Development of a CAMA System for Urban Property in Greece Based on Cadastral Data

Development of a CAMA System for Urban Property in Greece Based on Cadastral Data BY PANAGIOTIS ZENTELIS

T

o date, fiscal policy for real estate properties in Greece has not been— and could not have been—successful because of the lack of necessary information on the ownership, location, and value of property and the lack of management infrastructure at the government level. A consequence of the weakness in approximating real market values is that taxation is sustained at a high level. The government’s prime interest is in increasing revenues, rather than fairly distributing the tax burden. Until 1985, property values for tax purposes were assigned by Greek taxing authorities based on uncertain comparative data, resulting in mass complaints and appeals by citizens in the tax courts. In 1985, an objective method was instituted, in which the value is estimated through specific tables for each property category. The calculation of the taxable value, VTAX , is based on a specific value per region, readjusted every two to three years, and multiplied with several coefficients corresponding to the characteristics of every property. With this current methodology, appeals have stopped, but the assigned value, VTAX , has low correlation with the market value (V). Research has shown that, for the period from 1985 until today, in every readjustment of the

values estimated by using the objective method, the assigned value, VTAX , is systematically lower and rarely higher than the market value. In addition, the resulting percentage deviation, (V - VTAX)/V, is different for each property category in the same region; for example, for a house, the deviation can be 30 percent; for a store, 40 percent; and for a land plot, 60 percent. It is also different for the same category of property within the same region or from region to region; for example, for a house, the deviation can be 30 percent at one location and 50 percent at another. For these reasons, it can be concluded that: • The objective method cannot approximate market values, and moreover, periodical price increases often are made to serve political purposes (e.g., larger percentage increases within high-end areas). • The current common tax rate (e.g., 11 percent for property purchases and sales) varies according to the existing deviation, (V - VTAX)/V. As a result, for every citizen there is a different tax rate for the market value (e.g., one corresponds to

Panagiotis Zentelis is Assistant Professor of Cadastral Systems, Land Information Systems, and Real Estate at the National Technical University of Athens, Greece. Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

23

3 percent, another to 9 percent, and so on). Thus major inequities are created in the allocation of the tax burden.

Catalysts for Market Value Assessment Today, fortunately, certain needs that require the optimization of procedures for fiscal policy for property have become evident. These needs, demanding immediate solution, have evolved with the levying of the value-added tax on property on January 1, 2006. The basic categories of property taxes that created these needs are as follows: • The transaction tax on property that existed before January 1, 2006, the parental benefit tax, the inheritance tax, and similar taxes require knowledge of market value. • The large ownership tax requires knowledge of the sum of property values for every natural person or legal entity, when the accumulative value exceeds a pre-assigned limit.

more complicated procedure. For example, in some cases, differentiating the value of the land from the value of the construction on the land is required. In addition, the progressive creation of an e-cadastre system has contributed to the potential for more equitable market value assessment in Greece. The system is based on and was developed by using contemporary technology. Thus, the e-cadastre uses geographic information system (GIS) technology to provide geographical property depiction with multi-level information. Specifically, it offers the possibility of automated dispersal of property measurement data as well as qualitative information. This effort is furthered by the country’s continuing implementation of e-government procedures. The exploitation of e-cadastre information, combined with e-government capabilities, is succeeding because of the hierarchical development of a land information system (LIS). The development structure of a LIS-based e-cadastre system and the creation of a sub-LIS is shown in figure 1.

• The net-gain-in-value tax, imposed on January 1, 2006, applies to the unearned increment, which is the difference between the purchase and the sale price of a property. The coefficient of this tax varies according to the time period between the last purchase and the sale—from 20 percent for property purchased during the last 5 years to 5 percent for property purchased within the last 25 years. The applicable surplus value is determined from the purchase and the sale values on the contract records.

The basic sub-LISs concern the following: • The assessment of property value (sub.LIS.valuation). Its development requires the planning and operation of a computerassisted mass appraisal (CAMA) system, which can use the e-cadastre data and from which the estimated market value (V ) can be determined consistently, precisely, rapidly, and inexpensively in every instance. Estimated in this manner, the market value can be applied for every use and can serve the public and the private sector.

