USING A GEOGRAPHICAL INFORMATION SYSTEM FOR ... - CiteSeerX

USING A GEOGRAPHICAL INFORMATION SYSTEM FOR PROPERTY VALUATION

Author: Peter Wyatt Position:

Lecturer (currently seconded to the Valuation Office)

Address:

Department of Geography, Surveying and Construction University of Brighton.

Mithras House Lewes Road Brighton BN2 4AT United Kingdom

Phone: Fax: Email:

0272 642394 0273 642392 [email protected]

USING A GEOGRAPHICAL INFORMATION SYSTEM FOR PROPERTY VALUATION

SYNOPSIS This paper argues that property valuers in England and Wales face a difficult task; they must collect and assimilate data from a variety of sources that differ widely with regard to quality. Even within the Valuation Office, where valuers have access to privileged data, valuation has suffered due to the lack of a single comprehensive source of data and the effects of the recession lowering data throughput. There is thus a requirement for the computerised dissemination of property data for valuation purposes.

Furthermore, spatial referencing of such data is regarded as essential to

valuation and therefore the potential of value maps is examined. GIS is proposed as a solution to the problems inherent with manually produced value maps. A methodology is described which shows how a spatial analysis technique available on a GIS was used to examine the influence of accessibility on property value. Computer-generated value maps that illustrate spatial variations in value at the individual property level are possible in the UK because comprehensive and good quality data exists within the public and private sectors of the property market. However, it will take an initiative such as NLIS to ensure that such data is disseminated.

INTRODUCTION The Daily Mail reported that “if the citizens of Northern Ireland and Scotland - not to mention those of the Netherlands, France and Sweden - can find out what their home was worth last time it was on the market, it seems odd that those of us who live in England and Wales can’t” (1). Property data access in England and Wales is poor compared to other countries and this means that propertyrelated procedures such as conveyancing, commercial rent negotiations and valuations are made difficult.

The situation is improving, aided by the National Land Information System (NLIS)

Initiative, and data sharing is increasing. This paper argues that valuation in particular requires access to a myriad of data so that the valuer may make a prediction of value that has regard to all possible influences.

Locational influences on property value are widely regarded as the most important yet their incorporation into valuation methodology is often implicit. GIS-based value maps are introduced as a means of displaying variations in value at the individual property level.

Using commercial

property data collected for a study area and input into a GIS, a means of quantifying the influence of accessibility on retail property value is developed.

This is achieved by calculating a network

accessibility index for the properties in the study area. The analysis of spatial property data at the intra-urban level enhances the valuer's understanding of locational influences on value.

DATA ACCESS Valuers depend on access to comprehensive, reliable and timely evidence of property transactions in order to make informed predictions of value. Such data access does not currently exist within the property market of England and Wales. Despite a wide range of data sources, none can offer the detailed information that valuers require. This is due to legislative restrictions on data release to the public, confidentiality constraints and conservative attitudes. A literature review and questionnaire survey of local authority estate managers and valuers undertaken in 1992 showed that the majority expressed concern over poor access to comprehensive property transaction data.

Property transaction data has traditionally been available via a network of professional contacts. The 'jungle telegraph' can work well if the surveyor has established contacts throughout the property sector and in the area in which he or she practices. However, in many cases it can be difficult to access the required information especially when firms tend to be secretive or the market is sluggish. Other sources include in-house data, building society surveys, market reports, auction results and information published in the press. Respondents to the 1992 survey indicated that the quality and accuracy of these sources could be improved; many are too general, others are out-of-date or are published too infrequently. A solution may be a centralised, accurate, up-to-date and accessible source of property data. To this end, the survey also determined valuers' requirements from a information system that could supply property data on a national scale.

All valuation techniques rely on the collection and analysis of data; general data such as social, economic, planning and environmental attributes, and specific data including local market conditions, details of transactions such as location, physical and functional form and legal characteristics. "The validity of a final estimate of market value depends to a great extent on how well it can be supported by market data" (2).

Indeed, a valuer's experience includes stored

knowledge of past transactions (3). Due to legislative restrictions on the release of property data into the public domain, market data are often difficult to obtain and therefore valuers rely on their own knowledge and experience. When data are available they are often incomplete with legal or negotiating positions of the parties unknown. Confidentiality clauses and the inaccessibility of government data are two artificial barriers to data exchange within a property market that is naturally complex and diverse. A lack of data is thus a significant factor affecting the operation of valuation methodology.

Data requirements for property valuation The property market is unique; government intervention, a fixed supply of land, the long development period for new property and poor information availability distort its operation. A valuation should reflect the assimilation of all known data.

