Special taxing districts are all the rage. They are now the fastest growing form of government, even though they often operate out of sight and out of mind for most voters and taxpayers. These often unnoticed local governments have the power to tax and spend like municipalities and are set up to provide a specific service. They have contributed heavily to the overall growth of local government spending and debt loads, all while operating virtually behind the scenes and at a lower standard of accountability than traditional local governments. Many taxpayers don’t even realize they are paying taxes to these government-like entities.
The number of special districts has grown rapidly because special districts are not subject to many of the spending and debt limits municipal governments face. This has encouraged local governments to circumvent those limits by creating more special districts. Statistical analysis of states and cities across the nation shows that the share of local spending attributable to special districts correlates strongly with more overall per capita local government spending.
Special districts tend to have little accountability to voters. Special district elections are often timed for non-traditional periods, such as odd-numbered years or during the summer, when special interests can rally the troops to control the district’s creation and operation.
Moreover, multiple studies have shown that special districts provide services at a higher cost, and their services are no better or sometimes worse than those provided by municipal governments or the private sector.
To rein in this form of shadow government, the legislature should make special districts more accountable and limit the creation of special districts to necessary purposes. The caps placed on local borrowing and spending should be amended to apply to special districts as well. Additionally, more transparency measures should be implemented. Finally, special district tax and bond elections should take place alongside all other elections in November.
Special districts have long enjoyed loopholes that allow them to exist and operate outside normal limits on government power, resulting in “stackable Leviathans” that increase the tax burden. It is time to bring special districts into the light of day.
Introduction
In 2002, the Maricopa County Board of Supervisors in Arizona contemplated the future of the county-run Maricopa Medical Center hospital, which had become increasingly costly to operate. They would eventually spin the hospital and a few of the county’s health care programs into what is called a “special taxing district”—a layer of government that, as we’ll see in this study, is often less accountable to voters, less efficient than traditional county or city governments, and not subject to the same constitutional tax and expenditure caps that limit traditional local governments. Yet these districts are able to levy property taxes and spend money just like a county or city government.
The story of how the Maricopa hospital district came into existence is similar to hundreds of stories from hundreds of cities and counties across the nation over the past four decades. It’s also a story about how policymakers often use special districts to increase the size of government.
First, the Maricopa County Board of Supervisors convened a “citizen’s task force” to weigh the options. The task force was composed almost entirely of those who had a business interest or stake in the health care industry, as well as three board members of the Maricopa Integrated Healthcare System (MIHS), a system that included the Maricopa Medical Center.1 They met periodically between December 2002 and March 2003, after which they released their final report.
The task force report made clear that one of the main problems at the county hospital was not “excesses and inefficiencies” but, at least in part, “the effect of County expenditure limits on MIHS.”2
The task force report laid out the reform options and criteria they considered. In describing their criteria, the report noted that recommendations “with little or no prospect of legislative and public support [were] assiduously avoided.”3 Of the 10 options (which included both privatization and maintaining the status quo), four of them were to spin the health system off into a special district. Spending cuts at the hospital were dismissed early in the deliberations. The task force also dismissed the option of laying off some of the hospital’s 3,700 employees because it would “contribute to the state’s unemployment issues.”4
In making its final decision, the task force explicitly considered whether an option would be exempt from the spending limit.5 Task force members actually identified three statutes that outlined the special district arrangements under which they could do this with voter approval. A fourth statute required legislative action before Maricopa County could bring the creation of a new health services district to the voters, because the county exceeded the population limit in the statute. Any of these options would have allowed the county to move the hospital off of its constitutionally limited books and instead place it in a new district that would not have the same spending limits.
The task force finally recommended the type of special district that required legislative action, and the Board of Supervisors pursued that course of action. The Arizona Legislature expanded the eligibility of the statute the board wanted to use to create the special district. In August 2003, the board referred the creation of a new hospital district to the ballot.
The election materials the board distributed to voters stated that, “unlike law enforcement and criminal justice, the county is not required to operate a hospital.” But then they hastily noted that “closing the hospital would be devastating to the community,” even though doing so was far from the only option. For instance, the board’s own task force had listed other ideas, including the sale of the hospital to a private firm.
In a special election in November 2003, the voters passed the proposition that created the special hospital district. It was an off-year election, so voter turnout was light—around 13 percent, about a quarter of even-year November elections in the county—and the proposition won with 58 percent of the vote.6 That amounted to a total of 94,851 votes cast in favor of the proposition, a margin of victory of 28,501 votes. Or, to put it another way, roughly 2 percent of all registered voters in the county—fewer than one in 45 residents—created a special taxing district empowered to assess property taxes on every homeowner in the county.
Since 1994, Maricopa County had subsidized the medical center and its related health care programs to the tune of at least $178 million in total.7 Roughly $36 million of that subsidy came from the county general fund in 2003, and around $15.3 million had been authorized for fiscal year 2004. Moving the county health care system into the new district allowed it to spend roughly that amount of money on other programs.
This story is not unique, and neither are the consequences. Over the past 50 years, thousands of special districts have been created, many of them with the power to tax and spend. According to the Census Bureau, the number of special districts in the United States more than tripled between 1952 and 2012, burgeoning from 12,340 to 38,266. During that time, other types of local government increased by only 5 percent, from 37,061 to 38,910.
Meanwhile, the number of school districts actually declined dramatically due to a movement toward consolidation during this period, shrinking the number from 67,355 in 1952 to around 12,880 by 2012. Non-education special districts, on the other hand, account for over 40 percent of all state and local government entities today.
