1 Land Use, Tax Revenue and Community Services in Woodford County 1 David Freshwater, University of Kentucky Final Report, September 27, 2008 This report provides Woodford County with a Cost of Community Services (COCS) study that broadly follows the methodology developed by American Farmland Trust (AFT), who remain the main provider of this type of analysis. The results of the Woodford County COCS study are similar to those of virtually all similar studies. almost all these studies show that agricultural land generates more tax revenue than it costs local government to provide services to this particular land use. This is also true for commercial and industrial land. Conversely, residential land generates less tax revenue than the cost of providing public services to residential land. While there is some variability in the specific magnitudes of the ratios of property tax revenue to public expenditure on local services among studies, the broad results are identical. A particular feature of the report is that it tries to provide a context for interpreting COCS findings. Clearly COCS studies suggest that communities everywhere appear to be charging residential land users too little and agricultural and commercial land users too much. Because the results of COCS studies are so uniform it is important to determine why this practice occurs, and whether looking at tax revenue and expenditure on a land use basis provides a meaningful picture of the tax and expenditure decisions of a local government. A second extension of the work over standard approaches is a decomposition of the residential sector and the farm sector into a number of subcategories to assess the relative burden associated with different types of residential development and farm size in the county. This may provide local leaders with useful additional information as they determine where future residential development in the county will occur and in their ongoing efforts to maintain agricultural land. A third extension of the usual AFT methodology is to introduce a broader set of allocation rules for distributing revenues and expenditures across the various land uses. AFT relies mainly on the acreage share of each category for allocating expenditures when explicit external information is not available on actual sources of revenue and patterns of expenditure by land use. In this report expense and revenue items are also allocated on the basis of the number of tax paying units (farms, firms and households) and the assessed value associated with these tax paying units. 1 All opinions are those of the author. Data and advice was provided by a variety of local government agencies in Woodford County, Midway and Versailles. This assistance was instrumental in developing the report. Funding was provided by the Versailles‐Midway‐Woodford County Planning Commission.
Transcript
1
Land Use, Tax Revenue and Community Services in Woodford County1
David Freshwater, University of Kentucky Final Report, September 27, 2008
This report provides Woodford County with a Cost of Community Services (COCS) study that broadly follows the methodology developed by American Farmland Trust (AFT), who remain the main provider of this type of analysis. The results of the Woodford County COCS study are similar to those of virtually all similar studies. almost all these studies show that agricultural land generates more tax revenue than it costs local government to provide services to this particular land use. This is also true for commercial and industrial land. Conversely, residential land generates less tax revenue than the cost of providing public services to residential land. While there is some variability in the specific magnitudes of the ratios of property tax revenue to public expenditure on local services among studies, the broad results are identical.
A particular feature of the report is that it tries to provide a context for interpreting COCS findings. Clearly COCS studies suggest that communities everywhere appear to be charging residential land users too little and agricultural and commercial land users too much. Because the results of COCS studies are so uniform it is important to determine why this practice occurs, and whether looking at tax revenue and expenditure on a land use basis provides a meaningful picture of the tax and expenditure decisions of a local government.
A second extension of the work over standard approaches is a decomposition of the residential sector and the farm sector into a number of subcategories to assess the relative burden associated with different types of residential development and farm size in the county. This may provide local leaders with useful additional information as they determine where future residential development in the county will occur and in their ongoing efforts to maintain agricultural land.
A third extension of the usual AFT methodology is to introduce a broader set of allocation rules for distributing revenues and expenditures across the various land uses. AFT relies mainly on the acreage share of each category for allocating expenditures when explicit external information is not available on actual sources of revenue and patterns of expenditure by land use. In this report expense and revenue items are also allocated on the basis of the number of tax paying units (farms, firms and households) and the assessed value associated with these tax paying units.
1 All opinions are those of the author. Data and advice was provided by a variety of local government agencies in Woodford County, Midway and Versailles. This assistance was instrumental in developing the report. Funding was provided by the Versailles‐Midway‐Woodford County Planning Commission.
2
These additional allocation rules are introduced under the belief that they more accurately represent actual flows of funds.
COST OF COMMUNITY SERVICES STUDIES
Functions of COCS Studies
COCS studies serve two main functions. The first is to demonstrate that there are systematic transfers among owners of different parcels of land when one compares the amount of property tax raised per acre to the amount of public expenditure per acre on the basis of land use. The second function is to provide a rationale for preserving farmland. The results of COCS studies are often used to show that farmland generates more net revenue for a community on a per acre basis than do certain other land uses, particularly residential uses. From this observation it is often concluded that while converting farmland to another use may be profitable for the landowners involved, it is not in the interest of the local government.
These are two distinct issues. The first is best thought of as a question of tax equity. In general taxes always result in a redistribution of income, if only because ability to pay and the level of services received by any individual are rarely taken into account when setting tax rates. Moreover, redistribution can be an explicit objective of some taxes. While the property tax is generally not viewed as a tax that serves an explicit redistribution function, it is the case that high values of property are associated with a relatively high level of wealth, if not income.
Discussing tax equity in terms of land use is also problematic. Land per se has no standing in discussions of equity. It is the owner of the land that pays the taxes and the owner of land that receives the services. Clearly certain services are more closely associated with how land is used than are others. Public expenditure for schools is more closely aligned with residential land than with pasture. But the crucial point is that it is individuals who ultimately benefit from the provision of services and who pay taxes, not property, and considerations of fairness have to be judged in terms of individuals and not land.
Moreover the real issue in terms of tax equity requires looking at the total transfer to and from a tax payer. Local property taxes are only one of the taxes paid and municipal services are only one component of the services received by a taxpayer. For example, households typically pay a higher share of retail sales taxes that in turn can be used to fund state grants to local governments which benefit all households, including farmers. This suggests that while the difference between the levels of tax revenue generated per acre of land and the value of services provided is an interesting issue, it is not enough by itself to show that farmers are unfairly taxed.
