South Dakota property assessment: the past, the present, and Constitutional Amendment D.
South Dakota has depended on the property tax since territorial days. (1) For nearly 50 years, the state as well as local governments received most of their revenues from that source. State government abandoned its use in the late 1930s, but local governments still rely heavily on the property tax. Taxpayers and policy makers worry about overall property tax levels and how they are distributed among individual taxpayers, leading to questions about the quality of property assessment. This paper focuses on the latter concern by looking at assessment practices, especially recent practices and potential changes that could occur under Constitutional Amendment D.
State and local taxes
Despite changes in the state's tax structure, South Dakota governments continue to rely heavily on the property tax. In 2004, the latest year for which comprehensive data are available, the property tax accounted for 35 percent of all South Dakota state and local government tax revenues, somewhat above the 50-state average of 32 percent. (2) Among the 50 states, South Dakota ranked 16th in property tax as a percent of state and local taxes. Average property tax per person was $915 for South Dakota and $1,049 for the 50 states.
By contrast, the property tax made up 74 percent of South Dakota local taxes, only slightly above the 50-state average of 73 percent. South Dakota ranked 28th in this measure of property tax use. The difference in rankings arises from South Dakota's relatively heavy reliance on local taxes compared to its reliance on state taxes. South Dakota's local taxes per person were 86 percent of the 50-state average in 2004, but its state taxes were only 68 percent of the national average. As a result, local taxes accounted for 47 percent of state and local taxes in South Dakota, but only 42 percent of state and local taxes in the 50 states together. (Based on author's calculations from U. S. Census Bureau 2006) This is not a new phenomenon. In 1950, local taxes provided 64 percent of South Dakota's state and local taxes, compared to only 54 percent of state and local taxes in the 48 states. (Business Research Bureau 1952, 45) With this relatively heavy reliance on the property tax, it is important to understand how the tax is administered and examine the quality of property tax administration in South Dakota.
Defining the tax base
The property tax is nearly unique in how its base is determined. With most taxes, problems and disagreements most often arise in deciding what items or transactions to include in the tax base. For example, should the sales tax base include food? ... services? ... Internet sales? Once the composition is decided, the value or size of the base is usually clear: it is the value of the transactions when they occur. Defining base composition is often difficult but putting a value on it is fairly easy. With the property tax, on the other hand, deciding what is included in the tax base is fairly straightforward, especially when (as in South Dakota) personal property and financial assets are exempt. Some uncertainty remains, such as whether heavy factory equipment is real or personal. With land and buildings, the uncertainty is in knowing the value now, when the last market transaction occurred many years ago. What is the current market price of farmland bought 30 years ago? How much has a warehouse depreciated (or appreciated) in value since it was built 10 years ago? In sum, defining base composition is fairly straightforward, but determining its value is fraught with difficulties.
These innate uncertainties, especially when property tax burdens are high (or perceived to be so), can cause misunderstandings and hard feelings, spotlighting how property is assessed. We look now at criteria for evaluating property assessment, then turn to ways to measure assessment quality, and finally apply them to South Dakota.
TAX CRITERIA FOR EFFICIENCY AND EQUITY
Efficiency (as economists define it) dictates that taxes should not affect people's economic decisions. Pre-tax market prices in a well-working market do a good job of showing the true costs and benefits of goods and services, guiding economic decision makers toward the most efficient use of productive resources. Tax rate differences make some items relatively more expensive and others less expensive, causing decisions to be based on non-economic, tax considerations instead of exclusively on economic factors reflected in market prices. This leads to the first criterion, tax neutrality: none of the base should be excluded or exempted and all of the base should be subject to the same tax as a percent of value. (There may be policy-based reasons for deviating from this criterion.)
A non-property tax example may help clarify this criterion. Efficiency requires that businesses locate where their resource costs will be lowest, because that's where they use fewer and less valuable resources. Suppose that State A subjects capital equipment such as machinery and computers to its sales tax, but State B exempts equipment. Even if a firm's true costs of productive resources are lower in State A, the firm's decision makers might opt for State B if that saves them enough in sales tax. Including taxes in the decision is rational for owners maximizing their after-tax return but does not result in the most efficient use of resources. The same reasoning applies to the property tax: a firm might choose a state with lower property tax rates but higher input costs, even though it would be more efficient--from a wider perspective--to choose another location.
Different tax rates are consistent with tax neutrality if they reflect different costs of providing government services. Suppose that County C spends more per mile for road construction and maintenance than does County D, either because County C provides better services or because its input costs are higher. (Maybe County C is more urban, with higher land values and labor costs.) Higher taxes in that case lead business persons and homeowners to consider true resource differences in deciding where to locate. In this case, tax differences that reflect cost differences contribute to tax neutrality. In summary, efficiency calls for treating all property the same, except to reflect different costs of providing government services.