• The value-added tax (19 percent), instituted on January 1, 2006, requires, apart from the knowledge of market value, a

• The taxation of property value (sub. LIS.taxation). Its development has been achieved through the sub.LIS.valuation. The desired

24

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

Figure 1. Using an e-cadastre system to develop a LIS with hierarchical sub-LIS

consequences are the effective satisfaction of the before-mentioned needs and the reduction of tax rates. The result is an increase of revenues and the detection of property tax evasion. Social justice is preserved through the proportionality of tax burden allocation as well as the capability of more reasonable property policy.

Planning the CAMA System for Greece (CAMA.GR) The need to determine the market value of property using a unified, accurate, effective, fast, inexpensive, diachronic, and mass-use method, customized for Greece’s capabilities, led to the idea of developing a system for mass appraisal

of market value. Hereafter referred to as CAMA.GR, this system arose as a result of the dialectical relationship between theory and practice. In its initial design, CAMA.GR can be used for the usual property in urban areas, such as land plots, apartments, houses, offices, stores, and the like. However, CAMA.GR does not have the ability of being used in nonurban areas or for specific-use properties such as commercial centers, hospitals, hotels, and the like. The basic characteristics of CAMA.GR are described in the following paragraphs. Factors That Affect Market Value and Their Classification The uniqueness of each property in every category leads theoretically to an infinite number of factors that differentiate the value of different properties, now or in

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

25

the future. Depending on the number of factors being examined, we can approximate proportionally the real value of V. The contribution of these factors to the final modulation of V is difficult to determine, mainly because of their large volume, their different behavior in time and in place, and their dynamic correlation and evolution. The factors that affect V can be classified in different ways. For the purposes of the development study for CAMA.GR, these factors are classified by examining the V at different location levels. When the valuation of V at a location level is of interest, the factors that affect the value should be used according to the desired accuracy of estimation. Specifically, the factors that affect V are classified as follows: • Country level. This level includes factors with their original, secondary, etc. classification. A finite number of statistically important factors could explain the differentiation of V from country to country for the same

type of property. In other words, we could find the statistically important factors that affect or differentiate the V from country to country in order to have, for example, for every specific category of property in Greece, variation of prices from min V.GR to max V.GR (figure 2). • City level. This level includes factors with their original, secondary, etc. classification. A finite number of statistically important factors could explain the differentiation of V from town to town of a specific country for the same type of property. In other words, we could find the statistically important factors that affect or differentiate the V from town to town of a specific country in order to have, for example, for every specific category of property in Athens, variation of prices from min V.Athens to max V.Athens (figure 2).

Figure 2. A small “window” of factors Is enough to approach the buying value, Vi

26

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

• City parts level. This level includes factors with their original, secondary, etc. classification. A finite number of statistically important factors could explain the differentiation of V from part to part of a specific town of a specific country for the same type of property. In other words, we could find the statistically important factors that affect or differentiate the V from part to part of a specific town of a specific country in order to have, for example, for every specific category of property in the Gizi area (low-end suburb of Athens), variation of prices from min VGizi to max V-Gizi (figure 2). • Property Level. This level includes a finite number of statistically important factors that could explain the differentiation of V from property to property of the same type in specific parts of a town in a specific country. In other words, we could find the statistically important factors that affect or differentiate the V from property to property in a part of a specific city in a specific country in order to have, for example, for every specific category of property, an estimate of V for every specific property (PRi) at a specific location and corresponding to a combination of characteristics that affect its buying value, Vi (figure 2). The other factors, those of nonstatistical importance, that describe the uniqueness of every property are skipped because they do not affect the estimation of Vi . Basic Concepts Used in Development The basic concepts in developing CAMA. GR are as follows: • The classification of the factors mentioned above shows the

