Abstract
A key informational asymmetry in local public finance is the lack of information available to local residents regarding the financial status of the school districts and local governments in which they reside. Given that voters in many states must approve property and income tax increases for these local entities, the lack of full information on the financial status of these local entities may lead to sub-optimal voting decisions. State financial intervention systems have begun to make financial problems more salient to residents, potentially alleviating these informational asymmetries. This paper examines the effect of the Ohio fiscal stress labeling program on voting outcomes and the tax-setting behavior of local officials for school district and municipal government tax referendums. We use a difference-in-differences approach to examine data from over 3000 school district and 2300 municipality property tax elections from 2004 to 2012. While we find minimal evidence that the yes vote share changed for school district referendums following fiscal stress label receipt, we find very large increases (15 to 23 percentage points) in the likelihood of referendum passage for school districts following label receipt. We do not find much evidence of changes in the likelihood of passage or the yes vote share following label receipt for municipalities, but we do find that these voting outcomes rise following label removal. We also find that local officials do not appreciably change their tax-setting behavior in response to these labels, as the size and likelihood of property tax proposal are largely unchanged following label receipt or removal.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
In Ohio, approximately 38% of school district revenues and 45% of municipal government revenue are generated from local tax sources. Hence, any changes in voting outcomes for school district tax referendums will have critical effects on revenues of schools and municipalities.
Honadle (2003) and Kloha et al. (2005) found that 15 states had some fiscal health evaluation system in place and nearly a third more states are considering using these indicators. According to a 2013 Pew Charitable Trusts report (Trusts 2013), 19 states have a financial intervention system in place, but the structures of these policies vary widely. For example, the Michigan Emergency Manager law intervenes in local governments only when they experience severe financial problems (e.g., the City of Detroit), while other states, such as Massachusetts and Connecticut, take a more ad-hoc approach to financial interventions.
During non-presidential election years, Ohio holds elections in February, May, August and November. During presidential election years the presidential primary is held in March, and only three other election dates are scheduled—March, August, and November.
These labels are imposed on school districts and local governments for a variety of financial issues. For school districts, these labels largely are based on the five-year financial forecast that school districts are required to submit to the Ohio Department of Education each October. This forecast provides expenditure and revenue projections for the current and the next four fiscal years. Districts with projected current year deficits that exceed 2% of general fund revenue receive a fiscal oversight label. A school district receives a fiscal emergency label if the projected current year general fund deficit exceeds 15% of general fund revenue. Local governments can be subjected to fiscal oversight for a variety of reasons, including un-auditable financial records or general fund deficits. A local government receives a fiscal emergency label if the general fund deficit exceeds one-sixth of the general fund revenue or if it fails to meet payroll or debt payment obligations. For more detailed information regarding these different labels, see Thompson (2016a, b).
While Ohio defines three separate label categories—fiscal caution, fiscal watch and fiscal emergency, this paper follows Thompson (2016a, b) in combining fiscal caution and fiscal watch into one category, fiscal oversight. The financial recovery requirements of these two labels are similar—most notably, the school districts and municipalities themselves lead recoveries under both labels.
From the Ohio Secretary of State’s data, we collected the proposed property tax millage rate. This millage rate identifies the size of tax in mills, where one mill denotes that an individual would pay $1 in taxes for every $1000 of assessed property value. To make comparisons across districts/municipalities and across time, we convert this proposed property tax millage rate to an effective millage rate. We do this by multiplying the proposed millage rates by the median assessed-value-to-sales-price ratio of the particular school district or municipality in that election year (see Mikesell 2013). We collect the median assessed-value-to-sales-price ratio data from 2005 to 2012 for all municipalities and townships in the state (http://www.tax.ohio.gov/tax_analysis/tax_data_series/real_property/residential_sales_data.aspx) and then multiply this ratio by the proposed tax rate to get an effective millage rate for each municipality tax proposal. The State of Ohio does not report this ratio directly for school districts, however, and thus we approximate the ratio for school districts using a weighted average of the ratios for the overlapping municipalities and townships for each school district. We use as weights the fraction of registered voters of the school district residing in the specific municipality or township. Data on the number of registered voters comes from the Ohio Voter Files (http://www6.sos.state.oh.us/ords/f?p=111:1). It includes individual-level data on registered voters, including the voter’s school district-municipality or school district-township combination of residence. For some counties, the voter residential location information is missing and in those cases we examine precinct-level voting data and obtain aggregate precinct voter registration counts. We then match the precinct with the specific school district-municipality or school district-township combination to fill in the missing registered voter counts for those combinations. In the end, we get this weighted average assessment-to-sales-price ratio for each school district in each year from 2005 to 2012. We then multiply the proposed property tax millage rate by this weighted average ratio to obtain the “pseudo” effective millage rate for each tax referendum proposed by school districts.