The second issue of farmland preservation is also problematic. COCS studies do provide some support for use‐value, rather than market value, taxation of farmland because they demonstrate that farms place a relatively small burden on municipal services relative to the amount of taxes they pay. This is not the same as arguing that farmland should be kept in this particular use. The latter decision requires a
3
determination of the value of the flow of services from farmland relative to the value of the services that would be provided under other land uses. Simply looking at the stream of property taxes and the return flow of government services to farm property only provides a part of the total picture. The part it provides is the impact on local government revenues and outlays. While this is a valid concern for local government, the private benefits and costs of alternative land uses must also be considered. For example, what are the implications for housing availability and employment, from moving new residential development or commercial activity from certain parts of the county and into others?
Typical COCS Results
Because there have been numerous COCS studies it is useful to review previous results to establish a background for the Woodford County study. A simple search for “Cost of Community Service” studies on Google returns a large number of individual results. Rather than provide a long discussion of individual studies a few specific studies that are most relevant will be reviewed.
In Kentucky AFT completed a study of Shelby County in 2005 that generated the following results:
• the ratio of revenue to expenditure for residential land was $1.00/$1.21, • the ratio of revenue to expenditure for commercial/industrial land was
$1.00/$.24, and • the ratio of revenue to expenditure for farmland was $1.00/$.41.
Shelby County is particularly interesting because, like Woodford, it is adjacent to an urban center, and like Woodford it has a large number of horse farms and considerable development pressure. The AFT results show that the vast majority of the property tax base in Shelby county is from residential land (80%), but the tax base on that land fails to cover the cost of providing all the services consumed.
AFT also conducted a study in Oldham County in 2003 using the same methodology and here the results were:
• the ratio of revenue to expenditure for residential land was $1.00/$1.05, • the ratio of revenue to expenditure for commercial/industrial land was
$1.00/$.29, and • the ratio of revenue to expenditure for farmland was $1.00/$.44.
Oldham, like Shelby, is part of the Louisville metropolitan region where urbanization pressures are leading to the conversion of agricultural land to residential land. A third Kentucky study by AFT that was much more extensive in scope was completed for Kenton County in 2005. In this study AFT provided a typical COCS report as well as undertaking an examination of opportunities and threats to
4
agriculture in the county. Kenton county, while in Kentucky, is part of the Cincinnati metropolitan region and experiences development pressure from the larger multi‐state region. The ratios of revenue to expenditure from the Kenton COCS study were:
• the ratio of revenue to expenditure for residential land was $1.00/$1.19, • the ratio of revenue to expenditure for commercial/industrial land was
$1.00/$.19, and • the ratio of revenue to expenditure for farmland was $1.00/$.051.
The first COCS study in Kentucky was undertaken in Lexington‐Fayette county in 1999 by AFT. The basic conclusions parallel the more recent studies in the state, despite Lexington‐Fayette being a core urban county with a much higher population. The ratios of revenue to expenditure were:
• the ratio of revenue to expenditure for residential land was $1.00/$1.64, • the ratio of revenue to expenditure for commercial/industrial land was
$1.00/$.22, and • the ratio of revenue to expenditure for farmland was $1.00/$.93.
Variability across the four Kentucky counties can be explained in a number of ways. First each study is taken at a point in time and local government revenues and expenditures are known to vary considerably on a year to year basis. Second there are clear differences between Lexington‐Fayette and Kenton, Shelby and Oldham in terms of economic and demographic structure. While Woodford is part of the Lexington metropolitan region it is structurally much more like Oldham and Shelby counties. Kenton county has a much larger commercial/industrial base and a larger population than the other three outlying counties. Finally counties rely on the property tax base to varying degrees and provide a variety of services, which means the structure of their local finances can vary considerably. Deller provides a comprehensive review of COCS studies as part of his analysis of urban growth and rural land conversion (Deller, 2003). In Table 1 of his paper, reproduced here as Appendix 2, he lists 83 COCS studies conducted between 1986 and 2001 by a wide variety of organizations. These studies cover 19 states and address a variety of local government types including: counties, townships, cities and towns. In all studies residential land uses generates less revenue than the cost of services provided to them, with the value of the ratio ranging between a low of 1:1.02 and a high of 1:2.11. Both commercial and industrial land uses and farmland almost always have ratios of revenue to expenditure of less than 1, with commercial and industrial land generally, but not always, having a lower ratio of revenue to outlays than farming. For commercial and industrial land the ratios range from a low of 1:0.05 to a high of 1:1.04 and for farmland from 1:0.02 to 1:0.99. Economic Growth
5
Interest in COCS studies appears to be highest in local jurisdictions: experiencing rapid growth in population and economic activity, that are facing capacity limits on local infrastructure and have significant farmland conversion pressures. However, local governments everywhere face the challenge of defining an appropriate level of economic growth. While there is always an interest in expanding the size of the local economic base and diversifying income and employment opportunities there is a growing recognition that economic growth has associated costs. However few communities are prepared to argue that they have all the growth they want, or that they would prefer a smaller local economy, so the question is what types of growth and how much.
Woodford County has long had the enviable position within Kentucky of having one of the highest, if not highest, per capita incomes and the lowest level of unemployment. By these common standards the county appears to be performing admirably. It also performs well in terms of educational attainment and is recognized as having a high level of amenities. But the fact that Woodford is such an attractive place to live means that more people want to live there. For Woodford County a crucial question is how much of this potential growth in residential development should be accommodated and where in the county should it take place?
Increased residential growth has multiple implications. The first is the potential for eroding the current quality of life if too many people move into the county and congestion begins to make life more difficult. In addition there is the question of who bears the cost of accommodating new residential development if it leads to new outlays on: schools, water and sewer systems, solid waste disposal and other public services. Conversely more development may make the county more attractive to employers seeking additional workers. And a larger population could provide a large enough base to provide an opportunity for a broader array of retail shopping opportunities and new community services. Finally, if the county has surplus capacity in its basic infrastructure, more people can lead to being able to spread the fixed costs of paying for infrastructure across a larger tax base.
