Skip universal navigation

New York State Universal header

Skip to main content

ASSESSMENT EQUITY IN NEW YORK: RESULTS FROM THE 2013 MARKET VALUE SURVEY

Introduction

Distributional equity in real property taxation requires that properties of the same value be treated alike in terms of their assessments. New York State law (Section 305 of the Real Property Tax Law) stipulates that assessing jurisdictions must assess properties at a uniform percentage of value, and State courts have held that "value" means "market value." New York's two "special assessing units," New York City and Nassau County, must assess at a uniform percentage of market value within each of four specified property classes. This means that all parcels on the assessment roll (or, for special assessing units, within each class) should have the same, or at least very similar, assessment ratios (assessed value divided by market value). Each assessing unit has the right to choose the percentage of value to be used as an assessment standard.

The New York State Department of Taxation and Finance has statutory responsibility for regularly monitoring the equity of assessments. An analysis of assessment uniformity is carried out each time the Department's staff completes a market value survey.1 The present report documents findings from the 2013 market value survey.

Included herein is information for the State's 983 non-village assessing units, which consist of 2 counties, 61 cities and 920 towns.

2013 Market Value Survey Data and Estimation Methodology

For the 2013 market value survey, the level of assessment uniformity for each assessing unit was estimated using one of four possible approaches, as outlined below:

  1. Sales Ratio Study - This approach involves a systematic comparison of assessed values to sales prices for properties that have sold. It is used if there is a sufficient sample of sales to provide a reliable estimate of the ratio of assessed value to sales price. The sales prices may be time-adjusted to a common valuation date in order to establish value and assessment ratio as of that date. It is used for residential property only, and supplemented by appraisals for other property types.
  2. Computer Assisted Mass Appraisal Ratio Study (CAMA) - CAMA involves a systematic comparison of assessed values to market values generated by a statistical model. The model uses a multiple regression equation to predict the market value of residential parcels based on sales data and the physical inventory characteristics of the parcels. The CAMA approach is particularly useful in municipalities with few sales but good parcel inventory data. It is used for residential property only, and is supplemented by appraisals for other property types.
  3. Stratified Sample of Parcels Using Appraisals and/or Sales - A stratified sample was used where suitable data were available and the sales ratio and/or CAMA approaches were determined to be inappropriate. Direct use of time-adjusted sales in ratio determination was limited to residential property, and sample parcels were weighted based on the incidence of each type on the assessment roll.
  4. Review and Verification of a Recent Reassessment - The review and verification process was used in assessing units having conducted recent reassessments. It involves an audit-type analysis of the reassessment process and its results to verify that assessments adequately reflect current market conditions. As part of the review, an extensive audit document is completed (see example in Appendix A) in which the data and analytical processes used to determine the assessments are examined in relation to acceptable professional standards of practice.

Approaches 1 through 3 result in independent computations of the local level of assessment and the assessing units in question are referred to in this report as the "sampled assessing units." Classes of property comprising a very minor portion of the value on the assessment roll may not have been independently valued. The full value of these classes was estimated by applying the overall ratio of the sampled classes. In some cases, appraisals from a prior, recent market value survey may have been reused in the 2013 market value survey.

As mentioned above, assessing units having completed recent reassessments were not sampled, primarily to focus available resources on those with older assessment rolls. These units were handled through the fourth approach, "Review and Verification of a Recent Reassessment," and are referred to as "non-sampled assessing units". A recent reassessment means that the assessing unit has an equitable assessment roll, unless significant errors have been made in the reassessment process, in which case the roll would not have been accepted for review and verification in the survey. Therefore, in the statistical tabulations presented in this report concerning assessment equity, assessing units wherein the review of reassessment procedure was used and the reassessment was deemed acceptable have been counted as meeting equity standards.

It should be understood, however, that the designation of assessing units as having adequate equity does not necessarily imply that all properties in them are accurately assessed. Factors such as uniqueness of certain properties, rapidly changing neighborhood conditions and/or dynamic market segments in certain time periods, and other such factors, will necessarily mean that there will be inevitable imperfections on the roll, as assessment is not an exact science.

Statistical Measurement of Assessment Uniformity

The primary means of measuring assessment uniformity is a statistic known as the Coefficient of Dispersion (COD). The COD measures the extent to which the assessment ratios from a given roll exhibit dispersion around a midpoint. It is generally accepted that the median assessment ratio best serves as the midpoint or central tendency measure from which the average level of dispersion should be calculated.