We collected information on a total of 3286 school district and 3314 municipality property tax elections from 2004 to 2012 from the Ohio Secretary of State’s office. The smaller number of elections used in our analytic sample is explained by missing data for some of our control variables and variables included in the sensitivity analyses. This is particularly true for municipalities where we were unable to obtain a reliable measure of the number of registered voters who face fiscal stress labels in overlapping jurisdictions. These are often villages that do not have their own precinct separate from the local township and therefore we are unable to separate village voters from township voters in these cases to create the overlapping fiscal stress label control variable.
Given that income taxes differ considerably from property taxes in terms of their structure, especially given the different tax bases, the types of school districts and municipalities that use income taxes, and the fact that property taxes are more stable sources of revenue than income taxes, aggregating income tax and property tax referendums together in one analysis is problematic. Given that property taxes typically are used much more frequently by school districts and municipalities, throughout the remainder of our analysis we focus our attention on property tax referendums. For some of our regression analyses, we do estimate the regressions using only income tax referendums. The results of these analyses are available upon request.
For definitions of the variables used in this analysis see Table 5.
Column (1) provides descriptive statistics for the school districts or municipal governments that never receive a fiscal stress label during the time period examined in this study. Column (2) reports descriptive statistics for the labeled districts or municipalities that are observed both prior to receiving the label (Pre-Fiscal Stress) and during the time the label is applied to the district (Fiscal Stress). Column (3) displays the descriptive statistics for the labeled districts or municipalities that are observed both during the period the label is applied to the district and after the label is removed (Post-Fiscal Stress). While Columns (1–3) are mutually exclusive, districts that remained in fiscal stress duration during the full sample period are found in both Columns (2) and (3).
Instead of a referendum-level dataset, as we use in all of the other analyses, for this analysis we construct separate balanced panels of all Ohio school districts and municipal governments, with one observation per governmental unit for each election. We then create a new dichotomous indicator variable equal to 1 if a school district or municipality proposed at least one referendum during a given election, and equal to zero 0 if no referendum was proposed. We then estimate Eqs. (1) and (2) using this new dataset and the proposal indicator as the dependent variable.
We also estimate Eqs. (1) and (2) using proposed income tax rates and an indicator for the income tax referendum proposal as dependent variables. The results of this analysis are available upon request. For school districts, we find some evidence that the likelihood of an income tax proposal increased by 3.1 percentage points following the receipt of fiscal oversight. We find no evidence, however, that proposed income tax rates changed following fiscal stress label receipt or removal. For municipalities, we find very little evidence that the likelihood of an income tax referendum proposal changed following receipt or removal of the fiscal stress labels. We do find that the proposed income tax rate fell by 1.23 percentage points following notice of fiscal oversight, but rises by 0.83 percentage points following the declaration of fiscal emergency. The income tax rate continues to rise by 0.62 percentage points following label removal.
Given that more than half of label removals for municipalities comprise removals of fiscal emergency labels, we are unable to identify a separate effect of fiscal emergency removal from fiscal oversight removal for them. Thus, we include the \(endFS_{jt}\) variable in Eq. (2) and, for consistency in the results tables, we do the same for school districts. The results of Eq. (2) for school districts with disaggregated effects for removal of the different types of labels are available upon request.