However, the property market is

characterised by a lack of data on which to base investment decisions in comparison to alternative investment markets. Lizieri and Venmore-Rowland stated that "outside the property industry, there is widespread suspicion of the valuation process" and "valuers should seek to widen the set of information used to arrive at an appraisal and reduce their reliance and dependence on incomplete and often inconsistent property market data" (4). Brown, Morrell et al. commented that analytical techniques have been adapted from other investment markets to property but argue that adequate information provision is also required if the property market is to display consistency and compatibility with other investment markets (5). Feenan and Dixon note that the "ability to collect and compile data to analyse is sadly lacking ... partly due to secrecy problems and confidentiality and lack of motivation" (6) and a report by Currie and Scott highlighted the dichotomy between the complex property markets and the data available to analyse them (7). Millington refers to the relationship between property valuation and property data; he argues that the comparative method

is the most widely used valuation technique and that it is essential to have as much information as possible, yet this is often lacking in real world (8). A solution, therefore, would be to improve access to, and the dissemination of, property data.

Land Registry data was opened to public inspection in 1990. Prior to this only the owner or a person who had the owner's consent could inspect the register. Although this represents a significant step towards property data access on a national scale there are certain limitations. Firstly, approximately a third of all properties are unregistered, though most are rural properties and the Land Registry is seeking views on how all property can ultimately be registered. Secondly, the non-registration of tenancy agreements and leasehold interests of less than 21 years means that a significant number of property interests remain unregistered, thus making the Register of limited use for valuation. Thirdly, transaction data are not recorded at present, an omission that significantly reduces its usefulness with regard to valuation. The most definitive database of property transactions is held by the Valuation Office but this is not publicly available.

In summary, two factors complicate valuation; the absence of a central register of sales and confidentiality of information (9).

Institutional and government action should be taken to make

property data more readily available. Improved knowledge of the property market will increase mobility of people and funds and will allow better valuation methods to be developed (10). There are many property market characteristics that make valuation a difficult and often subjective task that is dependent on a high degree of experience and local knowledge, for example the heterogeneity of each interest, economic influences at the national, regional and local scales. These characteristics cannot be altered and indeed make valuation a challenging profession. However it would seem prudent to minimise valuation complexity by improving data accessibility and dissemination.

Information technology and property information Despite poor access to property data, developments in IT and data analysis techniques are developing rapidly. Early implementations of IT in the property market were merely computerised duplications of manual processes.

For example, property management software was based on

accounting programs and valuation software converted traditional valuation techniques into spreadsheet form. These programs have been developed and refined so that they now offer efficient alternatives to manual systems. With particular regard to valuation software, complex valuations and sensitivity analysis can be undertaken to a degree that would have been prohibitively timeconsuming manually. Advancements in IT have led to more analytical applications within the valuation profession; econometric modelling / regression analysis and other advanced statistical techniques have been used to automate valuations with varying degrees of success.

Artificial

intelligence in the form of expert systems, neural networks and, more recently, an object-oriented approach known as case-base reasoning, have also been examined. Having undertaken a detailed state-of-the-art evaluation of these applications it is clear that many depend on access to comprehensive data if they are to provide reliable results and none are able to manipulate spatial property information adequately. These two points are discussed below.

National Land Information System (NLIS) The NLIS initiative seeks to create a system that will coordinate and distribute comprehensive, accurate, up-to-date and regularly maintained information on land and property ownership, use and value for every land parcel in Britain, with links to planning, socio-economic, demographic, and environmental information. Historically, these data have not been publicly available for reasons which include confidentiality, secrecy and cost. Other hurdles include agreement on standards and the appointment of a controlling body.

It is thus appreciated that government intervention is

required in order to facilitate the creation of a NLIS. Yet the introduction of GIS technology in both the public and private sectors may accelerate the pace of change, and some of the problems are being addressed. Technology is no longer a barrier and referencing standards have been developed such as the National Street Gazetteer and National Land and Property Gazetteer. These are now British Standards and facilitate data integration. When operational a NLIS will offer the ability to link previously disparate datasets in new and geographical ways, thus giving the opportunity to reuse data that may have been collected for a specific purpose and hence add value to that data and reduce its net cost of collection.