capability of CAMA.GR at the property level, with properties placed in a specific part of a town. This part usually corresponds to the limits of a municipality (e.g., the urban complex of Athens comprises 52 municipalities). • Properties have substitute goods that can satisfy the same needs. As a result, the prices of property fluctuate within a limited range in a specific part of a city. • Most of the factors have already affected the formation of the known market value of property in an area. A small number of factors differentiate the market value between kindred properties, while other factors that differentiate the properties add small advantages or disadvantages. The market does not react smoothly to these variances but does create sales trends within a short time. • The market value of a specific property can be determined within a range of prices depending on the predefined effect of specific factors that differentiate the value according to the market data. These factors are taken into consideration for input, control, and usage of the model. In particular, the CAMA.GR system • assumes that, in the end, even with consecutive approximations, the highest and best use of property is achieved. • uses the reduced unit price of an area as the market value. • considers the market value as the dependent variable and the other characteristics of a property as independent variables.

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

27

Through multiple regression analysis, statistically important factors as well as their influence on the market value are determined. • reduces the few and relatively unvarying distributed known market values of property in new locations. The reduction is based on algorithms for property of every category that are verified from market data and that can be readjusted as necessary. • is populated with known market prices of property with known characteristics and satisfactory geographical distribution. These data are updated periodically. • is monitored for the accuracy of determined market values by comparisons with redundant known market values of property of known characteristics. • is used in a uniform way in order to determine the market value of every other property in an area of familiar characteristics. • can be used in any specific part of a city and thus can be used in every urban location in the country. Data Generated by CAMA.GR Following is a discussion of the basic factors that are taken into consideration, as well as the way in which each one affects the market value. The algorithms used in CAMA.GR construe in the best way the Hellenic real estate market data. The mathematical formulas and numeric prices incorporated in the algorithms were developed from known prices of appropriate property, when each factor was in effect separately (condition of partial balance, ceteris paribus) and then tested by the model’s performance as a whole. 28

Reduction of V at a Common Reference Base and Determination of VT In CAMA.GR, the reduction from time t of the known Vt to a common time base of reference T is required, as well as the determination of VT . For this price reduction, the diachronic change in prices follows the evolution of certain fiscal indexes. The most appropriate index for the time reduction of prices is the Housing Price Index, which is published every three months by the National Statistical Service of Greece and which indicates primarily the annual change of PRi . Also used was the Consumer Price Index, which is published monthly by the same organization and which indicates the change in the value of money. The time period in which the prices can be reduced is short (maximum five years), but increases when inflation is low. Reduction of V Due to the Size of the Property and Determination of VSTD The market value of property is influenced by its size. The differentiation of the unary prices is due to the corresponding differentiation of demand, which is maximized for a certain size. In this instance, algorithms were developed that describe the market satisfactorily for every category of property. Figure 3 is a graph of property value according to the size E, given of ESTD , of the corresponding VSTD and of the coefficient c, which has a different value for each category of property. ESTD is the size with the maximum demand. For example, for apartments, the maximum demand is for an area of 100 square meters, which can vary for every geographical area of the country. For land plots, the maximum demand is for an area equal to the adequate area, E0 , which is the minimum for which construction is allowed. For a land plot with an area multiple to E0 , the decrease of V results in ESTD being the whole multiple of E0 . In this way, we can determine instances of failure, when large areas are required to cover special uses (e.g., supermarkets). During

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

Figure 3. Reduction of property value due to size

the application of the model, special-use properties are excluded from the reduction because of their size. Reduction of V Due to the Dimensions of the Property and Determination of V0 A land plot with front b = 40 meters and depth d = 20 meters has a greater unit price than another land plot with front b = 20 meters and depth d = 40 meters. In general, V increases when for a constant average d, b increases. Conversely, V decreases when for a given front b, the average depth d increases, while all other factors are held constant. In other words, the market value of a property increases when the fraction of d/b decreases and vice versa. This can be explained if we assume an elementary area dE of a property that decreases in value as it diverges from the front of the plot. Instead of the fraction d/b, we can analyze the fraction of the equivalent rectangle, E/b2. To determine the change in market value attributable to the fraction E/b2, we must select a mathematic formula that can be verified by the market data for every category of property. Then the prices can be compared and reduced among properties of the same category with different dimensions, when the rest of the factors

