By N.Kalinina, S.Gulyaeva,

Introduction

1. Property tax reform in Russia – first steps

The current system of land and property taxes and its deficiencies

Reforming efforts and technical assistance support to those

A. Property tax experiment in Novgorod and Tver

tax in a municipality (what has been completed, with the project

support, in Novgorod and Tver)

2. Valuation aspects: theory and reality

1. General mass appraisal principles
2. What was completed

3. Preliminary results – valuation models

1. Valuation of apartments in multi-family buildings

3. Some problems arising when applying mass appraisal methods

1) Does the market exist ?– information aspects

2) Legitimacy issues, property rights and valuation formulas

4. What’s next – where to go

2. Equity aspects of property tax reform

Market value based tax and rent vs existing tax and rent levels

2. Time schedule for the experiment - preparation and implementation

5. Some Model’s Statistics

Introduction

This paper presents brief description of some results of the project supported by the U.S. Agency for International Development in order to support the development of a market value based property tax system in the Russian Federation. carried out by the team. The project started as part of land privatization technical assistance program, then was switched under the umbrella of tax reform program. Intellectual leadership of Joseph K.Eckert allowed to bring into the project at various stages of its implementation brilliant team of professionals who succeeded in design of a property tax system suitable for the peculiar environment of a transition economy. Not only the concept was developed but also practical implementation began and skills were transferred to Russian government and business professionals sufficient to introduce the new tax in Russian cities.

We are particularly thankful to J.Eckert, R.Denne, J.Malme, M.Montes, W.J.Brzesky, P.O’Connor,L.Brennan, J.Epstein, R.Frenzen for not only contributing their professional experience and knowledge to understand and help to adequately resolve various problems of complicated Russian reality, but also for their ability to share all “know-how secrets” and to train as many local professionals as possible.

This project would have failed unless straightforward will towards successful results in reforming existing tax system was demonstrated by our counterparts from municipalities of Velikiy Novgorod and Tver. A.Korsunov, S.Isaev and V.Antyufeev managed to push through integrated legal and fiscal cadastre development in situation of permanent fight between various governmental agencies for pieces of information, functions and control over “pieces” of real property. A.Ivakhnenko and L.Mishutkina being responsible for property tax departments in cities, tremendously supported the project by everyday implementation of various tasks and addressing its concerns.

Finally, we very much appreciate the team from Georgia State University lead by S.Wallace, J.Martines, R.Bahl and B.Edwards for careful attention and professional support to property tax reform efforts.

Paper presents thoughts and interpretations which were resulted from joint work, discussions and brainstorming of large team including staff of the Center for Real Estate Analysis. However, authors are completely responsible for all misinterpretations and mistakes (if any) of this paper.

Background

The current system of land and property taxes is as follows. Land tax rates are established in rubles per square meter, adjusted for inflation by multipliers set by federal budget law. Tax on assets of enterprises is based on balance sheet value of assets including real property, machinery and equipment, transportation vehicles, inventory and other non-monetary assets, rates are set by regions up to 2%. Tax on property of individuals is based on inventory value which reflects costs in 1969 prices adjusted by physical depreciation and inflation, tax rate is 0.1%. Land tax is shared between local (50%), regional (20%) and federal (30%) budgets, assets tax is 50% regional, 50% local, tax on property of individuals 100% goes to municipal budgets.

Some deficiencies of the existing system became recognized by a wide range of professionals and politicians. These perceived problems include:

(1) The existing system of property-related taxes for legal entities inhibits investments and discourages inefficient producers (as well as others) from making renovations of their so-called main assets. The situation with land-related assets is similarly problematic. Since land payments are unstable (both for taxes and rents, which may be changed every year), investment feasibility projections become extremely difficult, both for short-term and long-term investment projects. Also, the considerable difference between property taxes on natural persons and assets taxes on enterprises encourages the latter to reduce their tax burdens by re-registering their property in the name of natural persons.

(2) Local governments essentially have no revenue base of their own, and so have essentially no ability to make long-term development planning of their territories.