We also conduct an analysis that examines differential effects depending on the type of municipality—either city or village. The results of this analysis are available upon request. We find that cities change proposed effective millage rates in response to fiscal stress label receipt and removal, while no such statistically significant effect is found for villages. Cities reduce the effective millage rate by 1.89 mills following receipt of fiscal oversight, but reduce the effective millage rate by 2.01 mills following receipt of a fiscal emergency label. Cities also reduce proposed effective millage rates by 1.46 mills after label removal. We also find that villages are less likely to propose a referendum following label removal (9 to 9.6 percentage points less likely), but that cities exhibit no statistically significant effect.
In addition to the main specification presented here we conduct sensitivity analyses that include county-specific time trends, controls for the percentage of residents facing fiscal stress labels in other jurisdictions that overlap the school district or municipality, and controls for observable financial health information. The results of this analysis are available upon request. Those results indicate that while the estimated magnitudes of the point estimates change slightly, the overall conclusions from Table 2 are not very sensitive to changes in specification.
We also estimate Eqs. (1) and (2) using the income tax referendum yes vote share and an indicator for income tax referendum passage as dependent variables. The results of this analysis are available upon request. We generally find no statistically significant effects of label receipt or removal on the yes vote share or likelihood of passage for income tax referendums for either school districts or municipalities. The small number of income tax referendums that are in the sample—548 school district referendums and 449 municipality referendums—may be contributing to the weak power in this analysis.
The results of a Logit estimation of Eq. (2) are available upon request. The marginal effects show that school district referendums are 20.2 percentage points more likely to pass following fiscal stress label receipt. For municipal referendums, the yes vote share rises by between 3.9 and 5.3 percentage points following label removal, resulting in an increase in the likelihood of passage of between 21.1 and 23.5 percentage points.
These label removal effects are found only in property tax elections held by villages. For village referendums, the yes vote share rises by between 6.8 and 7.7 percentage points, resulting in an increase of between 24.1 and 26 percentage points in the likelihood of passage following fiscal stress label removal. We find no statistically significant effects that the yes vote share or likelihood of passage for property tax referendums proposed by cities change following label receipt or removal. The results of this analysis are available upon request.
We choose to use county-specific linear time trends in the main sensitivity analysis instead of unit-specific linear time trends, as unit-specific linear time trends are likely to be poorly identified for school districts and municipalities that have very few elections occurring over the sample period. In fact, 265 out of the 613 school districts and 465 out of the 932 municipalities held fewer than five property tax elections during the 2004–2012 period of our study. The issue with these poorly measured linear time trends is that they could bias the coefficients of our key explanatory variables. We do, however, estimate a specification with unit-specific time trends included for the sample of school districts and municipalities that conducted at least five property tax elections during the 2004–2012 period. We also estimate our baseline specification for this restricted sample to provide a comparison with the unit-specific time trend results (results of this analysis are available upon request). With school district-specific linear time trends included, some of the fiscal stress variables drop in magnitude and significance. The increase in likelihood of passage following receipt of fiscal emergency notices remains high, however, at 18.8 percentage points. With municipality-specific linear time trends included we find no statistically significant effects of label receipt or removal on the yes vote share or the likelihood of passage.
To create this variable, we use the individual voter files collected from the Ohio Secretary of State in 2013 and aggregate this individual-level data to obtain counts of total registered voters in each municipality, school district, and municipality-school district combination. As there is no retroactive data on these individual voter files by year, the information for 2013 gives us the best approximation of registered voters during the period of the study. For some counties, the municipality-school district counts are unavailable owing to missing observations. We fill in these missing values with counts of registered voters from the precinct-level elections collected from individual county boards of election websites. Since the registered voter counts do not vary across time—we have the counts for 2013 only—the number of registered voters is the same for all election dates during the 2004–2012 sample period. We estimate the variation in overlapping fiscal stress labels by interacting the fiscal stress label indicator (which varies across election dates, depending on the timing of label receipt and removal) with the registered voter counts for each municipality-school district combination. Then for each election date we aggregate, for both school districts and municipalities, the number of registered voters who are eligible to vote in an election while the overlapping jurisdiction holds a fiscal stress label. We then divide this count by the total registered voter count for the school district or municipality to calculate the fraction of registered voters eligible to make a voting decision with an overlapping fiscal stress label present.
