (zero excess return).

The 60/40 benchmark is arguably more appropriate than the S&P 500 since pension funds actually do mix stocks and bonds in their portfolios. But the results are still disappointing for pension fund management.

Eighty-four percent of public pension funds failed to beat a passive 60/40 portfolio over the past 20 years. While the 60/40 index delivered an average return of 7.9% over this period, public pension funds averaged 7.5%. The median pension plan has consistently underperformed this simple balanced portfolio, and the problem gets worse over longer time periods. Over 15- and 20-year horizons, the typical pension fund trails the 60/40 benchmark by approximately 1% to 2% annually.

This underperformance is particularly troubling because pension systems spend enormous resources on investment management. They hire consultants, pay management fees to external firms, and dedicate significant staff time to portfolio construction and oversight. Yet most would have served taxpayers and beneficiaries better by having invested more simply and passively.

Figure 12 compares the investment returns of public pension funds against a synthetic risk-adjusted benchmark, which accounts for each pension fund's specific risk and return profile over 5-year, 10-year, 15-year, and 20-year periods. The figure plots the excess average returns of 268 public pension plans, with the dots representing individual fund performances and the black dashed line indicating the synthetic benchmark (zero excess return).

Despite this tailored benchmark, many funds have not consistently met their risk-adjusted expectations. Over the past 20 years, most public pension funds still underperformed their personally tailored empirical benchmark—with 86% of funds underperforming the benchmark.