Professional Documents
Culture Documents
Cities
journal homepage: www.elsevier.com/locate/cities
A R T I C L E I N F O A B S T R A C T
Keywords: This paper introduces a principal-agent analysis of the public sector decisions made by a non-democratic
Planning applications planning body under public scrutiny. The central idea is that divergence between the interests of the principal
Principal-agent problem (citizens), whose objective is to maximize the total value of the land in Hong Kong, and the agent (planning
Approval risk
approval authority), whose concern is about Approval Risk, which is the agent’s risk of being accused of collusion
Logistic regression
Pseudo R2
or even bribery. This principal-agent analysis can be applied to any public-sector decision-making process in a
Land use society with the rule of law and freedom of speech. It was found by a logit analysis of 1440 sets of disaggregate
Community use data that planning applications by non-profit applicants for uses in Government/Institution/Community (GIC)
zones were more likely to be approved than those by for-profit private organizations, in line with the idea of a
lower perceived Approval Risk. Amongst the applications by private organizations, proposed high-value land
uses were less likely to be approved than those proposed for low-value uses, probably because the former results
in potentially higher economic gain for the applicants. The legislative change that took effect in July 2005 has
allowed a wider public participation in the planning application process. This policy change was used in this
study as a test condition for the effect on planning approvals of an increase in the principal’s involvement in the
decision-making process. It was found that the policy change did increase the success rate of planning appli
cations made by private organizations, though the impact was mainly on the applications for low-value land uses
only. The empirical result shows that the perceived Approval Risk for high-value land uses was too high to be
affected by the policy change.
* Corresponding author.
E-mail addresses: hrrbckw@hku.hk (K.W. Chau), lennonchoy@hku.hk (L.H.T. Choy), markchua@ln.edu.hk (M.H. Chua), wclai@hku.hk (L.W.C. Lai), esther.
yung@polyu.edu.hk (E.H.K. Yung).
https://doi.org/10.1016/j.cities.2023.104291
Received 26 August 2021; Received in revised form 5 March 2023; Accepted 9 March 2023
Available online 29 March 2023
0264-2751/© 2023 Elsevier Ltd. All rights reserved.
K.W. Chau et al. Cities 137 (2023) 104291
Osman (2017) noted that some revolts were aimed more at transferring Principal-agent problems have been used to analyze contractual ar
discretionary power of approving some exemption to developers from rangements in varied economic relationships: doctor-patient (Scott &
the politician to the people. Vick, 1999), owner-lessee (Marks-Bielska, 2021), insurer-insured (Sti
This distrust can be put into proper context by empirical work. For glitz, 1989), legislator-bureaucrat (Campbell & Turnbull, 2003; Con
instance, countering the usual belief that Hong Kong developers jacked nolly, 2017; Sager & Ravlum, 2005; Waterman & Meier, 1998), federal-
up prices of flats they sell, Wong et al. (2019) found that prices of first- local planning agencies (Chen, 2007), and school districts-school sites
hand flats sold by developers are at a discount compared to the prices of (Ferris, 1992).
the same flats sold in the second-hand market after a couple of In urban and planning studies, the principal-agent framework was
years—having adjusted for depreciation. Another study gave examples applied to examine central government restrictions on quasi-
wherein government policies have restrained giant developers by pitting autonomous municipal plans in different countries (Huang et al.,
them against each other (Wong, 2015). Nevertheless, the image of some 2019; Strand & Næss, 2017); to a federal transport planning agency
developers in the news print media and other housing experts has his designing incentives for local transport agencies to align to federal goals
torically been very negative if not hostile (Ley & Teo, 2014). (Chen, 2007); to analyze the interaction between elected officials and
In Hong Kong, redevelopment of church property by a developer into hired professional county managers (Campbell & Turnbull, 2003; Con
a residential block has experienced some community conflict despite the nolly, 2017) or professional experts like transport planners (Sager &
re-provision of a church in the lower portion of the building (Yip et al., Ravlum, 2005). It has also been used to analyze incentives for envi
2020). Acquisition and the eventual redevelopment of utilities like ronmental remediation (Zhang et al., 2018; Sheng et al., 2017; Permadi
power plants or bus depots into large private residential blocks is also et al., 2017; Weber, 2015); land utilization programmes (Marks-Bielska,
seen as a maneuver to acquire strategically located urban land by de 2021; Shi et al., 2018) and privatization schemes (Coats, 2002; Van
velopers (Poon, 2005). Researchers often neglect the point that de Slyke & Hammonds, 2003).