Spending by districts have been growing faster than traditional local government budgets and have become a major portion of government spending in many states. According to U.S. Census data, total real per capita local government spending (excluding education) grew 59 percent from 1977 to 2007. During that same period, real per capita special district spending more than doubled (118 percent). If special district spending were a line-item in an average local government’s ledger, it would be the third largest spending item, right behind welfare spending and education spending.
As we will see in this study, special districts have not only been the fastest growing form of government over the past two generations, they also have contributed heavily to the overall growth of local government spending and debt loads, all while operating at a lower standard of accountability than traditional local governments. This might be described not as a matter of Leviathan-like government growing out of bounds, but as a multitude of little Leviathans, all stacked on top of each other, both abetting and aiding the growth of the other. How this came to be and the institutional environment in which it occurred can reveal a great deal about the growth of government at the local level over the past 50 years. No strategy to rein in the spending appetites of state and local government would be complete without also tackling the growth of these multiple layers of government.
The Growth of Special Districts in the U.S.
Table 1 ranks the states based on the number of special districts per 100,000 people in a state. The number of special districts can vary by state based on characteristics unique to a state. For instance, rural states (like Wyoming and North Dakota) tend to have more special districts per capita than other states. Yet states with even a moderate level of special districts per capita can suffer from having a large number of governmental units overlapping particular geographic areas.
Table 1: Number of Special Districts (per 100,000 people), 2007
Source: Author’s calculations based on U.S. Census Bureau data.
Not all districts have the power to levy taxes—usually property taxes—but many do. (Other districts have only the power to raise money through fees.) The Census Bureau collects data on which districts have the specific ability to tax.9 The most recent published data show that the percentage of special districts with the power to tax can also vary greatly by state. Table 2 shows the percentage of special districts in each state with this power. Arizona leads the pack, with nearly every one of its special districts able to levy a tax of some sort.
Table 2: Percentage of Special Districts with Taxing Power
Source: Author’s calculations based on U.S. Census Bureau data.
The fact that many special districts have the power to tax is not simply an idle academic matter. As a practical matter, it can translate to wildly different property tax burdens for similarly situated homes. As reporter Michelle Lee of the Arizona Republic discovered, a home with the same property value as another home a mile away could pay dramatically more—sometimes up to 15 times more—in property tax add-ons as a result of special districts.10
Most special districts provide a specific service. In fact, “single purpose district” and “special taxing district” are usually synonymous in the literature and census descriptions. Special districts have had a reputation in the past as simply being a means of financing certain construction projects. That might have been true when special districts started to become common in states during the 1930s, but today that description isn’t accurate. Nationally, 60 percent of districts provide a program or service with their own employees, and 30 percent indirectly provide a program or service through contractual agreement with a private-sector company. Only about 22 percent construct public facilities.11 So while some districts fit the stereotype, most of them function like service-specific government agencies. Table 3 shows the top 10 categories of services that special districts undertake. The most common is fire protection. The top 10 functions listed here account for around 70 percent of all service-providing special districts.
Table 3: Special Districts by Top 10 Common Functions (percentage of total)
1. Fire protection 16.3
2. Water supply 9.7
3. Multi-function 9.1
4. Drainage 7.4
5. Soil and water conservation 7.1
6. Hospitals 6.2
7. Sewerage 5.7
8. Libraries 4.5
9. Irrigation 2.4
10. Highways 2.1
Source: Author’s calculations based on U.S. Census Bureau data.
Another attribute that people commonly associate with special districts is that they transcend normal political boundaries—meaning they straddle two counties or cities. A classic example is the irrigation district that follows the flow of a river. These types of districts are in the minority, however. Only about 10 percent of special districts cross the boundaries of two or more counties. As University of Chicago public policy professor Christopher Berry describes it, “For the most part, special districts are not fulfilling roles that existing governments are too geographically limited to fill. Between 1967 and 1992, in fact, the fastest-growing type of special district was one that shares exactly the same boundaries as an existing city or county. The multiplication of coterminous districts obviously cannot be explained by their greater territorial flexibility or the need for regional governance. Thus, while regional scope may be the rationale for some districts, it cannot be considered a general explanation for the growth in special districts.”12
Finally, a number of special districts exist mainly to serve as “conduits” for the financing of certain projects or services. The creation of the district would allow an expansion in the type of government-grade (i.e., tax-deductible on a taxpayer’s federal tax return) debt that can be issued. This is often the case with transportation-related districts, like road districts or air transportation districts. But all of these types of districts are in the clear minority when it comes to their share of total special districts in the United States today. In fact, only about 30 percent of special districts issue debt as a means of providing their services.
Where these types of districts may be especially attractive to municipal policymakers is in their ability to issue tax-free debt on behalf of a project that mainly enriches private interests. The Census Bureau collects data on what it describes as “public debt for private purposes.” The most recent data indicate that about 30 percent of all state and local public debt is of this sort. It’s easy to see that the federal subsidy to municipal debt through the tax code can indirectly subsidize these projects, often termed “community development” or “industrial development” projects. Indeed, a significant portion of special districts exist to finance projects of this sort.
Special district growth in Arizona
In Arizona, the number of special districts grew from 34 in 1952 to 326 in 2012. Most of that growth—more than 70 percent—occurred after 1980. The creation of special districts has grown even faster than the population. Figure 1 shows the trend in Arizona and measures it against population growth and the share of the population that resides in an urban area.
