UC Berkeley Study: Public Pension Costs Crowding Out Other Municipal Spending Priorities

Local government pension expenditures are rising, crowding out other spending priorities, according to the findings of a working paper authored by the University of California, Berkeley’s Sarah Anzia and presented at a recent conference co-sponsored by the Volcker Alliance and the Goldman School for Public Policy. To pension reform advocates this conclusion may seem unsurprising, but Anzia’s rigorously empirical approach is an important contribution to the largely anecdotal evidence of pension crowd-out.

To reach her findings, she collected Comprehensive Annual Financial Reports (CAFRs) for 219 cities and towns for the period 2005-2014 and then recorded the amount of each government’s annual pension contributions in a data set. Her stratified sample included municipal governments in eight size categories, ranging from cities with more than one million people to communities with fewer than 10,000 inhabitants.

Anzia found that pension costs rose for 85 percent of the cities in her sample over the 10-year period. The average increase, on an inflation-adjusted basis, was 69 percent; the median increase was 45 percent.

She also found a strong association between increased pension contributions and reduced government employment, suggesting that the need to make pension contributions deters cities from creating new positions, or, in some cases, even replacing existing workers. The employment impact of pension costs was especially pronounced for cities in states that require collective bargaining.

This finding points to an interesting conundrum for public employee unions. On the one hand, unions want to protect, and, if possible, expand employee pension benefits. But union dues revenue varies with the number of represented employees. So, as they advocate for their current members, unions may be losing future members and the dues they would pay.

Data availability issues and the unique focus of Anzia’s study led her to make some interesting choices in compiling her data set. She examined actual pension contributions rather than actuarially required amounts. This seems appropriate, given her interest in determining whether pension costs are crowding out other municipal expenditures.

She also did not consider Employer-Paid Member Contributions (EPMCs) due to difficulties in obtaining the data. Governments are not required to report the amount of member contributions they pay on behalf of their employees. The dataset also excludes interest payments on city-issued Pension Obligation Bonds because these amounts are usually not separately reported on municipal CAFRs. Other researchers could potentially enhance Anzia’s data set by collecting EPMC data from collective bargaining agreements and POB interest amounts from bond offering materials.

More controversially, Anzia excludes Other Post Employment Benefit costs. OPEB contributions are reported on CAFRs, and, like pension contributions, they potentially displace municipal service provision for the benefit of retired employees. I would argue that OPEB costs should be included, and, given the rapid pace of medical cost inflation, their addition would steepen the retirement expense slopes Anzia calculated.

That quibble aside, Anzia’s paper makes an important addition to our understanding of the budgetary impact of public employee retirement costs. By doing the hard work of locating and scraping relevant data from a large stratified sample of municipal CAFRs, she has provided academic validation of the intuitive—but hard to clearly and empirically demonstrate—phenomenon of service-level “crowd out” that is fundamental to many calls for reform of financially unsustainable public retirement systems.

Ironically, one of the nation’s worst cases of pension crowd-out is occurring in the city of Berkeley, the municipality in which Anzia’s campus is located.  As I reported previously, the city pays about 12 percent of its total revenue for pension contributions (the proportion of general fund revenue would likely be significantly higher, but the necessary data are not provided in municipal CAFRs). The city reported $536 million in net pension liabilities, and its contributions to CalPERS are projected to grow 85 percent (in nominal dollars) between now and fiscal year 2024-25.

Nonetheless, it would be trite to say that if Anzia was seeking evidence of pension crowd-out, she need go no further than her own City Hall. Robust academic findings about local government finance cannot be based on anecdotal evidence from one municipality. By studying financial statements from more than 200 cities, Anzia’s conclusions about pension crowd-out come with a much higher degree of confidence. Her findings lead to the conclusion that states will have to give their municipalities more tools to address escalating public employee retirement costs if they hope to avoid widespread service insolvency at the local level.