K-12 Referenda Passage Rates Fall

Focus #10 • May 2023


Referenda to allow school districts and local governments to exceed state tax limits passed at lower rates in April, but still in substantial numbers. An analysis of K-12 referenda going back to 2010 shows that larger shares of urban and declining enrollment districts have approved operating referenda. In addition, a larger portion of low-income districts have done so compared to high-income districts, raising concerns about which communities feel the most pressure to pursue these ballot measures.

After years of passing in record numbers and amounts, school referenda lost speed this spring, as voters approved them at the lowest rate since 2010.

Of the 83 referenda seeking to increase local property taxes for K-12 schools, 46 were approved by voters, or 55.4%. Referenda to finance capital projects passed at a higher rate than the two types of measures to finance districts’ ongoing operations. All three types, however, passed at lower rates than last year, and the overall passage rate represented a notable drop from last year’s approval rate of 80.1%.

At the same time, the majority of measures did pass and more of them were approved than in either 2019 or 2021. They will increase school spending by more than $600 million over the coming years.

With school referenda now done for the year, the Forum also took this moment to review voting on school measures in Wisconsin since 2010. Our review found referenda to support higher levels of operating spending have been passed in a larger share of districts with falling enrollment and low and declining levels of state aid. Referenda to support spending on capital needs have been approved in a larger share of districts with growing enrollment, higher levels of state aid, and high property values and household incomes.

All types of referenda also have been particularly likely to pass in urban districts. However, the analysis does not prove that these factors directly impact referendum proposals or passage rates. Capital referenda, for example, may be influenced by the age of district buildings, and urban districts simply may be more likely to have older facilities.

In addition, this brief will discuss the results of other April voter referenda to raise municipal and county property taxes in the state. The results from these ballot measures provide a limited window into how voters are thinking about local taxes, services, and spending as lawmakers debate the state’s policies in these areas as part of budget deliberations over the coming weeks.

April School District Referenda

Since the 1993-94 school year, K-12 school districts in Wisconsin have been subject to state-imposed limits on the amount of revenue they can raise per pupil from state general school aids and local property taxes combined. Since 2011, revenue limits have not kept pace with inflation, and in the current 2021-23 state budget they were frozen for both years.

In recent years, that has prompted a groundswell of school referenda in which voters have opted to increase district spending and their local property taxes by more than what the revenue limits would otherwise allow. This surge moderated somewhat this spring, however.

One reason may be that K-12 referenda in odd-numbered years take place in low-turnout spring elections and tend to pass at lower rates than in the high-turnout fall elections in even-numbered years. This year did see a drop in the passage rate from the 60.6% approved in calendar year 2021 – a difference that may reflect factors such as the specific measures on this year’s ballot and the impact of rising inflation on voters’ pocketbooks and sentiments.

Still, 69 Wisconsin school districts pursued referenda last month, including 14 districts that put two separate questions to voters, according to Department of Public Instruction (DPI) data. Forty-six of the referenda passed, which was more than voters approved in either 2021 (43) or 2019 (45). Of the 83 questions put to voters, referenda to increase districts’ operating budgets accounted for 54, or nearly two-thirds.

The most common kind of operating referenda is called non-recurring. It increases property taxes for only a limited period, typically several years. Half of the 44 non-recurring referenda, or 22, were approved, the lowest rate since 2010 (see Figure 1). Still, voters did increase taxes through this type of referenda by $257.9 million in the coming years. Voters passed six out of the 10 recurring referenda, permanently increasing property taxes by $19.8 million a year.

Meanwhile, voters approved 18 of 29 referenda, or 62.1%, to finance capital projects such as new construction or building renovations. That authorized $368.8 million in borrowing and, while it was a decrease from the 77% passage rate in 2022, was still higher than the 52.8% approval rate in 2021 and similar to the 61.5% success rate in 2019.

School Referenda Since 2010

As part of our analysis, we also looked at what share of districts had passed at least one referendum between 2010 and 2023, breaking down the districts by variables such as the change in their enrollment. These data were first gathered as part of a project for the Fair Aid Coalition, a group of high property value school districts in Wisconsin. Though the results do not prove that these factors cause referenda to be put before voters or to pass at higher or lower rates, they do show interesting correlations.

One noteworthy point is that over this time period most districts passed at least one referendum, so all communities whatever their type were relatively likely to have done so. Another important caveat is that our analysis did not extend back to the beginning of revenue limits, which to some degree embedded into state law the different spending levels present between districts as of 1993-94. Those varying starting levels of spending likely played a role in which districts have passed referenda over the years, but this question lies outside the time period examined in our analysis.

Most notably, our analysis looked at districts in which enrollment rose between the 2010-11 and 2021-22 school years, using a form of enrollment measure used drawn from the state’s revenue limit formula. Most districts – 306 to be exact – had a decline in enrollment over this period given the state’s relatively low birthrates and net migration and the impact of the pandemic. One hundred fifteen districts, however, did have an enrollment increase.

Voters in the great majority of districts passed at least one referendum over this period but the type of measure differed by group. Since 2010, 55.9% of declining enrollment districts have passed at least one operating referendum compared to 50.4% of increasing enrollment districts (see Table 1). This modest difference is perhaps not surprising given that declining enrollment districts multiply their per pupil limit by fewer students, yielding a lower revenue cap, tighter operating budget, and more reason to seek higher property taxes. Rising enrollment districts, on the other hand, multiply their pupil revenue limit by more students, boosting their revenues, easing budget strains, and lessening the potential need to seek an operating referendum.