velopers are interested in these urban sites (zoned GIC, if not Two elements are of particular importance in a principal-agent
Comprehensive Development Area (CDA)) as they are large, well-served relationship, (a) the information asymmetry between the parties; and
by roads and above all, usually under unitary ownership. (b) the possibility of conflict or agreement between them (Ross 1973;
Given these politically and socially sensitive issues regarding such Stiglitz, 1989; Waterman & Meier, 1998; Gailmard, 2012). In many
land use conversions, it is interesting to understand how non-elected scenarios, this information asymmetry may be described as a two-edged
Town Planning officials behave in such tricky cases. In a sense, both sword. Indeed, the principal engaged the agent precisely because the
researchers and the public tend to see GIC land as if they are food vac agent knows something the principal does not, such as the agent’s
uoles1 of greedy developers. expertise for example, in order to achieve the principal’s goal (Stiglitz,
This paper is a pioneer attempt to use the neo-institutional concept of 1989). On the other hand, the information asymmetry may also cause
the principal-agent problem to predict the behavior of Town Planning conflicts or misalignment of goals between the two parties, which un
Board (TPB), the planning approval authority in Hong Kong, regarding dermines the principal’s objective. For instance, the data requirements
decisions on applications to change GIC sites to other uses under for a regional policy system to work—principal-agency scenario—is an
different degrees of public participation. A GIC zone is intended to important area of study (e.g. McVittie & Swales, 2007).
provide facilities like schools, hospitals, public sports areas, and other In order to cover the breadth of this family of models (Gailmard,
government or community amenities to different districts or adjoining 2010), Waterman and Meier (1998) expanded and framed the analysis
districts. When planning applications are made for changes to any zone’s into eight scenarios wherein the principal and the agent have different
uses, the TPB is expected to have regard to public opinion and distrib combinations of information and conflict/consensus with respect to one
utive justice. another.
The paper is organized as follows: Section 2 presents a review of the In this study the citizen is the principal while the TPB is the agent. We
application of the principal-agent model to planning studies. Section 3 position this study after Gailmard (2010) in considering that the agent is
describes the town planning system in Hong Kong, as an institution, to better informed than the principal, and that the agent may or may not
provide a context to the study. Section 4 presents a principal agent have any motivation to offer better service, especially where the
model and Section 5 details the hypotheses and methodology. Section 6 approving body is not elected. On 10 June 2005, this relation was
presents and interprets the empirical results. Section 7 is a discussion further emphasized when planning applications were open for the first
and the last, Section 8, concludes the study with a few time to public viewing and comments, in line with the principal’s
recommendations. policing paradigm expounded by Mitnick’s (1975) seminal paper. Using
social media as a tool of communication between citizens and the gov
2. The application of the principal-agent concept to urban and ernment, Bonsón et al. (2017) reported that these interactions would
planning studies weaken the agency conflict between the two parties.
According to Nobel Prize winner Stiglitz (1989), the principal-agent 3. An outline of the Hong Kong planning application system
concept refers to the politico-economic dynamics between the principal
(e.g. an employer, a landlord) and an agent (e.g., employee, tenant) that Since its enactment in 1939, the Town Planning Ordinance has
act according to the principal’s interest. Stiglitz (1989) attributed the enabled the TPB to provide for the imposition of Hong Kong Outline
first use of the term principal-agent problem to Ross (1973). Jensen and Zoning Plans (OZPs) for “existing and potential urban areas”. These
Meckling (1976) illustrated the concept in one of the most cited papers plans, which began to emerge in the 1960s but were only legally valid
in finance by pointing out that the incentives of executives, who tar for those plans made after 1976, delimit what types of land uses are
getted short term profits and hence higher remuneration packages, always permitted or may be permitted on application to the TPB in the
might be incompatible with those of the owners, who aspired to long area they cover. Land parcels allotted for government or community
term growth of their companies. More than a theory, Gailmard (2012) facilities are normally designated as GIC zones.
describes principal-agent problems as a family of models wherein the Table 1 shows a sample table from the “Notes”2 of an Outline Zoning
principal and agent are identified to see how they affect each other.
