Residential Revaluation Report

Residential Revaluation Report 2010 Mass Appraisal of Region 8 for 2011 Property Taxes Prepared For Patricia Costello Thurston County Assessor

TABLE OF CONTENTS Page No. CERTIFICATE OF APPRAISAL...3 APPRAISAL TEAM...4 MASS APPRAISAL CONCLUSIONS...5 PREMISE OF THE APPRAISAL...6 Supporting Documents Used in the Mass Appraisal...6 CLIENT AND INTENDED USERS...6 ASSUMPTIONS AND LIMITING CONDITIONS...7 SPECIAL ASSUMPTIONS, LIMITING, AND HYPOTHETICAL CONDITIONS...7 JURISDICTIONAL EXCEPTION...7 PURPOSE AND INTENDED USE...8 TRUE AND FAIR VALUE...8 DATE OF APPRAISAL...8 PROPERTY RIGHTS APPRAISED...8 PERSONAL PROPERTY NOT INCLUDED IN THE APPRAISAL...8 MARKET AREA AND PROPERTIES APPRAISED...9 CITY AND NEIGHBORHOOD DESCRIPTION...9 ZONING...9 HIGHEST AND BEST USE...9 SCOPE OF THE APPRAISAL...10 REGION 8 MAP...11 1

NEIGHBORHOOD MAP...12 RESIDENTIAL VALUATION PROCESS...13 Land Model Specification...14 Land Model Calibration...14 Multiple Regression Analysis Assumptions...15 Validation of Region 8 Land Model...15 Building Cost Specification...19 Construction Cost Tables...19 Effective Age...20 Depreciation Rate Tables...20 Condition...21 Neighborhood Adjustment Model Specification...22 Neighborhood Calibration of Cost Model...22 Sales Ratio Validation of the Neighborhood Adjustment Process...23 Neighborhood Adjustment Model Validation...23 RECONCILIATION AND CONCLUSION...26 Summary of Inventory Statistics...26 APPENDIX...27 Neighborhood 09P1 Group1...27 Neighborhood 08N1 - Group 2...27 Neighborhood 08L1 Group 3...28 Neighborhood 08H1 Group 4...28 Neighborhood 09S1 Group 6...29 Overall Sales Ratio for Region 8...29 Multiple Regression Analysis Assumptions...30 2

CERTIFICATE OF APPRAISAL I certify that, to the best of my knowledge and belief: the statements of fact contained in this report are true and correct. the reported analyses, opinions, and conclusions are limited only by the reported assumptions and limiting conditions, and are my personal, impartial and unbiased professional analysis, opinions, and conclusions. I have no (or the specified) present or prospective interest in the property that is the subject of this report, and I have no (or the specified) personal interest with respect to the parties involved. I have no bias with respect to any property that is the subject of this report or to the parties involved with this assignment. my engagement in this assignment was not contingent upon developing or reporting predetermined results. my compensation for completing this assignment is not contingent upon the reporting of a predetermined value or direction in value that favors the cause of the client, the amount of the value opinion, the attainment of a stipulated result, or the occurrence of a subsequent event directly related to the intended use of this appraisal. my analyses, opinions, and conclusions were developed, and this report has been prepared, in conformity with the Uniform Standards of Professional Appraisal Practice. I have not personal inspected all of the property that is the subject of this report. Other appraisers involved in the review of property are listed on the following page. no one provided significant analytical assistance to the person(s) signing this certification. Appraiser # 013, Appraisal Analyst (signature on file) Date 3

APPRAISAL TEAM Often teams of appraisers complete one or more parts of a mass appraisal. Major contributors to this appraisal project include the following: Physical Inspection Team: Sales Validation: Sales Analysis and Model Building: Final Review: 028, Senior Appraiser 050, Senior Appraiser 030, Senior Appraiser 042, Senior Appraiser 029, Senior Appraiser 037, Senior Appraiser 007, Lead Appraiser 035, Appraiser Analyst 054, Appraiser Analyst 013, Appraiser Analyst 013, Appraiser Analyst 052, Chief Appraiser 4