One argument that proponents of special districts offer is that large urban centers can support having general-purpose municipal governments provide a number of basic services—like a fire department or a sewer system—but rural areas can’t do so as easily. Hence, special districts need to be created to provide these services and capture the “economies of scale” that come with serving disparate populations over long distances. As we’ve seen already in Table 1, areas with low urbanization do tend to have higher rates of special district creation. However, lack of urbanization is not likely to be a driver of special district creation for most states.
Figure 1: Special Districts, Population, and Urbanization in Arizona
Source: Author’s calculations based on U.S. Census Bureau data.
As Figure 1 shows, both population growth and the scope of urbanization have little to do with the pattern of growth in special districts in Arizona over the past 30 years. While there does appear to be a link between the two prior to the 1980s, there is little connection since then. Arizona population has been largely centered around urban areas since the 1950s, so levels of urbanization haven’t grown much—or, to put it another way, the population hasn’t become any more rural over time, and it cannot therefore be the main reason for the growth of special districts.
When viewed in the overall context of special district growth across the nation, this isn’t surprising. As political scientist Kathryn Foster points out in her book on special districts, “Contrary to predictions, population growth rates have an insignificant influence on the number of districts in metropolitan areas. Fast-growing metropolitan areas are not more likely than slow-growing areas to have special-purpose governments.”13
Why Are There So Many Special Districts?
So what can explain the growth? Institutional factors and changing public policies seem to matter most.
Nationally, the first period of special district growth occurred during the New Deal era of the 1930s. Many federal programs that originated then offered states and localities money tied to the creation of special districts. Of those, many were related to capital projects. This has created a general perception that special districts are mainly created to facilitate construction projects—a perception that doesn’t fit most special taxing districts today.
The second era of special district growth occurred in the 1980s. Just prior to this second era, voters in many states approved debt, tax, and expenditure limitations on local governments. Some of these initiatives were more comprehensive than others. All told, 36 states have some kind of limit on tax revenue or expenditure growth, and many of those states have additional limits on property tax rates and assessments.14
The spending, tax, and debt limits placed on local governments seem to be a driving force in the creation of new special taxing districts because local governments use the districts as a way to skirt the spending limits. Some of these limits have been in place since as early as 1875, but most were put in place in the 20th century, with the majority emerging in the 1970s and 1980s.15 Those limits generally apply to cities and counties, but usually not to special districts. Therefore, government officials free up room in their expenditure-limited general fund budgets by spinning off functions into special districts.
Studies conducted since the 1970s have verified this phenomenon.16 A comprehensive 1990 study included nearly 300 suburban metropolitan statistical areas and adjusted for population growth, population concentrations, institutional factors (such as whether cities have “home rule” authority) as well as the presence of tax limitations, particularly on property taxes.17 Of all these factors, the one that came out with the highest level of statistical significance as a way of explaining the growth of special districts was the presence of tax limitations. The author, Illinois State University economist Michael Nelson, concluded that tax limits, particularly on property taxes, “spur special district formation as local officials and/or citizens see increased use of this unit of government as an effective means to circumvent this regulation.”18
An analysis from 2000 included not just tax and expenditure limits, but also debt limits that apply to local governments. The author, Barbara McCabe, currently of the University of Texas – San Antonio, concluded that her results “generally substantiate the idea that state rules influence the creation of special districts. State restrictions on general-purpose local governments tend to lead to the formation of special districts. State limits on cities’ and counties’ fiscal authority—in the form of either tax and expenditure limits (TELs) or debt limits—generally increase the likelihood that more special districts will appear in the state.”19
Another study focused on special districts in Illinois arrived at the same conclusion but in a different way. Illinois often draws the attention of researchers studying the growth of local governments, because that state has among the largest number of special districts in the country. In this case, the study looked at just over 100 Illinois counties and discovered that those without tax limits had vastly fewer special districts than those with them.20 That is, counties that didn’t have tax limits to circumvent were less likely to spin off government functions into special districts. This result strengthens the case that special districts often serve the purpose of circumventing limits on government growth at the local level.
The limits that can be placed on county and city governments include general expenditure limits, overall property tax revenue limits, and overall property tax rate limits. According to data from the Advisory Council on Intergovernmental Relations, 25 states have at least one type of limit on lower levels of government, but presumably the limit does not also apply to special district governments. The average number of special districts per 100,000 people in the states without these limits is 12.9. For states in which there is at least one type of expenditure or property tax limit, the average is almost twice as high (25).21
This is not an argument against the need for tax and spending limits, however. Instead, it’s proof that current spending limits are too porous to truly restrain the growth of government, which is the reason to have them in the first place. As we’ll see next, the explosion in the number of special districts across the country is actually one of the drivers of local government growth and helps explain why some empirical studies have concluded that tax and expenditure limits are insufficient to limiting spending growth.
Special Districts Lead to Bigger Government
Substantial empirical evidence indicates that the proliferation of special districts leads to more government spending. The mechanism by which this works is highlighted by the Maricopa County hospital district example. To put the result simply, the more special (single-purpose) districts a locality has—and, for that matter, the larger the share of per capita government expenditures composed of special district spending—the larger the overall size of local government. This shouldn’t be surprising. It’s simple mathematics: When a special district is created to circumvent a local tax and spending limit, this new layer of “government” can spend money that is added to the total spending by traditional local governments. As a result of freeing up more budgetary resources in their general funds by creating special districts and off-loading government functions from the balance sheet, local governments can now spend even more than before.