In the case of capital referenda, 79.1% of increasing enrollment districts passed at least one measure compared to 67.7% of districts with shrinking enrollment. Again, this is logical: while both types of districts may have aging buildings to renovate or replace, increasing enrollment districts also have more students to accommodate, which puts pressure on them to add classrooms and carry out capital projects.

We also looked at the locale where districts are located. The state’s 17 urban districts (as classified by the National Center for Education Statistics), for example, have all passed at least one referendum since 2010. Nearly two-thirds, or 64.7%, have passed an operating referendum while only 39.0% of the state’s 77 suburban districts have done so. This seems logical in that urban districts generally are seen as having more budget pressures than more well-to-do suburban districts. Voters in urban districts also tend to be more liberal and may also be more likely to favor school funding.

Urban districts are also the most likely to pass capital referenda. However, it’s worth noting that one of the most likely reasons for a capital referendum is aging school buildings. No statewide data on the condition of K-12 buildings exist, but urban districts may be more likely to have aging buildings in need of major maintenance, renovations, or replacement than those in suburbs. So, it is possible that the high percentage of urban districts passing capital referenda is more a matter of correlation than causation.

The analysis also looked at median household income among school districts, dividing them into three groups that included the 105 districts (representing one-quarter of the total) with the highest household income, the one-quarter of districts with the lowest income, and the other 211 districts.

A larger share of high-income districts (77.1%) have passed a capital referendum since 2010 while only 58.1% of low-income districts have done so. Meanwhile, only 45.7% of high-income districts passed an operating referendum during the period whereas 55.2% of low-income and 58.3% of all other districts did so. These trends may reflect the fact high-income districts tend to have greater revenue limits and operating budgets to begin with – giving them less need for referenda. These districts may also enjoy greater support from their communities when they seek to upgrade their facilities.

Last, the analysis looked at high property value districts in the Fair Aid Coalition that receive little to no state general school aids (these are so-called no equalization or primary aid only districts). Compared to other districts, a smaller share of these districts has passed a capital referendum but a larger share has passed an operating referendum. That may reflect a greater need among these districts to focus property tax revenues on their operating budgets because of low state aid.

As noted earlier, these findings do not prove that the factors listed above cause more referenda of the various types or higher or lower passage rates, but they do outline notable trends that may merit further research and attention from policymakers.

Municipal and County Referenda

April’s K-12 referenda were not the only measures on the ballot to increase local property taxes beyond state-imposed caps. The state of Wisconsin also generally limits yearly county and municipal property tax increases used for operations to the rate of increase of net new construction within each jurisdiction. Voters have also been voting in increasing numbers in recent years on local referenda to exceed these limits. Here we report on these measures using data from the Wisconsin Elections Commission (WEC), the League of Wisconsin Municipalities (LWM), and local clerks.

Of the 25 referenda on the April ballot to increase levy limits for towns, villages, cities, and one county, 10 passed, or 40%, with one referendum in the village of Nelsonville in Portage County actually failing in a tie vote of 43 to 43. As Figure 2 shows, the number of successful referenda this year was a substantial decrease from the 29 approved in 2022 and 14 approved in 2018 but was still the third-most since levy limits were implemented in 2006 and the most in an odd-numbered year with no fall election.

In all, voters chose to increase their local taxes by a total of $4.5 million over the state limit for an average increase of 13.3% over the cap in those communities. Eighteen of the 25 referenda said the funds would be used at least in part for police, fire protection, and other public safety services, with emergency medical services and streets being the next two most common purposes. Notable votes included the successful referenda in the village of Caledonia, city of Jefferson, and village of West Salem, and the unsuccessful ones in the city of Marshfield and in St. Croix County.

In addition, voters in the city of Stevens Point approved two referenda to authorize street projects and voters in the town of Conover in Vilas County approved a room tax on short-term lodging. Voters in the city of Lake Geneva also rejected the possibility of imposing a sales tax on select tourism-related businesses and items known as the premier resort area tax.

Looking ahead

Though K-12 and other local referenda to raise property taxes passed at lower rates in 2023, they remain relatively common compared to a decade ago. That reflects tight school revenue limits and relatively flat state aid and levy limits for local governments.

For years, the Forum has warned that a reliance on referenda to fund local services could exacerbate disparities between wealthy communities whose residents can afford to approve them and other communities who cannot. In addition, the cost of a referendum may have a greater impact on those poorer communities who do approve them. Some of these concerns seem to be borne out in the differences between the shares of K-12 districts that since 2010 have approved various types of referenda in urban versus suburban areas, high-income versus low-income, and increasing enrollment areas versus those such as rural areas with declining numbers of students.

The Forum does not advocate for or against specific policies, but we have noted the state’s massive surplus represents a “golden opportunity” to tackle big issues, including the challenges facing schools and local governments. We have pointed out ways to provide new revenues to support local services as well as ways the state might incentivize greater local efficiencies through strategies such as shared services. The coming weeks could prove pivotal for these questions and state lawmakers may wish to give the continued widespread use of referenda special consideration as they draft the 2023-25 budget in the days to come.