Two distinct issues are involved in evaluating the benefits and costs of economic growth. The first is the total, or aggregate, costs and benefits. This assessment determines whether growth makes sense or not for the community. The second issue is the incidence of the benefits and costs, that is, who in the community receives the benefits and who bears the costs. Even if the aggregate benefit from additional growth is positive, a community may decide to forgo the opportunity if the incidence of benefits and costs is considered to be inappropriate. Or, the community may be able to find a way to alter the distribution of costs and benefits so that a “fairer” outcome can be achieved.
Financing Local Services
The taxing powers of local government are defined and regulated by states, since local governments are subsidiary to state governments. While there are
6
considerable differences among states in the types of fiscal powers assigned to local government the main source of local government revenue has been the property tax. Historically neither the federal nor state government have relied upon this tax, so it has essentially been reserved for local governments. However state governments have often imposed limits on the types of property that can be taxed and on the rates that can be charged.
The largest element of property value for all households is generally land and buildings, primarily the residence. For farm enterprises, farmland is by far the largest asset in terms of value. For commercial and industrial firms, land and buildings can be a significant share of total asset values, but may be less important in some cases. Generally, the aggregate value of land and buildings in most local jurisdictions provides a large tax base that can often finance most local government expenditures, even though it is typically supplemented by a variety of other revenue sources.
However, local government and local tax payers often find problems with their dependence upon the property tax. While property values are often relatively stable, there are intervals when values increase rapidly, and less commonly intervals when they decline rapidly. Instability in property values in turn leads to instability in local government revenue. Moreover property is a component of wealth, and high levels of property wealth may not be associated with a similarly high level of income. This is particularly true for individuals who have held their property for a long period of time and experienced capital appreciation. Unless their income has increased at the same rate they may have difficulty paying their tax bill.
In some jurisdictions this has led to caps on property taxes. In those jurisdictions where the caps are in place for an extended period the property tax may become inadequate to fund the desired level of local services. Farmers in particular argue that the property tax is unfair. Because agriculture relies on control of scarce farmland there is a tendency for increases in farm income to be rapidly capitalized into farmland prices. Moreover farmland in close proximity to urban areas typically has a market value that is driven mainly by nonfarm uses. Higher farmland prices lead to higher tax bills that can impact farm cash flows in ways that weaken the viability of the farm enterprise. For this reason many states have mandated that farmland be taxed at current use value rather than market value. While this provision significantly reduces the tax bill facing farmers, it can have a considerable impact on the local revenue base, especially if farmland is a major share of total property value for the local government.
Initially property taxes were used to raise money for general fund purposes, but increasingly a variety of dedicated mill rates have been added to the base tax rate to fund specific activities. The most common of these are local school taxes, but there is increasing use of dedicated tax revenues. These taxes raise money for a specific activities, for example the local contribution to Cooperative Extension services. This process assures that funding for specific activities does not have to compete with other priorities for general spending. While the process provides a stable flow of
7
funds, there is typically no matching of tax funding for an activity with those who use it. A library tax raises money from people in amounts proportionate to the value of their property, not to the value they place on libraries, nor the use they make of them.
In addition to property taxes most municipalities also employ fees to generate money. User fees are associated with a specific service or activity, and are in effect a user charge. The advantage of user charges is that only those individuals who utilize the service pay for it, and the amount they pay can be adjusted to reflect the amount of use and the cost of providing the service.
Increasingly local jurisdictions are moving to broaden their tax base by seeking authority to tax income as well as wealth. The most common of these taxes are local income taxes or occupational taxes on income earned within the jurisdiction, and local sales taxes on purchases made within a jurisdiction. The ability to introduce these taxes varies greatly by state and in some states only some local government have access to income and sales taxes. In Kentucky local governments can introduce an occupational tax, but not local income and sales taxes. In the communities where they have been introduced these taxes have allowed significant reduction in the role of the property tax in the local revenue base and these “income‐type” taxes often account for the largest share of revenue.
A broader tax base that raises funds from both income and wealth arguably provides greater stability to local government revenue and may even be considered “fairer” in the sense that sole reliance on income or wealth as the tax base can place a too high burden on some people and a too low one on others. However a broader tax base can make it more difficult to assess whether any particular group is being fairly taxed if only one aspect of the total tax base is being examined.
Cost of Services
Increasingly relevant questions for communities facing pressures to accommodate growth are, how will growth affect the demand for local public services and how will this increase in demand for services be funded? The problem mainly arises for those services that require: long term planning, large initial investments, and that have a fixed capacity. School buildings, water treatment plants, solid waste sites and other major infrastructure items fall into this category. Other public services, such as police and EMS services, are more incremental in terms of supply, in that additional capacity can be added relatively quickly and in an amount that closely matches any increase in demand.
Not only are there differences in the ability to match the supply of different public services to demand in a timely manner, there are also differences in the way municipal governments charge for different types of services. In the case of large infrastructure projects the main component of cost is typically the initial construction outlay, and while operating costs can be significant in terms of dollars involved, they are a small portion of the total expense. Typically communities set
8
prices for services so that each user covers an appropriate share of the capital and operating costs. In many cases this means each individual pays the average unit cost, where average cost equals total cost divided by number of users. This pricing approach may be modified by charging an average price for the capital cost and establishing a specific charge for operating costs based on actual use. For example a community can set water rates so there is a fixed amount levied per meter to cover capital costs, plus a variable amount based on actual water consumption.
Because these investments are: expensive to construct, have a long useful life, and have a fixed capacity, the common practice is to build more capacity than is currently needed at the time of construction (Figure 1). This means that for most of its life the investment is underutilized and the capital costs have to be distributed over “too small” a base of users. From this perspective the addition of new users is beneficial to the existing tax base. New users absorb some of the excess capacity so they lead to no additional cost, and their payments to the capital costs spread the fixed costs over a larger base, which allows lower charges for existing users. Of course each user should continue to pay the appropriate share of operating costs.