Equity or fairness criteria are also important in judging taxes. One such criterion is the benefit principle: the tax burden should reflect the value of benefits the taxpayer receives from services purchased with the tax revenue. People with more expensive homes or businesses probably benefit more from police and fire protection and other services, so it makes sense that they pay higher local taxes. Applying the same effective tax rate (tax as a percent of property value) to all property is consistent with the benefit principle in this case (although it is not clear that benefits increase in exact proportion to property value). Likewise, if one local jurisdiction provides more expensive services (better fire protection, for example), it makes sense under the benefit principle for persons in that jurisdiction to pay more taxes--and thus higher effective tax rates.
Another important equity criterion is horizontal equity, people with the same ability to pay should pay the same amount of tax (summarized as "equal treatment of equals"). If wealth in real property is deemed the appropriate measure of ability to pay--as seems to be the case, at least implicitly, with the property tax--then all real property should be taxed at the same effective rate.
A third equity criterion is vertical equity: taxpayers with more ability to pay should pay more tax ("unequal treatment of unequals"). This rather broad standard doesn't say how much higher taxes should be for people with higher ability to pay, but taxing all property at the same effective tax rate is consistent with vertical equity. If your house is worth twice as much as mine, you pay twice as much property tax, satisfying the vertical equity criterion. (3) Not everyone agrees with these criteria or on how to apply them. Some people dislike using real estate value to measure ability to pay.
Many states, including South Dakota, try to apply these criteria to all property of a particular type, but put different tax rates on different types or classes of property (e.g., residential, commercial, and farmland). They might impose the same rate on all farmland, for example, but use a different rate for residential property. This is consistent with the benefit principle if different property types receive different benefits from government services. It is consistent with horizontal and vertical equity if, for example, owners of residential property have a different ability to pay than do owners of commercial property with the same value.
Nearly all observers of the property tax agree that all parcels of the same type of property within the same taxing jurisdiction should face the same effective tax rate. To better understand the term "effective rate," we need to now consider the property tax process. The tax amount due for any piece of property is given by
Tax amount = assessed value x mill rate = AV x MR
Mill rate (MR) is the sum of mill rates of all local governments in which a particular property is located: county, school district, and, in some cases, city, township, and/or special district(s). (4) MR can also be thought of as the tax amount divided by assessed value, whereas effective tax rate (ETR) is the tax amount divided by market value:
ETR = tax amount / market value = (AV x MR) / market value = (AV / market value) x MR = ASR x MR
where ASR is the "assessment-sales ratio" (often referred to simply as the "sales ratio" or "assessment ratio").
To have the same ETR for each parcel within a local jurisdiction, all parcels must have the same ASR. Assessing all property at its market value (ASR = 100 percent, so ETR = MR for each parcel) is one way to accomplish this, but not the only way. As long as all parcels are under- or over-assessed by the same percentage (all have the same ASR), all have the same ETR and the tax amount depends only on a parcel's market value. "[A]n equitable distribution of the tax burden can be achieved only if assessments of value are uniform." (O'Sullivan, Sexton, and Sheffrin 1995, 35) To produce exactly equal ASRs the assessor would have to know the precise market value of every piece of property--an impossible task in a world of ever-changing market conditions. Nevertheless, assessors have developed techniques for estimating market values. For descriptions of these techniques--comparable (or comparative) sales, income (or capitalized value), and replacement cost--see Fisher 2007, 328-331 or Aronson and Schwarz 1987, 217-220.
Measuring assessment quality
Property tax professionals have several measures of the quality of property assessment for tax purposes. These researchers record actual prices at which properties sell in truly "arms length" market transactions, then compare each parcel's market (sales) value to its assessed value in the assessor's records immediately before each sale. Dividing the pre-sale assessed value by the market value yields the ASR for each property sold during a particular time period.
The median ASR (MASR) measures the general extent to which property is under- or over-assessed in the particular jurisdiction to which it applies. The median value is the one in the middle: half of all ASRs in the jurisdiction exceed MASR and half are less than MASR. (5) If MASR equals 100 percent, half the properties are under-assessed and half are over-assessed. MASR less than 100 percent means properties are generally under-assessed. While the MASR tells us the general assessment levels, it does not tell us how equally parcels are assessed: whether all parcels are over- or under-assessed to the same degree or there are wide variations in individual ASRs.
The coefficient of dispersion (COD) measures how equal (or unequal) assessments are within a jurisdiction: it is the average deviation of a jurisdiction's ASRs from its MASR. (6) If every parcel were assessed at its true market (sale) value, it would be true that ASR = 100 percent for each parcel, MASR = 100 percent for the taxing district, and COD = 0 for the taxing district. Tax neutrality and equity do not require that MASR = 100 percent, only that COD be low--ideally, zero--indicating that all property is under- or over-assessed to the same degree. There is no hard-and-fast criterion for what COD level constitutes good or even acceptable assessment, but most observers seem to consider a COD of 15 percent to 20 percent a good rule-of-thumb. Anything higher than that is usually thought to indicate poor quality assessment. (S. D. Department of Revenue and Regulation 2005, 9; Aronson and Hilley 1986, 133; Aronson and Schwarz 1987, 222)
The COD is not a good measure of vertical equity. Parcels could be assessed on average close to the MASR--hence presenting a low COD--yet still exhibit some undesirable assessment patterns. For example, more valuable parcels might be under-assessed more frequently or to a greater degree than are less valuable properties. If our goal is the same ETR for everyone in the jurisdiction, then any deviations should be due to random uncertainty arising from lack of information. We turn now to a closer look at the quality of property assessment in South Dakota, and how it has changed over time.