remain constant. Figure 4 is a graph of reduction of the Vx of the elementary area dEx of property, having the shape of an equivalent rectangle, where d is the corresponding average depth and b is the front or the sum of the front sides of the property. The reduction follows the equation Vx = V0 e-mx, where m is the coefficient of the value slope, which determines its reduction while x increases. If we set: V I= b  so V0   I =e-mb, so – ln I   m = b , so   Vx = V0 exp(ln I • xb) and average value is

  or

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

29

Figure 4. Alteration of property value according to dimensions

. When���������������������������������� dealing�������������������������� ��������������������������������� with��������������������� ������������������������� multiple������������ �������������������� -����������� front������ properties, the accretion�������������������� ����������������������������� ������������������� coefficient, c,���� ��� is� used as in

,   so

  Market research for several categories of property has indicated the values of ln I and the coefficient c. For example, the land plots are verified when ln I = -0.1, which corresponds to the value Vb /V0 = 0.90 and c = -0.05. From a given price of the market value of a property with specific dimensions, we can determine V0 with other factors remaining stable. The price, V0 , corresponds to an ideal value of a property, if every elementary part of its area is at the plot’s front. From this ideal value, we can calculate in a converse way the revised value of another property from its specific dimensions.

30

Value Change Due to Time and Determination of VN From research for the period 1960–2006 for every category of property in every geographical area of the country, the value in constant prices has increased at a larger or smaller rate. This constant increase in prices, at a pace faster than inflation, can be interpreted macroeconomically as mostly development-added value driven by the instability of the economy, by the lack (until recently) of substitute investment incentives, and by the lack of a developed market for property. This increase is more intense for land plots and less intense for other property categories (apartments, offices, and the like). Local fluctuations in the development curves of prices have been observed, and there is the possibility of prices decreasing due to the adverse effect of construction projects. In contrast to land, buildings can deteriorate over time and progressively decline in their ability to satisfy evolving needs. This decline is more intense in buildings constructed based on obsolete technology as well as those that lack armored concrete, anti-seismic protection, satisfactory heat insulation, a parking space, and so on. These buildings were constructed in that way because today’s effective regulations did not apply at the time.

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

The age of buildings is of interest because of the need to reflect the decrease in value resulting from the passing of time. This information is necessary to estimate today’s value, VN , of a property that was built N years ago and its deviation from the market value that corresponds to recent construction. The value, V, of a property is the sum of the value of land, Vland , and the value of the construction, Vcon ; the obsolescence of property comes from the antiquity of the construction. Figure 5 depicts the evolution of the value of land, f1(Vland , t), over time and the decrease in the value of construction, f2(Vcon, t), over time with the beginning of the construction (t = 0) as the starting point. The graph shows that for every specific property, we can determine the resulting decrease in the value compared to the value of new construction. Figure 5 also shows that for every specific property there are the following options: • Calculation of the net gain in value of a specific property over time, if, for example, its taxation is imperative. For f1, this model can use the Housing Price Index or the Consumer Price Index.

For longer periods of time, the model can use existing statistical data on the price development of the property in every category, which are available for every geographical area of the country. • Determination of the decrease in value that is created by the antiquity of the building, for example, to compare an old apartment with a new one. Specifically the function, f2, of the decrease in value due to antiquity of buildings has been assigned for 30 years. The corresponding algorithm includes coefficients with different prices for the high-end, the average-priced, and the lowend areas. Market data have shown that inexpensive buildings are constructed in low-land-value areas, and the loss in value is greater than that in high-land-value areas. In instances of usual construction (e.g., apartments, offices, and the like), the model uses coefficients of antiquity as follows: • If 0 < N ≤ 5 years, VN = V [1 - (N/100)]