(3) Property taxation of individuals has been inefficient from the fiscal point of view (since the tax collections are often less than the administration costs). Additionally, it does not provide for a fair distribution of the tax burden. The lion’s share of newly built valuable real estate is not currently taxed since current law does not establish the necessary rules for filing registrations of the properties themselves and the holders of rights to them. The method of determining the taxable values is long out of date; a better quality house in a prime location may have less residual taxable value than poorer quality prefabricated apartment buildings in remote locations incomplete contrast to their market values.

Reforming efforts and technical assistance support to those

These deficiencies of the existing tax system were considered by the Ministry of Finance of Russia when drafting the Tax Code, which envisaged a possible replacement of the property-related taxes and land payments with one consolidated real property tax at the discretion of local governments.

The US Agency for International Development since late 1995 supported the technical assistance project aimed at development of a market value based property tax system. Thus, this project provided an opportunity to design and test a model of a property tax system that would be suitable for the Russia legal, institutional, and socio-cultural context.

A. Property tax experiment in Novgorod and Tver

Legal and Institutional Aspects

In December 1995 the Novgorod Oblast approached the Russian government with a proposal to carry out a real property tax experiment in the city of Novgorod. The local government of Tver made a simultaneous proposal to the Russian government concerning the tax experiment. At the same time each city set up a local working group whose task was to make all necessary preparations for the experiment.

In response to those initiatives, the Russian government decreed in February 1996 (Decree #291-p of 28.02.1996 with amendments of 28.01.97 #123-p) the formation of an Inter-Ministerial Working Group to work out a consensus concept of the real property tax implementation, develop the legislative base and coordinate the experiment. This Working Group group consisted of representatives of the Ministry of Finance, Ministry of Justice, State Tax Service, Ministry of Economy, the State Committee on Land, State Committee on Construction, Ministry for State Property Management, representatives from the pilot cities and oblasts (recent up-date as of 06.03.98 # 312-p).

The property tax experiment became the main focus of the US AID funded project, and all project activities have been organized around preparing the two pilot cities for implementing a property tax there and rolling out the experiences gained in the pilot cities throughout the rest of Russia.

The Federal Law “On the implementation of the real property tax experiment in the cities of Novgorod and Tver” was approved by the Board of Government of the Russian Federation and submitted to the State Duma of the Russian Federation in December 1996, adopted in the third hearing by June of 1997, approved by the Council of the Federation in July, and signed by the President of Russia in 1997 August. The law granted authority to the two cities to establish a tax on real property and to determine the valuation methodology and the tax rates. The law defined all real property as taxable, except that exempted under foreign treaties, and left to the local Dumas the right to grant additional exemptions, although residential property could not be taxed at greater than fifty percent of its market value. The law also suspended the existing taxes on property in the cities when the local legislative body approved the new tax. The passage of this law made it possible for officials of the two cities to begin serious deliberations on tax policy and administration, in preparation for drafting the local regulations necessary to go forward with the “experiment”.

The law stipulated implementation of experiment in 1997-1998. However, consensus between the federal, regional and local governments with regard to intergovernmental fiscal relations for the period of experiment was reached only in 1998. Besides, tax impact studies prepared in both cities indicated that either residential tax burdens should become 60-100 times higher, or relatively high level of taxation for commercial and industrial property would be necessary to replace the revenues of the suspended taxes. Based on tax impact studies and intensive policy discussions, it was concluded that transition period should be introduced for phazing out replacement of existing taxes, in order to avoid significant changes of tax burden by groups of taxpayers. By May of 1999 the transition mechanism design was completed for both cities, and draft law “On Changes and amendments to the federal law “On implementation of the property tax experiment in the cities of Novgorod and Tver” was adopted by the State Duma in the first reading. The draft suggests that experiment timing is extended through 2003.

The drafting of the enabling legislation was done simultaneously with the development of a consensus on the tax experiment concept and the execution of the physical preparatory work in the pilot cities.