MASS APPRAISAL CONCLUSIONS Appraisal Date: January 1, 2010 Area Name / Number: Region 8 and corresponding neighborhoods Physical Inspection: Last inspected in 2009 Summary of Neighborhood Adjustments, Sales Ratios, and Assessed Value Changes: Sales Improved Valuation Change Region 8 Summary Statistics Nbhd Region Group Land Factor Bldg Factor New Land Adj New Bldg Adj # sales Mean Ratio Median Ratio Wgt. Mean Ratio PRD COD Avg. $ Change Med. $ Change Avg. % Change Med. % Change 09P1 1 0.92 0.93 0.92 0.93 12 0.91 0.88 0.91 1.00 0.09 $14,325 $6,324 4.97% 2.82% 08N1 06N1 2 1.00 1.00 0.92 0.85 3 0.94 0.89 0.95 0.99 0.06 -$30,534 -$26,426-9.22% -9.02% 08L1 3 0.97 1.00 0.91 0.93 8 0.90 0.89 0.91 0.99 0.05 -$14,115 -$13,251-5.15% -5.27% 08H1 4 0.82 0.85 0.75 0.75 3 0.87 0.90 0.87 1.01 0.03 -$14,524 -$15,505-8.51% -8.53% 09S1 6 0.90 1.00 0.98 0.93 8 0.87 0.88 0.88 1.00 0.12 $3,105 $2,247 1.77% 0.71% Overall 34 0.90 0.89 0.90 1.00 0.08 -$8,349 -$9,322-3.23% -3.86% Sales used in Analysis: Sales used in the analysis are validated following the guidelines laid out in the Sales Verification Procedure. Multi-parcel and multi-building sales are generally excluded as not being representative of this market area. Mobile home and condominium sales are also excluded from the analysis and valuation of standard single family residential construction. Mobile home and condominium sales are analyzed separately for the purpose of appraising these property types. Number of Parcels in the Population: The population of residential vacant land and standard single family residences within Region 8 equals approximately 3,400 parcels. Conclusion and Recommendation: Since the values recommended in this report improve uniformity, assessment level, and equity, we recommend posting them for the 2011 Tax Roll. 5

PREMISE OF THE APPRAISAL Supporting Documents Used in the Mass Appraisal "A mass appraisal is the process of valuing a universe of properties as of a given date using standard methodology, employing common data, and allowing for statistical testing." 1 A mass appraisal for ad valorem taxes is a complicated process involving large amounts of data, gathered and analyzed by teams of appraisers. We do not intend this document to be a self-contained documentation of the mass appraisal but to summarize our methods, data, and to guide the reader to other documents or files, upon which we relied. These documents may include the following: Individual property records maintained in a computer database Sales ratios and other statistical studies Market studies Model building documents Real estate sales database. Previous studies and reports filed in our office. Assessor s manuals for data collection analysis. Revaluation and sales verification manuals Property Tax Advisory Publications by the Washington State Dept. of Revenue. Title 84 RCW Property Tax Laws (Washington State Law) WAC 458 (Washington Administrative Code) The Appraisal Standards Board of the Appraisal Foundation annually publishes the Uniform Standards of Professional Appraisal Practice (USPAP). These standards are written by appraisers to regulate their profession and are the minimum standards for the conduct of property appraisal in the United States. They cover real, personal, and business property. We rely upon these standards in the development and reporting of our assessed values. CLIENT AND INTENDED USERS This report was prepared for Patricia Costello, Thurston County Assessor. Other intended users include the County Board of Equalization and the State Board of Tax Appeals. 1 USPAP, Appraisal Standards Board of the Appraisal Foundation, p. 3 6

ASSUMPTIONS AND LIMITING CONDITIONS The Appraisal Report, of which this statement is a part, is expressly subject to the following conditions: This revaluation is a mass appraisal assignment resulting in conclusions of market value. No one should rely on this study for any purpose other than administration and distribution of ad valorem taxation. The opinion of value on any parcel may not be applicable for any use other than ad valorem taxation. That the maps and drawings in this report are included to assist the reader in visualizing the property; however, no responsibility is assumed as to their exactness. That the legal description as given is assumed correct. No survey or search of title of the property has been made for this report, and no responsibility for legal matters is assumed. The report assumes good merchantable title and any liens or encumbrances that may exist have been disregarded. The opinions and values shown in the report apply to the subject parcels only. The assessors made no attempt to relate the conclusions of this report to any other revaluations, past, present, or future. The assumptions governing the use of multiple linear regression analysis have been met unless otherwise stated. Unless otherwise stated in this report, the existence of hazardous substances, including without limitation asbestos, polychlorinated biphenyl, petroleum leakage, or agricultural chemicals, which may or may not be present on the property, or other environmental conditions, were not called to the attention of nor did the appraiser become aware of such during the appraiser's inspection. The appraiser has no knowledge of the existence of such materials on or in the property unless otherwise stated. The appraiser, however, is not qualified to test such substances or conditions. If the presence of such substances, such as asbestos, urea formaldehyde foam insulation, or other hazardous substances or environmental conditions, may affect the value of the property, the value estimates is predicated on the assumption that there is no such condition on or in the property or in such proximity thereto that it would cause a loss in value. No responsibility is assumed for any such conditions, not for any expertise or engineering knowledge required to discover them. SPECIAL ASSUMPTIONS, LIMITING, AND HYPOTHETICAL CONDITIONS We assume that none of the subject land is contaminated or that any contamination would affect the value except as shown in individual property records or otherwise stated. Because of budget restraints, we have not inspected all comparable sales. We have inspected the interiors of only a small percentage of the properties. JURISDICTIONAL EXCEPTION Washington exempts all or a portion of the market value on specific types of property including open space, agricultural, forest, home improvement, and some low-income housing. 7