2
Expressly “to form part of the plan”, these Notes are distinct from the
1
https://biologydictionary.net/food-vacuole/. Here the term refers to land “Explanatory Statement” that expressly “does not form part of the plan.” Some,
banks of developers. who are confused, call the former “explanatory notes”.
2
K.W. Chau et al. Cities 137 (2023) 104291
3
Uses developed according to terms in the lease are “existing uses” under any
4
statutory plan imposed on the land after their development. https://news.rthk.hk/rthk/en/component/k2/1639277-20220316.htm.
3
K.W. Chau et al. Cities 137 (2023) 104291
which may attract adverse public attention. The outcomes of their de Risk for the TPB. Public or community land uses are therefore hypoth
cisions are transparent to the general public and can be easily monitored esized as more likely to be approved by the TPB, other things being
by the media. The risk facing the TPB members is that of being accused equal.
of collusion or even bribery when a decision appeared to have benefitted Hypothesis 2:
a private applicant, which is termed Approval Risk. The accusation may
H2. Applications for changes in use of land in Government, Institu
not have any legal consequences but could damage the reputations of
tional or Community (GIC) zones to low-value for-profit private uses are
the non-government TPB members, who have their own professional
more likely to be approved than those for change to high-value for-profit
careers or businesses but serving on the TPB on a part-time pro bono
private uses.
basis. One of the TPB’s objectives that is not in line with that of Hong
Kong’s citizens is that of minimizing the Approval Risk when making High value pro-profit private uses under Column 2 include colum
decisions on planning applications. We conjecture that this is the main barium, hotel, commercial (retail), office and residential use. Low value
source of the principal-agent problem that exists between the TPB as an private uses under Column 2 include storage, carpark and petroleum
agent and Hong Kong’s citizens as the principal. refilling station. Since the potential economic benefits of obtaining ap
Most TPB members are built-environment related professionals, they provals for high-value private use are much higher than those for the
have an edge in information and knowledge about the implications of low-value land uses, the Approval Risk should be higher for high-value
planning applications over ordinary citizens and can potentially make land uses and therefore less likely to be approved by the TPB, other
more informed decisions than ordinary Hong Kong citizens, especially in things being equal.
minimizing potential negative externalities. However, in practice, sub- Hypothesis 3:
optimal decisions for the citizens may be made due to the existence of
H3. Applications for change of land zoned for Government, Institu
the Approval Risk.
tional or Community (GIC) to for-profit private uses are more likely to be
The principal-agent problem is further aggravated by the fact that
approved with an increase in public participation.
the TPB members are appointed without any public consultation rather
than democratically elected. Given this non-democratic institutional An increase in the involvement of the public in the planning appli
setting in a society with rule of law and freedom of expression, the cation process through public participation can reduce the principal-
agent’s interest may not be aligned with that of the principal if the agent conflict since the principal can have more influence on the de
agent’s decisions to approve a planning application may increase the cisions by the TPB. As mentioned previously, there was a legislative
agent’s own risk (Approval Risk) of being publicly accused of collusion change after 9 June 2005, which has increased the scope of public
with the applicant for planning permission or even of corruption.5 participation substantially. This institutional change allows us to test the
The Approval Risk as perceived by the TPB is higher if their decision hypothesis empirically. As a result, the perceived Approval Risk should
can result in substantial potential gain in wealth for the applicant, even be lower after the policy change and therefore private uses be more
if the decision can achieve a Pareto gain (i.e. a win-win situation). We likely to be approved after 9 June 2005, other things being equal.