And because special districts are not usually bound by spending caps, they have grown faster than local government budgets, even after adjusting for inflation and population growth. To put this in perspective, average real per capita local government spending (minus education and special district spending) grew 25 percent between 1992 and 2007. But the average special district budget grew, in real per capita terms, by almost twice that—46 percent.
Arizona special district budgets were particularly expansionary. Real per capita local government spending grew by 13 percent. Special district spending, on the other hand, grew at more than triple that rate, by 42 percent.
A recent study by Florida Gulf Coast University economist Dean Stansel showed that when there are more special districts per capita, overall local government spending increases.22 His study was based on the U.S. Census Bureau’s Census of Governments between 1962 and 1992 for 314 metropolitan areas. After adjusting for a number of demographic and state-specific factors, he discovered that the number of special districts per 100,000 residents had a statistically significant and positive relationship to the size of total per capita local government spending. (Positive, in this context, means the variables move in the same direction, growing together.)
A working paper from Rutgers economist Christopher Goodman used a similar methodology for 1982 to 2002, but went further by adding a variable that measured the share of local spending at the special district level and by separating school districts from the special district totals. He discovered that the results got even stronger. Not only was the number of special districts more strongly positively related to the size of local government spending but so was the share of special district spending. He concluded that this clearly indicated “single purpose governments are driving the increase in per capita total expenditures.”23
That relationship, illustrated in Figure 2, shows there is some positive relationship between the share of special district spending and real per capita local government expenditures.
Figure 2: The Simple Relationship Between Special District Share of Spending and Real Per Capita Local Government Spending, 1992 – 2007
Empirical analysis can be done to see if this relationship holds up over time after being subjected to a number of control variables. The analysis conducted for this paper is a hybrid model based on the methodologies of the two studies cited above, but it uses the most recent data available from the Census of Governments.
The regression model, the full results and explanation of which appear in the appendix, includes the following variables. They were selected for inclusion since prior studies found a relationship between them and the growth of government:
- Special district spending as a percentage of total local government expenditures24
- The share of local spending that is financed by “intergovernmental revenue,” such as general revenue shared by the state government and federal grants to local governments25
- Real per capita personal income26
- The percentage of families living below the poverty line27
- The percentage of the state’s population with 16 or more years of schooling28
- The percentage of population living in an urban area29
- The size of a state’s population30
As expected, and in line with the other empirical studies on this subject, the share of special district spending was significantly and positively related to the size of real per capita local government spending. Essentially, each percentage-point increase in the size of a special district budget relative to the overall budget of all local governments translated into a $26 real per capita increase in the size of local government spending.31 This variable was the third-biggest driver of local government spending, behind overall growth of the population and the poverty level.
Special Districts = High Costs + Low Efficiency
Another reason that special districts drive up total local government spending is they tend to be more costly relative to alternative means of providing the same service. For example, some studies of special districts focus on fire protection districts and compare them with comparable fire departments run by a general municipal government. In each of these studies, the fire protection service provided by a special district was more expensive on a per capita basis than that provided by a city or county government. The result held even after adjusting for the costs of capital expenditure to build a firehouse and purchase fire trucks. The fire districts simply did not capture the “economies of scale” that their proponents often claimed was a beneficial feature of having them exist as special taxing district entities instead of as part of a general-purpose government.32
Higher costs do not necessarily indicate a higher level of service, nor do they translate into more customer satisfaction. A 1997 survey in Michigan revealed that citizens there were more satisfied with fire services provided by general-purpose local governments than those provided by special fire districts. In fact, a vast majority of those surveyed preferred services generally to be provided by a city or county government instead of a special district. Only 3 percent felt differently.33
Low efficiency levels and poor quality of service are a recurring, robust, and significant pattern in the studies that have been published, notes University of Chicago political science professor Christopher Berry. He concludes that the weight of the evidence suggests that as a general rule “special-function districts are less efficient and provide no better quality than general-purpose governments.”34
Costs of special water districts are another good example. A study of 26 rural Nevada water districts discovered that utilities offered by the special districts were significantly less efficient than those operated by city or county governments. The municipality-operated water utilities ran at 90 percent efficiency, just above the average level of 88 percent. The water district utilities operated at just 85 percent efficiency. Not surprisingly, neither type of public utility was more efficient than the utilities provided by private firms in that state (which were 91.3 percent efficient).35
There is a substantial amount of research on another type of special district that may seem mundane: park districts. However, these districts are no less prone to the influences that create higher costs but provide lower levels of service. A study of park districts in 10 Midwestern states found that, in per capita terms, they spent nearly three times the amount on average than was spent by parks and recreation departments of general purpose governments.36 Yet, as Berry describes it, “the average park district provides, on a per capita basis, fewer park facilities, fewer recreational programs, and fewer acres of parkland than the average parks and recreation department.”37
Library districts are another example of a high-cost, single-purpose government with bloated overhead and less desirable services than those offered by general-purpose governments. Berry and his co-author, Jacob Gerson of the University of Chicago Law School, analyzed more than 100,000 library districts across the nation. They concluded, “Library districts do spend significantly more than other types of libraries. But library districts have fewer books. Districts have a comparable number of employees, although their employees are less likely to be actual librarians, relative to libraries operated by general-purpose governments.”38 Using the number of visits as a proxy for demand of the libraries provided by these special districts, the authors conclude that, despite spending more, “library districts do not attract more visitors or have higher circulation than municipally run libraries.” So spending more does not produce a larger book collection, nor does it reflect higher utilization—it’s a classic case of government inefficiency. In fact, library districts spend about 10 percent more per patron visit than comparable municipal libraries.39
One potential explanation for the cost differentials in these studies is that these special districts may simply be more likely to have high capital costs for buildings and the like. After all, special districts have long been thought of as being primarily involved with capital projects or at least providing services that require higher capital costs and correspondingly higher debt levels to finance them. Yet analysis of the Census Bureau’s Census of Governments suggests this reputation is unwarranted. Studies that have used this data to pin down which categories constitute the lion’s share of special district spending conclude that the category of “current spending”—in other words, spending on providing services—is much larger than capital spending. The average per capita expenditure on current spending was $1,846 in the 2002 Census of Governments, while average capital spending was less than 15 percent of that ($267 per capita).40 It’s difficult to argue that these levels of capital costs are driving the observed phenomenon of increasingly costly special district services.