Obviously a problem arises when the excess capacity is exhausted. Now new users can only be accommodated with another major investment and the question is, who should pay for the creation of additional capacity. One argument is that new capacity should be paid for by those who created the increase in demand, so only new users would pay. However typically all users receive an upgrade in the quality, if not quantity of service if new capacity is added. For example adding a new water tower increases water pressure over the entire system and provides a secondary source of supply. This means that all users benefit. Further, simple equity suggests that if in the past groups of new users, who absorbed the previous
Explanation of Figure 1
With one school the amount of tax that has to be collected declines over time as enrollment increases and the debt burden is spread over a larger base. During this period additional students can be accommodated with no extra capital costs. And, each additional student (family) results in the fixed cost being spread over a larger base. However at some point the school has no more excess capacity and a new school has to be constructed.
New construction increases the total debt outstanding and every household experiences an increased debt burden. It is just prior to this point that citizens typically begin to question the benefit of additional growth in population. However once the new school is built it is again rational for growth to be encouraged, because once again more students lead to a larger tax base to generate the revenue to cover the new higher fixed capital costs.
Whenever there is a large amount of excess capacity there is a short term incentive for the community to seek additional growth, because the current debt burden can be spread over a larger number of households. A faster growth rate in enrollment allows debt per household (tax burden) to be reduced faster. What the community has to balance is the short term gain in reduced tax burden against the necessity to construct additional capacity sooner than would be needed with slower growth.
9
excess capacity, weren’t charged an impact fee, then it is unfair to charge those who happened to be one or two months late in arriving and found the excess capacity gone.
10
Moreover, the process repeats itself. New capacity well in excess of the current level of demand will be created with new investment, so there is effectively a transfer from residents of the community at the time of new construction to those who arrive later after some of the construction costs have already been retired. Because there is no practical pricing solution to this problem, given the uncertainty about the timing of demand growth, most communities continue to rely on average cost pricing for public services. While doing so involves significant transfers of welfare among residents, the approach has the virtue of being easy to implement.
Finally COCS studies are important simply because local governments rely to a large degree on property taxes, and the distribution of land uses plays a major role in defining the property tax revenue stream (Kotchen and Schulte). Because the demand for local services is unlikely to match the pattern of land use it is common for the distribution of property tax revenue not to match the distribution of local government expenditures. Further, changes in land use may either improve or exacerbate the mismatch between revenues and expenditures. By conducting a COCS study a local government gains information on the incidence of taxes and the incidence of expenditures by various land uses.
The Policy Implications of COCS Results
While there are often compelling reasons for one segment of society to provide transfers to other segments, this task is best accomplished by an explicit recognition of the need for the transfer and a particular policy or policies to accomplish it. The tax system is often used by the Federal and state governments to accomplish this function. Examples include, the mortgage interest deduction for home owners, progressive marginal tax rates on earned income, exclusion of some portion of capital gains from taxation, etc. At the local level similar transfers take place through the property tax, but there is far less strategic behavior involved; in that very few local governments explicitly choose to adjust tax rates on different classes of property to accomplish a transfer from one segment of the community to another.
Over the last decade there have been a large number of studies on the “Cost of Community Services (COCS), many performed by American Farmland Trust. All of the studies demonstrate that residential development is effectively subsidized both by agriculture and commercial and industrial property uses, if the standard for comparison is the share of property tax paid relative to the share of public outlays. While there are numerous criticisms of the interpretation of these studies and their implications for public policy, there is no denial of the basic fact that in every community studied the amount of tax revenue raised from the annual property tax is less than local expenditure on those residential acres. Conversely the amount of tax revenue raised from farmland and commercial and industrial land exceeds annual local government expenditures on these land uses.
COCS studies raise important public policy questions for local governments. The first one is, why do local governments appear to systematically subsidize the cost of delivering public services to residential areas through the property tax system? A
11
second question is, whether there are other revenue and tax decisions that either reinforce or offset this behavior, so that on a total tax and total outlay basis the subsidy to residential land use is reduced? A third question is, whether COCS analysis, with its focus on cross‐subsidies among land uses, is an appropriate measure of the soundness of local public policy? Two last questions are, whether the level of subsidy is appropriate, and whether the property tax is the appropriate mechanism for achieving the redistribution?
Because all COCS studies generate the same result, it is unlikely that the subsidy to residential land uses is a random outcome. However there does not appear to be any evidence that local governments explicitly set rates to generate this subsidy, so the answer must come from an indirect decision process. Clearly local tax rates are a major concern for community residents, and there is ample evidence that homeowners closely monitor their property tax burden. Two streams of thought can explain an incentive for local officials to limit residential property tax rates. The first is a simple political calculation that higher taxes can lead to unfavorable election outcomes for those associated with raising taxes, and since residential tax payers dominate the electorate they form a block of voters who should not be antagonized.
A second argument is that local governments compete for residents through a combination of tax rates and services they provide. If a significant portion of the local population is able to move to a new jurisdiction, then each locality has an incentive to deliver its particular set of services at as low a cost as is possible. This concept, first advanced by Tiebout, is most often used to explain why different localities will offer different sets of services. But it also can be used to explain why it may be common for local governments to subsidize homeowners. Neither farms nor businesses are as mobile as homeowners, nor do they have as great a concern with the mix of local services and their cost. This makes them relatively less sensitive to higher taxes.
While COCS studies show that the ratio of property taxes to expenditure favor residential uses, these studies only consider part of the picture. Property taxes are annual charges paid by owners to generate revenue for the community. But the typical community also has other local revenue sources, including: user fees, other taxes, and one time charges. Suppose there is no desire to subsidize any category of user then the objective of the community should be to collect revenue from each unit of demand that fully covers the cost of providing it with services.
Whether the community does this simply by matching property tax revenues to the cost of delivering services on an annual basis is less clear. Some users may be awarded higher initial outlays but have those costs recovered at a later date. For example, many industrial land uses receive local subsidies at the time firms first locate in the community or later expand their operations. Subsidies can include less than full cost recovery for infrastructure improvements, tax abatements etc. In this case a fiscal neutral tax system would have to charge these users higher rates at some future point in time to recover the initial subsidy. Other users may prepay a portion of their expenses. For example, if the community charges impact fees on
12
new development, then those households effectively prepay a portion of the cost of providing services. Third, some service users may pay a portion of their charges through another revenue stream. If the community has a local income tax or sales tax then residential users will provide a portion of the revenue needed to cover the cost of providing them with services through an alternative tax mechanism.
In principle one could argue that in an ideal system the charges associated with the consumption of any particular service to an individual user should be directly associated with the cost of providing that service to that user, in order to provide the most efficient level of each service. In practice it is more common to employ a limited number of charges, especially when it is hard to either monitor or collect charges for specific services at the time of use. In this case the best approximation is to try to ensure that aggregate revenues collected from each household or business match aggregate outlays to service that unit.
COCS and Agriculture
COCS studies use local government revenue and outlay by type and quantity of land use as their fundamental unit of measure. From a public policy basis this is a largely uninteresting approach. Local government exists to serve the interests of the community and community is measured in terms of individuals and households. While people hold land, they are rarely interested in the value of the services accruing to land. Rather, they are interested in the value of the service accruing to them as individuals, households and collectively as a community. Similarly businesses are mainly interested in the value and cost of the service that are provided to their enterprise. Agriculture is perhaps the only case where the notion of the value and cost of local government services can meaningfully be applied on a per acre basis. This largely reflects the fact that agriculture is inherently land using, so land is both the largest expense item in production and the largest asset; plus there is the fact that agriculture makes very little use of most government services, especially if the residential or household component of the farm is stripped out, so only the farm business is considered as being agriculture.
Many providers of COCS analysis, once they demonstrate the existence of the subsidy to residential land uses, argue that it leads to an unfair tax on agriculture. The subsidy, it is argued, leads to undue conversion of farmland to other uses, particularly residential development. As an advocacy group for the preservation of farmland this is an appropriate point for American Farmland Trust to make. Further Brueckner and Kim analyze the effect of property taxes on city size. They find that higher property tax rates tend to reduce sprawl by increasing the implicit cost of housing for home owners causing them to favor smaller homes (p.7). However possibly offsetting the effecting on homeowners is the opposite effect on developers, where higher tax rates on buildings lead to less intensive development, or lower density per acre, which encourages sprawl (p.6). Thus while property tax rates can be shown to affect sprawl, the direction of the effect is ambiguous.
13
However there are other possible arguments about the impact of the property tax. The most compelling of which is the recognition that property taxes are only one of a number of ways that local government generates revenue, and therefore an asymmetry between property tax charges and services provided may not imply that a subsidy between land uses exists. Second, it is often the case that tax rates are often set on the ability to pay, not simply to recover the actual cost of services provided to citizens. Redistribution is an inherent function of government and this task can be accomplished through differences in tax charges and through different expenditure patterns. Of all activities agriculture has the least ability to relocate to another jurisdiction so it inevitably becomes a target for taxes. And agriculture is perhaps the single activity with the least potential to be exposed to other sources of revenue/taxes, than the property tax. This may make its share of property taxes higher than might be expected. However the prevalence of “use value tax rates” rather than “market value rates” that are used for other land uses provides agriculture with a considerable tax shelter.
An important consideration that is often ignored is that the property tax is really a tax on a particular type of wealth, property, not on income or financial assets. Where an individual’s level of income and property value are strongly correlated the distinction is not especially important. However there are examples of individuals where there is only a weak relationship between income and property wealth. Farmers are an obvious example, as are retired people. In both cases the net worth of the individual is dominated by property and an individual’s net worth can be high relative to annual income. For these individuals property taxes can be a large financial burden.
A possibly important caveat for COCS studies is the use of average costs to determine the relationship between revenues and outlays for different land uses. Commonly only three land uses are considered, agriculture, residential and commercial and industrial. More recently the common practice is to remove the farm household from agriculture and make it part of residential. To do this property taxes associated with the farm residence are assigned to the residential category and farm households are treated in the same way as other residences for expenditure shares.
While this approach recognizes that the household is not formally part of the farm enterprise it can have an important impact on results. Typically the cost of providing the same service to a farm household is higher than providing it to an urban residence. Urban children may walk to school while farm children are bussed. The cost of providing water lines to a house in a compact subdivision is higher than providing water to farm houses that are a mile apart. The cost of EMS services is less for urban households than farm households both because distance is shorter and because farming has a high incidence of accidents. The same argument can be made about differentials in costs for urban, suburban and exurban residences. For this reason it is desirable to break residential down into a number of sub‐categories in
14
order to get a better sense for how the type of residence affects the ratio of expenditures to revenue.
For the community a difficult choice results. It can be argued that taxes should not be a crucial factor in leading individuals to sell their assets. Under this argument those with low incomes relative to their tax bill might be granted some reduction. However another and equally valid argument is that wealth is simply a way to store income and that equity demands that society consider both income and wealth in determining the ability to pay taxes.
WOODFORD COUNTY ANALYSIS
The main objective of the project is to develop a Cost of Community Service analysis for Woodford County to determine the relative contributions of different land use to county revenues and the relative share of county expenditures that flow to these land uses. An approach that generally follows the standard AFT methodology will be followed, but because of specific conditions in Woodford County some modifications are introduced. In terms of order of importance they are:
1. The inclusion of occupational tax revenue as well as property tax revenue in county revenue.
2. Integrating the budgets of the cities of Versailles and Midway with the Woodford County budget to give a combined local government analysis.
3. Segmentation of residential land use into multiple spatial categories. 4. Use of two distinct farm sizes. 5. The use of a number of assignment rules for allocating revenue and expenses
beyond the standard share of acreage.
Explanations for these changes are provided below.
Occupational Tax
Some local governments, including Woodford County and the cities of Midway and Versailles, impose an occupational tax on workers. The occupational tax is equivalent to a local income tax charged on income earned within the jurisdiction in that is a tax on workers. In many communities the occupational tax is a more important source of revenue than the property tax. The advantage of an occupational tax is that it is a tax on income, rather than on wealth, so there is a more direct relationship between the tax and current ability to pay. Moreover a combination of a tax on income and a tax on wealth allows the local government to tax both elements of economic wellbeing: which is arguably fairer, provides a more stable tax base than reliance upon either instrument alone, and may make it simpler for those burdened by taxes to meet their obligations. Like all taxes the burden of the tax is shared. In the case of an occupational tax, even though it is nominally a payroll tax paid by workers, some portion of the tax burden will fall upon employers
15
because they experience a somewhat higher cost of labor and consequently operate with a lower level of employment. However, because of the tax workers experience a reduced take‐home wage and thus workers typically absorb most of the cost of the tax.
More than half of Woodford County revenue comes from an occupational tax on employers that is based on the amount of wages paid to workers employed in the county. Two things are important about the tax. The first is that it is primarily a tax on workers because it raises the effective wage paid by each employer, but workers do not receive this increment to wages (Figure 2). The second point is that it is a tax levied at the place of employment, not the place of residence, so workers who live outside the county also have to pay the tax.
Because it is a tax on employment and effectively reduces the take home wage, the revenue raised by the tax should be assigned to the residential land use sector. This means that residential land generates both property tax revenue and occupational tax revenue. However the amount of occupational tax attributed to residential property has to be reduced by the amount generated from workers who reside out of the county. This portion can best be treated as a lump sum transfer to the local government from local employers.
Unified Local Government Budgets
There are two incorporated places in Woodford County, the cities of Versailles and Midway. Each has its own independent government that
provides additional services to city residents beyond those generally available in the county. Moreover the city of Versailles provides a number of services that extend well beyond its boundaries, some of which cover the entire county. Non‐city recipients pay for these services either through fees or through transfers from the Woodford County and Midway governments to Versailles. In particular Versailles
Explanation of Figure 2
Each employer has a demand for labor and will employ more workers as the prevailing wage falls. Similarly, higher wages lead to more people being willing to work. Before the introduction of an occupational tax the prevailing wage is Wo and Eo workers are employed. The effect of the tax is to raise the cost of each worker to the employer. This effectively shifts the cost schedule for employers up even though workers still have the same take‐home wage.
Employers see the new prevailing wage as Wt, which is made up of the amount they actually pay their employees plus the amount they have to pay local government in occupational tax. The result is a reduction in the total number of workers employed, corresponding to Et. Because employment is reduced, the burden of an occupational tax largely falls on workers.
16
provides police and EMS dispatch services for the county and it is the source of piped water for much of the county2.
Versailles also provides other public services to a significant share of county residents who live in the urban service area. This land is not part of the city but businesses and residents are able to obtain Versailles city services. Because residential units in the county are segmented into city and non‐city categories, it is possible to allocate the Midway and Versailles budgets to the appropriate
2 Versailles provides treated water to local water districts in the county that in turn resell the water to customers in their service region. Versailles is responsible for the cost of the water treatment facility and establishes charges to recover the portion of the cost associated with providing water for non‐city residents, but whether these charges exactly match the cost of provision is unknown.
17
residential categories. The result is that city residents incur higher taxes in the form of an additional occupational tax and additional property taxes but also have access to additional services. City residents pay a somewhat lower county mill rate to reflect the fact that the city provides them with services that the county would otherwise have to provide. In addition there is a small number of city farms that are also exposed to city taxes and services.
Property Tax
The approach used in Woodford county extends the basic AFT methodology to include a number of different residential land uses. This is important in Woodford county because the county has experimented with a number of development strategies to more effectively manage growth. As a result it is possible that different types/locations of residential development generate different demands for local services. While there is no expectation that any residential land use category will generate sufficient property tax revenue to cover the full cost of the services provided to residential land, it may be useful to see if there are significant differences among different types of housing development. Other land uses conform to the standard AFT approach, except apartment buildings of any size are considered residential.
The residential land uses considered are:
• Within city limits – dwellings with the city limits of Versailles and Midway. • Within the urban service boundary ‐ dwellings found within the Versailles
and Midway urban service boundaries but outside the city limits. • Rural cluster housing – dwellings found in specific parcels of land in the
county that were subdivided and developed at a higher density. • Rural housing – dwellings found in the rural part of the county, including all
housing on less than 10 acres of land and the homesteads removed from farm parcels.
Farm Size Differences
Farms in the county are roughly divided by size on a north‐south basis with large Thoroughbred farms concentrated in the northern half of the county and smaller mixed farms found in the southern half. Given the bimodal nature of farm size within the county and the geographic concentration of the two farm sizes, two categories of farm are used. Larger farms are defined as those with 50 or more acres of land, and smaller farms have more than 10 acres but less than 50 acres of farmland. As usual, the value of the farm homestead is allocated to residential use.
It is important to note that in the small farm category the amount of property tax paid by the farm is typically very small. This reflects the fact that the farm has less than fifty acres and the assessed value of the land is approximately $600.00 per acre with a mill rate of $8.085 per $1,000.00 of assessed value. Thus the property tax on 25 acres of land would be $121.28. For these farms the value of the homestead
18
typically greatly exceeds the value of the land in terms of generating tax revenue for the county. While most public services are provided to farm households, and are captured in the residential category, a significant number still accrue to the farm business and hence to farmland. These would include extension, roads, EMS (for farm related calls) and general administrative expenses of local government. In this sense the farm is no different than any other business in the county.
Allocation Rules
The use of allocation rules is important only to the extent that there is a mismatch between the allocation of revenue and expense items. If revenues and expenses are allocated using the same rule the nature of the rule is largely irrelevant because everything is distributed the same way. In this situation the ratio of expenses to revenue will only vary by the amount that revenue exceeds expenses, or vice versa. In the case of a segregated fund that has a narrow tax base and whose revenues are used to deliver services to the same population that provides funds the choice of an allocation rule is not particularly important because the same rule is applied to both revenues and outlays. However if there is a mismatch between how revenues are generated and expenditures are made in terms of categories of provider/recipient then the allocation rules are important.
This is clearly the case for the General Fund, which obtains revenue from a variety of sources, some of them tied to the quantity of land, some of them tied to the number of tax filers, some of them tied to the value of the assets each taxpayer holds or their income. Similarly the General Fund involves a variety of expenditures, some of which vary by the size and type of land, others by the number of households, and others by some other criterion. This suggests that it might be appropriate to have more than one rule for allocating revenue and expense items across categories.
AFT analysis relies upon the share of assessed value in each category of land use to allocate revenue and expense items for which no detailed information on actual patterns is available. It is the consistent use of this specific allocation rule that leads to the association of COCS studies with land use. In principle COCS studies can be conducted using any of number of categories, such as, income classes, subareas within a jurisdiction or different types of household. It is not clear that the share of assessed value is always eh best default allocation rule. For example, to the extent that road expenses are a function of the number of miles of highway serving different land uses, then the share of the land base is a useful way to allocate road expenses. But, if road expenses are more of a function of the number of users, then allocating a large share of expenses to the small share of users who happen to own a large share of the land, but account for a small share of the miles driven would be inappropriate.
In the analysis two additional allocation rules are introduced. The first is to count the total number of tax‐paying entities in the county. This is the total of residences, businesses and farms. For farms two units are identified. The first is the farm household which receives all the services that any other household receives and the
19
second is the farm business which receives services appropriate to the enterprise. This approach results in more tax units than are found on the actual tax rolls. The second allocation rule uses the share of total acres in each land use. It may be appropriate where the size of the land base influences revenue or expenses.
The table shows that the choice of allocation rule can make a large difference in how any particular line item in the budgets is allocated. Since there is limited information on the distribution of expenses or revenues for many items in a budget, the choice of allocation rules could play a large role in determining the final ratio of revenues to expenditure for the various land use categories. For example, if acreage is used for allocation then large farms would receive 75% of the amount of expenditure or revenue. But if the budget line item is allocated on the basis of assessed value, then large farms would only receive 10% of the total amount.
RESULTS
The resulting ratios of revenue to expenditure are reported below. The table contains two rows. The first row provides the actual ratios calculated for the various land use categories based upon the allocation of revenue and expense items in 2006‐2007 budgets. The second row provides a sensitivity analysis that adjusts the calculated ratios by 10%. The adjustment reduces the ratio of revenue to expenditures by 10% for land uses that have a ratio greater than 1.0, and it increases the ratio by 10% for land uses that have a calculated ratio larger than 1.0. The effect is to bring all ratios closer to the balance value of 1.0. This adjustment is provided to account for errors in allocation that might lead to an overstating of the relative share of expenditures and revenue in each land use category. Moreover the adjustment is intended to correct specific types of systematic error that could lead to an overstating of the extent to which the ratio of revenues and expenditure diverge from each other. The second row can be interpreted as a conservative estimate of the degree of balance between revenues and expenditures for each category of land use. In using these results it is important to recognize that COCS studies rely upon a snap shot of revenues and expenses at a specific point in time and are not capable of examining how revenues and expenses move over time.
20
The results are broadly consisistent with other COCS studies, in that residential land uses generate less revenue than the amount of public expenditure, while Commercial and Industrial land uses and farming, in aggregate, generate more revenue than expenditure. Compared to other COCS studies in Kentucky carried out by AFT the values are not greatly different. For residential land uses the ratios fall below those for Lexington‐Fayette but are above Kenton, Oldham and Shelby. The ratio for commercial and industrial uses is slightly less than for other Kentucky counties and the aggregate ratio for farmland is below Fayette but above the others. When compared to the 83 studies compiled by Deller, the results for Woodford County fall clearly in the mid‐range of the various ratios. When farming is disaggregated into large and small size classes a clear difference is found. Large farms generate considerabley more revenue than the cost of providing them with services, whereas small farms result in higher expenditures than revenue. The result for smll farms can be readily explained. Recall that only farm related land and buildings is included in the farm category. For small farms with less than 50 acres of land that has an assessed value of approximately $600.00 per acre the amount of revenue generated in property taxes is too small to cover the public service costs for that parcel of land. Consider a 30 acre farm with an assessed value per acre of $620.00 and farm buildings assessed at $15,000.00. The total assessed value is $33,600.00, and the associated mill rate is $8.085 per thousand dollars of assessed value. This results in $271.66 in taxes paid. This has to cover the farm’s share of general administration costs, access to fire, police and emergency measures, plus roads and cooperative extension. Since most of these expenses do not vary much by size of farm, the cost of providing them to a small farm is roughly the same as providing them to a large farm, but the tax revenue stream is much lower. Disaggregating residential land use into various categories results in similar differences. Notably the cost of service provision ratios are higher for Midway and Versailles and the Urban Service Area than for other residences in the county. This is despite additional tax revenues collected directly by city government in the two cities. It suggests that the additional services provided to residents in the two cities are not fully funded by tax revenues on an annual basis. In the case of the Urban Service Area there is no additional level of government collecting revenues, but residents pay fees for the extra services they receive. These fees are included in the calculation as part of the revenue and expenditure stream for the various city enterprises. For residences within in the county, but outside the urban service area, a much lower ratio of expenditure to revenue is found. This largely reflects the fact that they
21