Assessing units with good assessing practices have low CODs, showing little deviation of individual assessment ratios from the median ratio. For example, if the median ratio for the parcels sampled in a given assessing unit is 50 percent, a house with a market value of $100,000 should be assessed at $50,000, a commercial property valued at $400,000 should be assessed at $200,000, and a $2,000,000 industrial parcel should be assessed at $1,000,000. If all other sampled parcels were similarly assessed at 50 percent of market value, the median ratio would also be 50 percent and the average deviation, as measured by the COD, would be zero. Conversely, an assessing unit with little assessment uniformity would have widely varying assessment ratios among the sampled parcels, resulting in high dispersion around the median ratio and, therefore, a high COD. Widely varying ratios result in unequal tax bills for properties of equal value.

Examples 1 and 2 show two hypothetical assessing units, each attempting to assess properties at 80 percent of market value. In Example 1, the assessed values range from 52 percent to 120 percent of market value, indicating a relatively high level of dispersion and poor assessment practices. Assessments such as these would result in an inequitable distribution of local taxes between property owners.

Example 1. Coefficient of Dispersion of 30 Percent: Low Uniformity
Parcel # Assessed Value Market Value AV/MV Ratio Absolute
Deviation
from Median
1. $120,000 $100,000 1.20 .40
2. $110,000 $100,000 1.10 .30
Median 3. $80,000 $100,000 .80 .00
4. $58,000 $100,000 .58 .22
5. $52,000 $100,000 .52 .28
Total Deviation 1.20
Total Deviation

No. Parcels

=
1.20

5

=

.24 average deviation from median

COD =
Avg. Deviation

Median Ratio

=
.24

.80

=

30 percent

Example 2 shows a hypothetical case where assessments are more uniform. The assessment ratios range from 64 percent to 92 percent, and are closer to the target ratio of 80 percent, showing substantially less dispersion than is evident in Example 1. While some dispersion is indeed present, it is significantly lower than in the previous example and within an acceptable range when factors such as measurement error and valuation uncertainty are taken into account.1

Example 2. Coefficient of Dispersion of 10 Percent: Acceptable Uniformity
Parcel
Number
Assessed
Value
Market
Value
AV/MV
Ratio
Absolute
Deviation
from Median
1. $92,000 $100,000 .92 .12
2. 88,000 $100,000 .88 .08
Median 3. 80,000 $100,000 .80 .00
4. 76,000 $100,000 .76 .04
5. 64,000 $100,000 .64 .16
Total Deviation .40
Total Deviation

No. Parcels

=
.40

5

=

.08 average deviation from median
COD
=
Avg. Deviation

Median Ratio

=
.08

.80

10 percent

A second statistical measure of assessment uniformity, called the price-related differential (PRD)2 is also used in the current report for assessing units with no recent reassessment equity. The PRD is used to determine if there is a bias on an assessment roll toward systematic over-assessment of either high- or low-value properties in comparison to the average property. In computing the PRD, the simple mean of the assessment ratios is divided by the value-weighted mean ratio. If no bias exists, the two ratios should be close to each other, and the PRD should be near 1.00. This is referred to as "neutral" assessment practice, i.e., no price-related bias. However, if the simple mean ratio is considerably lower than the value-weighted mean, a low PRD results (less than 1.00). In this case, there is said to be a bias toward "progressivity," that is, higher-value properties are being over-assessed and lower-value properties are being under-assessed. In the opposite situation, where the PRD is high (greater than 1.00), "regressive" assessing is evident. In other words, lower-value properties are being relatively over-assessed and higher-value properties are being relatively under-assessed. The International Association of Assessing Officers (IAAO) has established a range for the PRD which denotes uniform practices, i.e., neutral assessing: the PRD must fall in the range 0.98 to 1.03 to be considered acceptable.

Coefficient of Dispersion Standards

Upon the completion of a market value survey, two coefficients of dispersion are calculated for the assessing units with no recent reassessment activity, one for residential property alone and one for all property classes combined. To evaluate the CODs calculated in this process, they must be compared to accepted guidelines for assessment uniformity. The International Association of Assessing Officers, in its publication entitled Standard on Ratio Studies (January 2010), has recognized that the ability of an assessing unit to attain uniformity is affected by several factors, such as the types of property it contains, community size, population density, the degree of diversity of properties, market activity levels, and the relative ages of structures. The IAAO recommends a range of acceptable COD values, based on these categories and neighborhood characteristics, as well as the increased difficulty experienced in assessing classes of property other than residential. IAAO standards are summarized in Table 1.