What likely explains these results are instead political and institutional realities that sabotage or discourage accountability, transparency, and pressure to be more efficient, such as having to compete with other agencies for limited general funds in the context of a traditional budget-making process by a traditionally elected city council or county council. That’s the explanation to which we turn next.
Special Districts Are Harder for Taxpayers to Monitor and Heavily Influenced by Special Interests
Just as there is an institutional explanation for the proliferation of special districts—porous tax and expenditure limits—there is a simple, institutional explanation for the increased costs that researchers uncover time and time again with respect to special district operations. The answer is that multiple levels of government, all of which have the power to tax the same pool of property and people, can be harder for taxpayers to consistently monitor than one or two general-purpose governments, such as a city or county. As a consequence, the special districts can be heavily influenced by special interests, like labor unions, that can dominate low-turnout elections and that frequently have the ear of the district boards, with little countervailing pushback or feedback from voters.
Accountability becomes further reduced when a special district’s decision makers are appointed instead of elected. And in the cases when the policymakers are elected or their decisions require voter approval, deliberate disenfranchisement often follows in the form of elections that occur in off years or at unusual times of the year.
The idea that special district governments are generally less accountable has been a theme in the published analysis since at least the 1950s, shortly after they began to multiply quickly in the wake of New Deal programs that encouraged their creation with the promise of federal money. J.C. Bollens, one of the earliest political scientists to study special districts, noted in his seminal work on this topic in 1957, that the proliferation of these districts can create “confusion, misunderstanding, and indifference.” He noted the tendency for special districts to be headed by those who were appointed, not elected, which made them “at best, twice removed from the voters.” He even went so far to describe the decision-making process in special districts as a “multi-ring circus.”41
Now, more than 55 years later, we have substantial empirical evidence that these insights were basically correct. Data show that more overlapping governments makes it harder for voters and taxpayers to monitor the activities of those governments. Thus, the tax base on which all of these overlapping governments feed is a “common pool” from which small amounts are extracted from less-suspecting (or less active) taxpayers by more engaged and better-organized interest groups.
One good way to test whether an increasing number of districts leads to higher levels of spending and taxation comes from Charles Berry’s seminal and comprehensive analysis based on an “overlap index” he created using the Census Bureau’s Census of Governments data. For all counties in the United States, he measured the number of jurisdictions that overlap the average city within a county. He then conducted statistical analysis to determine whether this overlap had any influence on the tax burden within the county.
His analysis showed that more overlap leads to higher spending and tax burdens, even after adjusting for “inherent attributes of counties” that might make them prone to higher government spending, such as demographics and location. “To put the result in perspective,” he writes, “increasing from a relatively low level of overlap (the 25th percentile) to a relatively high level (the 75th percentile) results in an increase of $130 in locally raised [per capita] revenue,” which Berry notes is equivalent to an 11 percent increase above the average.42 This is entirely consistent with the behavior you would expect from a government that is less accountable, more opaque, and which taxpayers have less cognizance of than a general-purpose government.
Advocates of special districts often argue that those districts are undertaking roles that the municipal government can’t or shouldn’t, and therefore the tax burdens need to be higher to support the additional activities. Yet the published analysis indicates that’s not true either. As Berry writes, the results that he and others have generated over the years “strongly suggest that the effect of jurisdictional overlap on local budgets is not simply an artifact of changes in local functional performance, but rather reflects increased spending on the same functions that would otherwise be provided by general-purpose governments.”43
All of these results are consistent with the theory that special districts are created as a means to circumvent spending and tax limits. They also imply that the subsequent growth in government taxation and spending is the result of institutional factors that put average taxpayers and voters at a severe disadvantage relative to the interests that have a stake in the creation and operation of a special district.
Voter turnout and non-traditional elections
One factor that makes special districts less accountable to voters is the fact that they generally have low voter turnouts in the elections that create them or that pertain to their operations, such as levying a tax or issuing debt. The stories that researchers have collected from across the country make the story of the creation of the Maricopa County hospital district look like the norm instead of the exception.