LOCATIONAL INFLUENCES ON PROPERTY VALUE Assessment of the financial impact that physical and legal factors have on property value, using the comparison method of valuation, is relatively straightforward if sufficient data are available. This is not true for locational influences and the development of a methodology that attempts to measure the impact that location has on value is an important addition to valuation theory. GIS provides a technological platform on which to base such an analysis and an initial stage is the spatial representation of property information in the form of value maps, discussed below. Surveyors base values on locational criteria yet the potential of GIS for such an operation remains largely undiscovered (11). Early GIS consisted of a spatially referenced database, simple data manipulation functions and geographical display.

However, many GIS now incorporate spatial analysis

techniques as 'modules' which can be used to examine data that has been recorded in a geographically referenced database. Such techniques can only be relied upon when the data are more accurate, timely, and comprehensive. Therefore, the NLIS initiative is timely.

Value maps A value map is "a cartographic or spatial representation of statistical data which reflects the value of property" (12). Value maps are rare in this country but they are well known in other countries, many of which use them for property taxation purposes. Research that uses value maps is also rare although interest in the geographical representation of property values has increased since GIS have been widely available. Potential uses of value maps include; -

planning (ensure optimum use of land and property, aid compulsory purchase and

compensation) -

taxation, in Denmark the public can check their assessments by viewing value maps and transaction details (Clarke 1965)

-

property management and asset valuation

-

identification of areas of high value for investment and development purposes

-

enable the public sector to purchase land at a low price

-

show changes in value over time, perhaps due to blight, green belts or zoning

-

an aid to valuation.

Value maps are powerful decision-making tools that illustrate the geography of property values over space and time in a way that most property professionals find intuitive, namely in map form. They were first used in Britain by Anstey (13) who produced value contour maps to show how property development would alter the pattern of land values. Figure 1 provides an example of one of Anstey’s value contour maps. Value map research continued with the work of Howes (14) who investigated examples of value maps that had been produced in this country and abroad. Figures 2 and 3 are two examples of value maps that were produced by planning departments of local authorities in this country.

All of the maps were generalised, related to one land use only and a specific point in time. Problems included choice of scale, presentation, availability of data and valuation difficulties such as out-ofdate values or anomalies. These problems are typical of traditional map production and have the effect of dissuading valuers and planners from producing value maps. It appeared that each map was produced in isolation - underlining the need for data sharing and, indeed, according to Howes, many chief planning officers felt that confidentiality restricted their decision-making. Traditional (non-GIS) value maps are out-of-date by the time they are produced and only relate to a single point in time. Often they are too general to be of use for valuation purposes where individual property details are required. They only display the pattern of values and not the causes, and data collection and map production are time consuming and expensive. Despite these drawbacks, Howes believed that value maps were useful aids and investigated different methods of displaying property value data geographically. Figures 4 illustrates one of his methods. It is now possible to approximate floor areas from OS digital maps using a GIS, a valuable facility now that floor-space statistics are no longer published by the Department of the Environment. This allows the presentation of individual property values per unit area rather than using grid squares which bear no relation to property boundaries.

GIS and value maps GIS are now able to address some of the problems inherent with traditional value maps by producing them efficiently and as part of a wider suite of data analysis techniques. Howes acknowledged the

limitation of a value map with regard to its information content but the functionality of GIS means that a value map need not be a static display but can form part of a more comprehensive analytical process.

Advances in IT mean that computer-based maps allow frequent updating, quicker

production time and greater analytical potential. A GIS is quicker, cheaper, more versatile and adds value to data collected for other uses. Problems in the past related to lack of staff resources, the production of expensive paper maps and the complex calculations involved. A GIS can manage the data input, calculations and presentation of the results in a graphical format. Dixon (15) highlighted the value GIS might be to valuers by giving a spatial dimension to their information.

A SPATIAL PROPERTY INFORMATION SYSTEM FOR VALUATION The comparison method of valuation is the most widely adopted in practice and involves the assessment of value by comparing it with evidence of transactions for similar properties. Every method of valuation has recourse to sales comparison to a certain degree. It is regarded in English and North American valuation theory as the most reliable method (16). Richmond (17) states that "this procedure is widely adopted in practice, but requires the keeping of adequate records of transactions."

Using the comparison method of valuation, physical, legal and market factors can be compared directly but every property is spatially unique, so spatial factors require an alternative method of adjustment. Fraser (18) says "of dominant importance in understanding the demand for any urban property is its location, both in a regional as well as a local sense." There is an adage, common within the property industry, that says there are three crucial factors which determine the success of a property; location, location and location (19). It has been handled subjectively up until now, for example, Millington (20) writes "a skilled valuer will be able to quantify the differences in value caused by a different geographical situation." The aim of the research at the University of Brighton was to develop a spatial property information system for a study area in order to assess the potential that a NLIS holds for valuation and to devise a more explicit approach to the analysis of spatial influences on value.