Concept for the Experiment

The objectives of the real property tax implementation in Novgorod and Tver may be grouped into:

Tactical objectives: to develop an applicable real-property-taxation model in the Russian context and to test its viability for the purpose of amending the draft Tax Code, as necessary; and

Strategic objectives which stem from the desire that the tax system should promote the development of federalism and democratic local government institutions in order to facilitate economic growth:

Short-term objectives – to create a favorable investment climate through:

1. Creating stable real property tax rates and land rents, fixing their amounts as a percentage of (assessed) market values;
2. removing active assets from the tax base, which would support technological modernization of production and result in the creation of new jobs;

Long-term objectives: to increase step-by-step the tax burden on real property owners; to finance municipal development (for example long-term road construction projects, infrastructure development projects, and improvement of public services) from current municipal funds and by tapping the bond-backing potential of a growing tax base. In addition the placing of holding costs on value land will increase land being put to its highest and best use.

Social and economic constraints to be accounted for at the preparatory stage of the experiment are:

• a probability of missing revenue due to possible reductions in the tax base during the transition to real property taxation of enterprises (in lieu of their current taxes), inasmuch as the buildings and structures of the enterprises account for only half of the currently taxable assets, the rest being machinery, equipment, raw materials, inventory stock, and intangible assets;
• the possible tax burden increase on owners of dachas, housing, and garages should be limited by the capability of the owners to pay the tax;
• A need to correlate the tax burden on owners and the financial costs of tenants (users) so as not to impede the process of privatization.

The benchmarks for the tax policy:

• to determine the tax base and a set of real property tax rates and rental rates such that the total revenues to be raised would be not less than the revenues from the replaced taxes shared among the budgets of the various levels of government;
• to expand the tax base by including state and municipally owned property (used under right of daily management or right of economic maintenance);
• To achieve a well-balanced transition to the new tax by avoiding abrupt changes in the tax burden on the owners of garages and housing as well as those of industrial property.

The Annex 2 tables are designed to demonstrate equity aspects of the market value based real property tax. Table 1 presents comparisons between the current tax burdens for individuals, and tax liabilities when the property tax is introduced. First three lines compare current inventory value and market value of three similar cottages. One, located in the downtown has estimated market value double of that of one located in outskirts. The inventory value relationship does not consider location and therefor is inequitable. The same is true for apartments. Only the negligible range of the tax keeps taxpayers form revolt.

Table 2 demonstrates how land tax and land rent levels, shown as percentage of market value, vary by location and type of business. Land rent for retail shops is 7-8 times higher than for industrial enterprises in downtown area and 4 times more than in edge of the city of Novgorod.

Annex 3 outlines the time schedule for the experiment. The conceptual framework for the property tax reform effort in conjunction with the market development, market value tax base increase, and city infrastructure investment financing is shown in the Annex 4 diagram.

B. Preparatory work for the implementation of the real property tax in a municipality (what has been completed, with the project support, in Novgorod and Tver)

1. The legislative and regulatory framework for the tax experiment encompassed drafting and securing passage of the following laws and regulations:

• On the procedure of the implementation of the real property tax experiment in the city of Novgorod (Tver) (to be adopted by the respective Oblast Legislative Assembly);
• On the construction and maintenance of the real property tax register (fiscal cadastre) for the purpose of taxation (to be adopted by the respective representative authority of the local government – municipal Duma of Council);
• On the real property tax (to be adopted by the respective representative authority of the local government), with an amendment on transition period;
• Regulation on valuation of real property for taxation (Methodology of market value assessment for taxation - to be adopted by the respective representative authority of the local government);

1. Provision of an effective local organizational structure in support of the real property tax reform entailed the creation of a local working group to work on preparing for the implementation of the real property tax and the establishment of a Tax Administration Office in the city administration (for recording and valuing real property for the purpose of taxation).
2. Technical, material and human resources were developed. These included provision of necessary hardware and software and training of staff in the following areas: real property register (fiscal cadastre) maintenance, valuation of real property, tax impact analysis, and public relations.
3. Software was developed in support of a Property Tax Information Management System ( PTIMS); this included the creation of a data base structure appropriate for the administration of the tax at the level of a city or rural rayon.
4. A real property register suitable for taxation was developed, which involved re-recording of real property data and mass scale collection and entry of data on real properties, clarification of property rights associated with them and the holders of those rights.
5. Market analysis and mass-valuation of real property were performed so as to be able to base assessments for taxation purposes on estimates of market value.
6. A variety of calculations were performed for the purpose of tax impact analysis. These showed, by groups of real property, the consequences of alternative tax policy choices in the areas of defining the tax base, setting tax rates, and developing exemption policies.
7. Work was done to coordinated tax policy with ancillary policies. This involved a determination of rent rates for land, non-residential premises, and housing (occupancy fee) based on market values so that the tax burden on property owners would not exceed payments for the right to use the property (if it were not owned).
8. Public relations work was done to explain the tax experiment work, to study public opinions, and develop a campaign in support of the reform.

