Estimation of the stock of land in OECD countries

1. Estimation of the stock of land in OECD countries Working Party on National Accounts 14-16 October 2008 Young-Hwan Kim, OECD

2. Introduction • Survey on estimates of the stock of land in OECD member countries with a few simple questionnaires in 2007 • Preliminary results were presented at 2007 WPNA meeting • Since the 2007 WPNA meeting, more detailed information has been provided

3. Progress • 15 countries responded, of which 10 countries derive estimates of the stock of land - Australia, Czech Republic, France, Japan, Korea, Slovak Republic, Canada, Germany(preliminary), Finland(Intermittently), New Zealand(rural land) • The Netherlands, Denmark provides their experiences while deriving experimental, unofficial estimates

4. Purpose of estimating the stock of land • To make a complete national balance sheet • Land is the most valuable non- produced asset for most countries • To measure estimates of the capital service provided by land • Land is also considered as a factor of production • Land underlying dwelling attracts much attention for analytical purposes

5. Types of land • According to SNA 93, land (AN.211) is categorized by four types: • Land underlying buildings and structures • Land under cultivation • Recreational land and associated surface water • Other land and associated surface water • Land underlying buildings and structures • Dwellings and non-residential buildings

6. Sectoral breakdown • Five institutional sectors by SNA 93 • Non-financial corporations • Financial corporations • General government • Households • Non-profit institutions serving households

7. Types of land and Sectoral breakdown • Most countries follow the SNA classifications in general • New Zealand publishes estimates of rural land (land under cultivation) only • Canada publishes estimates of only the total of land surrounding residential and non- residential buildings and agricultural land

8. Data sources • Administrative sources • Cadastre maintained by a land registry office, tax authorities or land information centre • Direct sources • Population and housing census, Business survey, Other ad hoc survey

9. Difficulties of valuation • Lack of resources, lack of data, and low priority • Two major issues • How to obtain land prices corresponding with land types • How to separate land values from the total combined value of buildings and structures and land

10. Several methods • The survey has revealed four principal methods for estimating the stock of land: • Direct surveys • Using cadastres • Deriving land residually • Applying land-to-structure ratios

11. Direct survey method • Method of obtaining land values directly from respondents • Including population and housing censuses, business surveys, national wealth surveys including housing wealth surveys, and other ad hoc surveys • Czech Republic, Slovak Republic

12. Direct survey method • An alternative method when volume and price data are unavailable or incomplete • Difficulty in collecting accurate market values • Other data sources needed for better estimation

13. Using Cadastres • Many countries that derive estimates of the stock of land use cadastral data either fully (volumes and prices) or partly (volumes only) • Cadastres are maintained to record the physical status and legal ownership of land, but are often used for taxation purposes • They contain comprehensive information on land such as location, size, use, ownership, value and so on

14. Using Cadastres • Japan, Korea and Finland rely mainly on cadastral data (prices and volumes) • Australia uses cadastral data to estimate some land types, but only uses them to a limited extent in deriving estimates of land under residential buildings • France uses cadastral data in conjunction with other data to obtain its estimates

15. Using Cadastres • Simple, if complete cadastres exist • Problems in identifying the value of land from transaction values that include both land and structures • Details and quality can vary across regions within a country • The possibility of underestimating the value of the stock of land • The levels of the official price usually lag or understate the real market transaction price

16. Land residual method • Subtracting the value of buildings and structures from the total value of buildings and structures including land • It is mainly applicable to urban land, especially land underlying dwellings • The value of the stock of buildings and structures is mainly estimated using the perpetual inventory method (PIM)

17. Land residual method • Australia, France, Canada • The following information is required: •  Total value of dwellings including land • A price index of dwellings including land (House price index) • Net capital stock of dwellings excluding land at current prices • Comprehensive volume information • Certain indicators for sectoral breakdown

18. Land residual method • Netherlands considers using this method to estimate the stock of land underlying dwellings but has not proceeded to publish its estimates • Denmark used land residual method to derive unpublished, experimental estimates of residential land for the period 1995 to 2002 • The estimates of the value of land were negative for some years

19. Applying land-to-structure ratio • Canada uses LSR method to estimate the stock of residential land and non-residential land. • LSR = (APV – BPV) / BPV • BPV : building permit value • APV: absorption price value, which is the sale value of the total residential real estate unit

20. Applying land-to-structure ratio • However, Canada does not exclusively rely on LSR for the residential land • Other sources are used to supplement the LSR methodology. • the LSR approach is labour-intensive and APV come in with a delay • Australia also uses this method on allocating the aggregate estimates of the stock of land, such as commercial and rural land, by institutional sector.

21. Conclusion • These methods are not mutually exclusive • It is evident that a country’s choice of method(s) is largely determined by what data are available in that country for the different types of land • As a result, no general method can be recommended at this stage

22. Conclusion • The best approach is to begin by making an assessment of the available data • Administrative data, particularly cadastres, can provide a good basis for developing estimates • Nevertheless, they should not be used without first being satisfied as to their quality

23. Conclusion • Estimates and methods need to be tested over a reasonable time span, including the ups and downs of the economic cycle, before proceeding to publication

24. Conclusion • Cross-checking the estimates with data from other sources, if possible • Use of (even infrequent) household asset and debt surveys • Comparisons of the (1) ratios of land to structure, or (2) proportion of land to wealth for countries with largely similar characteristics in the real estate market (e.g., Canada and U.S.; France and Germany. • Analysis of the trend-cycle of real estate prices.