Physical analysis and geographical display of property values Figure 6 provides an overview of the methodology. The physical analysis (1) occurs through a dialogue with an interface (a). A general profile of the physical characteristics of the property to be valued (the subject property) is elicited from the user (b), after which a database of properties (including the address, detailed description and price / value information) is searched and all those that match the profile of the subject are selected and displayed in turn (c). This display stage provides the user with an indication of the type of comparable that is being searched. More detailed questions are then asked about the subject property in order to make quantitative and qualitative assessments of factors considered to influence value (d). The answers given are used to measure differences between value factors of the subject property and each comparable property in turn. If differences are large the comparable is deselected because it is fundamentally different from the subject property. If differences are within predetermined tolerance limits then adjustments are made to the values of the comparable properties to take these differences into account (e). If there are no differences then no adjustment is made; such properties would be good comparable evidence, although rare in practice. Upon completion of this stage the result is a set of comparable properties that are written to a new data file (f) whose values have been adjusted to take into account all physical differences between them and the subject property. Consequently, variations in the values of the comparable properties are explained solely by locational factors and therefore these values are referred to as 'locational values' henceforth. The locational values provide the foundation upon which the spatial analysis (2) is based.

Using the GIS (g and h) the comparable properties are displayed on a map and shaded according to their locational values (i), figure 7 is an example. This provides a visual representation of the spatial distribution of comparable properties and the magnitude of their locational values. The geographical display of the selected comparable properties also aids spatial analysis (j, k and l). Depending on the location of the subject property it is possible to assess its value and also determine which properties provide the best comparable evidence, not only in terms of physical value factors but also in terms of location.

The research highlighted the issues that must be addressed when recording and analysing locational value factors, for example, the best method of spatially representing property interests that exist on different floors of the same building; a feature common to commercial property.

Locational analysis of property values Accessibility evaluates the ease with which contacts may be made in terms of net economic benefits derivable from a particular location. The more accessible a location the greater the benefits and/or the lower the costs. Accessibility is a relative quality accruing from a location's relationship to the transport system and affects volume of business (turnover and revenue) and therefore the rent or price paid for a location. The level of demand for an accessible location depends on the type of property and activity that is proposed for that location. Those activities that rely on the movement of customers benefit the most from maximum accessibility.

Consequently, retail activities tend to

outbid competing land uses for the most accessible locations in the central urban area. Further, specific retail trades compete with one another for the most accessible sites such is the strength of association between accessibility and turnover. Therefore, retail location depends on customer access and traders bid the highest prices for those sites with maximum accessibility.

An accessibility index is an aggregate measure of how reachable a location is from other locations. It is a relative measure for comparing locations. The accessibility index of an origin is computed by adding the effects of all other locations. This technique is advocated by the Appraisal Institute (21) who comment that the quality of a property's location can be quantified by calculating the time-distance relationships or linkages between the property and all possible origins and destinations of customers coming to and going from that property. The influence of accessibility on value was investigated by constructing a network 'accessibility' index using a gravity modelling technique available on Arc/Info_ GIS and calculating the relative accessibility of each shop in the study area. The index values were found to correlate significantly with the locational values of the properties.

The development of a spatial property information system allows a more explicit approach to the quantification of the effect of location when reconciling property values as part of the comparison method of valuation. The influence of location on value is regarded by many as the most important value factor, yet spatial analysis of comparable evidence is implicit when valuers use their local knowledge and experience to financially adjust the values of properties that are in different locations. This research makes spatial adjustment more explicit by proposing that a statistically significant proportion of the spatial element of retail value can be attributed to each property's relative accessibility.

In a typical valuation the values of selected comparable properties are adjusted both physically and spatially so that they provide valuation evidence for properties that are notionally the same as the subject. The adjusted values of the comparable properties provide a 'value range' within which the predicted value of the subject may lie.

The comparison method of valuation relies on the

identification, quantification and adjustment of value factors which is usually a straightforward procedure for physical characteristics after the property has been inspected, but relies on a qualitative assessment for location unless the property is analysed with regard to its spatial context. The methodology introduced above emphasises the analysis of the spatial element of value and argues that an index of accessibility at the intra-urban level can provide a measure of relative spatial value.

GIS provides a technological platform on which to base such an analysis, indeed, the

Appraisal Institute of America believes that "GIS is expected to have a strong impact on traditional location analysis just as computerised discounted cash flow modelling has dramatically influenced financial analysis in the past two decades" (22).