Pursuant to the Law of the Russian Federation “On the implementation of the real property tax experiment in the cities of Novgorod and Tver,” work has progressed as described below:

1. Real property registers were created for tax administration purposes with the data on the real properties, including location, physical characteristics, designated and actual uses, market value, rights, subjects of rights, and tax exemptions. Real properties were assigned identification numbers based on the land cadastre number. The data in the Land Committee, BTI, Registration Office and real Property department as well as other bodies were used for this purpose. Field reviews of non-residential properties were performed, and cadastre numbers assigned to those land parcels that had not been inventoried before. Currently, the registers have data on 200 000 real property objects in Tver and 155 000 objects in Novgorod

• The registers do not yet include underground and similar storage structures or transmission lines;
• The inventory of rights is incomplete (the determination of the legal status of real property and identification of right owners is problematic in Novgorod in respect to garages, while in Tver other types of real property are also of concern);
• The data on natural persons is incomplete – tenants have not been verified; no ID numbers have been assigned to taxpayers who are natural persons;
• The Regulation on the real property register maintenance is being amended so as to reflect recent changes in federal law.

1. A Property Tax Information Management System was created within the city information network to provide information support to the processes of real property taxation. Computer-based modules have been developed for the following: data entry, storage and correction; mass appraisal module, assessment module; and statistical analysis module, appeals module (being completed).
2. A methodology for real property valuation was developed based on mass appraisal techniques. The created mathematical models are capable of generating values close to market values for apartments, detached houses, industrial and commercial real properties.

The valuation method needs to be refined; this could be achieved through field review and quality control exercises scheduled for the 3^nd quarter of 1999. After this the valuation method will be submitted to the local Dumas for consideration and approval.

3. Fiscal impact analyses were undertaken both in terms of total budget revenues and possible changes in the tax burden on various categories of taxpayers. As better data became increasingly available many scenarios of tax rates, tax base and exemptions (tax benefits) were analyzed. Data sufficient for various types of economic analysis, enabled the development of policy options suitable for each city. In Novgorod, the tax burden shift from legal entities which own land to individuals who own non-residential real estate (and to a smaller extent, residential) would encourage industrial production increase and business development while keeping budget revenue neutrality. Some of the examples of tax burden shifts are presented in Annex 2, Table 1 and Table 3. For Tver, the property tax will first replace land tax and tax on property of individuals, some revenue increase projected for the first year of the new tax implemetation.

2. Valuation aspects: theory and reality

Mass appraisal, like single-property appraisal, is applied economic analysis. In appraisal, factors affecting the supply and demand of real estate are expressed in valuation models (formulas). Using statistical methods and computer-assisted mass appraisal (CAMA) systems, property tax administrators can produce appraisals that take into account more supply and demand factors than usually are considered in conventional single-property appraisals. Consequently, mass appraisals can be highly accurate.

Mass appraisal models typically are based on three broad “approaches to value,” (1) the sales comparison approach, (2) the income approach and (3) the cost approach. Valuing properties for the experiment, sales comparison approach was used. Cost approach model is being finalized for industrial buildings and non-building structures. Special attention is devoted to valuation of land. Final results for Tver would be obtained later on this year.

In mass appraisal, the sales comparison models generally have the following general form:

V=f(X₁,X₂,.....X_N),

where V is estimated sale price, f stands for “a function of”, and (X₁...X_n) are property characteristics and location characteristics. Models may be additive, multiplicative, or “hybrid” in form. Multiple regression analysis or the adaptive estimation procedure may be used to calibrate the models.

The first step in mass appraisal modeling is data collection. Data are to be collected on those basic characteristics of the valued real properties which are most significant to real estate brokers in pricing, and could be measured (found in archives) by property inventory bodies. Data may be categorized as either qualitative (e.g., wall ,material) or quantitative (e.g., size). Qualitative data are based on discrete, predefined categories.