further conjecture that the perceived Approval Risk arising from this
principle-agent problem can be lowered by increasing participation by 5.2. Data
the citizens in the planning application process, since this can empower
the TPB to make decisions that are beneficial to the citizens despite such Non-aggregate data of planning applications from 1975 to 1999 were
decisions may also bring finance gain to the applicant. This view is manually collected from the Planning Department and those from 2000
contrary to the recent government idea of curtailing public to 30 June 2020 from its website. These data cover the location, size, use
participation. and status of the projects. A total of 4403 planning applications for
The above principle-agent analysis is used to predict the behavior of changes in use on sites under full or partial GIC zoning were identified.
the TPB in making decisions on planning applications for change of land These applications include extensions of previous permissions or some
zoned for GIC use to other uses. These predictions are stated as testable reviews of past decisions.
hypotheses in the next section. It was discovered, as in the case of CDA zoning (Lai et al., 2015,
2016), multiple s.16 applications were made for many projects. Each
5. Hypotheses and methodology application may be linked to a preceding and a subsequent application,
as the case may be. The related applications normally form a project
This section presents three refutable hypotheses derived from the “application web”, and we have depicted each project as a chain of
principle-agent analysis presented above and the method of testing the linked applications.
hypotheses. All in all, a total of 625 sites were found from the data 1975 to 2020.
Of the 625 projects, 148 have been approved. With their addresses
5.1. Hypotheses disclosed in the application forms, the names of the existing buildings
were identified.
The three empirical hypotheses and their rationales are as follows: To the extent possible, the name of the developer and owner for each
Hypothesis 1: site were traced through real estate websites and the Land Registry. One
difficulty here was to link the names to the ultimate owners. When big
H1. Applications for changes in use of land in Government, Institu developers carry out a construction project in Hong Kong, they normally
tional or Community (GIC) zones to public or community uses are more establish a new subsidiary company (Yip et al., 2020).
likely to be approved than those for change to for-profit private uses. After eliminating applications for changes in use to more than one
Public or community uses are typically applied by NGOs or non- type of use and the “Others” category, applications around the policy
profit making organizations for erecting hospitals, schools, churches change date (to avoid potential sampling bias) a total of 1440 usable
etc. It is postulated that public uses do not result in private economic observations were left for the purpose of testing the three hypotheses.
gain by the applicants and therefore incur substantially lower Approval The applications (land use types) are categorized as shown in Table 2.
4
K.W. Chau et al. Cities 137 (2023) 104291
Table 2 some applications may have dragged on for a long time, so the TPB
Categorization of GIC uses in terms of value. members may have changed. Since we do not have information on the
Use category in Use category in Type of land use applied for members judging each case, we had to assume that there was no change
Hypothesis 1 Hypothesis 2 of membership or that change of membership had no effect on the
PRIVATE HIGH-VALUE Columbarium or funeral related uses decisions.
Hotel or hostel or holiday camp
Commercial use 6. Empirical results and interpretation
Residential use
Office use
LOW-VALUE Storage or logistic storage or container
Table 3 shows the result of testing Hypothesis 1. All observations are
parking or recycling areas used in estimating the logit model. The size effect is captured by the
Light and private vehicle parking control variable GFA which is the gross floor area in m2. Smaller de
Petroleum refilling stations velopments were found to be more easily approved. However, the
PUBLIC PUBLIC Regular school up to secondary level or
relationship was not linear in GFA. The variable is significant at the 1 %
special, technical and vocational schools
Church, temple or other religious after natural logarithmic transformation indicating the marginal effect is
functions smaller for larger GFA which suggest diminishing returns.