PURPOSE AND INTENDED USE The intended use of this appraisal is for administration of ad valorem taxation. After certification by the Assessor, these values will be used as the basis for assessment of real estate taxes payable in 2011. We do not intend the values to be used for or relied upon for any other purpose. This report serves as a record of the revaluation which is subject to review and change by the County Board of Equalization, the Washington State Board of Tax Appeals, and the courts. TRUE AND FAIR VALUE The basis of all assessments is the true and fair value of property. True and fair value means market value (Spokane etc. R. Company v. Spokane County, 75 Wash. 72 (1913): Mason County, 62 Wn. 2d (1963); AGO 57-58, No. 1/8/57; AGO 65-66, No. 65, 12/31/65) The true and fair value of a property in money for property tax valuation purposes is its "market value" or amount of money a buyer willing but not obligated to buy would pay for it to a seller willing but not obligated to sell. In arriving at a determination of such value, the assessing officer can consider only those factors which can within reason be said to affect the price in negotiations between a willing purchaser and a willing seller, and he must consider all of such factors. (AGO 65,66, No. 65, 12/31/65) Properties are appraised as of January 1, 2010. This report was completed October 26, 2010. DATE OF APPRAISAL PROPERTY RIGHTS APPRAISED This appraisal is of the fee simple interest in the real property. The fee simple estate is the absolute ownership unencumbered by any other interest or estate, subject only to the limitations imposed by the governmental powers of taxation, eminent domain, police power, and escheat. 2 PERSONAL PROPERTY NOT INCLUDED IN THE APPRAISAL No personal property was included in the value. Fixtures are generally accepted as real property. Business value is intangible personal property and it is not appraised. 2 The Dictionary of Real Estate Appraisal. 3d ed. Appraisal Institute, p.140 8

MARKET AREA AND PROPERTIES APPRAISED The subject of this mass appraisal is the residential property (excluding mobile homes and condominiums) contained in the market area designated as Region 8. Regions are generally influenced by the same broad market trends. This area includes approximately 3,400 properties and is shown on the map on page 11 of this report. Our property records contain photographs, sketches, legal descriptions and other characteristics of land and buildings on each property. CITY AND NEIGHBORHOOD DESCRIPTION Region 8 includes northwest Thurston County, south of HWY 8 and Hwy 101, west of Blacklake Blvd and north of Rochester. This region is further broken into 6 residential neighborhoods that are designed to reflect similar land and building characteristics and neighborhood amenities. The neighborhoods and their codes are shown on page 12. They are all considered to be stable in terms of the life cycle of a neighborhood. ZONING Thurston County exercises jurisdiction over land use and community planning. The regulations for use and development can be found in its ordinances. We show property zoning as a land characteristic on our digital maps. HIGHEST AND BEST USE True and fair value -- Highest and best use. Highest and best use is the most profitable, likely use to which a property can be put. It is the use which will yield the highest return on the owner's investment. Any reasonable use to which the property may be put may be taken into consideration and if it is peculiarly adapted to some particular use, that fact may be taken into consideration. Uses that are within the realm of possibility, but not reasonably probable of occurrence, shall not be considered in valuing property at its highest and best use. [WAC 458-07-30 (3)] The highest and best use concept is based upon traditional appraisal theory and reflects the attitudes of typical buyers and sellers. The market sets the highest and best use based on the theory of wealth maximization for the owner with consideration given to community goals. To estimate highest and best use, four elements are considered: 1. Possible use. What uses of the site in question are physically possible? 2. Permissible legal use. What uses of the site are permitted by zoning and deed restrictions? 3. Feasible use. Which possible and permissible uses will produce a net return to the owner of the site? 4. Highest and best use. Among the feasible uses, the use which will produce the highest net return or the highest present worth? 9

The highest and best use of the land or site if vacant and available for use may be different from the highest and best use of the improved property. This is true when the improvement is not an appropriate use, but it contributes to the total property value. For the purpose of this appraisal the highest and best use of all vacant and improved property is considered to be single family residential or related to a single family residential use. SCOPE OF THE APPRAISAL Under state law, the assessor receives a copy of each Real Estate Excise Tax Affidavit and is therefore privy to the sale price, date, and description of all real estate sales. Our staff compiles and verifies this data into our sales database as explained in our sales verification procedure. Thurston County is on a six-year revaluation cycle. Every property is revalued annually. At least once each six years, each property is inspected and its data refreshed. The assessor collects property characteristic data as discussed in our Residential Data Standards Manual. Other than new construction, the last physical inspection of residential property in Region 8 was during the first half of 2009. A region map is included on next page followed by a map of the neighborhoods within the region. The appraisal considers the cost approaches to value with sales used to calibrate the model to a specific neighborhood. Neighborhood adjustments are widely used to adjust for time and location and are a normal and standard part of the cost approach to value. The Marshall Swift cost manual provides what they call current cost multipliers and local area multipliers to adjust for time and location. Because this is a national valuation service, we fine tune their cost rates even further to consider differences between neighborhoods and local market trends. Whether we make these adjustments to the raw land and cost rates or to the preliminary cost values, does not impact the mathematical calculation and does not affect the final result. It is more convenient to apply the time and location adjustments to the preliminary cost values, because it makes the statistical updating of values from year to year much easier. A market model (strict sales approach) has not been developed for 2010 due to time and budget limitations. The use of an income approach was not considered to be applicable because homes in this area are not typically purchased for their income potential. The flow chart on page 13 describes the land model developed as part of the mass appraisal process and how it is used in the sales adjusted cost approach. The model is discussed in more detail starting on page 14. 10