Figure 5. Depreciation of property due to construction age

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

31

• If 5 < N ≤ 25 years, VN = V [0.95 - ((N - 5)/100) • ci] • If 25 < N ≤ 30 years, VN = V [0.95 - (20/100) • ci - (N - 25)/100] where N is the number of years of antiquity and ci is a coefficient with a value of c1 = 22/20 for high-end areas, c2 = 36/20 for average-priced areas, and c3 = 46/20 for low-end areas. The decrease in value due to antiquity cannot be used on • properties that have been repaired because of damage or properties that have been systematically refurbished (e.g., stores). • old buildings that are being listed because of their architectural value (these buildings are treated in a unique way). Value Change Due to Height and Determination of Vi If we assume a building of regular use (apartments, offices, and so forth.) with n floors above ground, the unary value increases toward the higher floors (figure 6). Market data show that the change in value by height can be considered linear, without large variation when the maxi-

mum number of floors is eight or nine. This maximum number of floors covers most buildings in the country, given the fact that high buildings are not generally allowed with a few exceptions. In these rare instances (n + 1 ≥ 10), the line AB tends to become horizontal, changing the increase in the price from linear to an S curve. For the usual cases, the linear model of the figure provides a price change from floor to floor (segment AB) except for the first floor (segment CA) and for the last one (segment BD), where the change in value is broader (φ2 > φ1 and φ3 > φ1). The change in value is represented by the line CABD. The algorithms used are as follows:   

where i ⇒ 0 for the first floor (floor = 1), i ⇒ i for intermediate floors (1 < floor < n), i ⇒ i + 1 for the top floor (floor = n). These algorithms determine the value, Vi, of the i floor as a function of the average value V floor , where   u = (Vmax -Vmin)/V floor. Market data on a building with a variety of floors provided values of the coefficient u from 0.217 to 0.238. For CAMA.GR, u = 0.22.

Figure 6. Property value per floor in a non-special-use building

32

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

In the case of property that is placed at the ground level, • If at the ground level there is a kindred use with the use of floors above (e.g., an apartment at the ground level in an apartment building), then the Vground is driven from the previous algorithm if i ⇒ -1. • If the building is outside of the central business district area and there is a different use of the ground floor (e.g., a store in an apartment building), then    Vground = V floor • (Rground /R ),

where R represents the relative annuities. For the usual cases,

   Rground = R

and the market is verified for

   Vground = V floor . • If the building is within the central business district area, then a coefficient of commerciality is set:   e =Vground/ V floor

and therefore

   Vground = e • V floor . Because the average value of the building is   V = (n • V floor + e • V floor)(n + 1), we can derive   V floor= V • (n + 1)/(n + e) and   Vground = V • e • (n + 1)/(n + e). These algorithms are used in every instance from which V or V floor has been derived from any previous reduction of other factors of value change. Reduction of Land Value Based on Quid Pro Quo and Determination of Vrel.land In Greece, for the last 50 years, buildings can be constructed using quid pro

quo p%. This means that the landowner provides 100 •(1 - p)% of the land plot to the developer and, as a reward, receives the 100 • p% of the building. Because of the long and frequent use of this method in the market, it is easy to indicate p for every region. Based on ���������������� the equation of the value being exchanged between the landowner and the developer, the following is the formula for the unary price of the land plot (see figure 7):   Vland = [p/(1 - p)] • k • l • r where • k is the construction cost of the unary price of a useful area • r is the effective floor area ratio, which is defined as the number that, when multiplied by the area of the land plot E, shows the total construction area allowed to be built • l is the derogatory coefficient as a function of the deviation of the constructed area to the useful area (e.g., l = 0.93 if the common-use areas of the building are 7 percent). The graph of Vland = f(p) is shown in figure 7, in which the exploitable part is marked. If the coefficient p is known, the unary price of the land can be determined and vice versa. From this formula, we derive the value of the relative land, Vrel.land , for the construction of 1 square meter of useful area. The necessary area is   Erel.land = 1/(l • r) and its value is   Vrel.land = V/(l • r) = k • p/(1 - p). Other Algorithms of the Model The CAMA.GR model includes other algorithms as well. The most important ones are used to determine the value • of auxiliary underground areas (storage areas, parking spaces)

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

33

Figure 7. The value of land according to p (percentage)

• of buildings characterized as listed that have or have not fully used the effective floor area ratio • of semifinished construction as a function of the building category and the construction phase of the finished building • of the special ownership rights, wherever partial value assessment is required (usufruct value, bare ownership value).