Table 1. Ratio Study Uniformity Standards Indicating Acceptable General Quality, IAAO*
Type of property - General Type of property - Specific COD Range**

Single-family residential(including residential condominiums)

Newer or more homogenous areas 5.0 to 10.0
Single-family residential Older or more heterogeneous areas 5.0 to 15.0
Other residential Rural, seasonal, recreational, manufactured housing, 2-4 unit 5.0 to 20.0
Income-producing properties Larger areas represented by large sample 5.0 to 15.0
Income-producing properties Smaller areas represented by smaller sample 5.0 to 20.0
Vacant land --- 5.0 to 25.0
Other real and personal property --- Varies with local conditions
*These types of property are provided for general guidance only and may not represent jurisdictional requirements.
**CODs lower than 5.0 may indicate sales chasing or non-representative samples.

In its work with various types of assessing units, the Department's staff has also found that the more rural areas, where there are relatively few sales and properties are more heterogeneous, pose greater difficulty in establishing accurate assessments and market values. Thus, in measuring assessment uniformity, staff has taken the view that somewhat higher COD levels would be acceptable in areas with rural characteristics as contrasted with urban and suburban areas. The guidelines established are summarized in Table 2 and were applied in determining the number of assessing units achieving equity for purposes of this report.

Table 2. Department Guidelines for Assessment Uniformity
Population Density (persons/sq. mile) Coefficient of Dispersion, All Property
100 or less < 20
101 to 400 < 17
401 or more < 15

Coefficient of Dispersion Results

For the 2013 market survey, the median residential COD among the sampled assessing units was 17.57, and the median for all property classes combined was 19.67.3 In other words, half the sampled assessing units achieved greater uniformity than indicated by these median values, and half achieved less. The range in the all-property COD was 6.67 to 65.74. For the residential COD, the range among assessing units was 5.48 to 60.32.

The COD results presented herein are point estimates. If the estimation were replicated using an alternative data set, it is likely that somewhat different figures would be obtained due to sampling error. Gloudemans, an expert in the field, has proposed a confidence interval approach to recognize the problem of sampling error. His approach results in a range within which the COD estimate will fall with a known probability. However, the approach does not obviate the need for making point estimates of the COD.4

Table 3 summarizes the 2013 COD information according to type of assessing unit, as measured by population density, and the COD guidelines shown in Table 2. A total of 41 percent of the sampled assessing units had 2013 CODs that reflected uniform assessing practices for the entire roll. In the case of residential property only, 21 percent of the sampled units had uniform rolls.

Table 3. Summary of COD Values for Sampled Assessing Units, by Degree of Urbanization
(2013 Market Value Survey)
Population
Density
(Per Sq. Mi.)
No. of
Assessing
Units
Uniform COD Level Percent of Assessing Units
Achieving Uniform Level
Residential All Property Residential All Property
< 100 231 15 20 22% 37%
< 100 - < 400 106 12 17 25% 43%
> 400 96 10 15 17% 46%
TOTAL 433 -- -- 21% 41%
Note: Data for the individual municipalities within a Coordinated Assessment Program (CAP) are reported.

Table 4 shows the combined results for sampled and non-sampled assessing units. When the non-sampled units -- those for which a recent reassessment program was reviewed and verified -- are combined with sampled units achieving satisfactory uniformity, a total of 726 (73.9%) of the state's assessing units had uniform assessment rolls. This represents an increase of two assessing units over the number found to be equitable in the 2012 survey analysis.

Table 4. Assessment Uniformity, Sampled and Non-Sampled Assessing Units*
(2013 Market Value Survey)
Total Number with Uniformity
Residential All Property
Sampled 433 93 176
Non-Sampled 550 550 550
Total 983 643 (65.4%) 726 (73.9%)
*Data for the individual municipalities within a Coordinated Assessment Program (CAP) are reported.

The geographic distribution of equitable assessing is shown in Figure 1. It can be observed that, in many parts of New York, all or nearly all the municipalities in a county have uniform rolls. On the other hand, clusters of rural assessing units with inequitable rolls are to be found in several areas, including the Northern Catskills, Southern Tier, Adirondacks and Southwestern New York. Some of the densely populated units in the Lower Hudson Valley and Long Island areas also have not achieved equitable rolls.