REGION 8 MAP 11

NEIGHBORHOOD MAP 12

RESIDENTIAL VALUATION PROCESS Cost Approach Land Model Base Land Rates (applied within PI area based on Market Area and lot size) Adjustment Rates (applied within PI area based on land characteristics) Cost Land Value Bldg Model Cost Rates (applied countywide to building characteristics, updated annually) Cost Building Value Depreciation Rates (rcnld) (applied countywide based on condition and effective age, updated as needed) Statistical Update of Update Model Cost Land Value Final Land Value Cost Approach by Nbhd (all areas updated annually) Cost Building Value Final Building Value (all areas updated annually) Sales Approach Sales Model Final Land Value from Final Land Value Statistically Updated Cost Approach (updated annually) Residual Bldg Value Final Bldg Residual Value (updated annually) 13

COST APPROACH Land Model Specification A multiplicative model format is used in the development of base land rates and adjustment rates. Land Model Format: LV = b 0 X SQFT b1 X LINVIEW b2 X b 3 LI3 X b 4 LI4 X b 5 LI5 X... Where: Continuous Variables = SQFT, LINVIEW Binary Variables LI3, LI4, LI5... = Land Influences (i.e. region, view, wetlands, etc.) b 0, b 1, b 2, b 3, b 4, b 5... = Regression Coefficients Land Model Calibration Multiplicative model calibrated using log-linear MRA Logarithms are used to convert a multiplicative equation to a linear form. Standard Multiplicative form: SP/SQFT = a * SQFT b * c NBHD *... Log Linear form: LN(SP/SQFT) = LN(a) + (b * LN(SQFT)) + (LN(c) * NBHD) +... Log Linear form has the same form as a standard linear equation: Linear equation: Y = a + (b * X) + (c * Z) We can then calibrate the Log-Linear form using standard multiple regression analysis. The calibrated model is then converted back to its Standard Multiplicative form by applying the antilog function. EXP[LN(SP/SQFT)] = EXP[LN(a) + (b * LN(SQFT))] Region 8 Land Model see Region 8 work files for model coefficients and other output. 14

Multiple Regression Analysis Assumptions Multiple regression analysis is based on several assumptions regarding the data going into the model and the output from the calibration process. These assumptions are validated to determine the accuracy of the model and identify any limitations that may exist. A detailed discussion of the MRA assumptions is included in the Appendix. Validation of Region 8 Land Model Normal Distribution of the Residual Errors Total number of sales = 899 (from 1/1/06 12/31/08 trended to 1/1/09) Region 8 sales = 23 The residual errors are for the most part normally distributed. While the frequency distribution illustrates output from the square foot land model, similar results were obtained for the acreage model. 15

Constant Variance of the Residual Errors The residual errors are for the most part are distributed evenly along the range of values. Similar results were obtained for the acreage model. Comparison of Predicted and Actual Sale Price per Sq. Ft. The values predicted by the model accurately reflects actual trended sale prices. Similar results were obtained for the acreage model. 16

