The single most important economic story of the moment is inflation. Unfortunately, we know little about its long-term effects on Colorado’s public pensions, a situation that it unlikely to change in the near future.
First, the good news.
For public pensions around the country, the main immediate effect of inflation has been to boost their Cost of Living Adjustments (COLAs). Most pensions’ benefits are tied to inflation, and with prices rising, that means more money being paid out right now to beneficiaries. Those cash flows out also come at a time when markets are down. What’s more many pensions base their contributions on employees’ salaries, which have not increased yet to match rising prices.
Senate Bill 18-200, the most recent legislative effort reforming the finances for the state’s Public Employees Retirement Association (PERA), placed limits and curbs on the annual increases. The bill set a target date of 2048 for full funding of the system, and until that target is reached, COLAs fluctuate between 0.5% and 2.0% per year. If PERA is ahead of schedule, COLAs will increase by 0.25%; if it is behind schedule, COLAs will decrease by 0.25%. However, they can never drop below 0.5% or rise above 2.0% until PERA is fully funded.
Whatever effects PERA will feel from inflation, a sudden boost in cash outflows from unrestricted COLAs isn’t one of them. There may be a serious political effect from this, as current retirees see their benefits lose ground at an accelerated pace. There may be calls to revisit or revise the SB18-200 COLA formula, which would also revisit the political dynamic that got us here in the first place – satisfy employees and retirees today, and let tomorrow and its taxpayers and beneficiaries look after themselves.
What we don’t know
The bad news is that that’s about all we know, and for two reasons it’s all we’re likely to know. First, there’s the decision-making timeline that PERA uses. Second, there are the limitations of the modeling tool PERA uses to look into the future.
We don’t know the effects on benefits. We don’t know the effects on returns. We don’t know the effects on contributions, and whether they’d keep pace with benefits.
Currently, PERA assumes a long-term inflation rate of 2.3%. This fall, PERA’s board will hear the results of an actuarial audit which will include an evaluation of its inflation assumption. However, the board has decided that it will take no action on the inflation assumption until it hears from its next experience study in the fall of 2024. These studies look at how employers and employees actually behave in terms of staffing, salaries, and retirement.
For instance, the state government might choose to increase salaries but hire fewer people. Schools might choose to hire more people but at a lower salary. Those decisions will affect both contributions and benefits differently. Since PERA sees itself as reflecting those decisions, and reporting the effects of those decisions, it won’t change its models to predict what they might do.
Here is where the limitations of the modeling tool come into play. Certain variables, such as the state’s annual $225 million line-item contribution or the rate of return, can be varied individually, on an annual basis, up to 15 years in the future.
Other variables, such as inflation, staffing growth and salary growth need to be programmed in at specific values. Their relative inflexibility, the inability to easily conduct sensitivity analyses with these variables, and the fact that staffing and salary growth decisions will probably be taken in response to inflation, make it difficult to use the model to judge their effects.
Even the COLA good news is unclear. COLAs and contributions may be adjusted each year, depending on whether PERA is behind or ahead of schedule. This is done automatically, according to a formula. But the models don’t take that into account. Right now, COLAs are at 1.0%, and because of strong returns over the last three years, they may be raised again. Certain funds can see the date when their contribution levels are reduced. But the models don’t take that into account. What are known as Monte Carlo models, which run thousands of random tests simulating three or four decades, don’t take that into account, which means they may be overstating the effects of the good investment news of the last few years.
An incurious PERA board
Which raises the question – why isn’t the board even considering asking the programmers of the model, Segal and Associates, to turn those variables into annual parameters, or to conduct a three-way sensitivity analysis? Inflation, as we know, is the economic issue of our time right now. The issue was raised last year, at the 2021 PERA Subcommittee hearings, and even that was a year after producer prices started to rise dramatically.
And yet, the PERA board has decided to take no action on its models until 2024, nearly four years after inflation started to cloud the horizon. Even if some of the results are the consequences of decisions that employers have not yet made, PERA’s models could be changed to help inform decision-makers of those consequences. This level of incuriosity was irresponsible two years ago. It’s inexcusable now.
PERA’s accounting policies do have one ironic effect. Along with almost all American public pension systems, PERA discounts its liabilities at its expected rate of return. We’ve previously noted the perverse incentives this creates in terms of both that assumption and the investment policy it encourages. By contrast, private pensions typically use the returns on high-quality corporate bonds as the discount rate. If borrowing costs increase, then the value of those future benefits, measured in today’s dollars, decreases.
For years, Independence Institute has advocated that public pensions use the state’s cost of borrowing as the discount rate. Had PERA followed that advice, it would now be seeing the value of those liabilities fall, and its funded levels increase.
Joshua Sharf is a senior fellow in fiscal policy at the Independence Institute, a free market think tank in Denver, who has written on PERA extensively.