The methodology does not assume, as traditional urban location theories do, that the point of maximum accessibility is at the centre of the urban area. Rather it simply correlates the accessibility index value for each property to locational values. This is more realistic as the point of maximum accessibility can no longer always be regarded as the geographical centre of an urban area. Accessibility advantages in small urban areas can also be measured whereas traditional urban economic theories find this difficult.

CONCLUSIONS Two recent developments have occurred that emphasise the need for this research. Firstly, the NLIS initiative and the potential release of comprehensive property information in England and Wales. Such data release has been requested by valuers for many years and will aid the assimilation of comparable evidence for valuation purposes. Secondly, the development of GIS which facilitates the unified storage, manipulation and analysis of property data, both spatially and aspatially, in ways that previously required many time-consuming operations. Spatial influences on value are often examined implicitly in traditional valuation methodology because of difficulties in spatial data manipulation.

There are still many hurdles to overcome; with regard to the private sector,

confidentiality and secrecy and their consequent effect on data sharing are major issues and in the public sector legislation is required to release transaction data into the public domain. This raises the issue of whether this should occur at the individual property level or aggregated to some degree. Advantages for valuers include access to comprehensive disaggregated data in a format that is intuitive (map form), regularly updated and facilitates the investigation of the spatial element of property value.

LIST OF REFERENCES 1.

Stone, M. Why the price secrets of a house should be revealed to everyone. Daily mail. 1994, August 10; 25.

2.

Appraisal Institute. (1992). The Appraisal of Real Estate. Chicago, The Appraisal Institute.

3.

Anstey, B. An introduction to the theory of value: land and property valuation. Building. 1970, January 16; 63-66.

4.

Liziei, C. and Venmore-Rowland, P. Valuation accuracy: a contribution to the debate. Journal of Property Research. 1991, vol 8; 115-122.

5.

Brown, S., Morrell, G., Unsworth, R. and Ward, C. Property Indices Research Report. Society of Property Researchers. 1984, Technical Paper Number 1.

6.

Feenan, R. and Dixon, T. Information and Technology Applications in Commercial Property. Macmillan, London 1992.

7.

Currie, D. and Scott, A. (1991). The place of commercial property in the UK economy. London Business School.

8.

Millington, A. An Introduction to Property valuation. Estates Gazette, London 1990.

9.

Accuracy in valuation. Estates Gazette. 1985, April 20, 274; 295.

10. Harvey, J. The Economics of Real Property. Macmillan, London 1981. 11. Dixon, T. IT Applications in Property. Mapping Awareness and GIS in Europe. 1992 vol 6 no 2; 49-51. 12. Howes, C. Value maps: Aspects of Land and Property Values. Geo Books, Norwich 1980. 13. Anstey, B. Value contour maps: a new tool for planners. Planning Outlook. 1949, Iss 2; 29-39. 14. Howes, C. Value maps: Aspects of Land and Property Values. Geo Books, Norwich 1980. 15. Dixon, T. IT Applications in Property. Mapping Awareness and GIS in Europe. 1992 vol 6 no 2; 49-51. 16. Turner, D. An Approach to Land Values. Geographical Publications Ltd., Cambridge 1977. 17. Richmond, D. Introduction to Valuation. Macmillan, Hong Kong 1986. 18. Fraser, W. Principles of Property Investment and Pricing. Macmillan, London 1991. 19. Britton, W., Davies, K. and Johnson, T. Modern Methods of Valuation of Land Houses and Buildings. Estates Gazette, London 1989. 20. Millington, A. An Introduction to Property valuation. Estates Gazette, London 1990. 21. Appraisal Institute. (1992). The Appraisal of Real Estate. Chicago, The Appraisal Institute. 22. Appraisal Institute. (1992). The Appraisal of Real Estate. Chicago, The Appraisal Institute.

Figure 1 - Barbican, overlaid by isovals

Figure 2 - relative shop values in Lewes

Figure 3 - rateable values per street block

Figure 4 - aggregate rateable value projection (Howes 1980).

general profile (b)

collect comparables (c)

Stage one: selection and specific profile (d) PHYSICAL ANALYSIS (1) physical reconciliatio of comparab adjust comparable values (e) write selected and properties adjusted comparables to a new data file (f)

interface dialogue (a)

measure effect of accessibility on value (l)

accessibility index (k)

join to geographical data (g)

SPATIAL ANALYSIS (2)

spatial analysis (j)

GIS (h)

value map display (i)

Figure 5 - overview of operation.

Stage two: spatial analys of the selecte comparable properties