For successful modeling, data should be thoroughly analyzed. Measures of central tendency and dispersion such as mean, median and coefficient of variation (COV), coefficient of dispersion (COD), and standard deviation should be calculated for quantitative data. Qualitative data should be analyzed using frequency distribution, histograms and correlation analysis.

At the next stage, data should be “cleaned up” by using multiple regression analysis (MRA). Which identifies outlyer properties that should be eliminated from further analisis.

The second step in real estate market modeling is linearization of qualitative variables. Linearization of variables implies transformation of qualitative variables which were initially coded as quantitative values, into quantitative values reflecting market preferences.

Step 3 in the modeling process is an analysis of how various factors affect the real property price and interact with each other. The purpose of this stage is to: first, identify the factors which contribute the most to the price building, and, second, to identify relationships between the descriptive variables. The basic tool for relationship analysis is multiple regression analysis (MRA). At this stage, there should be an analysis of variables into the total determination coefficient in order to identify the significance of the variables.

The location value response surface analysis provides a means of adjusting for location in econometric modeling. In the first stage, value influence centers are determined. To identify location influence centers, a linear model has to be built and should include all the factors available, for the moment, as independent variables, except for the variables describing location of a property. Based on this model, estimated values and residuals will be computed as the ratio between actual price and estimated value. To identify the location influence centers a contour plot should be built which transforms the relationship between the residuals (coefficients) and the coordinates. By analyzing the concentration of level lines on the graph, one can identify location influence centers and the radius of their influence.

The next stage of location analysis is Global Response Surface design. The Global Response Surface is a three dimension surface showing the price per sq. m of a property with regard to its location. To build this surface, a non-linear regression method may be used with the variables describing the distance of a property as independent variables, and price per sq. m as the dependent variable.

Models can be classified into two broad categories, linear and non-linear.

Linear Regression Model is the simplest type of mass appraisal models.

To build a linear real property valuation model one would build a linear regression as a function of all the qualitative and quantitative characteristics of a property. Then a thorough statistical analysis has to be made in order to identify significant variables.

To refine the model, non-linear regression techniques can be used. The most convenient tool for nonlinear model design is adaptive estimation procedure (AEP or feedback). Which allows qualitative factors to the computed as percentage adjustments and additive adjustments such a square factor living area to computed as ruble adjustments per square foot.

The main criterion of model quality is the percentage of dispersion explained by this model. The coefficient of determination is such a criterion for regression analysis. This coefficient is usually identified as R-squared coefficient in scientific works.

Another important indication of valuation quality is a coefficient of dispersion (COD). The COD is the major statistic used to determine quality in real estate appraisal.

where R_i are ratios of the actual prices of the objects to the modeled values; are median values of these ratios, N is the number of observation in the estimated sample. In addition to being applicable to judging the quality of non linear models with reference to the model-building data set, the COD is also widely used in judging the quality of models as time passes and new data on sales prices and predicted values become available.

Valuation quality control

The initial application of a mass appraisal model is straightforward. However, the property value estimates that result from the application of a mass appraisal model should be reviewed before the estimates are used to determine tax obligations. The value review process considers data quality, the appropriateness of the valuation model in question, and the success of the calibration of the model.

An essential step in the model application and review process is to apply the model to a test group of properties. The test group should not include any of the properties used to develop the model. Based on testing, model deficiencies could be discovered, caused by inadequate representation of property classes, types and locations. Then more data should be added, until model quality control measurements for the testing sample become acceptable.

After the final model is applied, the property tax administration should review each value estimate. Appropriateness of model, consistency of value estimates should be checked. For those properties which have special features, adjustments to the model estimate could be developed reflecting value influence of object’ peculiarities. The review process should be documented, adjustments should be described in the manual. For Novgorod and Tver, some adjustments developed based on the field review of values, were included in the valuation methodology, to be approved by the city legislators.