The geographic distribution of equitable assessing is shown in Figure 1

Another view of the equity of assessment rolls can be obtained from analysis of the level of assessment reflected on the roll, as contrasted with the degree of uniformity. Table 5 shows the distribution of 2013 equalization rates, which reflect the average percentage of market value used in assessing. The data indicate that over 71 percent of all assessing units now have assessments that are at least 75 percent of current market value. Although there are some assessing units with assessments that are well below market value but which are still uniform and equitable, experience has demonstrated that current market assessments are strongly correlated with equity, and the level of assessment findings thus support the Table 4 data on assessment uniformity.

The 89 assessing units having a level of assessment of 25% or less probably have decades-old rolls, i.e., no general reassessment program has likely been conducted in a great many years. The additional 48 having a level of assessment in the 25% to 50% range are also very outdated, even in areas of the state where real estate markets experienced rapid appreciation during certain periods.

Table 5. Level of Assessment, as Measured by 2013 State Equalization Rate

Level of Assessment

Number of Assessing Units*
0.00 - 10.00 56 (5.7%)
10.01 - 25.00 33 (3.4%)
25.01 - 50.00 48 (4.9%)
50.01 - 75.00 143 (14.6%)
75.01 - 100.00 667 (68.0%)
Greater than 100.00 34 (3.4%)
Total 981 (100%)
*Data for the individual municipalities within a Coordinated Assessment Program (CAP) are reported.
Data for special assessing units of Nassau County and New York City are excluded.

Figure 2 shows the trend in assessment uniformity among New York assessing units since 1980. In the 1980s, only about 10 percent of all assessing units had acceptable uniformity. Dramatic improvement occurred in the early 1990s however, and by 2004, over 80 percent of all assessing units were assessing uniformly. There were modest declines thereafter, in a context of turbulent real estate market conditions in some areas, which persist to the present day. It is believed that real estate market volatility is the primary reason for the noticeable lack of progress in achievement of greater uniformity over the past decade.

 Figure 2 shows the trend in assessment uniformity among New York assessing units since 1980

Price-Related Differential Results

As indicated earlier, another important summary statistic for assessment performance is the price-related differential (PRD). The PRD is calculated by dividing the simple mean assessment ratio by the weighted mean ratio, where the weighted mean is the sum of assessed values divided by the sum of sales prices and/or appraised values. The simple mean counts the ratio of each property equally, regardless of the property's value, whereas the weighted mean counts each ratio differently, weighting ratios of higher-value properties more heavily, in proportion to their dollar value. If no assessment bias exists, the two mean ratios should be equal, producing an index of 1.00. Where there is evidence of a bias in favor of under-assessing the higher-value properties relative to the lower-value ones, the simple mean ratio will be higher than the value-weighted mean ratio, producing an index greater than 1.00 (regressivity). The reverse will be true in cases of over-assessment of high-value properties relative to those of low-value (progressivity). IAAO suggests that the PRD have a value between .98 and 1.03 for neutral assessing.

Table 6 summarizes the extent of value-related equity as measured by the PRD for the sampled assessing units. About 42 percent of the sampled assessing units assessed residential property in a neutral manner, i.e., they generally did not tend to favor either high- or low-value properties. However, 57 percent tended to over-assess low-value homes relative to high-value homes, while only two units tended to do the reverse. These results are similar to those found in the prior market value survey.

Table 6. Value-Related Bias in Assessing, Sampled Assessing Units,
2013 Market Value Survey
Residential Class          All Property Classes
Price-Related
Differential
Number of
Assessing Units
Percent Number of
Assessing Units
Percent
Progressive 2 1% 73 17%
Neutral 184 42% 168 39%
Regressive 247 57% 192 44%
TOTAL 433 100% 433 100%
Note: Data for the individual municipalities within a Coordinated Assessment Program (CAP) are reported.

When all property classes are combined, the situation changes significantly. Table 6 shows that 17 percent of the sampled assessing units use assessing practices that are biased toward over-assessment of higher-value properties, indicating over-assessment of some non-residential classes (generally industrial, commercial and utility property). About 44 percent demonstrate the opposite behavior, regressive assessing, meaning that they tend to overvalue the lower-priced properties (generally vacant land or low-value residential parcels). The remaining 39 percent of the assessing units assess in a neutral manner with respect to value when all property classes are considered together.

