STANDARD ON RATIO STUDIES—2013
24
is essential. The general goal is to identify areas in which
the assessment levels are too low or lack uniformity and
property groups for which additional reappraisal work
may be required. In such cases, it also is highly desirable
to stratify on the basis of more than one characteristic
simultaneously.
Stratification can help identify differences in level of ap-
praisal between property groups. In large jurisdictions,
stratification by market areas is generally more appropriate
for residential properties, while stratification of com-
mercial properties by either geographic area or property
subtypes (e.g., office, retail, and warehouse/industrial)
can be more effective.
3.5 Matching Appraisal Data and Market Data
The physical and legal characteristics of each property
used in the ratio study must be the same when appraised
for tax purposes and when sold. This implies two essential
steps. First, the property description for the sold parcel
must match the appraised parcel. If a parcel is split between
the appraisal date and the sale date, a sale of any of its
parts should not be used in the ratio study.
Second, the property rights transferred, permitted use,
and physical characteristics of the property on the date of
assessment must be the same as those on the date of sale.
Properties with significant differences in these factors
should be excluded from the ratio study.
When statutory constraints are imposed on appraisal
methods, the resulting assessment may be less than market
value. In such cases a sales ratio study may not provide
useful performance information. Constraints typically
apply to land that qualifies for agricultural-use value,
subsidized housing, mineral land, and timberland.
Sales may include property of a type other than the type for
which the ratio study analyses is intended. However, sales
including more than minimal values of secondary categories
are unlikely to be representative, even with adjustment.
For example, a property that is predominantly commercial
may include residential components. This sale can be
included as representative of the commercial category. In
this case, the numerator in the ratio calculation would be
the total appraised value including the value of both the
commercial and residential components.
In a second example, for a ratio study of vacant land, the
numerator in the ratio should reflect only the appraised
value of the land. The sale price should be adjusted for
the contributory value of the improvements or the sample
should be excluded from further analysis.
3.5.1 Stratification for Equalization Studies
Oversight agencies generally should define the strata
prior to acquiring and compiling data for the ratio study.
Predefined stratification is more transparent and enhances
cooperation between the oversight agency and the juris-
diction appraising the property subject to equalization. In
general, oversight agencies should not redefine the strata
once they have been defined for equalization purposes,
especially in the case of direct equalization. It is appropri-
ate, however, to collapse strata to compensate for otherwise
inadequate samples sizes. In addition, a reappraisal or
equalization order can be targeted for specific problem
areas that cause noncompliance at a broader level of ag-
gregation. If value stratification is necessary, predefined
strata may not be practical.
3.5.2 Stratification for Direct Equalization
Strata should be chosen consistent with operational re-
quirements for the required level of equalization. Statistical
issues in the determination of strata include the size of the
population and resulting strata and the likely variability
of the ratios in each stratum. Care must be taken not to
over-stratify, that is, to create strata that are too small to
achieve statistical reliability (see section 6, Sample Size”
in part 1 and Sherrill and Whorton [1991]). No conclu-
sion about stratum level or uniformity should be made
from stratum samples that are unreliably small (resulting
in unacceptably large margins of error). Ultimately, the
degree of stratification is determined largely by available
sales data, unless it is cost-effective and practical to add
sufficient independent appraisals. If sufficient sales or ap-
praisals are not available for a given stratum, it should be
combined with similar strata. When strata are combined,
provided there is no reason to suspect dissimilar ratios as
evidenced by different level or uniformity measures, such
combinations permit broader applicability of ratio study
results and prevent ratio study analysis from becoming too
focused on substrata with few sales or appraisals. When
jurisdiction or category wide equalization actions are re-
quired, reliability of component strata is not an issue.
3.5.3 Stratification for Indirect Equalization
Indirect equalization develops an estimate of full market
value, but assessed values of individual properties are
not altered. Such studies can use a substantially different
approach to stratification than ratio studies intended for
performance evaluation or direct equalization. The purpose
of stratification in this case is to minimize distortions due
to different assessment levels, which can vary by property
type, value range, geographic area, and other factors. If
stratification creates a more representative sample, equal-
ization decisions may be based on results from individual
stratum. . If the overall sample is representative of the
population then equalization decisions should be based
on overall sample results. A reasonable number of strata
with small samples and larger margins of error can increase
overall representativeness and may reduce the margin of
error for the overall jurisdiction-wide sample.