do not have sewer connections and in some cases do not have treated water supplies. However these residents incur additional costs associated with wells and septic systems that they have to fund privately. The higher ratio of public services expenditure to revenue calculated for cluster developments compared to dispersed housing is largely a reflection of a considerably higher revenue stream from dispersed housing that in turn reflects higher assessed values. Simply put, the houses on dispersed acreages tend to be more expensive than the houses in cluster developments. While it may cost the county less to deliver services to cluster homes, their smaller tax base more than offsets the cost savings. The use of multiple allocation rules does not appear to have had a large impact on the results. This suggests that COCS results are not especially sensitive to the choice of default allocation rules. One reason for the lack of sensitivity may be that particular line items of revenue and expenditure that are closely linked tend to have the same allocation rule. This means that the shares of revenue and expenditure asssigned to each land use are somewhat related. If this is the case then it is more likely to be the case that the main factor driving the consistent results of COCS studies does not come from a redistribution among various revenue and spending subcategories, as much as a mismatch in the amounts of revenue and expenditure for certain items, such as education and other investment type functions. Summary The results of the study are largely consistent with all other Cost of Community Service studies. Residential development returns far less in tax revenue than it costs to service it. Commercial and industrial land and farming, on the other hand, generate more revenue than their cost of service. This largely reflects the fact that many more public services are provided to the people who live in houses than to the businesses that occupy commercial, industrial and farm land. The use of a mix of allocation rules does not appear to influence the general results relative to other COCS studies that adopt the AFT approach. An impliction of this outcome is that the AFT results are not simply an artifact of the particular allocation rule that the organization imposes. In this study a number of allocation rules were used, as were specific allocations for certain line item revenues and outlays where additional information was available. Despite this significant difference in approach the results were largely consisitent with AFT findings. Even with a large share of the local tax base coming from an occupational tax, which is a tax on labor and not property, Woodford County still falls in the middle of the general distribution of COCS study results. To some extent this suggests that the county has the right balance of tax sources relative to its expenditures. Because it is similar to other jurisdictions in terms of cost of service ratios it is difficult to argue that local government in Woodford is behaving inappropriately. Essentially what all local governements do is raise revenue from those able to pay and distribute it in
22
the form of services to the general public. There is a redistribution function in this process, but since the pattern in Woodford is largely consistent with that of other jurisdictions it is hard to argue that there is an inherent inequity. The sensitivity analysis in the results suggests that even with a 10% adjustment in the ratios that nothing major changes. Small farms become roughly in balance in terms of revenues and expenditure, which suggests that currently they may be roughly paying a “fair share” of the costs of operating local government. County residents outside the urban service area may be the form of residence most disadvanteged by the current process, since they incur personal expenses to offset more limited public services that other residences do not. Yet their situation is not all that clear. Residents out in the county place higher costs on local government in terms of greater school bus transportation distances, longer travel for police, fire and EMS service calls, and more expensive road maintenance charges. These issues all point to the difficulty of simply using COCS results to set public policy. While COCS studies provide a pespective on the collection and distribution of local government revenues, it is from a specific perspective. And, as noted at the begining of the study, the perspective is based on the use of land, which is arguably a relativley minor factor in making public policy decisions. Most importantly it is crucial to remember that a central function of government is redistribution. In the provision of services there is always likely to be some level of subsidy and this subsidy may make us collectively better off. COCS studies point out that property taxes involve a subsidy from farmland to residential land. But it is impossible from just using a COCS study to determine if this is desirable public policy. More information is required about other aspects of the local budget and about the preferences of the community. The simple fact that all places that have undertaken a COCS study find the same result suggests that the process is either largely seen to be fair, or that the results need a larger context to be meaningful.
23
Bibliography American Farmland Trust. 2003. Cost of Community Services Study: Skagit County, Washington. AFT: Washington D.C. available: http://www.skagitonians.org/upload_pubs/aft=spf_.pdf American Farmland Trust. 2003. Cost of Community Services Study: Oldham County, Kentucky. AFT: Washington D.C. available: http://www.oldhamahead.org/OldhamCOCSfinal.pdf American Farmland Trust. 2005. Cost of Community Services Study: Shelby County, Kentucky. AFT: Washington D.C. available: http://more‐ky.org/current‐news/cost‐of‐community‐services‐for‐shelby‐county/shelbyfinal6‐01‐05.pdf American Farmland Trust. 2005. A Profile of Agriculture in Kenton County, Kentucky. AFT: Washington D.C. available: http://www.kentoncounty.org/conservation_district/docs/20050401_American_Farmland_Trust_Report.pdf American Farmland Trust: Farmland Information Center. 2007. Cost of Community Services Studies Fact Sheet. mimeo. available: http://www.farmlandinfo.org/documents/27757/COCS_09‐2007.pdf Brueckner, Jan and Hyun‐A Kim. 2003. Urban Sprawl and the Property Tax. International Tax and Public Finance. vol.10, pp 5‐23. Deller, Steven. 2003. Urban Growth, rural Land Conversion and the Fiscal Well‐Being of Local Municipalities. Staff Paper 461, Agricultural and Applied Economics University of Wisconsin‐Madison. Available: http://ageconsearch.umn.edu/handle/12687
Edwards, Mary. 2004. The Cost of Community Services in the Towns of Gibraltar and Nasewaupee, Door County, WI.. available: http://www.doorcountylandtrust.org/
Edwards, Mary. 2000. Community Guide to Development Impact Analysis. University of Wisconsin Extension. available: http://www.lic.wisc.edu/shapingdane/facilitation/all_resources/impacts/analysis_background.htm
Kotchen, Matthew and Stacey Schulte. 2008. A Meta‐Analysis of Cost of Community Service Studies. paper presented at a Colloquium in Honor of Roger Bolton, Williams College, April, 2008. available: http://www2.bren.ucsb.edu/~kotchen/links/COCS.pdf
Prindle, Allen and Thomas Blaine. 1998. Costs of Community Services, CDFS‐1260‐98 Ohio State University Fact Sheet, Community Development. available: http://ohioline.osu.edu/cd‐fact/1260.htm
24
Appendix 1: Allocation of Revenues and Expenditures
25
26
Table 1. Cost of Community Service Studies Residential Combined Farmland, (with farm Commercial Forest and homes) & Industrial Open Lands Study Source Connecticut Bolton 1 : 1.05 1 : 0.23 1 : 0.50 Geisler, 1998 Durham 1 : 1.07 1 : 0.27 1 : 0.23 Southern New England
Forest Consortium, 1995
Farmington 1 : 1.33 1 : 0.32 1 : 0.31 Southern New England Forest Consortium, 1995
Table 1. Cost of Community Service Studies Residential Combined Farmland, (with farm Commercial Forest and homes) & Industrial Open Lands Study Source Kinderhook 1 : 1.05 1 : 0.21 1 : 0.17 Concerned Citizens of
Table 1. Cost of Community Service Studies Residential Combined Farmland, (with farm Commercial Forest and homes) & Industrial Open Lands Study Source Utah Cache County 1 : 1.27 1 : 0.25 1 : 0.57 Snyder and Ferguson,
1994 Sevier County 1 : 1.11 1 : 0.31 1 : 0.99 Snyder and Ferguson,
1994 Utah County 1 : 1.23 1 : 0.26 1 : 0.82 Snyder and Ferguson,
1994 Virginia Augusta County 1 : 1.22 1 : 0.20 1 : 0.80 Valley Conservation