Assessment in South Dakota
Assessment quality has vexed South Dakotans for as long as they have had a property tax. One writer stated that in the 1880s "many of the people of the time were of the opinion that real estate was assessed at not more than 50% of its true value" even though "[a]ll property was to be assessed at its true cash value." (Schneider 1918, 8, 11) Since at least 1919, state law has provided that "[all] property shall be assessed at its true and full value in money," where "true and full value [means] the usual cash selling price at the place where the property to which the term is applied shall be at the time of assessment." (Serr 1932, 33, citing South Dakota Revised Code of 1919, Sections 6700 and 6666) The definition has remained unchanged to the present. (SDCL 10-6-1 ; also Compiled Laws of South Dakota 1929, Sections 6700 and 6666; SDCL 10-6-1.3 contains another definition.)
Evidence of assessment quality is quite scarce before the 1950s. A 1936 ASR study, probably South Dakota's earliest, examined 1,758 farm sales in 20 counties during 1925-1934. "Counties from all sections of the state were chosen so as to.. give an approximate picture of the situation in all parts of the state." "[A]ll family transfers, trades, forced sales, and sales which seemed not to have the full consideration in the deed were eliminated." (Westbrook and Strand 1936, 10) This study found a MASR of approximately 95 percent; 53 percent of ASRs were below 100 percent, 47 percent were above. (Westbrook and Strand 1936, 148)
While the MASR was close to 100 percent, ASRs were by no means uniform, ranging from 10 percent to 350 percent. Only 308 ASRs (17.5 percent of the total) were between 90 percent and 110 percent. (Westbrook and Strand 1936, 15) Considering each county separately, average ASRs (median ASR is not reported for counties) ranged from 69 percent to 127 percent and CODs ranged from 19 percent to 59 percent. Only one COD was below 20 percent and only five were between 20 and 29 percent. Counties with average ASR closer to 100 percent tended to have smaller CODs, (Westbrook and Strand 1936, 22) although the correlation was not strong.
The 1936 study also reports how ASRs varied with land values per acre and with total farm values, finding that lower-valued property tended to be over-assessed and higher valued property tended to be under-assessed (Westbrook and Strand 1936, 16-19),
which means simply that those who own the farms of higher value pay a smaller proportion of taxes than they rightfully should, while those who probably can less well afford to pay taxes are forced to do so on valuations which are set at more than their farms are worth. In all counties studied, farms of least value are assessed at a higher percentage of sales value. (Westbrook and Strand 1936, 23)
This pattern clearly violates the vertical equity criterion. Overall, the study's authors concluded, "There can be little doubt that the work of assessment is poorly done in South Dakota." (Westbrook and Strand 1936, 21)
"Prior to 1957, South Dakota law specified that all property be assessed at true and full value; the Thirty-fifth Legislature amended the provision to require a 60 per cent assessment level." (S. D. State Legislative Research Council 1964, 14) Historically, many states have lowered their target ASRs closer to the actual ASR. The rationale seems to be that moving the statutory requirement closer to actual assessment would give assessors a more attainable goal and would lead to more equal assessments. Assessors often responded with lower assessments, frustrating legislators' efforts. South Dakota was no exception.
In 1958, the Department of Revenue began performing an annual study of property assessment. As Chart 1 shows, ASRs hovered around 45 percent, about 75 percent of the statutory target, for the first few years after the 1957 change, then gradually declined for the next 15 years, reaching barely half the target that was supposed to be legally required. Chart 1 also shows that assessors tended to assess urban property higher than rural property relative to market values. These early studies do not report ASRs for agricultural and non-agricultural properties, but later experience suggests that the rural/ urban discrepancy is primarily due to differences between agricultural and non-agricultural ASRs.
In 1977, the Legislature restored the 100 percent ASR goal, but the county commissioners set a "taxable percentage" of 60 percent or less, to be applied to the new (presumably higher) assessed values. The mill levy was then applied to the new "taxable value" instead of directly to the assessed value. (Ring 1989, 3) This new arrangement allowed assessors to more directly target market value, but the commissioners could choose a taxable percentage that left the taxable value close to the old assessed value, presumably easing the minds of taxpayers fearing that more accurate assessments would raise their taxes. As Chart 1 shows, statewide ASRs more than doubled (rural ASRs nearly tripled) between 1976 and 1977 and never again reached the very low levels of the early 1970s.