Recent Reassessment Activity Subsequent to the 2013 Market Survey

Approximately 6 percent (26) of the 433 assessing units for which CODs and PRDs were calculated are conducting or plan to conduct a reassessment on a roll subsequent to the one that was utilized in the 2013 survey.  For these assessing units, the COD and PRD estimates contained in this report are a measure of past assessment equity only; the level of uniformity on the newer roll may well be significantly improved from the level on the roll evaluated. Of the 550 assessing units for which recent reassessment projects were reviewed for the 2013 market survey, 353 have scheduled a subsequent reassessment project.  With the advent of the Cyclical Reassessment Aid Program, and its requirement to complete a reappraisal at least once in every four years, these subsequent reassessments will be performed on a roll completed between 2014 and 2017, depending on the schedule of the municipality's planned participation.  Thus, nearly 65 percent of those that have already taken steps to maintain equitable assessing practices are projecting that they will reassess again in the next one to four years.

Figure 3 indicates the number of municipalities that have submitted plans for the Cyclical Reassessment Aid Program.  As of December 9, 2013, some 308 assessing units are committed to updating their assessment rolls on a regular basis, thus ensuring that equity will be maintained as market conditions change. It should be stressed that Figure 3 signifies a snapshot in time and that the data it represents are subject to continuous change.

 Figure 3 indicates the number of municipalities that have submitted plans for the Cyclical Reassessment Aid Program.

Conclusions and Recommendations

Assessment equity in New York improved dramatically over the past thirty years, reached a peak in 2004 and, since then, has hovered around 70 to 80 percent of assessing units. Following the large gains in earlier years, there seems to have been a moderate decline in the second half of the past decade, and more than one quarter of assessing units currently have very outdated assessments. It appears that existing aid programs for quality assessment administration are no longer attracting significant numbers of new participants. Thus, it may be worthwhile for policymakers to consider additional tools to bring about greater assessment equity. Other states employ a number of such tools, and virtually all states have more requirements directed at maintenance of assessment uniformity than New York has.

Standard of Assessment

All states other than New York, New Jersey, and Pennsylvania require that a common level of assessment (most frequently, 100 percent of current market value) be applied in all assessing jurisdictions. This approach has many advantages, including, equitable tax treatment of like properties, greater taxpayer understanding of the basis of the tax, and both fair and cost-effective apportionment of taxes and state aid among local governments. It is also consistent with recommendations of the IAAO concerning the most appropriate state policies with respect to assessment standards.

Reassessment Cycle

In recognition of the fact that real estate markets are constantly changing, many states require that assessments be updated periodically. While annual updating is the ideal, actual practice usually involves cycles that generally range from two to six years, in recognition of the time that is generally needed to accumulate sufficient market data (especially in small, rural assessing units), the amount of work required, including data verification, valuation, outreach to property owners, and assessment appeals, and the cost of the entire process. While New York's aid incentive program for reassessment now provides for a four-year reassessment cycle, it must be understood that the program is not mandatory.

Direct Equalization

States using this policy tool directly change assessments on the rolls prepared by local governments, in contrast to the indirect equalization used in states such as New York, where the portion of a county or school tax levy to be borne by a particular municipality is adjusted in recognition of that municipality's overall level of assessment. The main advantage of direct equalization is that it changes the assessment, so that property owners can then compare their assessments to local market prices. A limitation, however, is that the state-level adjustments may well be quite crude factors that are applied to an entire property class or assessment jurisdiction, especially when adjustments must be made for many assessing units.

Withholding of State Payments

Rather than attempt to make rough adjustments to assessments in order to bring them to the correct market level, many states instead use monetary sanctions. These generally involve the withholding of monies that local governments would ordinarily receive from the state on an annual basis, such as state aid payments, a local share of certain state taxes, or fees that are collected by the state and sent to local governments, in whole or in part. The payments are restored only when the assessments in question are brought up to state standards.

Ordering a Reassessment

Some states simply direct a local government with faulty assessments to conduct a reassessment and, if the locality in question fails to comply with the directive, they hire a contractor to do the work and charge the local government for the cost of the project. Massachusetts is a nearby state where such action is the final remedy for outdated or inequitable assessments.


1This analysis is required by Section 1200 of the Real Property Tax Law.

2This statistic is sometimes referred to as the Index of Regressivity.

3The special assessing units of New York City and Nassau County are excluded in calculating the median COD because they use a classified assessing system.

4See Robert J. Gloudemans, "Confidence Intervals for the Coefficient of Dispersion: Limitations and Solutions," Assessment Journal, Nov./Dec. 2001.

Updated: