Does the planning system affect housing prices? Theory and with evidence from Hong Kong

ARTICLE IN PRESS

Habitat International 27 (2003) 339–359

Does the planning system affect housing prices? Theory and with evidence from Hong Kong Eddie Chi-man Hui*, Vivian Sze-mun Ho Department of Building and Real Estate, The Hong Kong Polytechnic University, HungHom, Kowloon, Hong Kong Received 16 April 2002; received in revised form 16 August 2002; accepted 19 August 2002

Abstract Hong Kong has been faced with property market ups and downs for many reasons over the past 10 years or so. It once experienced shortages of housing supply, leading to prolonged escalating prices. This was followed by a downturn of housing prices since the end of 1997 and in the wake of the financial turmoil in Asia. Hong Kong people have been very concerned with the housing market situation, particularly housing prices, and consider that the problem needs to be addressed to prevent its recurrence in the future. The Hong Kong government uses its land-use planning system to address this ‘‘market failure’’. Also, such a system imposes constraints on land supply and development, through various means by altering the supply of housing land. As a result, the residential property has been affected, and so is the housing price. This paper attempts to explore the impact of the land-use planning system (and land supply) on housing prices in the territory. It starts with the preliminary introduction of the planning system and a review on the previous literature on the impact of land-use planning system on housing prices in Hong Kong. An analytical model is established, followed by the discussion of the empirical findings of the impacts of the land-use regulation on land supply, housing supply and housing prices in the territory. The results in Section 7 show that the planning indicators: (i) the approval rates of planning applications; (ii) the residential floor area under planning applications; and (iii) the area of greenbelt and open space zones, have significant impact on housing prices. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Land-use planning system; Housing supply; Housing prices; Planning applications

*Corresponding author. Tel: +852-2766-5881; fax: +852-2764-5131. E-mail address: [email protected] (E. Chi-man Hui). 0197-3975/03/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 1 9 7 - 3 9 7 5 ( 0 2 ) 0 0 0 4 2 - 5

ARTICLE IN PRESS 340

E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

1. Introduction Hong Kong faces a shortage of housing supply, primarily due to limited developable or redevelopable land relative to the keen demand for different types of land uses by its sizable population. Hong Kong has an area of 1098 km2, of which 184 km2 are developable lands (Census and Statistics Department, 2000a, b). Residential land accounts for only 32% of the developable land in Hong Kong. With a total population of 6,720,2001 (the number of households in Hong Kong is about 2,050,800), there is a persistent demand of land supply for housing. This has led to the high population density, reaching an average population density of 63102 persons per square kilometer, compared with other cosmopolitan cities such as London (4483 persons per square kilometer), Singapore (6088 persons per square kilometer) and Tokyo (5384 persons per square kilometer). On the other hand, the housing shortage results in prolonged escalating prices, partly fueled by speculation. Loh (1998) criticized that the Government controls land supply and its policies ‘‘have not prevented a housing shortage and have led to extremely high land and property prices’’. As land is so scarce in Hong Kong, the appropriate use of land to meet different land-use demand becomes an important issue. The government uses its land-use planning system to regulate or impose constraints on land supply and development—to address the problems arising from ‘‘market failure’’. The planning system in Hong Kong has been subjected to most criticisms. For instance, developers have complained that ‘‘land approval procedures were slow, bureaucratic and inherently anti-development’’ (South China Morning Post, 6 June 1997). Lai (1997) pointed out that the reasons put forward by the Town Planning Board in rejecting planning applications are often so vague and in general, there is no assured method by which an applicant can revise the original submission in order to get through the approval process in a new application or review. This agrees with Staley (1994) that the planning application process increased uncertainty in the development process, since public administrators had discretion over determining the types, pace and pattern of development on district level. The government itself is dissatisfied with the fluctuating property market in the past with property values either rising too rapidly or falling drastically, which leads to many undesirable social problems (Ho, 2001). Given the de facto long-term constraints on land supply, the Government recognises the need to increase supply to provide affordable housing for households. The Housing Branch (1997) has also acknowledged that the ‘‘smaller the gap between the supply and demand for private housing, the lower the pressure on domestic property prices’’. The paramount questions for the Government are to resolve the dilemma between (1) the policy of providing adequate and affordable housing against the shortage of housing supply and (2) the land-use regulations for control on residential development. All these questions relate to the landuse planning system. Land-use planning is regarded as a mechanism for the government to exercise its control on the urban development process. There are theories, which advocate that land-use regulations (such as zoning and growth controls) may affect property market by constraining supply and increasing 1

The total population in 1999 was 6,720,200 (as quoted from the Hong Kong Annual Digest of Statistics, 2000). The population density in 2000 was 6310 persons per square kilometer (as quoted from the Census and Statistics Department’s webpage). 2

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

341

demand. Gerald (1992) states that the planning system restricts land supply in four main ways: (1) restricting the total quantity of housing land made available; (2) restricting the location of land that is made available; (3) restricting the way that the available land is developed; and (4) changing the timing of development. A limited number of empirical researches have attempted to test this relationship in order to capture the real impact of government regulation on housing supply. Mayo and Sheppard (1996) undertook a comparative study on the impact of regulatory control on housing supply in Malaysia, Thailand and Korea. These countries adopt a different degree of development control. Korea has a relatively strict development control while Thailand has little effective regulation of development. Malaysia offers an intermediate case. Empirical findings confirm that countries with more restrictive planning systems have reduced supply elasticities. The paper suggests that the ‘‘stochastic development control’’ (this refers to a situation which exists when governments regulate and limit the nature, timing, or extent of housing development in a way which cannot be forecast or known with certainty) introduced by the planning bureaucracies may alter the structure of housing supply. Subsequent to Mayo and Sheppard’s study, Mayer and Somerville (1999) established a structural model describing the relationship between land-use regulation and residential market supply. They considered two types of land-use regulations, namely, those that impose explicit financial costs on builders (development or impact fees) and those that delay or lengthen the development process. The model shows that in the absence of input prices for land, housing starts are properly specified as a function of changes in the level of house prices, and not as a direct function of the level itself. Based on the quarterly data on a panel of 44 US metro areas, the authors found that land-use regulations have a significant effect on both the steady-state level of new construction and the responsiveness of local supply to price shocks. In order to identify the key variables of land and housing price determination for the purpose of formulating the analytical model, it is necessary to draw upon the overseas literature due to limited local studies. There are a considerable amount of studies done in the 1970s and 1980s. A regression study on New Jersey communities by Sagalyn and Sternlieb (1973) concluded that zoning increases housing costs. Seidel (1978) argued that the increase in housing prices is caused by the widespread adoption of local ordinances restricting the rate and form of development. Buttler (1981) investigated the impact of land-use zoning on housing rent, land rent, the design parameters of a building and population density in a residential city. A theoretical land-use model was used. The study shows how both land-use zoning and housing attributes determine housing rent, and rent and design parameters in equilibrium. Based on the previous researches, a number of refined studies have been undertaken in the 1990s. Bramley (1993) studied the impact of land-use planning and tax subsidies on the housing supply and price in Britain. He performed a cross-sectional analysis at the inter-urban level to express the degree of variation in the supply elasticity and in the impact of policy measures on local housing markets. A model was developed to represent the effects of planning policies on the supply elasticity. Evans (1996) subsequently raised questions concerning Bramley’s method of using cross-sectional data for time series analysis. One of the main flaws, which Evans raises, is that the data collected for Bramley’s analysis were in the year when England’s housing prices were at their peak, thus creating a moderately biased environment. In response to the comments, Bramley (1998) further examined the indicators of planning constraint and its impact on housing land supply. A range of quantitative and qualitative measures of planning restraint developed in

ARTICLE IN PRESS 342

E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

the context of cross-sectional modelling of housing supply in England were examined. These indicators were assessed, first, in a priori terms, second, in terms of their interrelationships, and, third, in terms of their performance in statistical explanation of variations in a number of outcomes relating to the supply of land with planning permission for housing, new housebuilding, the share of urban land, density and house prices. It is found that certain quantitative measures are crucial in modelling land supply and housing development outcomes. The structure-plan housing provision ‘numbers’ and the amount of land zoned in local plans are important for the flow of land into the system, whereas the stock of land with permission is the key driver of a new building in conjunction with demand and/or prices. Moreover, four main dimensions of planning policy and constraint were identified by means of factor analysis, namely, ‘the dimension of land area which is or is not subject to formal constraints of existing built-up area, greenbelt or AONB’; ‘informal constraints’; ‘reduced provision’; and ‘environmental capacity and/or subjective restraint’. Hannah, Kim, and Mills (1993) also analysed the effects of land-use controls over land supply on housing in Korea. Through a case study of five Seoul development projects, it is found that a substantial part of the rise in housing prices has resulted from the government’s tendency to underallocate land for residential use. Monk and Whitehead (1996) also undertook a similar study on the impact of land-use planning controls on the supply, price and type of dwellings provided. The study reveals that planning raises the land and housing prices by affecting land supply in different locations, densities, type and mix of houses built in different localities, and speculative behaviour and volatility. To a large extent, most of the overseas literature suggests a relationship between land-use regulation and housing supply/price. Land-use planning affects the property market by restraining the location and the density of supply, thus increasing price. Empirically, land-use regulation has an effect on housing supply elasticity, though the degree of impact will vary in different contexts. Land availability tends to be an important variable in affecting the property market. However, in Hong Kong, there has never been any research adopting econometric model to study the effects of the land-use planning system on the housing market. Only Tang and Choy (2000) tried to use a logistic regression analysis in office development applications in urban Kowloon of Hong Kong. Most researchers tend to adopt a quantitative approach in analysing the relationship between land supply, housing demand and housing supply, disregarding the planning issues (Peng & Wheaton, 1994; Tse, 1998; Tse, Ho, & Ganesan, 1999; Hui & Lui, 2002). For example, Tse (1998) conducted a study to examine the impact of land supply on housing prices in Hong Kong. The Granger causality test was used in the study, fitted with annual data from 1976 to 1995, to test whether land supply affects housing prices. The results show that there is no causality between land supply and housing prices in Hong Kong. A later study by Tse et al. (1999) focused on the determinants of house prices and investment demand for residential property. Specifically, it examined the role of population growth, transaction volume, inflation and interest rate in determining house prices. A reduced-form equilibrium model to explain change in house prices in Hong Kong was developed. For qualitative study, there are numerous studies discussing various aspects of the planning system (Bristow, 1984; Planning Department, 1995; Lai, 1996, 1997, 2000; Tang & Leung, 1998; Yeung, 1998). All this concerns the debates of the planning policy and the statutory planning enforcement mechanism.

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

343

In view of the above, housing prices are affected by economic, demographic and other supply and demand factors. Overseas literature considers that planning also has a significant impact on housing market by affecting locations, densities and uses of land supply. Given below is a summary of the critical planning factors that have been mentioned in previous paragraphs: * * * * *

the the the the the

adoption of local ordinances restricting the rate and form of development (Seidel, 1978); housing provision numbers noted on the structure plan (Bramley, 1998); amount of land zoned (Bramley, 1998); stock of land with planning permission (Bramley, 1998); and government’s tendency to underallocate land for residential use (Hannah et al., 1993).

These can act as guidelines for the establishment of the planning indicators which will be mentioned in our model specification. Although the housing problem in Hong Kong has attracted considerable scholarly attention, there appears to be limited research on the effects of the land-use planning system on the land and housing markets except a few overseas studies (Bramley, 1998; Gerald, 1992; Monk & Whitehead, 1996). What are the impacts of land-use planning on the housing market in Hong Kong? As the government adopts its land-use planning system to regulate and impose constraints on land supply and development, it is important to provide sufficient housing against the shortage of housing supply. As a backdrop, the mechanism of the planning system and the performance of the planning department in terms of the zoned area and the planning application approval rate will be discussed in Section 2. In order to fill this gap of knowledge, this study attempts to investigate the relationship between the land-use planning system, land supply and housing prices in Hong Kong; how the ‘‘land-use planning system and constraints’’ influence land supply in general and the residential property market in particular. Based on the findings of the empirical study, it is hoped to draw some insights for policy direction in the land-use planning and the residential market in Hong Kong.

2. Hong Kong’s planning system Planning activities mainly fall into three broad categories: forward planning, which mainly involves the preparation and updating of regional strategies, county structure plans, and local plans; development control, which involves local planning authorities responding to applications for the right to undertake specific developments from land owners or developers; implementation, which involves direct industrial, commercial or housing development by or involving public authorities (Bramley, Barlett, & Lambert, 1995). As mentioned by Dale and McLaughlin (1999), there are two basic approaches to regulating how land is developed and used. This can be by way of legislation applying to all properties uniformly, or by way of a permit system in which a property owner must make application at the time of a proposed development. Nnkya (1998) puts forward that ‘‘Land-use planning can be defined as a programme of state intervention in land use and environmental change to mediate conflicts of interests over how land should be used, developed, and coordinate individual activities which if left to proceed otherwise would lead to an environment for living that is characterized by negative externalities, inefficient use of land and services, inequity and unfair distribution of resources’’.

ARTICLE IN PRESS 344

E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

The planning application system is one of the key features of the administrative framework of land-use planning in Hong Kong. Although the administrative procedures associated with the processing of planning applications are clearly stipulated within the Town Planning Ordinance, with the right of applicants to seek a review and an appeal of the Town Planning Board’s decisions being clearly specified, the arbitrary nature of planning in Hong Kong as compared to other professional disciplines, and the uncertainty associated with the planning application process within the property development process are usually criticised. Land-use planning is carried out at three levels, namely, territorial, sub-regional and district planning. This is known as the three-tier planning system, which comprises the territorial development strategy, the sub-regional development strategies and the district plans. Guiding the preparation of these plans is the Hong Kong Planning Standards and Guidelines, which is a Government manual of criteria for determining the scale, location and site requirements of various land uses and facilities (Hong Kong Government, 1994). Statutory plans are prepared by the Town Planning Board (TPB) under the Town Planning Ordinance (Chapter 131). There are two types of statutory plans, namely, Outline Zoning Plans (OZPs) and Development Permission Area (DPA) Plans. The OZPs, which are prepared under Section 3(1)(a) of the Town Planning Ordinance, show the proposed land uses and major road systems of individual planning scheme areas. Areas covered by such plans are zoned for such uses as residential, commercial, industrial, open space, government, institution and community uses, green belt or other specific purposes (Section 4 of the Town Planning Ordinance). Attached to each OZP is a Schedule of Notes showing the uses which are always permitted (Column 1 uses) in a particular zone and other uses for which the TPB’s permission must be sought (Column 2 uses) (Information Services Department, 2000). Indeed, enforcement of town planning proposals under OZPs is largely a matter of control under Buildings Ordinance (through building plan approval process) and Government leases. Bristow (1984) asserted that ‘neither is wholly effective, and neither is fully under the control of planners—a source of some dissatisfaction’. 2.1. The planning application system As mentioned above, attached to each OZP/DPA is a Schedule of Notes showing the uses which are always permitted (Column 1 uses) in a particular zone and other uses for which the TPB’s permission must be sought (Column 2 uses). Section 16 of the Town Planning Ordinance enables the TPB to grant permission for uses under Column 2 of the Notes. The TPB shall consider applications within 2 months and may approve the application with or without conditions or refuse to grant permissions applied for. If an applicant is aggrieved by the decision of the TPB, he has the right to seek a review of the TPB’s decision and has the right to a hearing by the TPB under Section 17 of the Town Planning Ordinance. In the event that an applicant is again aggrieved by the TPB’s decision upon review, the applicant has the right to appeal to an independent Appeal Board under Section 17B of the Town Planning Ordinance. If a proposed use does not fall within Column 1 nor Column 2 in a particular zone, an application for rezoning, which is not provided for under the provision of the Town Planning Ordinance, must be submitted to the Town Planning Board for consideration.

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

345

The existing use of building or land is permitted to continue without any planning permission until redevelopment or a change of use takes place. Redevelopment or change of use may only be carried out if it conforms to the plan, or if required, after planning permission has been obtained. In considering a planning application, the TPB would usually take into account such factors as the planning intention and Government policies, social, economic and environmental impacts of the development on the wider area, traffic and infrastructural implications and compatibility of land uses (Planning Department, 1995). Lai (1997) criticised the arbitrary nature of planning in Hong Kong as compared to other professional disciplines like engineering or building and suggested that this is best illustrated by a comparison between a planning application and a building application. He pointed out that the reasons put forward by the TPB in rejecting planning applications are often so vague and general that there is no assured method by which an applicant can revise the original submission in order to get through the approval process in a new application or review. Consistent with the contention put forward by Lai (1997), Staley (1994) asserted that the planning application process increases uncertainty in the development process, since public administrators have discretion over determining the types, pace and pattern of development on district level. As the Hong Kong’s planning department has its own discretion in approving the planning applications, it controls the approval rate. Fig. 1 shows the number of approved and nonapproved residential planning applications from 1988 to 2000. These are the total number of planning applications applied for residential use under Sections 16, 17 and renewal cases. The non-approved cases include the rejected, invalid, no need to reply and wait for updating cases. It can be seen that the pattern of approval cases is similar to the pattern of the total number of planning applications. It means the number of planning applications is proportionate to that of approved cases. From 1988 to 1993, the number of non-approved cases was always higher than the number of approved cases. However, the trend from 1993 to 2000 demonstrates that the planning department had a tendency to approve more cases when the number of planning applications increased. The number of approved cases was always higher than the number of nonapproved cases during that period. The average approval rate between 1988 and 1993 was about 42%, whereas the average approval rate between 1993 and 2000 reached 63% (see Fig. 2). This increase in approval rate may be due to two reasons. First, the planning department seems to have been more efficient in processing planning applications and more lenient when making decisions over time. Both planners and applicants have become more knowledgeable about the application system, which leads to higher success rate. This is echoed by Dale and McLaughlin (1999) that the role of planners and administrators is gradually changing from that of arbiter of the public interest to one of facilitating amongst the range of interest groups that collectively define the public interest. Second, the planning department has taken into consideration more the increasing demands in society than before. From 1993 onwards, the planning department has been facilitating more urban development and/or redevelopment in the territory. It has been suggested that a higher approval rate is in response to the economic, social, and environmental demands being placed upon the land and to involve individual citizens and the community as a whole in the planning process (see Dale & McLaughlin, 1999). Not only does the planning department control the approval rate as discussed, but also the approved residential floor area which has a bearing on the density of development. Fig. 3 shows the patterns of the number of approved cases and the residential gross floor area (GFA) under

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

346

Number of approved and non-approved residential planning applications 180 160

No. of planning applications

140 120 100 80 60 40

0

1988 Q1 Q2 Q3 Q4 1989 Q1 Q2 Q3 Q4 1990 Q1 Q2 Q3 Q4 1991 Q1 Q2 Q3 Q4 1992 Q1 Q2 Q3 Q4 1993 Q1 Q2 Q3 Q4 1994 Q1 Q2 Q3 Q4 1995 Q1 Q2 Q3 Q4 1996 Q1 Q2 Q3 Q4 1997 Q1 Q2 Q3 Q4 1998 Q1 Q2 Q3 Q4 1999 Q1 Q2 Q3 Q4 2000 Q1 Q2 Q3 Q4

20

Year approval cases

non-approval cases

total number of applications

Fig. 1. Number of approved and non-approved residential planning applications.

The approval and disapproval rate of planning application cases 100%

Percentage

80% 60% 40% 20% 0%

88 89 90 91 92 93 94 95 96 97 98 99 00 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 2 3 4 1 1 1 1 1 1 1 1 1 1 1 1 1

of approval cases

44 33 60 60 33 56 33 55 22 41 46 50 32 40 21 34 38 33 35 41 40 45 56 61 58 50 56 51 56 60 57 54 64 49 70 75 74 65 59 56 54 57 60 79 72 79 73 67 71 58 67 75

of disapproval cases

56 67 40 40 67 44 67 45 78 59 54 50 68 60 79 66 62 67 65 59 60 55 44 39 42 50 44 49 44 40 43 46 36 51 30 25 26 35 41 44 46 43 40 21 28 21 27 33 29 42 33 25

Year of approval cases

of disapproval cases

Fig. 2. The approval and disapproval rate of planning application cases.

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

347

7000000

120

6000000

100

5000000

80

4000000

60

3000000

40

2000000

20

1000000

0

square metre

140

1988 Q1 Q2 Q3 Q4 1989 Q1 Q2 Q3 Q4 1990 Q1 Q2 Q3 Q4 1991 Q1 Q2 Q3 Q4 1992 Q1 Q2 Q3 Q4 1993 Q1 Q2 Q3 Q4 1994 Q1 Q2 Q3 Q4 1995 Q1 Q2 Q3 Q4 1996 Q1 Q2 Q3 Q4 1997 Q1 Q2 Q3 Q4 1998 Q1 Q2 Q3 Q4 1999 Q1 Q2 Q3 Q4 2000 Q1 Q2 Q3 Q4

Number

Total number of approved cases and the residential GFA under application

0

Year approval cases

residential GFA under planning applications

Fig. 3. Total number of approved cases and the residential GFA under application.

application from 1988 to 2000. The patterns shown are irregular. For example, in 1998 Q4, when residential floor area under application decreased substantially, the number of approved cases increased. In 1999 Q4, however, the trends of both patterns decreased. The residential floor area under application reflects the market demand. The patterns illustrate the planning application approval situation under the fluctuating market demand. In fact, the planning department does not impose a constant limit on development density and approval rates. The department assesses each unique case independently based on its individual merits and its fulfilment of the statutory requirements other than the market sentiment. 2.2. Zoning Zoning entails the delineation of a community into zones within which certain uses are permitted and others are prohibited. It provides a set of standards with respect to the uses to which a parcel of land may be put, and the size, type, and placement of improvements on that parcel (Dale & McLaughlin, 1999). This offers the construction industry with certainty: developers know where they can build, what they can build and how much developable land is available (Gallent & Kwang, 2001). Fig. 4 illustrates the percentage of residential zoning and greenbelt and open space zoning of the total planning scheme area, from 1988 to 2000. The percentage of residential zoning falls within 10–15%, whereas the greenbelt and open space zoning remains at about 30%. Based on the TPB Guidelines, there is a general presumption against development (other than redevelopment) in greenbelt zoned areas. Usually, only a limited scale of

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

348

% of residential and greenbelt and openspace zoning 50% 45% 40%

Percentage

35% 30% 25% 20% 15% 10%

0%

1988 Q1 Q2 Q3 Q4 1989 Q1 Q2 Q3 Q4 1990 Q1 Q2 Q3 Q4 1991 Q1 Q2 Q3 Q4 1992 Q1 Q2 Q3 Q4 1993 Q1 Q2 Q3 Q4 1994 Q1 Q2 Q3 Q4 1995 Q1 Q2 Q3 Q4 1996 Q1 Q2 Q3 Q4 1997 Q1 Q2 Q3 Q4 1998 Q1 Q2 Q3 Q4 1999 Q1 Q2 Q3 Q4 2000 Q1 Q2 Q3 Q4

5%

Year % of greenbelt and openspace zoning

% of zoning involves residential use

Fig. 4. % of residential and greenbelt and open space zoning.

development will be approved on a supportive planning ground. This is, however, considered to be similar to the open space zone. The TPB Guidelines have not covered the open space zone. Therefore, the greenbelt and open space zones are usually designated as landscaping areas or recreational uses. It is essential to maintain at a certain percentage as it serves as a buffer between and within urban areas (Town Planning Board Guidelines, 2001).

3. The model Section 2 has discussed all the determinants that affect housing prices, including planning and economics factors. These determinants apply to Hong Kong and also the region. This section aims to establish a model for analysing the relationship between housing price and these factors. In order to study the effect of planning system on the housing demand and housing supply, we attempt to integrate planning indicators into the econometric analysis of the housing market. Political changes may have a bearing on the housing market; and the return of Hong Kong’s sovereignty to China might have affected housing prices. However, that was not the case. Housing prices remained steadily upwards in 1997 until the impact of the Asian financial turmoil. This has caused subsequent effects. First, there was an acute downturn of economic performance of the territory. That led to a dramatic decline in prices and transaction volume. All these changes can be proxied by the economic performance. We adopt time series regression to identify the relationship between housing price, land supply and land-use planning. The literature review section has suggested that the performance of the

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

349

housing market is determined by the interaction of both demand and supply factors. Eqs. (1) and (2) are first established to structurally present the housing demand and housing supply situations in Hong Kong. They draw out the overall picture of the housing market. In view of the previous studies, the availability of data in Hong Kong and the situation of Hong Kong, we have selected and incorporated some of the relevant demand and supply factors into the model. The demand factors we have chosen include income, employment, demographic and economic factors. For the supply factors, they include the amount of land supply and the planning constraints. The detailed description of the data will be given in the following section. The core assumptions of the regression model are as follows: (a) The model is of medium flow equilibrium with lags. Under market equilibrium, quantity demanded equals quantity supplied. (b) There may be time-lag effects of some explanatory variables so that the market functions in a sequential way. (c) Planning policies are in part responsive to market demand and in part autonomous. Given these assumptions, we may therefore write out expressions for the supply of and demand for housing units of the Hong Kong housing market in the following manner. The quantity demand function is QD ¼ a1 HP þ a2 I þ a3 GDP þ a4 POP þ a5 UE þ a6 ASP

ð1Þ

and the quantity supply function is QS ¼ b1 LSt3 þ b2 UFAt3 þ b3 GBt3 þ b4 ARt3 þ b5 RZt3 þ b6 GAt3 þ b7 HP;

ð2Þ

where HP is the average housing price, I the average household income, GDP the gross domestic product, POP the population, UE the unemployment rate, ASP the agreement for sale and purchase, LS the land supply includes auctions, tenders, private treaty grants, letter A/B, UFA the residential usable floor area completed, GB the area of greenbelt and open space zoning, AR the approval rates of planning applications, RZ the area of residential zoning, GA the total residential floor area under planning applications, t  3 the 3 years time lag. The above are the structural equations for the supply of and demand for housing, respectively. As there is no available information about quantity demanded, we are unable to separately identify Eqs. (1) and (2), but we can estimate the reduced form. As the quantity demanded (QD ) equals the quantity supplied (QS ) under market equilibrium, i.e. QD ¼ QS :

ð3Þ

We can get the reduced form by re-arranging the equations; then HP ¼ g þ a2 I þ a3 GDP þ a4 POP þ a5 UE þ a6 ASP þ b1 LSt3 þ b2 UFA þ b3 GBt3 þ b4 ARt3 þ b5 RZt3 þ b6 GAt3 :

ð4Þ

After re-arranging and reparameterising the equations, the factors which affect housing price significantly can be identified under model testing.

ARTICLE IN PRESS 350

E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

4. Data Following model specification, this section discusses the data sources; and explains each variable and its hypothesized impact. Relevant data are extracted from various issues of Hong Kong Property Review, issued by the Rating and Valuation (R&V) Department, and the Monthly Statistics, issued by the Census and Statistics (C&S) Department. The planning data are collected and modified from the Planning Department. The sample quarterly data cover from 1988 Q1 to 2000 Q4 and consist of 52 observations. 4.1. Dependent variable Housing price (HP): Quarterly average HPs are derived from the transaction records published in Hong Kong Property Review by the R&V Department. The data adopted are the average HPs of five classes in four districts. The five classes range from Classes A to E with saleable area not exceeding 39.9, 40–69.9, 70–99.9, 100–159.9 m2 and area of at least 160 m2, respectively. The three districts include Kowloon, New Kowloon and New Territories. As the quarterly average HP includes all classes across the whole territory for each quarter, it is a proxy of the aggregate HP level. 4.2. Independent variables 4.2.1. Indicators of economic factors Average household income (I): It is the quarterly median household income obtained from the Quarterly Report on General Household Survey issued by the Census and Statistics Department. Ganesan and Ho (1994) suggest that the ability to pay for home ownership is supported by current household income, estimated future household incomes and current and future assets of the households. An increase in income level leads to higher affordability to pay for home ownership if the HP keeps constant. Gross domestic product (GDP): It is always an indicator of the performance of the local economy. An increase in GDP would mean an improvement in wealth and income levels among local people. The increase in household incomes would normally raise housing demand and housing prices. The data are from Hong Kong Monthly Digest of Statistics. 4.2.2. Indicators of demographic factors Population (POP): Population growth puts direct pressure on housing demand, particularly when the growth stems from the affordable group with home buying needs. A high level of population growth tends to cause higher future housing prices when the supply of housing units cannot meet the demand. However, there are no official quarterly population statistics available in Hong Kong. Therefore, the half-yearly population statistics cited from Hong Kong Annual Digest of Statistics are transformed to quarterly data and the missing data are estimated by taking the average of two nearby values. Unemployment rate (UE): The unemployment rate figures are obtained from the Hong Kong Annual Digest of Statistics. The data include all age and sex groups. The higher the unemployment rate, the lesser the people have the affordability to puchase housing units, which

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

351

leads to decreases in demand. As Green and Hendershott (1996) explain, the region is more likely experiencing recession when people lose their jobs. It is this recession which can drive down the price of the house. 4.2.3. Indicator of market factors The agreement for sale and purchase (ASP): This is the agreement for sales and purchase of property that are registered with the Land Registry, since the transactions may involve short-term resale of property for anticipation of capital gains, which is generally labeled as speculations. The transaction volume may reflect the speculative demand for property. Tse et al. (1999) apply the Granger causality tests on the annual residential property prices from 1975 to 1995. Their findings suggest that the transaction volume is ‘‘Granger causing’’ to the housing prices in Hong Kong. Speculative demand is expected to contribute to price increases to some extent. Since there is no official quarterly data on transaction volume, and then the quarterly transaction volume is calculated by summing up of all the monthly transactions within that quarter from Hong Kong Annual Digest of Statistics (see Hui & Lui, 2002). 4.2.4. Indicators of land and housing supply Usable floor area (UFA): This is the UFA of newly completed residential buildings in Hong Kong. The figures include buildings of the Hong Kong Housing Society, Private Sector Participation Scheme of the Hong Kong Housing Authority and private buildings. The data are collected from the Annual Digest of Statistics issued by the Census and Statistics Department. This figure represents the new housing supply. The increase in the provision of usable floor area will decrease the housing price, ceteris paribus. Land supply (LS): This is the overall disposal of government land by means of public auction, public tender, private treaty grants, letter A=B in both Urban Areas and the New Territories. Land uses include pure residential use and composite residential and commercial use. The data are collected from the Annual Digest of Statistics. After buying a site, developer needs to produce a certain amount of residential floor area within a few years in accordance with the lease conditions. The developers at least take several years to turn the site into housing units. It is expected that an increase in housing production in future will decrease future housing price. As the land supplied by the government is readily developable, it requires 6–12 months further for the building plan approval and 2–3 years for the construction period, it is therefore assumed that the average development period after obtaining the land is 3 years. Therefore, 3 years time lag of land supply is being tested. 4.2.5. Indicators of planning constraints In Hong Kong, there is a lack of readily available and easily interpretable measures of planning constraints. Having examined the planning policy and the massive and scattered information, the extent of the policy constraints are considered to be reflected by the approved residential gross floor area of planning applications, the area of residential and restricted zones and the approval rates of planning applications, respectively. Residential gross floor area under planning applications (GA): This is the total residential gross floor area of all the planning applications which have been assessed by the Planning Department. The data can only be obtained on the computer in the office of the Planning Department. The

ARTICLE IN PRESS 352

E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

residential gross floor area of the planning applications has to be checked case by case. This indicator implies the concept of development density control. The more the residential gross floor area has been approved, the more the relaxation is on the development density. The higher development density will finally lead to an increase in housing supply and a decrease in housing prices. Residential zones (RZ) and greenbelt/open space zones (GB): There are no readily available data provided by government departments. For the area of RZ, it is the total area of all zones which involve residential use. These zones include Residential Group (A), Residential Group (B), Residential Group (C), Residential Group (D), Residential Group (E), Commercial/Residential (C/R), Village (V) and Comprehensive Development Area (CDA). The area of greenbelt and open space zones reflect the planning constraints. The area figures were taken off from OZPs of all districts in Hong Kong. Mayer and Somerville (1999) have built a model describing the relationship between land-use regulation and new residential construction. They conclude that land-use regulations such as zoning and growth control have significant effects on both the steadystate level of new construction and the responsiveness of local supply to price shocks. The increase in the area of residential zoning will relax the constraints on housing supply, which leads to a decrease in housing price. However, as no residential developments are allowed in greenbelt and open space zoning, the increase in greenbelt and open space zonings will further constrain the housing supply, which causes the decrease in housing price. Approval rate of planning applications (AR): This is the percentage of the number of approved planning applications involving residential use against the total number of planning applications involved residential use. The data can only be obtained from the computers provided in the office of the Planning Department. The approval rate of planning applications is one of the indicators of planning contraints. The higher the approval rate of planning applications, the more the housing units will be produced. This will in turn decrease the housing price in future, ceteris paribus. All the three planning indicators are assumed to have 3 years time lag effect on the housing market. It is assumed that no lease modification is required after zoning or planning approval. After zoning or obtaining planning approval, it still needs 2–3 years for development. Thus, these 3 factors with 3 years time lag effect are being tested. The descriptive statistics for the data are put forth in Table 1.

5. Expected results The above expectations of the impact of various factors on the housing price are shown in Table 2. The economic condition of Hong Kong is reflected by the average household income and the gross domestic value. It is expected that a better economic condition will generate a greater demand for housing and the housing price will eventually rise. For the demographic factor, it is represented by the population and unemployment rate. The increase in population will bring more demand for housing units; however, the increase in unemployment rate will decrease the purchasing power of people and the housing price will drop eventually due to falling demand. The agreement for sale and purchase is an indicator of market atmosphere. The hot market sentiment will steer up housing price. The residential gross floor area under planning applications reflects the

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

353

Table 1 Descriptive statistics

I—Average household income (HK$) GDP—Gross domestic product (fixed price at 1980) (HK$Mn) POP—Population UE—Unemployment rate (seasonally adjusted %) ASP—No. of agreements for sale and purchase of building units LS—Land supply includes auctions, tenders, private treaty grants, letter A/B with 3 years time lag (m2) GA—Total residential gross floor area under planning applications with 3 years time lag (m2) GB—Area of greenbelt and open space zoning with 3 years time lag (ha) RZ—Area of residential zoning with 3 years time lag (m2) AR—Approval rate of planning applications with 3 years time lag (%) UFA—Residential usable floor area completed (’000 m2) Valid N (listwise)

N

Minimum

Maximum

Mean

Std. deviation

52 52

8797 55,041

26,884 103,106

18,433 78,285

5480 12,274

52 52

5,621,450 1.30

6,974,800 6.20

6,171,397 2.86

440,562 1.47

52

11,743

62,843

29,022.96

12,222.78

52

1025

475,817

102,463

120,819

52

52,054

3,237,502

836,819

863,634

52

7946

15,597

12,252

2614

52

3393

16,140

7842

4936

52

21

80

50

12.89

52 52

18

249

100.19

45.28

Table 2 Expected results Independent variables

Expected relationship with housing price

I—Average household income GDP—Gross domestic product POP—Population UE—Unemployment rate ASP—Agreement for sale and purchase LS—Land supply includes auctions, tenders, private treaty grants, letter A/B UFA—Residential usable floor area completed GB—Area of greenbelt zoning and open space zoning AR—Approval rates of planning applications RZ—Area of residential zoning GA—Total residential gross floor area under planning applications

+ + +  +   +   

ARTICLE IN PRESS 354

E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

demand for residential use. The greater demand for residential use will finally lead to an increase in housing and a decrease in housing prices. For the supply side factors, the land supply through auctions, tenders, private treaty grants and letter A/B, and the newly built residential usable floor area reflect the housing production. It is expected that the increase in the supply of housing units relative to demand, will lead to the decrease in housing price. The planning constraint factors are the area of residential zonings and approval rates. Such factors measure the effects of planning constraints on the housing price. For example, a higher approval rate would possibly result in more housing land for developing, followed by an increase in the volume of housing production. An increase in housing supply as such would have a negative effect on housing price, ceteris paribus. This holds true for the other two planning constraints. The looser the constraints, such as higher approval rates, more area of residential zonings and lesser area of greenbelt and open space zonings, the more the housing production will be. This will eventually lead to lower housing price in future.

6. Empirical results ‘‘Stepwise regression’’ is adopted so as to avoid working with more variables than are necessary while improving the equation at every stage. It first starts by selecting an equation with the best independent variable and then attempts to build up with subsequent additions of variable (Draper & Smith, 1998). Variables are scrutinised for removal or entry at each step. The default criteria for the probability of F-to-enter equals or is smaller than 0.05, whereas the criteria for the probability of F-to-remove equals or is smaller than 0.1. We have tested 1-, 2- and 3-year time lag for the variables GB, AR, GA, LS and RZ. The results show that only the 3-year time lag has more reasonable and significant outcomes. The stepwise regression has established eight models with three variables being excluded eventually. The adjusted R2 is 0.924. The variables which are identified as significant include average household income, population, unemployment rate, agreement for sales and purchase, residential usable floor area completed, area of greenbelt and open space zonings, approval rates of planning application and total residential gross floor area under planning applications. The final equation arrived at after running the stepwise regression is HP ¼ 0:957I þ 4:248E  02 POP  8902:076UE þ 0:248ASP  54:266UFA þ 1:983GBt3  17465:426ARt3  4:045E  03GAt3  225884:6:

ð5Þ

Table 3 reports estimation results for regression of housing price on various demand, supply and planning variables. The coefficient g is a mixture of the parameters from Eqs. (1) and (2). The adjusted R2 is quite high, indicating that the model we have chosen fits the actual data quite well. All of the above explanatory variables are statistically significant at the 10% level. Three variables that are excluded from the model setting are GDP, LSt3 and RZt3. The estimated coefficients of all explanatory variables are coherent with the hypothesized signs. The multicollinearity of the variables have been tested. Within the eight variables, which have been selected by the stepwise regression, only three of them have tolerances less than 0.1. They are I, POP and GBt3 with tolerances of 0.08, 0.036 and 0.066, respectively.

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

355

Table 3 Statistical results of Eq. (4) Explanatory variables

Coefficient

t

Sig

Correct sign

I POP UE ASP UFA GBt3 ARt3 GAt3 g (constant)

0.957 4.248E02 8902.076 0.248 54.266 1.983 17,465.426 4.045E03 225,884.605

1.944 4.617 6.325 3.289 2.803 1.743 2.066 2.647 5.369

0.058* 0.000** 0.000** 0.002** 0.008** 0.089* 0.045** 0.011** 0.000**

Yes Yes Yes Yes Yes Yes Yes Yes

Adjusted R2 ¼ 924 F Value : 78.354 Degrees of freedom: 43 *Significant at 10% significance level. **Significant at 5% significance level.

The model that has been established by the ‘‘stepwise’’ regression is shown as follows: The results confirm with Bramley (1993) that house prices are determined by economic, demographic, supply and other variables. He suggested that general planning policies for land release (especially structural plans) had quite a substantial effect on housing output, although it was far from being a one-for-one relationship. He further asserted that the effect of local land release targets and policies on local house prices in general is very weak indeed due to the openness of local markets in Britain. However, this research proves that the planning system, which restricts the location, density and amount of land supply, has a statistically significant effect on housing price. This agrees with Peng and Wheaton (1994) that supply restrictions in Hong Kong have caused higher housing prices, though they mainly focus on the impact of restricted land sales supply on the housing market. Both the land supply and the area of the residential zoning have been dropped out in the stepwise regression. In accordance with Tse (1998), there is no casual relationship between land supply and housing prices. This is mainly due to the land-banking behaviour of the developers. The developers’ land banks tend to decrease when market interest rates increase. For the residential zoning factor, different residential zones impose different plot ratio restrictions on the development. The increase in residential zones does not necessarily lead to an increase in housing production. The developers develop or redevelop the site only if there is enough incentive for them, such as much higher allowable plot ratio than the existing plot ratio. Therefore, it is not surprising that the area of residential zoning does not have a significant effect on the housing production as well as the housing prices. In our study, the approval rate of planning applications adopted in the regression model reflects the magnitude of planning delays which affects development cost and production time. As Hong Kong’s high-rise development involves considerable construction time, the approval

ARTICLE IN PRESS 356

E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

rate is estimated to have 3 years time lag significant effect on the housing price. The significance of approval rate of the planning applications in our result also supports the view of Mayer and Sheppard (1996) that the stochastic development control alters the structure of housing supply in a way which may change the equilibrium price of housing by an amount far in excess of what would be expected from the costs imposed by the expected duration of planning delay. Other than approval rate, the planning system may also influence the housing market by controlling the density and use of the future development. The explanatory variable of the total residential gross floor area under planning applications shows that it is statistically significant with 3 years time lag effect. It means that the residential gross floor area will turn into the real supply after few years, which will eventually lead to the decrease in housing price. However, Hannah et al. (1993) stated that a substantial part of the rise in house prices has resulted from the government’s tendency to underallocate land to urban residential use. In view of the above, the housing price is influenced by the development constraints and other demand and supply factors. Our result is also in line with what Chestire and Sheppard (2000) studied. They estimated the effects of the planning system on housing prices by analysing the housing markets in Darlington and Reading in 1984. It was found that the low price of houses in Darlington did not only reflect looser constraints on development, but also a falling demand relative to supply of houses. In reviewing the results presented, the extent to which the existence of relationship agrees with the expected outcomes is outlined in Section 6. It is observed that the planning system in terms of its approval rate and total residential gross floor area under planning applications, has a statistically significant effect on the housing market in Hong Kong. It should be noted that there is a discrepancy between the planning system and the housing market in Hong Kong as mentioned in the previous literature review. The planning authority may give more careful attention to these two particular planning conditions when considering alterations to planning policies.

7. Conclusions The government adopts its land-use planning system to regulate and impose constraints on land supply and development. Section 2 has studied the planning system in Hong Kong. It analysed the approval rate of residential planning applications and the extent of residential zones and greenbelt and open space zones in the territory. There exists a similar pattern between the number of planning applications and the approved cases. This suggests that the government will adjust its leniency of the approval procedure in view of the high market demand in the recent years. The recent increase in the approval rate facilitates more urban developments and meets the balance between the economic, social and environmental demands. Despite the approval rate, the government also controls the density of development. The patterns of the number of approved cases and the residential floor area under planning applications reflect the planning approval situation under fluctuating market demand. In fact, the approval strategy is flexible to meet the changing environment and demand. Apart from the planning application system, the planning department also controls the land use through the zoning system. For residential zone, one can develop residential development or higher-density development through Section 16 applications.

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

357

However, there is a general presumption against development in the greenbelt and open space zones. This serves as a buffer between and within urban areas. The whole land-use planning system is essential for it maintains the environment sustainable. This study also focuses on how planning impinges on the supply of land for housing and affects the housing price as a result. The ‘‘stepwise’’ regression results demonstrate that the planning system has significant impact on the housing market in Hong Kong. The analysis demonstrates that most of the planning variables affect housing prices statistically, getting into the stepwise equation in the first few steps. In order to control the housing supply in future, the government can adjust its planning control by restricting uses, the densities, areas and approval rates of the developments. The implications of this study are as follows: In view of the above, the Hong Kong government should focus on its planning policy so as to achieve a stable housing market in the long term. The government should be aware of the impact of existing planning policy on the existing and future housing market. For example, if the government foresees a demand for housing, it can streamline its planning application procedures, increase the residential zoning areas and increase the allowable development density in order to meet the future demand. Usually planning policies do not take immediate effects on housing market as developers require substantial time to develop the site into high-rise residential buildings. Among all planning controls, the zoning method is considered to be the least effective. This is because existing use of building or land is permitted to continue without any planning permission until redevelopment or a change of use takes place, despite change of zonings. The government may consider providing some incentives for the developers to convert the existing use of land when there is a need. In order to integrate the fundamental role markets play in allocating resources in a market economy, urban planning and land-use regulation must adopt market-oriented principles and concepts before these processes can realize the goals and objectives of planners. It allows market conditions to determine the pace and scale of development. Planners in Hong Kong mainly concern the implementation and enforcement of plan; they rely too heavily on a static concept of community and adopt a resource allocation process that presumes the future can be determined reliably and controlled by the government (Staley & Scarlett, 1998). In order to achieve the target of ‘‘adequate and affordable housing’’, the government should allow flexibility in its planning policy so as to adjust to the changing economic and demographic environment. Therefore, we recommend the government to set up a forecasting model so as to help estimating the effect of planning policy. As our model is the integration of housing market factors and planning factors, we believe that it can resolve the dilemma between (1) the policy of providing adequate and affordable housing against the shortage of housing supply and (2) the land-use regulations for control on residential development.

Acknowledgements This research is funded by The Hong Kong Polytechnic University Central Research Grant BQ364 and GYD04.

ARTICLE IN PRESS 358

E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

References Bramley, G. (1993). The impact of land use planning and tax subsidies on the supply and price of housing in Britain. Urban Studies, 30, 5–30. Bramley, G. (1998). Measuring planning: Indicators of planning restraint and its impact on the housing market. Environment and Planning B: Planning & Design, 25(1), 31–57. Bramley, G., Barlett, W., & Lambert, C. (1995). Planning the market and private housebuilding. London: UCL Press. Bristow, R. (1984). Land-use planning in Hong Kong: History, policies and procedures. Hong Kong: Oxford University Press. Buttler, H. J. (1981). Equilibrium of a residential city, attributes of housing, and land-use zoning. Urban Studies, 18, 23–39. Census and Statistics Department (2000a). Hong Kong Monthly Digest of Statistics. Hong Kong: Government Printer. Census and Statistics Department (2000b). Hong Kong Annual Digest of Statistics. Hong Kong: Government Printer. Chestire, P. C., & Sheppard, S. (2000). Building on Brown Fields: The long term price we pay. Planning in London, 33, 34–36. Dale, P., & McLaughlin, J. (1999). Land administration. Oxford: Oxford University Press. Draper, N. R., & Smith, H. (1998). Applied regression analysis. Canada: Wiley-Interscience. Evans, A. W. (1996). The impact of land use planning and tax subsidies on the supply and price of housing in Britain: A comment. Urban Studies, 33, 581–585. Gallent, N., & Kwang, S. K. (2001). Land zoning and local discretion in the Korean planning system. Land Use Policy, 18, 233–243. Ganesan, S., Ho, K. O. (1994). International Association for People—environment studies. The urban experience, affordability housing designs for urban development: A case study of Hong Kong. Department of Architecture, University of Hong Kong, July. Gerald, E. (1992). The relationship between house prices and land supply. London: Department of the Environment Planning Research Programme, HMSO. Green, R. K., & Hendershott, P. H. (1996). Age, housing demand, and real house prices. Regional Science Urban Economics, 26(5), 465–480. Hannah, L., Kim, K. H., & Mills, E. S. (1993). Land use controls and housing prices in Korea. Urban Studies, 30, 147–156. Ho, S. M. (2001). The land administration system in Hong Kong. Unpublished M.Phil. dissertation, The Hong Kong Polytechnic University. Hong Kong Government (1994). Hong Kong planning standards and guidelines. Hong Kong: Government Printer. Housing Branch (1997). Homes for Hong Kong People: The way forward—long term housing strategy review consultative document. Hong Kong: Housing Branch, Government Secretariat. Hui, C. M., & Lui, A. (2002). Rational expectations and market fundamentals: With evidence from an Asian City’s boom and bust cycle. Journal of Property Investment and Finance, 20(1), 9–22. Information Services Department (2000). Hong Kong: The facts, town planning. Hong Kong: Information Services Department. Lai, W. C. L. (1996). Zoning and property rights: A Hong Kong case study. Hong Kong: Hong Kong University Press. Lai, W. C. L. (1997). Town planning in Hong Kong: A critical review. Hong Kong: City University of Hong Kong Press. Lai, W. C. L. (2000). Town planning practice: Context, procedures and statistics for Hong Kong. Hong Kong: Hong Kong University Press. Loh, C. (1998). Stop this flip-flop housing policy. South China Morning Post, 9 February 1998. Mayer, C. J., & Somerville, C. T. (1999). Land use regulation and new construction. Regional Science and Urban Economics, 30, 639–662. Mayo, S., & Sheppard, S. (1996). Housing supply under rapid economic growth and varying regulatory stringency: An international comparison. Journal of Housing Economics, 5, 274–289. Monk, S., & Whitehead, C. M. E. (1996). Land supply and housing: A case-study. Housing Studies, 11(3), 407–423. Nnkya, T. J. (1998). Land use planning practices under the public landownership policy in Tanzania. Habitat International, 23(1), 135–155.

ARTICLE IN PRESS E. Chi-man Hui, V. Sze-mun Ho / Habitat International 27 (2003) 339–359

359

Peng, R., & Wheaton, W. C. (1994). Effects of restrictive land supply on housing in Hong Kong: An econometric analysis. Journal of Housing Research, 5(2), 263–291. Planning Department (1995). Town planning in Hong Kong: A quick reference. Hong Kong: Government Printer. Planning Department (2001). Town Planning Board guidelines. Hong Kong: Government Printer. Rating and Valuation Department. (1988–2001). Property review. Hong Kong: Government Printer. Sagalyn, L. B., & Sternlieb, G. (1973). Zoning and housing costs. New Brunswick, NJ: Rutgers Center for Urban Policy Research. Seidel, S. (1978). Housing costs and governmental regulations. New Brunswick, NJ: Center for Urban Policy Research, Rutgeres University. South China Morning Post (6 June 1997). Hong Kong. Staley, S. R. (1994). Planning rules and urban economic performance: The case of Hong Kong. Hong Kong: The Chinese University Press. Staley, S. R., & Scarlett L. (1998). Market-oriented planning: Principles and tools for the 21st century. Planning & Markets, 1(1), 1–7. http://www-pam.usc.edu/volume1/vlila5sr.html. Tang, B. S., & Choy, L. H. T. (2000). Modelling planning control decisions: A logistic regression analysis on office development applications in urban Kowloon, Hong Kong. Cities, 17(3), 219–225. Tang, B. S., & Leung, H. F. (1998). Planning enforcement in Hong Kong: Implementing new planning law before the change of sovereignty. Town Planning Review, 69(2), 153–169. Tse, R. Y. C. (1998). Housing price, land supply and revenue from land sales. Urban Studies, 35(8), 1377–1392. Tse, R. Y. C., Ho, C. W., & Ganesan, S. (1999). Matching housing supply and demand: An empirical study of Hong Kong’s market. Construction Management and Economics, 17, 625–633. Yeung, C. W. (1998). Town planning control in Hong Kong. In T. N. T. Poon, & E. H. T. Chan (Eds.), Real estate development in Hong Kong (pp. 31–50). Hong Kong: Pace Publishing.

Further reading Case, F. E., & Gale, J. (1981). The impact of housing costs on the california coastal zone conservation act. AREUEA Journal, 9, 345–366. Dowall, D. E. (1984). The suburban squeeze: Land conversion and regulation in the San Francisco Bay Area. Berkeley: University of California Press. Dowall, D., & Landis, J. D. (1982). Land use controls and housing costs: An examination of San Francisco Bay Area communities. AREUEA Journal, 10, 67–93. Elliott, M. (1981). The impact of growth control regulations on housing prices in California. AREUEA Journal, 9, 115–133. Frieden, B. J. (1979). The environmental protection hustle. Cambridge, MA: MIT Press. Hui, C. M., Chan, P. C., Wong, K. W., Wong, K. C., Leung, Y. P. (2000). The supply of land for housing in Hong Kong. Research Monograph. Department of Building and Real Estate, Hong Kong Polytechnic University. Pollakowski, H., & Wachter, S. M. (1990). The effects of land use constraints on land values. Land Economics, 66, 315–324. Pryce, G. (1999). Construction elasticities and land availability: A two-stage least-squares model of housing supply using the variable elasticity approach. Urban Studies, 36(13), 2283–2304. Stanford Environmental Law Society (1989). Land use and housing on the San Francisco Peninsula. Stanford Environmental Law Annual, 4, 535–557. Urban Land Institute and Gruen Associates (1977). Effects of regulation on housing costs: Two case studies. Research #27. Urban Land Institute.