Region 8 Square Foot Rate Table Land Flag Square Feet Square Foot Value Base Rate Rate Group S.F. Adj Group Size Adj. Ratio S.F. Adj Factor 1100 2,000 $18.92 $3.32 3707 707 0.092 5.697 1100 2,500 $16.08 $3.32 3707 707 0.115 4.842 1100 3,000 $14.08 $3.32 3707 707 0.138 4.240 1100 3,500 $12.59 $3.32 3707 707 0.161 3.789 1100 4,000 $11.42 $3.32 3707 707 0.184 3.438 1100 4,500 $10.48 $3.32 3707 707 0.207 3.155 1100 5,000 $9.71 $3.32 3707 707 0.230 2.922 1100 5,500 $9.05 $3.32 3707 707 0.253 2.726 1100 6,000 $8.50 $3.32 3707 707 0.275 2.559 1100 6,500 $8.02 $3.32 3707 707 0.298 2.414 1100 7,000 $7.60 $3.32 3707 707 0.321 2.287 1100 7,500 $7.22 $3.32 3707 707 0.344 2.175 1100 8,000 $6.89 $3.32 3707 707 0.367 2.075 1100 9,000 $6.32 $3.32 3707 707 0.413 1.904 1100 10,000 $5.86 $3.32 3707 707 0.459 1.763 1100 12,000 $5.13 $3.32 3707 707 0.551 1.544 1100 14,000 $4.58 $3.32 3707 707 0.643 1.380 1100 16,000 $4.16 $3.32 3707 707 0.735 1.252 1100 18,000 $3.82 $3.32 3707 707 0.826 1.149 base size> 1100 21,780 $3.32 $3.32 3707 707 1.000 1.000 1100 24,000 $3.09 $3.32 3707 707 1.102 0.932 1100 27,000 $2.84 $3.32 3707 707 1.240 0.855 1100 30,000 $2.63 $3.32 3707 707 1.377 0.792 1100 35,000 $2.35 $3.32 3707 707 1.607 0.708 1100 40,000 $2.13 $3.32 3707 707 1.837 0.642 1100 43,560 $2.00 $3.32 3707 707 2.000 0.603 1100 50,000 $1.81 $3.32 3707 707 2.296 0.546 1100 55,000 $1.69 $3.32 3707 707 2.525 0.509 1100 65,000 $1.50 $3.32 3707 707 2.984 0.451 1100 75,000 $1.35 $3.32 3707 707 3.444 0.406 1100 87,120 $1.21 $3.32 3707 707 4.000 0.364 Land Influence Adjustments Easement Fair Nbhd Good Nbhd Shape Steep Restricted No Road Dirt Road Multiplier: 0.50-0.95 0.80 1.25 0.85 0.30-0.85 0.50-0.85 0.50-0.85 0.85-0.90 0.90-0.95 Gravel Road Golf Course Golf Com. Avg Lake <Avg Lake 20% Wet 40% Wet 60% Wet 80% Wet 100% Wet Multiplier: 1.15 1.05 2.50 1.50 0.90 0.70 0.55 0.45 0.30 Economic Prelim Plat High Traffic Med Traffic Unbldable Timber Multiplier: 0.75-0.95 2.00-4.00 0.85 0.95 0.30 1.20 No View Limited View Good View VGd View Exc View Multiplier: 1.00 1.10 1.25 1.45 1.60 17

Region 8 Acre Rate Table Land Flag Acres Per Acre Value Acre Rate Rate Group Acre Adj Group Size Adj. Ratio Acre Adj Factor 9150 0.75 $107,741 $27,022 2850 850 0.150 3.987 9150 1.00 $87,356 $27,022 2850 850 0.200 3.233 9150 1.25 $74,241 $27,022 2850 850 0.250 2.747 9150 1.50 $65,000 $27,022 2850 850 0.300 2.405 9150 1.75 $58,091 $27,022 2850 850 0.350 2.150 9150 2.00 $52,702 $27,022 2850 850 0.400 1.950 9150 2.25 $48,366 $27,022 2850 850 0.450 1.790 9150 2.50 $44,790 $27,022 2850 850 0.500 1.658 9150 3.00 $39,215 $27,022 2850 850 0.600 1.451 9150 3.50 $35,047 $27,022 2850 850 0.700 1.297 9150 4.00 $31,796 $27,022 2850 850 0.800 1.177 base size> 9150 5.00 $27,022 $27,022 2850 850 1.000 1.000 9150 6.00 $23,659 $27,022 2850 850 1.200 0.876 9150 7.00 $21,144 $27,022 2850 850 1.400 0.782 9150 8.50 $18,353 $27,022 2850 850 1.700 0.679 9150 10.00 $16,302 $27,022 2850 850 2.000 0.603 9150 12.00 $14,273 $27,022 2850 850 2.400 0.528 9150 14.00 $12,756 $27,022 2850 850 2.800 0.472 9150 16.00 $11,573 $27,022 2850 850 3.200 0.428 9150 20.00 $10,350 $27,022 2850 850 4.000 0.383 9150 25.00 $9,585 $27,022 2850 850 5.000 0.355 9150 30.00 $8,775 $27,022 2850 850 6.000 0.325 9150 40.00 $7,740 $27,022 2850 850 8.000 0.286 9150 50.00 $6,570 $27,022 2850 850 10.000 0.243 9150 75.00 $4,860 $27,022 2850 850 15.000 0.180 9150 100.00 $3,960 $27,022 2850 850 20.000 0.147 9150 200.00 $2,385 $27,022 2850 850 40.000 0.088 Land Influence Adjustments Easement Fair Nbhd Good Nbhd Shape Steep Restricted No Road Dirt Road Multiplier: 0.50-0.95 0.80 1.25 0.85 0.30-0.85 0.50-0.85 0.50-0.85 0.85-0.90 0.90-0.95 Gravel Road Golf Course Golf Com. Avg Lake <Avg Lake 20% Wet 40% Wet 60% Wet 80% Wet 100% Wet Multiplier: 1.15 1.05 2.50 1.50 0.90 0.70 0.55 0.45 0.30 Variable: Economic Prelim Plat High Traffic Med Traffic Unbldable Timber Multiplier: 0.75-0.95 2.00-4.00 0.85 0.95 0.30 1.20 Variable: No View Limited View Good View VGd View Exc View Multiplier: 1.00 1.10 1.25 1.45 1.60 18