References
Balsdon, E., Brunner, E. J., & Rueben, K. (2003). Private demands for public capital: Evidence from school bond referenda. Journal of Urban Economics, 54(3), 610–638.
Beauliew, N. D. (2002). Quality Information and consumer health plan choices. Journal of Health Economics, 21(1), 43–63.
Berkman, M. B., & Plutzer, E. (2005). Ten thousand democracies: Politics and public opinion in America’s school districts. Washington, D.C.: Georgetown University Press.
Besley, T., & Case, A. (1995). Incumbent behavior: Vote seeking, tax setting, and yardstick competition. American Economic Review, 85(1), 25–45.
Bowers, A. J., & Chen, J. (2015). Ask and ye shall receive?: Automated text mining of Michigan capital facility finance bond election proposals to identify which topics are associated with bond passage and voter turnout. Journal of Education Finance, 41(2), 164–196.
Bowers, A. J., & Lee, J. (2013). Carried or defeated? Examining the factors associated with passing school district bond elections in Texas, 1997–2009. Educational Administration Quarterly, 49(5), 732–767.
Bowers, Alex J., Metzger, S. A., & Militello, M. (2010a). Knowing the odds: Parameters that predict passing or failing school district bonds. Educational Policy, 24(2), 398–420.
Bowers, A. J., Metzger, S. A., & Militello, M. (2010b). Knowing what matters: An expanded study of school bond elections in Michigan: 1998–2006. Journal of Education and Finance, 35(4), 374–396.
Brender, A. (2003). The effect of fiscal performance on local government election results in Israel: 1989–1998. Journal of Public Economics, 87(9), 2187–2205.
Brunner, E. J., & Ross, S. L. (2010). Is the median voter decisive? Evidence from referenda voting patterns. Journal of Public Economics, 94(11), 898–910.
Brusca, I., & Montesinos, V. (2006). Are citizens significant users of government financial information? Public Money and Management, 26(4), 205–209.
Button, J. W. (1992). A sign of generational conflict: The impact of Florida’s aging voters on local school and tax referenda. Social Science Quarterly, 73(4), 786–797.
Chernew, M., Gowrisankaran, H., & Scanlon, D. P. (2008). Learning the value of information: Evidence from health plan report cards. Journal of Econometrics, 144(1), 156–174.
Coe, C. K. (2007). Preventing local government fiscal crises: The North Carolina approach. Public Budgeting and Finance, 27(3), 39–49.
Colburn, C. B., & Horowitz, J. B. (2003). Local politics and the demand for public education. Urban Studies, 40(4), 797–807.
Dafney, L., & Dravnoe, D. (2008). Do report cards tell consumers anything they don’t already know? The case of Medicare HMOs. RAND Journal of Economics, 39(3), 790–821.
DeBartolo, G., & Fortune, P. (1982). The demand for public services: inferences from municipal bond referenda. National Tax Journal, 35(1), 55–67.
Dunne, S., Reed, W. R., & Wilbanks, J. (1997). Endogenizing the median voter: Public choice goes to school. Public Choice, 93(1–2), 99–118.
Ehrenberg, R. G., Ehrenberg, R. A., Smith, C. L., & Zhang, L. (2004). Why do school district budget referenda fail? Educational Evaluation and Policy Analysis, 26(2), 111–125.
Figlio, D. N., & Lucas, M. E. (2004). What’s in a grade? School report cards and the housing market. American Economic Review, 94(3), 591–604.
Fiva, J. F., & Kirkebøen, L. J. (2010). Information shocks and the dynamics of the housing market. Scandinavian Journal of Economics, 113(3), 525–552.
Fletcher, D., & Kenny, L. (2008). The influence of the elderly on school spending in a median voter framework. Education Finance and Policy, 3(3), 283–315.
Gradstein, M., & Kaganovich, M. (2004). Aging population and education finance. Journal of Public Economics, 88(12), 2469–2485.
Holcombe, R. G., & Kenny, L. W. (2007). Evidence on voter preferences from unrestricted choice referendums. Public Choice, 131(1–2), 197–215.