In the next major statutory change, the 1989 Legislature eliminated the 1977 taxable value provision, returning to the older, simpler system of applying mill rates directly to assessed values. (Ring 1989, 4-6) Perhaps this change made the statutory requirement that property "be assessed at its true and full value in money" more clear, helping ensure more accurate assessment. The Legislature also provided additional incentives for local assessors to take their job more seriously and stipulated that, beginning in 1993, a county's MASR had to be between 85 percent and 100 percent and its COD less than 30 percent. (Ring 1989, 21) The statutory maximum COD is now 25 percent. (SDCL 10-6-33.9)
Chart 1 shows that statewide ASRs generally trended upward after 1991 and have hovered around 87 percent to 90 percent since 2001. Property tax assessments are finally hitting the statutory targets, at least since about 2000, indicating improved assessment quality. However, appearances can be deceiving.
Excluding data from comparative sales
Because most property sells infrequently, assessors must find ways to estimate each parcel's market value if they are to hit the MASR and COD targets specified in the law. Perhaps the most common method used for making these estimates is comparable sales, under which assessors use the prices of recently-sold parcels to estimate the values of comparable parcels (nearby location, similar size and age of residence, similar soil productivity, etc.) that have not sold recently. If external factors push up sale prices, then assessments and taxes for comparable properties will also rise. Two examples will illustrate.
Developers who plan to convert land to residential or commercial use as an urban area expands will pay more than the agricultural value of that land. Nearby land with potential for similar conversion also rises in value as landowners anticipate further expansion, and the comparable sales method raises assessments accordingly. If property value is the proper measure of ability to pay--as seems to be the assumption with the property tax--then it makes sense that property taxes on this land should also rise. Unfortunately, potential for future development does not increase current agricultural productivity, so farmers operating on the urban fringe may have particular difficulty paying their taxes, and might even be forced to sell out.
Retirees and persons wanting vacation homes may be willing to pay high prices for sites in a scenic area such as the Black Hills. Nearby sites, whose owners have lived there for a long time, will also rise in value. Assessment based on comparable sales will reflect those rising values, leading to higher assessments and taxes for nearby property. Again this is consistent with horizontal equity and the assumption that property value measures ability to pay, but taxes may grow beyond the taxpaying ability of a person who purchased a site several years before.
In 1998 and 1999, the Legislature changed the way property is assessed and the way ASR studies are reported, in an attempt to keep a few out of line sales from inordinately influencing assessments of other property. SDCL 10-6-74, enacted in 1999, reads as follows:
Any real property which sells for more than one hundred fifty percent of its assessed value, may not be used for the purpose of valuing other real property. The sale of any real property which is not used for the purpose of valuing other real property pursuant to this section may not be used in any sales ratio study.
Thus, any sale resulting in an ASR of less than 66.7 percent (the reciprocal of a sales-to-assessment ratio greater than 150 percent) is automatically excluded from the comparable sales used for updating assessments and from data used for calculating a jurisdiction's MASR and COD.
Some idea of the magnitude of this exclusion can be gleaned from the sales ratio frequency distribution table published in each year's South Dakota ASR study (South Dakota Department of Revenue and Regulation 2005). This table shows the number of sales with ASR values in varying ranges (Under 40 percent, 40-50 percent, etc., through 150 percent and over). Tables for recent years show no sales with ASR below 60 percent and relatively few in the 60-70 percent range. Tables published prior to enactment of these exclusions suggest that roughly 10 percent of sales statewide, and as much as 20-25 percent of sales in some counties, resulted in ASRs below 60 percent. This does not include ASRs in the 60-70 percent range, although about two-thirds of them will also be excluded. There may be some later indications that these exclusions have grown in importance.
A parcel that sells for 50 percent more than its assessed value might really be an outlier that sold at an exorbitant price and should not be included in comparable sales used for updating assessments. On the other hand, the outlier might have sold at its market value but was seriously under-assessed before the sale; this is especially likely when property values are growing rapidly and assessments lag behind. In that case, to mandate that sales be excluded is to mandate underassessment and takes discretion from assessors who are expected to assess as accurately as possible. A truly voluntary, arms-length transaction between knowledgeable parties should usually reflect the true market value. With certain sales excluded from use for updating assessments, it might make sense to leave them out of the estimates of assessment accuracy, but it has the potential for biasing ASR and COD estimates. Screening out only the most extreme cases of underassessment will overstate MASRs and understate CODs as compared to the methods used in typical ASR studies and used in South Dakota before 1999. It is probably not a coincidence that statewide MASR figures appear to have improved dramatically right after the data exclusions began, but there is probably no way to know how much of the apparent improvement is real and how much can be attributed to required data screening. Any comparison of recent ASR studies in South Dakota to earlier studies or to studies from other states is at best questionable and may be nearly meaningless.
The equity implications of this data screening could be serious. To the extent that a high degree of underassessment can be hidden, resulting in reported CODs being lower than actual CODs, the horizontal equity criterion could be violated to a much greater extent than published CODs suggest. In particular, those parcels of property that are under-assessed to the greatest degree (i.e., sales price more than 50 percent above the assessed value that determines tax liability) and thus receiving the greatest degree of preferential treatment are also the ones hidden by statutory screening. The potential for abuse seems obvious.