Building Cost Specification Model Format for RCNLD: BV = [(c 1 X Q 1 ) + (c 2 X Q 2 ) + (c 3 X Q 3 ) +... ] X Pct. Good Where: Building Components = Q 1, Q 2, Q 3... Costs per unit = c 1, c 2, c 3... Construction Cost Tables Marshall Swift cost rates, adjusted to the current year and local area, are used to determine the replacement cost of each residential improvement. Adjustments can also be made for various structure types and for other building components based on locally advertised building costs. The complete set of rate tables is too lengthy to include here. However, an example of the rates for the main floor level of a residence by quality grade is shown below. The complete set of rate tables is stored within the Sigma CAMA System. 19

Effective Age The effective age of a building is largely based on its overall condition. It is a measure of how old a building looks and not how old it actually is. As a result, any type of maintenance, repair, remodel, or renovation will tend to reduce the effective age. The more extensive the maintenance or repair work the more the effective age is reduced. This concept suggests that a very old building can be brought back to almost new condition, thereby reducing the effective age to a level that is typical of much newer construction. Depreciation Rate Tables Periodically, the depreciation tables are calibrated using residential sales representing all years of construction. The most recent estimates of the land values are subtracted from the sale prices to determine the residual building values. These values are compared to the replacement cost new to arrive at an estimate of the percent good, which is then correlated with the effective age of the building to produce a set of depreciation tables. An example table for a stick built house is show below. The depreciation rates are expressed as a percent good. DEPRECIATION TABLE 1 (2011DEP) 20

The graph below shows the relationship between the percent good, actual age, and effective age. 120 Percent Good 100 80 60 40 20 0 0 10 20 30 40 50 60 70 80 Eff. Age 0 10 42 87 103 114 123 130 140 Actual Age Condition Because many properties are in better or worse condition than what is typical for their age, we need a method to adjust the depreciation rate accordingly. There are two ways to accomplish this. One is to adjust the effective age and the other is to adjust the condition rating to raise or lower the amount of depreciation that is applied. Adjusting the effective age would involve a fairly complex set of instructions and calculations for different situations that may be encountered. Minor remodels, major renovations, and building additions would require different adjustment techniques. Even with these procedures in place, there would be substantial appraiser judgment involved that would open the door for inconsistencies in the way effective age is determined and depreciation is applied. A better method is to establish guidelines for determining the condition rating to apply to each property. In general, if an improvement to a parcel of land is typical for its age and has received average maintenance, it would be considered to be in average condition. If the improvement has had less than average maintenance, it will be in less than average condition. If the improvement has received better than average maintenance, it will be in better than average condition. The following graph shows the effect that the condition rating has on the percent good curve. It summarizes the relationship between effective age, building condition, and the rate of depreciation. 21

120 100 80 60 40 20 0 Percent Good by Condtion Rating 0 10 20 30 40 50 60 70 80 Effective Age VPr % Good Pr % Good Fr % Good Av % Good Gd % Good VGd % Good Exc % Good Neighborhood Adjustment Model Specification The equation for the neighborhood adjustment has an additive model format but without the constant term. V = b 1 (LV) + b 2 (BV) Where: b 1 and b 2 are based on a combination of regression analysis and appraiser judgment Neighborhood Calibration of Cost Model Initially regression coefficients are developed to apply to both land (b 1 ) and building (b 2 ) values within each neighborhood. A preliminary adjustment to the neighborhood land values is determined first by considering only available vacant land sales within the region. After making the initial adjustment to the land value, the coefficient for the building value (rcnld) can be determined. This again produces a preliminary adjustment or starting point for determining the final neighborhood building trend. Next, each neighborhood within the region is analyzed to consider its unique characteristics, amenities, and market conditions. This final adjustment to the neighborhood land and building values is largely based on the appraiser s analysis of individual sales ratios guided by the region wide sales analysis. An iterative process of adjusting the initial coefficients is applied to each neighborhood to reach the desired level of assessment, PRD, and COD. As an example, final adjustments for neighborhood 09P1 are shown below. Final Neighborhood adjustments for 09P1: o o b 1 = 0.92 land value adjustment b 2 = 0.93 building value adjustment Final Ratios for 09P1: 22

Mean 0.91 Median 0.88 Weighted Mean 0.91 Price Related Differential 1.00 Coefficient of Dispersion 0.09 The sales ratio analysis of each neighborhood in Region 8 is included in the appendix along with the list of the sales that were used in the analysis. Sales Ratio Validation of the Neighborhood Adjustment Process Neighborhood trends were calibrated using 34 sales that took place between 1/1/09 to 3/31/10 trended to 1/1/10. Region 08 Ratio Statistics (New Value / Trended Sale Price) Mean 0.90 Median 0.89 Weighted Mean 0.90 Price Related Differential 1.00 Coefficient of Dispersion 0.08 Neighborhood Adjustment Model Validation Assessment Uniformity by Neighborhood 23

Assessment Uniformity by Quality Grade Assessment Uniformity by Building Style 24

Assessment Uniformity by Year Built Assessment Uniformity by Square Feet of Living Area 25

RECONCILIATION AND CONCLUSION Considering the quantity and quality of data and the reliability of the various models as shown in the performance tests above, we have concluded that the Sales Adjusted Cost Approach produces an accurate estimate of market value. Summary of Inventory Statistics Region 8 Inventory Statistics Stats Final Value Chgamt Chgperc (%) Chglnd (%) Chgimp (%) 08H1 Mean 162980-60780 -41.08-7.98-14.61 Median 191650-32175 -10.47-8.53-10.17 08L1 Mean 228442-37125 -27.72-6.11-8.31 Median 250475-18050 -5.58-6.18-4.86 08N1 Mean 196356-60417 -41.27-8.05-16.54 Median 222451-41375 -11.70-8.00-12.92 09P1 Mean 221551-3658 -19.17 0.00 3.83 Median 240250 5800 2.24 0.00 3.80 09S1 Mean 211596-31537 -32.40 9.79-11.11 Median 201800-5400 -1.85 8.89-5.12 Total Mean 202660-53068 -32.89-1.60-8.34 Median 225050-20450 -8.75-6.16-5.29 26

APPENDIX Neighborhood 09P1 Group1 Group Parcel ID Nbhd Acres Style SFLA Sale Date Sale Price Trnd SP 01 42380006400 09P1.590 RN 2731 07/01/2009 $550000 $538621 44010011200 09P1.400 RN 2257 01/17/2009 $320000 $306759 58150010400 09P1.230 RN 1336 09/18/2009 $222000 $218938 58150011300 09P1.230 RN 1397 10/01/2009 $203000 $200900 58150011800 09P1.240 RN 1002 06/04/2009 $189000 $184438 31330008200 09P1 1.25 SL 2043 03/24/2009 $257000 $248138 58150009800 09P1.370 RN 1814 07/14/2009 $259000 $253641 82860000901 09P1.440 TS 2614 01/09/2009 $350000 $335517 82860001000 09P1 1.27 TS 2604 08/27/2009 $360000 $353793 44010010000 09P1.610 RN 1504 05/01/2009 $240000 $233379 44010010200 09P1.890 OS 1602 11/12/2009 $275000 $273103 58150009000 09P1.230 RN 1184 08/14/2009 $222500 $218664 Sales Ratios for Neighborhood Group 1 (Based on New Value/Trended Sale Price) Mean 0.91 Median 0.88 Weighted Mean 0.91 Price Related Differential 1.00 Coefficient of Dispersion 0.09 Neighborhood 08N1 - Group 2 Group Parcel ID Nbhd Acres Style SFLA Sale Date Sale Price Trnd SP 02 09500009001 08N1 2.00 TS 2160 09/30/2009 $282600 $278702 13722320200 08N1 5.15 RN 2112 11/18/2009 $389900 $387211 13701120404 08N1 1.98 SE 1902 06/22/2009 $290000 $283000 Sales Ratios for Neighborhood Group 2 (Based on New Value/Trended Sale Price) Mean 0.94 Median 0.89 Weighted Mean 0.95 Price Related Differential 0.99 Coefficient of Dispersion 0.06 27

Neighborhood 08L1 Group 3 Group Parcel ID Nbhd Acres Style SFLA Sale Date Sale Price Trnd SP 03 13723320801 08L1 5.14.00 08/12/2009 $150000 $142500 13726410700 08L1 3.78 OS 1536 09/11/2009 $272500 $268741 13723120000 08L1 1.11 CA 1152 12/11/2009 $175160 $174556 13727445200 08L1 5.00 TS 2765 09/17/2009 $425000 $419138 13723120503 08L1 2.50 RN 1852 12/21/2009 $325000 $323879 13724220701 08L1 2.50 RN 1678 04/03/2009 $275000 $266466 13726120300 08L1 4.77 RN 1796 07/24/2009 $375000 $367241 13726120610 08L1 5.00 RN 1508 02/17/2010 $361500 $361500 Sales Ratios for Neighborhood Group 3 (Based on New Value/Trended Sale Price) Mean 0.90 Median 0.89 Weighted Mean 0.91 Price Related Differential 1.00 Coefficient of Dispersion 0.05 Neighborhood 08H1 Group 4 Group Parcel ID Nbhd Acres Style SFLA Sale Date Sale Price Trnd SP 04 09650005200 08H1 6.75.00 11/13/2009 $119000 $116620 35850001000 08H1 1.11 RN 1544 02/09/2010 $219900 $219900 35850001600 08H1.850 RN 1444 05/11/2009 $211335 $205505 Sales Ratios for Neighborhood Group 4 (Based on New Value/Trended Sale Price) Mean 0.87 Median 0.90 Weighted Mean 0.87 Price Related Differential 1.01 Coefficient of Dispersion 0.03 28

Neighborhood 09S1 Group 6 Group Parcel ID Nbhd Acres Style SFLA Sale Date Sale Price Trnd SP 06 13824210101 09S1 4.48.00 04/15/2009 $192000 174720 13824430101 09S1 4.91 RN 2357 02/03/2009 $421000 $405031 13824410604 09S1 2.74 RN 1704 06/01/2009 $300000 $292759 12819330100 09S1 9.00 RN 1712 09/08/2009 $462000 $455628 13814131101 09S1 6.30 SL 1480 05/08/2009 $339000 $329648 13823140104 09S1 5.03 TS 2573 11/04/2009 $350000 $347586 13823140300 09S1 1.24 RN 1632 10/21/2009 $305000 $301845 13823441200 09S1 3.58 RN 2216 02/23/2009 $345000 $331914 Sales Ratios for Neighborhood Group 6 (Based on New Value/Trended Sale Price) Mean 0.87 Median 0.88 Weighted Mean 0.88 Price Related Differential 1.00 Coefficient of Dispersion 0.12 Overall Sales Ratio for Region 8 Mean 0.90 Median 0.89 Weighted Mean 0.90 Price Related Differential 1.00 Coefficient of Dispersion 0.08 29

Multiple Regression Analysis Assumptions Complete and Accurate Data: Data definitions and standards have been developed to ensure our data is as complete and accurate as possible. A procedure has been established to ensure sales are properly verified. Annual training is conducted to remind appraisers of the standards that have been developed. Representativeness: It is assumed that the sale sample adequately represents variables in the model. Violation of this assumption may affect the accuracy of the model in predicting the value of properties that are under-represented. For example, if there are no sales of Excellent view, the model would make no distinction from the typical Average view and an Excellent view. Using scalar or linearized variables in the model has mitigated this potential problem. Linearity: It is assumed that the marginal contribution of a variable is constant over the range of values for the variable. Each additional unit of size or quantity adds equally to the value. The assumption is violated when economies of scale or other non-linear relationships are present. Developing a multiplicative land model has helped to create linear relationships between the dependent variable and independent variables. For example, using the natural logarithm of the lot size (acres) addresses the decreasing marginal utility of adding additional units of land. See example below. Total Value 120000 100000 80000 60000 40000 20000 0 0 10 20 30 40 50 Acres LN(Value) 11.600 11.400 11.200 11.000 10.800 10.600 10.400 0.000 1.000 2.000 3.000 4.000 LN(Acres) Additivity: It is assumed that the marginal contribution of one independent variable is not affected by the changes in other variables. The assumption is violated when one impendent variable interacts with another. This assumption generally does not hold for land models o Land characteristics are often interactive. For example, the adjustment for view may be influenced by the size or topography of the land parcel. A multiplicative model helps to address this issue but converting the format to log-linear terms. No Correlation between Independent Variables: It is assumed that there is no correlation between independent variables. This assumption is addressed by reviewing the correlation matrix and by either eliminating one of the correlated variables or combining the highly correlated variables. 30

Normal Distribution of Residual Errors: Violation of this assumption affects the interpretation of the SEE, COV, and t-statistics. With large samples and proper screening of the sales, this assumption is typically not a problem. The assumption is verified by examining a histogram of residual errors. See example below. Histogram Dependent Variable: trndadjsp 250 200 Frequency 150 100 50 Mean = -3.13E-15 Std. Dev. = 0.995 N = 2,138 0-4 -2 0 2 4 6 Regression Standardized Residual Constant Variance of the Error Term (homoscedasticity): The residual errors should be consistent as prices increase. Violation of this assumption implies the residual errors are not evenly distributed (heteroscedasticity). As a result the model will chase high priced sales that may not be representative of the market. Sales have been properly screened to ensure accuracy of the data, and outliers have been removed to reduce the likelihood of this problem. Expressing the sale price (dependent variable) in per square foot or per acre terms has also helped to minimize this potential problem. Verified by examining a scatter diagram comparing residual errors to corresponding predicted values. See scatter diagram below as an example. The horizontal line-of-best-fit indicates that the residual errors are evenly distributed among the predicted values. Scatterplot Dependent Variable: trndadjsp Regression Standardized Residual 4 2 0-2 -4-2 0 2 4 6 Regression Standardized Predicted Value 31