An extensive 2007 study of special district elections in Nassau County, New York, discovered that voter turnout was far lower than that of general elections. The 2002 gubernatorial election turnout was 44 percent in that county. But the turnout in the succeeding special district election ranged from 2 percent to 14 percent.44 Other studies, although not as comprehensive, find similar examples. One study of Michigan school board elections found that turnout among registered voters averaged 7.8 percent. Three-quarters of the districts had turnout of less than 10 percent.45 Another example comes from a Texas utility district in which a $14 million bond was issued with the turnout of only 32 voters.46
Some of these examples may seem like outliers, but low voter turnout in special district elections is commonly observed by scholars who have studied the issue. There is even general consensus regarding one of the main reasons for low turnout: these elections usually occur in off-years or at unusual times of the year. The Nassau County Executive office, which conducted the study of its special district elections, found that those elections were held on 24 different days of the year. It also found that there was at least one special district election in 11 of the 12 months of the year. Yet not a single one of the special district elections was held on the same day as the general election in November. The report concluded: “The scattered nature of the special district elections, and the extreme difficulty involved in obtaining some of their election dates, place an undue burden on voters. Even the most conscientious of Nassau residents would find it difficult to vote in all of the elections that concern them.”47 Additionally, the Michigan school elections studied in the examples cited here were all in June.
In California, the Little Hoover Commission found that turnout was significantly lower in special district elections than the turnout for other types of local government elections throughout the state largely because of non-traditional election schedules. The report also outlined the outcome of an intriguing natural experiment: Sacramento County required its special districts to hold elections in even-numbered years. This simple change increased voter turnout to 45 percent of registered voters from where it was initially at 17 percent. A 1994 study of school districts in California’s Riverside County found a similar pattern when it changed elections to even-numbered years.48
A study of Cook County, Illinois, by the University of Chicago’s Christopher Barry found that voter turnout in special district elections was “almost universally lower than in general-purpose elections.” The special district elections—for school, library, park, fire, and community colleges—occurred in April of odd-numbered years. Turnout for those averaged 21 percent. By comparison, the turnout for the November 2004 election was nearly 50 percent of registered voters. In addition, the overlap of the many jurisdictions meant there was an overwhelming number of elections for voters to keep track of. Barry estimated that “a typical resident of Cook County is represented by 70 different elected officials and would have to go to the polls on six different dates to vote for all of them.”49
For many local policymakers, low voter turnout is a feature, not a bug. When an election is in the middle of the summer in Phoenix, for instance, when vastly fewer people are around—and the election date is chosen purposely by local officials because of that fact—it is much more likely that these elections don’t accurately represent the normal preferences of the general public. Those most motivated to influence the election have the wherewithal, resources, and inside information about these contests that may seem like a low-stakes contest for the average voter. In other words, while most general elections are about voters choosing elected officials, most special district elections that are intentionally placed on unusual dates are usually more about the elected officials choosing which voters will participate.
That was the conclusion of a 1997 study that compared all school and municipal bond elections in Oklahoma from 1988 through 1992, all school and municipal bond elections in Ohio from 1963 to 1987, and a subset of all school bond elections held in 20 states during 1994. The study concluded that local officials systematically avoid holding bond elections at the same time as the general election and that bond referenda are more likely to be approved when there is comparatively lower voter turnout.50
Control over the election can also sometimes extend to the locations in which those elections will take place. A fire district in Colorado was roundly criticized by the media there for a 2006 referendum on a $43 million bond issue and tax increase in May of that year. All 16 polling stations were at fire stations. As a Rocky Mountain News editorial writer mused, “If you’re opposed, all you need is the courage to show up at any given station and ask a burly fireman where you go to vote against the tax increase.”51
A vivid example of how those with a greater personal stake in the outcome of a special district election are more likely to vote comes from the research of Terry Moe at Stanford University. He examined 14 school district elections in California between 1997 and 1999. For each of these, he also gathered data on the names and ZIP codes of school district employees and was able to match these names to county voter files. This allowed him to track each employee’s voting history. During his research, he found that many teachers lived in a district different from the one that employed them. As might be expected, the teachers who lived in the same districts in which they were employed had a much higher stake in those school district elections than those who didn’t live in the districts that employed them—and that’s exactly what Moe found. The voting turnout for teachers who lived in a district different than the one where they were employed was 18 percent. But the turnout for a teacher who worked in the same district in which she lived was over twice as high, at 46 percent.52
It’s easy enough to see the intangible costs of often-blatant attempts to disenfranchise all voters except for those who have a strong stake in the outcome. As the Advisory Council on Intergovernmental Relations concluded in its 1987 report that a single special district in a geographic area might be held accountable relatively easily, “multiplying the number of single-function units increases the costs of accountability overall.” There are examples, however, of explicit costs of government being higher as a direct result of voter disenfranchisement. The University of Chicago’s Christopher Berry and Jacob Gersen found that school districts that held their elections in odd-numbered years negotiated significantly more generous contracts with teachers than did districts that held their elections in even-numbered years coincident with state and national elections. “Comparing teachers at an equal salary ‘step,’” the authors conclude, “those working in districts with odd-year elections earn roughly $2,000 (or about 4 percent) more than those working in districts with even-year elections. Moreover, we find evident that turnout goes up and salaries go down within a district when its election date is changed from odd to even years.”53
Recommendations
The proliferation of multiple layers of government that can speed the growth of local government spending, abetted by a lack of transparency and captured by particular political interest groups, is a trend that must be stopped. Local policymakers are not likely to unilaterally undertake such reforms, so it is incumbent upon state legislators to put into place taxpayer safeguards that not only make these layers of government more accountable, but also over time reduce their number.
Some of the proposals outlined here can be undertaken incrementally. Some require federal action. But they all point to a new institutional framework that aligns the interests of taxpayers and government officials in a way that is absent in the current policy environment.
1.) Make tax and spending limits apply to all forms of local government, including special districts. If a state has a tax and expenditure limitation on localities, then special districts should be brought under this limit as well. This will require the services provided by the special district to compete with other general fund budget services in a way they don’t have to now. This could also facilitate privatization of some services. While it’s important to make sure that the districts included in the calculation are those special districts that were created with the direct approval of the relevant local government—not one mandated by the state or federal government—it’s likely that most special districts will be included.
State policymakers can require all new special districts that are created within the boundaries of a county or city to be counted as counting toward the applicable constitutional or statutory spending limit if one exits.
Alternatively, state policymakers can forbid localities from creating new special districts until the total aggregate tax burden or spending of all local governments combined, including special districts, equals no more than the relevant tax or expenditure limit. A transition period with a deadline can be allowed to avoid disruption in certain services. But over time, localities should be forced to deal comprehensively with the scattered array of special districts simultaneously operating within a city or county boundary.
Localities could also be given the power, if they don’t currently have it, to review and terminate special districts that no longer serve a useful purpose or are shown to be less efficient than the alternatives. This could occur via a vote by a city council or county board to dissolve a special district. It could also mean that special districts, as a general rule, are subject to “sunset” provisions the way other state agencies that provide a service are. Thus, the burden of proof to keep the special district operating after the sunset date is on the district itself. To that end, state legislators could also require that localities routinely perform a cost-benefit analysis of existing special districts to make sure they are efficiently providing the services they were created to provide, and honestly appraise whether those services could more effectively be provided by the general-purpose county or city government, or even by the private sector.
2.) Make special districts more transparent. Special districts have many of the powers that traditional local governments have. Therefore, they should be required to meet the same standards of transparency as those governments as well. Debt levels, spending levels, and revenue collection data should all be reported to either a relevant county or state agency for inclusion in an annual report or searchable database available to the public.
Frequent and regular audits by an objective auditor, such as the state auditor general office, should be a common practice. State legislators could require these by law. In addition, the creation of new special districts could be strictly forbidden in a county or city until all of the existing and functional special districts operating within their territory have an audit no older than five years. If a district audit uncovers a need for reform of a specific district’s operations such as their accounting systems and practices, a report showing substantial progress on these reforms needs to be furnished to and verified by the state auditor’s office before a special district can be created.
A “boundary transparency” database could also be constructed and administered by the state government that would make it easy for voters to enter their address and see how many special districts overlap their property. It could include the name of the district, the purpose of the district, and the tax rate applied. Integration with existing budget transparency databases could also provide taxpayers with links to spending and debt data.
3.) Make special districts more accountable. A number of policy reforms can be enacted to make special districts more accountable. To minimize the incentive of a special district to embark on a project that enriches a specific private interest or special political interest, policymakers should enact a requirement that only elected boards may issue government-grade debt. This would likely reduce the attractiveness to use a special district mainly as a mechanism to finance a project that may not pass muster with voters.
Special districts have a long way to go to repair the disenfranchisement they have practiced for decades. To both discourage the scheduling of election dates that are engineered to ensure low voter turnout, and to help policymakers determine which special districts are the most useful and important to voters, a “voter turnout trigger” can be enacted.54 If voter turnout falls below a specific level for two subsequent elections, the special district would be dissolved. This proposal could also be a complement to the aforementioned power given to local governments (county and/or city) or the state to dissolve special districts that have outlived their usefulness, are inefficient, or even corrupt.
Finally, and perhaps most important, special districts should be required to consolidate their elections with all other traditional government elections in November. This will increase voter turnout and also diminish the likelihood that special districts will be captured by political interests that might be arrayed against the best interests of voters or taxpayers.
Conclusion
Getting a handle on special districts is the missing piece of the state tax revolution that began in the 1970s. That period of history gave us many state tax and expenditure limits that were meant to finally put a leash on the spending appetites of Leviathan at the state level. Special districts have long enjoyed loopholes that allow them to exist and operate outside these limits. Policymakers should look toward the goal of eliminating those loopholes as a means of finally and effectively restraining the growth of government at the local level.
Appendix
The empirical estimates cited in this paper come from an ordinary least squares (OLS) regression model designed to gauge the influence of a number of demographic and policy variables on the change in real per capita local government expenditures. The model is cross-sectional and uses data from 48 states for the period of 1992-2007. It excludes Alaska and Hawaii, as do many fiscal studies of this sort, because these states are outliers in the data due to their unique geographic characteristics. In the case of Alaska, the singular fiscal reality that most of its state and local revenue is derived from oil royalties rather than taxes also makes it incomparable with other states. Heteroskedasticity has been adjusted for by using Eicker-White standard errors. The results are below.
References
1. The task force’s final report includes the names of the members as well as the final recommendations: http://www.maricopa.gov/agendas/pdf/taskforcerpt.pdf.
2. Ibid, 6.
3. Ibid, 3.
4. Ibid, 16.
5. Ibid., appendix, Exhibit 7.
6. Author’s calculations based on data from the Maricopa County Recorder/Election Office, http://recorder.maricopa.gov/electionresults.
7. State of Arizona, Office of the Auditor General, Financial Statement Audit of the Maricopa County Medical Center, Fiscal Year 2003, January 9, 2004, http://www.auditorgen.state.az.us/Reports/Counties/Maricopa/Financial_Audits/Medical_Center/FinancialAudit_June_30_2003/Medical_Center_June_30_2003_Highlights.pdf.