Unequal treatment might occur in yet another way. Suppose that School Districts A and B have equal published (i.e., with data excluded) MASRs, but District A's true MASR (i.e., calculated with all data) is lower because more of its property is so under-assessed that it is excluded from published figures. Both districts have the required MASR of 85 percent, but one in fact has substantially lower assessments compared to actual market value. Because the South Dakota general state aid formula bases state aid on a school district's assessed value--higher assessed value per pupil means the district collects more revenue per pupil from local property taxes and receives less per pupil from the state--District A would receive more state aid and District B less than if property were more equally assessed. The statutory data screening would in this case require and hide what most citizens probably consider to be inequitable treatment. Intentional under-assessment--if it's egregious enough--is rewarded with more state aid for local schools.
Constitutional Amendment D
Notwithstanding these attempts to slow assessment growth by excluding sale prices deemed to be out of line, South Dakota citizens have proposed an even more restrictive alternative. Constitutional Amendment D, which was placed on the ballot by initiative and will be decided by South Dakota voters on November 7, 2006, reads as follows:
The legislature is empowered to establish a procedure for each class of property detailing the methods that shall be used to value property. The value of property, except centrally assessed property, shall be based on the acquisition value of the property and shall apply to any real property sales occurring after January 1, 2007. The assessed value of such property may be annually adjusted by a factor not to exceed three percent as determined by the legislature. The annual adjustment provided in this section in the value of such property shall be applicable each year beginning with the 2003 assessed property valuation, which is hereby established as the base year for this valuation procedure. However, the value of property may be further adjusted if there is a change in use or classification or to account for any addition, improvement, or destruction to the property. (S. D. Secretary of State 2006)
It is clear that adjustments begin with 2003 assessments, but it is not clear--at least to this author--when annual adjustments begin. Is 2003 also the base year for beginning the annual adjustments, so that they can be adjusted upward by three percent (or less, at the Legislature's discretion) per year to get the assessments to be used for 2007 taxes payable in 2008? Alternatively, since parcels sold after January 1, 2007, are assessed at their "acquisition value," is 2007 also the year in which three percent (or lower) annual adjustments begin, so that 2003 assessed values will determine 2007 taxes payable in 2008?
Assessments of individual parcels could not rise by more than three percent per year, but it is not clear whether this factor must be applied uniformly to all property. Could the Legislature decide, for example, that farmland assessments could be raised by one percent while non-agricultural assessments were raised by three percent? Could each parcel's assessment be changed by a different amount (e.g., the amount by which the assessor believes each parcel's full and true value changed during the year, as ideally happens now), with a three percent limit?
In exploring some of the possible effects of this Amendment, we assume that annual adjustments would start with 2003 and would be for the full three percent per year; that is, 2007 assessments would be 12.55 percent (three percent compounded annually) above 2003 assessments. These assumptions produce the smallest changes that Amendment D would likely require, compared to what would occur under current practices and constitutional provisions.
Table 1 shows selected years' statewide assessed values for the three largest property classifications. Starting with assessed values for 2005, the latest year for which assessment data are available, we assume that, without Amendment B, assessments would grow during 2006 and 2007 at the average annual growth rate between 1999 and 2005. (7) The estimated four-year growth in actual assessments is 31 percent for agricultural property, 39 percent for owner-occupied property, and 32 percent for other (primarily commercial and industrial) property, compared to 12.55 percent growth assumed to occur under Amendment D. In other words, under Amendment D agricultural property would be assessed at about 86 percent of what it would be under current practice, residential property at about 81 percent of current practice assessments, and business property at about 85 percent of current practice levels. We assume for comparison purposes that the statewide average AS R was 90 percent in 2003 for all three types of property. Under that assumption, Amendment D would force the statewide average ASR to drop from 90 percent in 2003 to about 72 percent for owner-occupied residences and about 77 percent for the other two property types. This estimate unrealistically assumes no reassessments in 2007. If we assume that five percent of property sells or experiences "addition, improvement, or destruction" in 2007 (8) and is reassessed at its "acquisition value," the average statewide ASRs would be slightly higher at about 74 percent for owner-occupied residences, 78 percent otherwise, still down considerably from 90 percent assumed for 2003.
With estimated 2007 assessments under Amendment D below 2005 and 2006 actual assessments, government officials could raise mill rates, since Amendment D limits assessments but not taxes. Currently, the Legislature each year establishes mill rates that each school district must impose. This produces a certain amount of local revenue for the district; the state then adds enough aid to enable each district to spend a specified amount per pupil. Chart 2 shows that these rates have declined in most years, even though spending per pupil rose. With slower assessment growth under Amendment D, the Legislature could continue the same increases in spending by adjusting the required mill rate accordingly. Thus, Amendment D would not, of itself, lower total property tax increases.