The development of valuation methodology which is suitable for taxation purposes started from analyzing the market. Secondary transactions were collected, asking prices being used to supplement actual transaction prices. For non-residential objects, it was necessary to supplement transaction prices by representative samples of fee appraisals. The preliminary modeling showed that at least three types of objects should be distinguished, based on significant differences in characteristics influencing values for two classes of residential properties (multistoried apartment buildings and detached houses ), and non-residential objects including commercial and industrial use.

Altogether, samples of initial information used to build models included:

for Novgorod, 638 apartments, 66 single family houses and 80 non-residential objects; for Tver - 1512 apartments, 115 single family houses and 159 non-residential objects.

NCSS and feedback software tools were used for model calibration purposes. All three models were designed as hybrid models (presented in the next section).

Test samples were selected from fiscal cadastres representing various combination of property characteristics and location. After the models were applied to control samples, field review was performed, and additional characteristics were defined, necessary to improve quality of prediction. Namely, size adjustment variable was introduced into non-residential model to reflect decrease of value for large-size objects, differentiated by use type of objects. Besides, quality of building proved to be important, in addition to the year of construction, which reflects difference in maintenance and physical appearance of buildings. This variable in order to be applicable to all properties, required adding building quality description to the set of characteristics available from BTI files. For this purpose, all non-residential objects were inspected and described using specially developed manual.

Testing single family housing model, it was detected that amenities are not picked up by the model adequately. Therefor, additional variable called “utilities in place” was introduced to the model which reflects availability of main utilities in the area.

The federal law enabling the experiment provides that real property valuation for taxation purposes has been implemented by the governments of pilot cities in accordance with valuation methodology approved by the municipal legislators.

Some deficiencies of models were related to location factors, as initial samples appeared to not represent major differences of neighborhood quality.

In order to refine models it proved to be necessary to add more transactions (appraisals) to samples, and special efforts were devoted to resolution of this issue.

After up-dated models were recalibrated to meet the model quality criteria, they were applied to universe of properties in cadastre.

Valuation quality control consisted of office review and field review of “problematic” values/objects. Based on this field review, several sets of adjustments were developed to be included into valuation methodology to reflect those features of objects which were difficult to be measured based on statistical analysis. Those adjustments “for atypical property” include:

- for land: complete or partial unsuitability of a land plot for development (in case of guilles, wetlands etc.), encumbrance with servitude by the owner of another plot, and lack of utilities on the particular parcel under residential development with utilities available within the area where this parcel is situated;

- for buildings – location at the main streets, luxury (atypically high) quality, et al.

Those adjustments range provides guidance to the property tax department on how to react to appeals if the property is overestimated or under estimated.

The valuation methodology drafts were completed and are expected to be approved as soon as the federal law enabling the experiment is amended (extended).

The methodology outlines general provisions (legal definitions of real properties, methodology application and appeals procedures), valuation principles, models and coefficients, adjustment for atypical objects and application rules.

Valuation principles, among other things, requires refinements of models based on market monitoring and sales ratio control, and outlines quality control procedures.

After sales comparison approach model was applied to industrial properties, it became clear that it did not pick up in same cases technology specific features of industrial properties. For example, current market demand for non-residential use objects does not distinguish between reinforced walls and high premises with special storage equipment and ordinary buildings suitable to store boxes with cigarettes as long as it relates to warehouses. At least, it is not possible to draw statistically reliable results based on fisting market information. Sales comparison model were not applicable for valuing non-building structures (tanks, pipelines etc) as there were no sales of such objects. Therefor, it was decided to develop cost approach model to value industrial properties. For special purpose properties, the cost model is being developed to be introduced to the methodology later on this year.

3. Valuation results

The below listed formulas as well as base values of one square meter and tables of coefficients were computed based on application of multiple regression analysis and adaptive estimation procedure to above described samples of initial market data. Those formulas as well as tables of coefficients are provided in the Methodology for Real Property Valuation for Taxation Purposes, which is supposed to be adopted by city legislators upon federal legislation enabling the experiment to be extended.

Independent variables like number of floors, wall material, total area, living area, number of rooms, use type, utilities, other property characteristics needed for the purpose of market value estimate calculation, should be taken out of fiscal cadastre which is a legitimate data base for taxation purposes. Questions regarding accuracy of information on property in the cadastre should be addressed to Bureau for Technical Inventory with regard to building and premise characteristics, to the Land Committee with regard to land parcel characteristics, to the title registration agency with regard to information on legal status and right holder of a property object. This appeal regarding property description in the cadastre could be channeled to corresponding agency through the property tax department.

