It’s common nowadays to read critiques of public-sector pension plan investment assumptions. State and local pension plans assume an average investment return of 8 percent, which, in this economy, just *feels* too rosy. The mainstream media has picked up on it, running articles pointing out that the 8 percent mark has not been met over the last decade, and there’s an entire cottage industry in the accounting world arguing the unfunded liabilities should be evaluated on much lower investment assumptions.

Are they right?

To find out, I compared long-term real investment returns with the real rate of return assumed by state teacher pension funds*. The word “real” in this context means investment returns minus the effects of inflation. The investment return data calculates the real return of a conservative portfolio invested 25 percent in the S&P 500, 25 percent in small US stock, 25 percent in long-term US corporate bonds, and 25 percent in an equal split of 30 day treasury bills, intermediate-term treasury bonds, and long-term treasury bonds**. I calculated the assumed real rates of return of state teacher pension plans by subtracting their inflation assumption from their investment return assumption.

The average one-year real return for the balanced portfolio has been 6.72 percent since 1926, a common starting place for the modern stock market. For comparison’s sake, the average state teacher pension plan assumes a real rate of return of 4.59 percent. In other words, they’re relatively conservative in relation to historical averages.

Averages, of course, can mask highs and lows. What’s more important is to see how various extended *periods *of time performed compared to the state assumptions. To do that, I used 30-year rolling averages of real returns. So, for example, 1926 to 1956 was one period and 1927 to 1957 was another. The series ended with the 30-year period from 1989 to 2009. Thirty-year periods also matter in the real world, because that is the time horizon states are given to calculate how much they would have to invest today in order to be fully funded in the future.

The chart below graphs these 30-year periods against state assumptions. The blue bar represents the real rate of return assumed by 49 out of 50 states (New Jersey does not publish an inflation rate). The most common assumption, used by 19 states, was a real return between 4.5 and 4.99 percent. States could get there in any number of ways. They could assume an 8 percent investment return and a 3.5 percent inflation rate, or they could assume a 4.75 investment return with a 0 percent inflation rate. Most states tended towards the former than the latter, but all that matters is the difference between these two figures.

The red bars indicate the frequency that real rates of return were observed over 30-year periods. The most common occurrence put investments minus inflation between 5 and 5.49 percent.

The most important thing to understand from this chart is that states are generally conservative in their investment assumptions***. Every state is below the 83-year average. There were five 30-year periods where many states would have assumed a rate in excess of what they actually returned, but, for the most part, state assumptions are more conservative than historical averages.

*This being an education blog, I chose to focus only on plans that include teachers, but teachers are the largest group of workers covered in state retirement plans, and the plans including teachers are very similar to plans covering other government employees.

**The data come from the Shwartz Center for Economic Policy Analysis, an older version of which appeared in this paper.

***Remember that the returns used in this chart are significantly more conservative than the average state investment portfolio. That means two things. One, the riskier investments actually born by states would produce higher highs and lower lows, and, two, the real rate of return on riskier investments like equities, despite greater volatility, would actually be higher over time.

**Update: **Chris Tessone takes issue with my assumptions, but he misses several key points. One, it’s true that past performance is no guarantee of future results, but that doesn’t mean it should be casually disregarded, either. It’s not like we’re talking about a small sample size: My data use 83 years of stock and bond performance, which includes many dramatic boom and bust eras. Two, his statement that pension funds are “heavily weighted in the direction of private capital investments” is just not true. Most pension funds have a small portion, say 10 percent, invested in private capital investments like real estate or commodities. The funds mainly invest in equities and bonds.

Most importantly, my post was solely about the financial elements behind public sector pensions. I have been quite strong in my criticism of the *structure* and the incentives for workers in the plans. But the structure and the financial assumptions are separate issues. For more, read our *Better Benefits* report.