Under current law, other local governments may increase property tax revenues annually by three percent or the rate of inflation, whichever is less. (SDCL 10-13-35, 10-13-38) When assessments rise faster than these tax limits, local officials compensate by lowering the mill rate. Chart 3 shows that median effective property tax rates fell between 1994 and 1997 as the property tax relief program was put in place. Effective rates continued to fall after that as assessments rose faster than did property taxes subject to the limits. If assessments rise less rapidly under Amendment D, as seems likely, local officials could compensate by setting mill rates higher than they would be without assessment limits.
While Amendment D would not by itself force a reduction in overall property taxes, it would shift their distribution. Property owners who experienced more rapid appreciation would pay proportionately less taxes under Amendment D than they would if their assessments rose with property values. In this scenario, with owner-occupied residences appreciating more rapidly than are other types of property (as occurred between 1999 and 2005), owner-occupied residences would bear a smaller share of total property taxes than they would under the current law and constitution; agricultural and business properties would bear a larger share. Burdens would also be shifted onto purchasers of any type of property in 2007, since they would be assessed at their acquisition price, while assessments on old property were held down by Amendment D.
To better understand possible longer-range implications of adopting Amendment D, we now examine a 10-year scenario using the following assumptions for each year from 2007 to 2016: properties grow in market value at the 1999-2005 average annual rate (see Table 1); five percent of property sells each year and is reassessed to its sale value; after property is sold, its assessed value grows three percent per year; and property that does not sell experiences assessment growth of three percent per year. In 2016, after 10 years with Amendment D, assessments would be about 62 percent to 72 percent of what they would be under more accurate assessments. The statewide average ASR would drop from 90 percent in 2003 to about 56 percent for owner-occupied residences and 65 percent for business and agricultural property in 2016. Taxes would be shifted even more from residential property to agricultural and business property as compared to what would happen under current provisions, assuming that residences appreciate more rapidly. Different relative growth rates would yield different tax shifts.
Of particular concern for horizontal equity, effective tax rates within each property classification could be much different on similar properties that sold at different times. A farm or business that sold in 2016 would be assessed at about 150 percent of the assessed value of an identical entity that did not sell between 2007 and 2016. A residence that sold in 2016 would be assessed at about twice the assessment of an identical house that did not sell during that time. In other words, effective tax rates on some properties could be double those on other properties, depending on when each property sold. More unequal assessments mean that coefficients of dispersion would be higher under Amendment D than they are now and higher than they would be in 2016 under current practices.
Other states that adopted assessment limits have experienced more unequal assessments. California's Proposition 13, adopted in 1978, rolled assessments back to 1975 levels and limited growth to two percent per year unless the property sells; it also limited total property tax rates to 1 percent of assessed value. One study found a COD of 58 percent in 1984 and another found a COD of 53 percent in 1982, both well above the 20 percent standard. Four studies found CODs of 9 percent to 16 percent in the few years before Proposition 13. (Fisher 2007, 335, 337, citing Wiseman 1986 and U. S. Bureau of the Census 1984) One author stated that "in 1978 California voters chose to sacrifice equity in property taxation for certainty regarding year-to-year changes in tax liability." (Wiseman 1986, 31, quoted in Fisher 2007, 337)
Another California study found that by 1991 there were "dramatically large disparities in property taxation among virtually identical properties." In two counties, "for those households [that stayed in their homes from 1975 to 1991], their effective tax rate is approximately 0.2% compared to 1% for recent purchasers of property. To put it another way, if an individual who has owned his house since 1975 suddenly purchased an identical property next door, property tax payments would increase by a factor of five." (O'Sullivan, Sexton, and Sheffrin 1995, 137)
In 1992, Florida limited assessment growth for owner-occupied homes to three percent per year unless the property sells; "similar Floridians can live in identical houses but receive different tax levies." A reporter found "condominium neighbors where the first pays $5,700 in countywide taxes while the second pays $2,300." (Hawkins 2006, 8, 10)
In 1994 Michigan limited assessment growth to the lesser of five percent or inflation unless the property sells. By 2002 farmland had "taxable values less than 60 percent of what they would be without the limit. Taxable values of residential property ... were nearly 80 percent of what they would be otherwise. The limit likely had a greater effect on agricultural than residential property because ownership changes less often for agricultural property." (Fisher 2007, 339, citing Feldman, Courant, and Drake 2003)
These studies found other predictable effects of assessment limits. If one sells property she has owned for some time and buys another property, she gives up a favorable assessment for a higher assessment based on market value. This tax penalty for moving lowers household mobility. (Hawkins 2006, 9; O'Sullivan, Sexton, and Sheffrin 1995, 50, 52, 83, 84) This system also has a disproportionately high effect on persons or businesses that tend to sell property more frequently. A study of Proposition 13 found that lower income persons would tend to benefit from assessment limits, because (at least in the California cities studied) lower income persons tend to remain in the same place for longer. It also found that older homeowners would benefit while younger homeowners would pay higher effective tax rates because the latter are more likely to have purchased a home recently and more likely to want to trade up. (O'Sullivan, Sexton, and Sheffrin 1995, 73-78)
These findings suggest that limiting assessments in South Dakota is likely to place disproportionate property tax burdens on young people buying their first home or first farm, and anyone starting a business. It might also discourage economic development because a firm buying land and/or buildings would see assessments immediately raised above those of competitors. One observer states the following: "Newcomers and young people are the lifeblood of communities. By rewarding a small group of old-timers--indeed, by cementing them in place--we make it harder to attract the ones we need most." (White 2006)
SUMMARY AND CONCLUSION
Property tax assessment is a difficult and uncertain task. "A unique feature of the property tax is that its base, usually the market value of property, must be estimated, as opposed to being determined by the outcome of a specific transaction." (O'Sullivan, Sexton, and Sheffrin 1995, 35) Uncertainty and subjectivity inevitably bring doubt, dissent, and sometimes protest to an act that no one likes under the best of circumstances: paying taxes.