Updating and Using CAMA.GR To update CAMA.GR, a few known values of property in any category are used that correspond to a specific time and that can periodically be replaced with new known prices of other property. From the largest part of the known values, specific coefficients of CAMA.GR are calculated. Based on these coefficients and using the algorithms of the model, we can determine the value of other properties, kindred or not, when their characteristics are known. From the redundant known values of the property, we can check the accuracy of the estima34

tions derived from CAMA.GR for these properties. After that, the model can be used for mass appraisal. Indication of the Ideal Figures (V id , pid) When the model is updated through this procedure for every known value of property in every category, the necessary reductions are made according to the algorithms of the model, and an ideal value, Vid , of the relative land that is necessary for the production/construction of 1 square meter of useful constructed area is determined. For each specified value, Vid , a relative pid is assessed according to the formula that connects these two figures. The pair of ideal figures (Vid , pid) are the last data for updating the model that corresponds to every geographical location. The determination of Vid and pid requires the following two reductions for each instance: • reduction of the value of an empty land plot because of r. As already mentioned,   Vid = Vrel.land = V/(r • l) = k • pid /(1 - pid)

or   pid = Vid /(Vid + k)

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3



where usually l = 0.93 and the construction cost, k, is specific for every area and form of property. At this point, we must be careful in the instance in which the allowed floor area ratio, r, cannot be developed fully because of the plot’s geometry or because of lack of static endurance of the existing construction.

• reduction of the value of property with construction to the value of land. In this instance, given the value, V, and the return on investment, q (e.g., for a return of 25 percent, q = 1.25), we conclude that   Vid = (V/q) - k

or   V = q • (Vid + k)

In the instance in which the property with the construction is within the central business district area and given the fact that   Vid.ground = ε • Vfloor = ε • Vid.floor , the Vid is analyzed in two components:   Vid.floor = (Vfloor /q) - k,   Vid.ground = (Vground /q) - ε • k From Vid , we indicate the relative pid . These formulas approximate the market better as it approaches the highest and best use (e.g., the market behavior is not to build an expensive cottage in a low-end area). In addition, the marginal minimum values (minVid and minpid) are 0.33k and 0.25, respectively. From these marginal values, we derive a minimum sale price of the property, and thus can test whether the system q is adaptive to the market data. Separation of a Specific Area From the effective town planning data and the existing situation, an area is segmented

• in parts U1, U2, …, of the same floor area ratio and the same allowable uses (houses or offices, and so forth) • in parts CBD1, CBD2, …, where development for commercial use exists or is allowed, which parts can have the form of a linear zone. During the updating of the model with specific prices of property of specific characteristics and specific geographical location, we indicate the relative pair (Vid, pid) for every part (U1, U2, …, CBD1, CBD2, …) of the whole area. In this way, a triangular network is created from which, during the use of the model, we calculate the relative pair of ideal prices for every location of the partial area by insertion. Note that in properly chosen areas (e.g., same factor r, same use, and so on), the ideal prices (Vid , pid) present insignificant differences. Thus there is usually a representative pair of prices for the whole area. This is expected because the unary land value changes according to the ratio p/(1 - p), which is relatively constant for an unvarying area. Also, when buildings of different use coexist after the division of the area, we must use balanced data for every category of property in order to indicate the typical pair of values. Updating, Testing, and Renewing Prices The known prices of the market that are used to update the model must be reliable, representative, and complete. Prices due to urgent causes (e.g., auctions) or transactions between relatives must be rejected. To better update the model, the basic categories of property (land plots, apartments, offices, and the like) are used, while certain prices for auxiliary areas (storage areas, parking spaces, and such) are excluded. (Table 1 shows this updating; the right arrows feed the system.)