Honadle, B. W. (2003). The states’ role in U.S. local government fiscal crises: A theoretical model and results of a national survey. International Journal of Public Administration, 26(2), 1431–1472.
Ingle, W. K., Johnson, P. A., Givens, M. R., & Rampelt, J. (2013). Campaign expenditures in school levy referenda and their relationship to voter approval: Evidence from Ohio, 2007-2010. Leadership and Policy in Schools, 12(1), 41–76.
Ingram, R. W., & Copeland, R. M. (1981). Municipal accounting information and voting behavior. Accounting Review, 56(4), 830–843.
Jin, G. Z., & Sorensen, A. T. (2006). Information and consumer choice: the value of publicized health plan ratings. Journal of Health Economics, 25(2), 248–275.
Kloha, P., Weissert, C. S., & Kleine, R. (2005). Someone to watch over me: State monitoring of local fiscal conditions. The American Review of Public Administration, 35(3), 236–255.
Kogan, V., Lavertu, S., & Peskowitz, Z. (2016). Performance federalism and local democracy: Theory and evidence from school tax referenda. American Journal of Political Science, 60(2), 418–435.
Lentz, C. (1999). Predicting school referenda outcomes: Answers from Illinois. Journal of Education Finance, 24(4), 459–482.
Mikesell, J. (2013). Fiscal administration (9th ed.). Boston, MA: Cengage Learning.
Pope, D. G. (2009). Reacting to rankings: evidence from “America’s Best Hospitals”. Journal of Health Economics, 28(6), 1154–1165.
Rivenbark, W. C., & Roenigk, D. J. (2011). Implementation of financial condition analysis in local government. Public Administration Quarterly, 35(2), 241–267.
Rosenbaum, W. A., & Button, J. W. (1989). Is there a gray peril?: Retirement politics in Florida. Gerontologist, 29(3), 300–306.
Shober, A. F. (2011). Attracting capital: Magnets, charters, and school referendum success. Journal of School Choice, 5(2), 205–223.
Sielke, C. C. (1998). Michigan school facilities, equity issues, and voter response to bond issues following finance reform. Journal of Education Finance, 23(3), 309–322.
Spreen, T. L., & Cheek, C. M. (2016). Does monitoring local government fiscal conditions affect outcomes? Evidence from Michigan. Public Finance Review, 44(6), 722–745.
Tedin, K. L., Matland, R. E., & Weiher, G. R. (2001). Age, race, self-interest, and financing public schools through referenda. The Journal of Politics, 63(1), 270–294.
Thompson, P. N. (2016a). School district and housing price responses to fiscal stress labels: Evidence from Ohio. Journal of Urban Economics, 94, 54–72.
Thompson, P. N. (2016b). Effects of fiscal stress labels on municipal government finances, housing prices, and the quality of public services: Evidence from Ohio, working paper.
Trusts, P. C. (2013). The state role in local government financial distress. Washington: DC, Pew Charitable Trusts.
Vermeir, J., & Heyndels, B. (2006). Tax policy and yardstick voting in Flemish municipal elections. Applied Economics, 38(19), 2285–2298.
Zimmer, R., & Jones, J. T. (2005). Unintended consequence of centralized public school funding in Michigan education. Southern Economic Journal, 71(3), 534–544.
Zimmer, R., Buddin, R., Jones, J., & Liu, N. (2011). What types of school capital projects are voters willing to support? Public Budgeting and Finance, 31(1), 37–55.
Acknowledgements
The authors are grateful to Mike Conlin, Alison Johnston, and Carol Tremblay for their comments and suggestions. A special thanks to Kate Epstein for providing expert copy-editing services on the final version of the paper. We also want to thank the participants of the 41st Annual Conference of the Association for Education Finance and Policy for their helpful comments and discussion.
Author information
Authors and Affiliations
Corresponding author
Appendix
Rights and permissions
About this article
Cite this article
Thompson, P.N., Whitley, J. The effect of school district and municipal government financial health information on local tax election outcomes: evidence from fiscal stress labels in Ohio. Public Choice 170, 265–288 (2017). https://doi.org/10.1007/s11127-016-0395-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11127-016-0395-7