Rapidly rising property values can add to the uncertainty and doubt as assessors try to keep assessments in line with property values, as state law requires. Taxpayers may doubt that prices are really increasing so rapidly; if they accept that prices are rising, they may doubt that property "could really be worth that much," or blame outsiders for driving prices too high. This is not just an academic exercise--or restaurant discussion over coffee--that excites the participants but has little personal impact. It affects every property owner's pocketbook, often to the tune of thousands of dollars. Some taxpayers' liabilities may be rising faster than their ability to pay. Thence come dissent and protest.
The Legislature has responded by mandating that assessors must ignore certain sales that might be considered out of line with market values; to be consistent, they also mandated that these sales be excluded from measures of the quality of property tax assessment. While assessment quality seems to have improved considerably in the past 10-15 years, data screening casts doubt on that improvement.
Believing that these exclusions do not do enough to reduce assessment increases, South Dakota citizens have placed Amendment D on this fall's ballot, which would limit assessment increases to three percent per year, except that property will be reassessed at its acquisition value when it sells. Because this amendment limits assessments but not tax revenue, it will not reduce overall property taxes without additional legislation. Some property owners will experience lower assessments than they would under current practice, but those lower assessments may be largely offset by higher mill rates. Owners who purchase property are likely to have essentially the same assessments as under current practices and higher mill rates--giving them higher tax burdens.
Amendment D could have quite serious effects on assessment quality. It will almost certainly force assessments to continually fall farther below actual values. It will require that different properties be subject to vastly different effective tax rates, leading to serious inequities in property tax burdens. Each of these disparities will grow over time.
Note: This article does not necessarily represent the viewpoint of the Business Research Bureau, the School of Business or The University of South Dakota.
(1) For a brief discussion of the property tax during the early days of statehood, see Business Research Bureau 1952, 42-43 and Whitlow 1942, 107-109.
(2) Here, "50 states" and similar expressions include the District of Columbia.
(3) If you paid only 50 percent more than I paid, vertical equity would still be satisfied. More specific measures of vertical equity are beyond the scope of this paper--and the source of wide disagreements.
(4) The terms "tax amount" and "mill rate" are used here for clarity. In South Dakota law, "revenue" seems to be synonymous with "tax amount" as used here. "Tax levy" and "mill levy" seem to be synonymous with "mill rate" as used here (SDCL 10-6-1.1), although "tax levy" and "levy" sometimes mean "revenue" or "tax amount" (e.g., .SDCL 10-12-7, 10-12-8, 1-13-35)." The term, one mill, wherever it is used in this code means 'one dollar per thousand dollars of taxable valuation.'" (SDCL 10-6-1.2)
(5) In this case statisticians prefer the median to the mean (or average) as a representative value, because the median is less affected by a few very high values that are not representative of the overall sample.
(6) This calculation can be thought of as having three steps. First, the difference between MASR and each parcel's ASR is calculated. Second, the differences are summed and the sum divided by the number of sales. Finally, the result of the second step is divided by MASR.
(7) Assessment data prior to 1999 are not strictly comparable to later figures because of changes in assessment practices from introduction of the owner-occupied category in the 1990s and changes due to exclusion of sales prices greater than 150 percent of assessed value.
(8) The assumption that five percent of property sells each year is chosen arbitrarily for comparison purposes and used for all classifications of property for simplicity. It is likely that farmland and business property sell less frequently than do owner-occupied residences. If property sells more frequently than once every 20 years on average, the estimates of Amendment D's effects are too high; if property sells less frequently, estimates are too low.
Aronson, J. Richard and John L. Hilley. 1986. Financing State and Local Governments, 4th ed. Washington, D. C.: Brookings Institution, Studies in Government Finance.
Aronson, J. Richard and Eli Schwartz. 1987. Management Policies in Local Government Finance, 3rd ed. Washington, D. C.: International City Management Association.
Business Research Bureau. 1952. The South Dakota Economy at Mid-century: 1900-1950. Vermillion, S. D.: University of South Dakota, School of Business Administration, Business Research Bureau, Bulletin 26 (June).