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

35

The model can be updated with new known prices of property periodically or constantly. Recent market data replace previous data and lead to new values (Vid , pid), which are valid in the new relative locations. The periodicity of the model is a function of market activity and inflation level. Before CAMA.GR is used for mass appraisal, the accuracy of the model is tested by using redundant known prices of property with known characteristics. The model is populated and tested by using prices of different categories of property for several locations (high-end, average-priced, low-end) and a different combination of town planning data. This technique produces satisfactory results, which are verified by market behavior. In general, the accuracy of the model increases when the demanded prices correspond to kindred properties that have been used to populate the model. Specifically, the observed maximum deviation between certain test prices and prices that emerged from the use of the model was of the order of ±4.3 percent for land plots, apartments, and houses; ±5.2 percent for stores within or outside of the central business district; 6.9 percent for listed buildings; and 6.3 percent for properties of auxiliary use (mostly underground areas).

The Use of CAMA.GR After the model has been updated and tested, it is ready to be used for mass appraisal of every property. The use of the system for the basic categories of property as well as for cases of partial calculation of V is based on the stated procedure for

each instance and is shown in table 1 (left arrows verify the system). Except for the valuation of the market value of every property in CAMA.GR, the system provides additional information on each and every property, as well as thematic maps with combined information. A few examples are provided in figures 8 through 12.

Conclusions From the discussion in this article, it can be concluded for the country of Greece that: • The need to assess market value for every use and for the service of the public sector has not been met. • The capability of developing an e-cadastre system in combination with e-government procedures allows the development of sub.LIS.valuation based on CAMA.GR, with the assistance of GIS technology. • CAMA.GR, which has been developed by using Greek data, provides confirmed data on the market value of urban property in each category, except specialuse property. • The structure of CAMA.GR allows its development in all urban areas and ensures the capability of parallel or independent function with the e-cadastre system being compiled. • CAMA.GR can be extended

Table 1. Analytical system operation for consecutive reductions of given data

36

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

Figure 8. Land use of the study area

Figure 9. A building characterized as “Listed” due to its special architectural value

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

37

Figure 10. Cadastral maps and tables

Figure 11. Civil planning data and restrictions in the study area

38

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

Figure 12. Differentiation of market values according to location

to special-use property as well as to non-urban property, in order to progressively develop a complete, common, and reliable model of market value assessment in real time and for all property in Greece.

Sources Aluko, B.T. 2005. Building urban local governance fiscal autonomy through the property taxation financing option. Journal of Property Tax Assessment & Administration 2 (4): 17–32. Craymer, M., and M. Craymer. 1997. Collection of selected articles on geodesy, surveying and land information systems. Surveying and Land Information Systems: Journal of American Congress on Surveying and Mapping 54 (4): 261.

Ball, P. 2001. Recent politics and administration. Choromidis, K. Law of civil-planning (in Greek). Hague, R., and M. Harrop. 2005. Comparative Politics and Administration. Huxhold, W.E. 1991. An introduction to urban geographic information systems. New York: Oxford University Press. International Association of Assessing Officers. 2004. Standard on property tax policy. Chicago: International Association of Assessing Officers. Kousoulas, C. 2001. Cadastre law (in Greek). Larsson, G. 1991. Land registration and cadastral systems: Tools for land information and management.

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3

39

Romaliadis, A. 1991. Urban planning property regulations within or outside the city plans (in Greek). Vettas, E. 2005. Ground: Case-law relevant to land management (in Greek). Zentelis, N., T. Labropoulos, and P. Zentelis. 2000. Development of an urban information system based on cadastral data. Paper presented at the

40

Integrating GIS & CAMA 2000, Urban and Regional Information Systems Association (URISA), Miami Beach, FL, April 16–19. Z e n t e l i s , P. 2 0 0 1 . R e a l e s t a t e : Valuation–appraisals–development–investments–management (in Greek). Zentelis, P. 2003. Cadastre and land information systems (in Greek).

Journal of Property Tax Assessment & Administration • Volume 4, Issue 3