8. The Census Bureau, through its five-year Census of Governments, continues to collect this data. However, they have not published the data since 1992, so that year’s is the most recent available.
9. Michelle Ye Hee Lee, “Taxation Vexation: Special districts often lead to wildly disparate tax bills among neighbors,” Arizona Republic, June 19, 2003.
10. U.S. Census Bureau data cited in Christopher R. Berry, Imperfect Union: Representation and Taxation in Multilevel Governments (New York: Cambridge University Press, 2009), 35.
11. Berry, 41-42.
12. Kathryn Foster, The Political Economy of Special Purpose Governments (Washington, D.C.: Georgetown University Press, 1997), 133.
13. Michael Bell, David Brunori, Joseph Cordes, and Bing Yuan, “Tax and Expenditure Limitations and Their Effects on Public Finance” (working paper, George Washington University, August 2007) 10-12, http://www.gwu.edu/~gwipp/lincoln/Yuan_Cordes_Brunori_Bell.pdf.
14. Advisory Commission on Intergovernmental Relations, “Tax and Expenditure Limits on Local Governments,” March 1995, http://www.library.unt.edu/gpo/acir/Reports/information/M-194.pdf.
15. For examples and literature reviews, see Susan A. MacManus, “Special District Governments: A Note on Their Use as Property Tax Relief Mechanisms,” Journal of Politics, vol. 43 (1981): 438-444; and Foster (1997).
16. Michael A. Nelson, “Decentralization of the Subnational Public Sector: An Empirical Analysis of the Determinants of Local Government Structure in Metropolitan Areas in the U.S.,” Southern Economic Journal 57, no. 2 (October 1990): 443-457.
17. Ibid, 456.
18. Barbara Coyle McCabe, “Special-District Formation among the States,” State and Local Government Review 32, no. 2 (Spring 2000): 121-131.
19. David L. Chicoine and Norman Walzer, Governmental Structure and Local Public Finance (Boston: Oelgeschlager, Gunn & Hain, 1986).
20. Author’s calculations based on data from the Census Bureau and Advisory Commission on Intergovernmental Relations, “Tax and Expenditure Limits on Intergovernmental Relations,” March 1995. The averages exclude North Dakota and Wyoming since they are clearly outliers in that they have an unusually high number of special districts per 100,000, and most of their special district growth can be explained by the fact that they are generally very rural states.
21. Dean Stansel, “Interjurisdictional Competition and Local Government Spending in U.S. Metropolitan Areas,” Public Finance Review 34, no. 2 (March 2006): 173-194.
22. Christopher B. Goodman, “Local Government Fragmentation & the Local Public Sector: A Panel Data Analysis,” November 2012, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2170783.
23. U.S. Census Bureau, Census of Governments, various years.
24. Ibid.
25. U.S. Bureau of Economic Analysis, U.S. Department of Commerce, various years.
26. U.S. Census Bureau, various years.
27. Ibid.
28. Ibid.
29. Ibid.
30. This relationship was significant at the 99 percent level of confidence. The regression model outlined here explains 55 percent of the phenomenon observed. Or, for those familiar with statistics, the R-squared was 0.55. See the appendix for details.
31. See studies cited in Foster, Table 3-2, p. 74.
32. Lyke Thompson, “Citizen Attitudes about Service Delivery Modes,” Journal of Urban Affairs 19, no. 3 (1997): 291-302.
33. Berry, 146.
34. Arunava Bhattacharyya, Thomas R. Harris, Rangesan Narayanan, and Kambiz Raffiee, “Technical Efficiency of Rural Water Utilities,” Journal of Agriculture and Resource Economics 20, no. 2 (1995): 373-391, cited in Berry (2009).
35. David Emanuelson, “A Comparative Analysis of Illinois, Ohio, Colorado and South Dakota Park Districts and Parks and Recreation Departments to Wisconsin, Iowa, Missouri, Kansas, Indiana, and Michigan Parks and Recreation Departments,” paper presented at the annual meeting of the Midwest Political Science Association, April 2008, cited in Berry (2009).
36. Berry, 146.
37. Ibid., 145.
38. Ibid.
39. Ibid., 113.
40. J.C. Bollens, “Special District Governments in the United States,” (Berkeley: University of California Press, 1957), 253.
41. Berry, 181.
42. Berry, 103-104.
43. Thomas R. Suozzi, “Special District Election Date Study: A Crazy Quilt,” Nassau County Executive, 2007, cited in Berry, 64.
44. Debbi Weimer, “Data Brief 5,” Education Policy Center, Michigan State University, 2001, cited in Berry, 65.
45. Berry, 65.
46. Suozzi, cited in Berry, 64.
47. Cited in Berry, 65.
48. Berry, 65.
49. Stephanie Dunne, W. Robert Reed, and James Wilbanks, “Endogenizing the Median Voter: Public choice goes to school,” Public Choice 93 (1997): 99-118.
50. Cited in Berry, 68.
51. Terry Moe, “Political Control and the Power of the Agent,” Journal of Law, Economics, and Organization 22, no. 1 (2006): 1-29.
52. Berry, 68.
53. For more information, see Nick Dranias, “A New Charter for America’s Cities,” Goldwater Institute, March 2009, http://www.goldwaterinstitute.org/article/new-charter-american-cities-10-rights-restrain-government-and-protect-freedom.