Location coefficients are distinguished by city land value zones, defined by the city regulation for the purposes of collecting payments for land (tax and rent).

1. Valuation of apartments in multi-family buildings

Assessed value = K1 * K2 * K3 * K4 * K5 * K6 * (K7 * B1 + A2 * B2 + A3 * B3 + A4 *B4 + K8 * B5)

where:

K1 is a number of floors coefficient;

K2 is an apartment location floor coefficient;

K3 is a building wall material coefficient;

K4 is a last floor location coefficient;

K5 is a building location coefficient;

K6 is a building wear coefficient (year of capital repair or reconstruction could be used instead if available);

K7 is a number of rooms coefficient;

B1 is a base value of the number of rooms coefficient;

A2 is total area of the apartment;

B2 is a base value of one square meter of apartment total area;

A3 is living area of the apartment;

B3 is a base value of one square meter of apartment living area;

A4 is kitchen area;

B4 is a base value of one square meter of the kitchen area;

K8 is number of balconies/loggias coefficient;

B5 is a base value of balcony/loggia availability.

Few examples of values of coefficients together with the base values of square meter of land and building are given in Annex 5.1.

The statistical quality of this model is as follows, based on the model application to the control sample: coefficient of dispersion 10.6% (see Annex 5.4), R-squared is 92% (see Annex 5, mutliple regression report for apartments). In other words, the model explains 92% of variation of the dependent variable, and difference between the model prediction and actual value varies around 11% of median.

2) Valuation of single-family detached houses, apartments in detached single –family houses, residential attachments and assigned land plots

The valuation of single-family detached houses, apartments in detached single –family houses, residential attachments and assigned land plots shall be done by the formula:

Assessed value = K1 * K2 * K3 * (A1 * B1 * K4 * K5 + A2 * B2)

Where:

K1 is a location impact coefficient;

K2 is a utilities availability coefficient;

K3 an undeveloped land coefficient (when the house total area us zero; that is the valuation object was not appraised);

A1 is the total area of the valuation object;

B1 is a base value of one square meter of the total area of the valuation object;

K4 is a wall material coefficient;

K5 is a depreciation factor of the valuation object;

A2 is a land area;

B2 is a base value of one square meter of the land area.

The input of each such characteristic into value of a property is expressed in values of coefficients (weights). Few examples of values of coefficients together with the base values of square meter of land and building are given in Annex 5.2.

The statistical quality of this model is as follows, based on the model application to the control sample: coefficient of dispersion 18.9% (see Annex 5.5).

The object for valuation for commercial and industrial property shall be buildings or premises, or a share in the building (premise). Valuation of a building, premise, a share and related land plots shall be made according to the formula:

Assessed value = K1 * [K2 * K3 * K4 * K5 * K6 * A1 * B1 + K7 * B2 * A2],

where

K1 is a location coefficient of the valuation object;

K2 is a depreciation coefficient;

K3 is a use type coefficient of the valuation object;

K4 is a wall material coefficient of the valuation object;

K5 is valuation object size and type coefficient;

K6 is a quality construction coefficient;

A1 is the total of the valuation object;

B1 is a base value of one square meter of the valuation object;

K7 is a land use type coefficient;

A2 is land area;

B2 is a base value of one square meter of the land;

Few examples of values of coefficients are given in Annex 5.3.

The statistical quality of this model is as follows, based on the model application to the control sample: coefficient of dispersion 19.8% (see Annex 5.6).

3. What’s next – where to go

The following literature was used when working on this paper:

Property Appraisal and Assessment Administration. IAAO Text book, Russian Edition

Roll-out Manual, Chapter VIII – Valuation, Property Tax Project deliverable, Manuskript, 1997

Methodology for Real Property Valuation for Taxation Property for the City of Novgorod, draft, Manuscript, 1999

Package of Documents Developed for the State Duma of the Russian Federation, for the first reading of enabling legislative amendment, Manuscript, 1999