Fisher, Ronald C. 2007. State and Local Public Finance, 3rd ed. Mason, Ohio: Thompson South-Western.
Hawkins, Richard. 2007. Four Easy Steps to a Fiscal Train Wreck: The Florida How-to Guide. Atlanta, Ga.: Georgia State University, Andrew Young School of Policy Studies, Fiscal Research Center, FRC Report No 132 (August).
O'Sullivan, Arthur O., Terri A. Sexton and Steven M. Sheffrin. 1995. Property Taxes and Tax Revolts: The Legacy of Proposition 13. New York: Cambridge University Press.
Rackham, John. nd. "Design and Use of Assessment Ratio Studies by Local Assessors." Paper presented at the 37th Annual International Conference of Assessment Administration, Boston, Mass.
Ring, Raymond J. 1989. South Dakota's Property Tax System: Description, Evaluation, Potential Reforms, Prepared for Commission on Tax Fairness and Government Cost Effectiveness (July).
Schneider, Eugene C. 1918. "Taxation in Dakota Territory." Masters Thesis, University of South Dakota.
Serr, Harold A. 1932. "The Tax System of South Dakota, 1919-1932." Masters Thesis, University of South Dakota.
South Dakota. Secretary of State. 2006. Constitutional Amendment D-Initiated Constitutional Amendment to establish a procedure for each class of property detailing the methods that shall be used to value property, http://www.sdsos.gov/electionsvotereqistration/ upcomingelectionballotquestionstatus.shtm [4 September 2006]
South Dakota. State Legislative Research Council. 1964. Some Approaches to Tax Relief and Equalization. Staff Background Memorandum. Pierre, S. D.: State Legislative Research Council.
South Dakota. Department of Revenue and Regulation, Property Tax Division. 2005. Assessment and Sales Information 2004. Pierre, S. D.: Author. http://www.state.sd.us/drr2/publicat.htm#PROPERTYTAX/SPECIALTAXES [17 July 2006]
South Dakota. Department of Revenue and Regulation. 2005. Annual Report 2005. Pierre, S. D.: Author. http://www.state.sd.us/drr2/ publicat.htm#Annual
United States. Census Bureau. 2006. State and Local Government Finances: 2003-2004, State and Local Summary Tables by Level of Government. http://www.census.gov/govs/www/estimate04.html [3 July 2006]
Westbrook, R. B. and N. V. Strand. 1936. Inequalities Arising from the Assessment of Farm Real Estate in South Dakota. Agricultural Economics Department Bulletin 300. Brookings, S. D.: South Dakota State College of Agriculture and Mechanic Arts, Agricultural Economics Department.
White, Otis. 2006. "How Tax Limitations Are Like Rent Control." Governing.com (September 29) http://www.com/notebook.htm [2 Oct 2006]
Whitlow, C. J. 1942. The Property Tax in South Dakota. Ph. D. dissertation, University of Michigan.
Wiseman, Michael. 1986. "Proposition 13 and Effective Property Tax Rates in San Francisco." Research Papers in Economics, no. 86-8. Berkeley: University of California at Berkeley (February).
About the author:
Raymond J. Ring is Professor of Economics in the School of Business, The University of South Dakota, Vermillion, South Dakota. Dr. Ring teaches Public Finance, Money and Banking, and History of Economic Thought. He has studied South Dakota's tax system as well as the sales tax in all 45 states, and has written reports and journal articles on both.
Table 1 Projected effects of Amendment D, 2007 and 2016 Agri- Owner- Other non- Year cultural Occupied Agricultural Assessed values 1999 $10.758 $10.401 $7.072 (billions) 2001 11.550 11.903 7.722 2003 13.089 13.993 8.75 2005 15.252 16.661 10.194 Ave. annual percent 1999-2001 3.6% 7.0% 4.5% change 2001-2003 6.5% 8.4% 6.4% 2003-2005 7.9% 9.1% 7.9% 1999-2005 6.0% 8.2% 6.3% 2007 projections Est. actual assessed value 17.13 19.49 11.52 Est. AV with Amend. D, no sales 14.73 15.75 9.85 Est. AV with Amend. D, 5% sales 14.85 15.94 9.93 Amend. D, no sales/actual 0.86 0.81 0.86 Amend. D, 5% sales/actual 0.87 0.82 0.86 2016 projections Est. actual assessed value 28.92 39.52 19.93 Est. AV with Amend. D, 5% sales 20.92 24.31 14.12 each year Amend. D, 5% sales/actual 0.72 0.62 0.71 Source: 2005: S. D. Department of Revenue and Regulation, Annual Report 2005, p. 24 Other years: S. D. Department of Revenue and Regulation, Property Tax Statistical Report, various years. Chart 3 Median effective property tax rates Median effective rate (percent) Year 1994 1997 2005 Residential 2.87 2.16 1.87 Commercial 2.92 2.82 2.42 Agricultural 1.77 1.38 1.23 Source: S. D. Department of Revenue and Regulation, Annual Report, various years. Note: Table made from bar graph.