Other public utilities like transmission Another variable control variable a location dummy variable which
lines, and reservoirs equals one if the site is location on the Hong Kong Island (HKI). Since
Social welfare facilities such as hospitals
and elderly centres
Hong Kong Island is the core urban area with limited supply of devel
Public sports, cultural and other opable urban land. Land on Hong Kong Island is also less substitutable
recreational facilities by sub-urban areas since it is isolated from Kowloon and the New Ter
ritories by the sea (Victoria Harbour). The results show that HKI is not
significant, however, its interaction with GFA is significant at the 5 %
previously (Tang et al., 2005, 2007; Tang & Choy, 2000; Tang & Tang,
level. Since GFA cannot be equal to zero, the result suggests that GIC
1999; Yu & Hui, 2017). This model is similar to probit regression model
sites on Hong Kong Island are less likely to be approved for development
in similar studies (see for instance Butsic et al., 2011; Lai & Ho, 2001;
compared with those not on Hong Kong Island and the different in
Chau & Lai, 2004). Logit and other discrete choice models have been
probability of being rejected for development increases with GFA.
used to study agency conflicts (e.g.) and different aspects of the
The first variable of interest is the dummy variable PRIVATE which is
principal-agent conflicts in many areas (Scott & Vick, 1999; Connolly,
equal to 1 if the land use application is for-profit private uses (See
2017; Mustafa et al., 2017; Daneshvary & Clauretie, 2013; Longarela-
Table 2). Its coefficient is negative and significant at the 1 % level, which
Ares et al., 2020; Horiuchi & Shimizu, 2001; Nguyen, 2011; Hamman
is consistent with Hypothesis 1.
et al., 2010; Bhat, 1997; McGuire & Ohsfeldt, 1989). We have chosen to
The marginal effects are calculated at the sample means of the in
estimate the logit instead of probit model since logit model since the
dependent variables as well as mean value of all sample observations
goodness of fit measure (the pseudo r squared) is more extensively
using the methods described in Anderson and Newell (2003) and Car
studied due to the simplicity of the logistic functional form. We have also
levaro and Sénégas (2006). The marginal effect of a dummy variable
estimated the probit model and the results are very similar. In order not
represents the change in probability of success when the dummy vari
to make the paper unnecessarily long, the probit results are not pre
able switches from zero to one. Therefore, the probability of success in
sented but are available upon request.
the application for change of GIC land use to for-profit private land use is
We have estimated the following logit model to test the three hy
24.12 % lower compared with public and community use.
potheses using the GIC planning application data we have collected:
Table 4 Shows the result of testing Hypothesis 2. Since we are testing
( )
Pr(PA = 1|X) = F XT β (1) the difference between high and low value private use, we only include
applications for private use (i.e. PRIVATE = 1) in estimating the Logit
where PA is a binary dummy that equals one if the planning application model. All variables are of the expected sign. The variable HIGH-VALUE
is successful and zero otherwise; Pr is the probability that the binary is equal to 1 if the application is for high value use (Table 2) and zero
dummy variable Y equals 1 conditioned on a vector X of variables that otherwise. Its coefficient is negative and significant at the 1 % level
affect the TPB’s decision, F is the cumulative distribution function (CDF) meaning that high value private use is less likely to be approved as
of the logistic distribution. The vector of parameters β were estimated by predicted by Hypothesis 2. On average, for-profit high value private land
maximum likelihood method. use is 21.067 % less likely to be approved compared with the low-value
The vector of variables X includes the category of applications, policy
dummies which equals one after the policy date (i.e. 9 June 2005) and
Table 3
zero otherwise and control variables.
Result of testing Hypothesis 1.
Variable Coefficient ME1
ME2
5.4. Limitations
Constant 4.4550 (0.3880) *** 0.5847
0.6258
Despite our efforts to collect as much data as possible and ensure its
ln(GFA) − 0.1575 *** − 0.0207
accuracy, there are still a few limitations in the data used to test the (0.0347) − 0.0221
hypotheses. Firstly, a large number of applications have incomplete PRIVATE − 1.8379 *** − 0.2412
information, so we had to exclude these observations from our analysis. (0.2813) − 0.2582
This may lead to sampling bias if the incomplete records are not random. HKI − 0.2534 − 0.0333
(0.1795) − 0.0356
Secondly, the size of the application project was measured using the HKI × GFA − 1.28 × 10− 5 ** − 1.68 × 10− 6
gross floor area of the completed project. However, some projects had no (5.64 × 10− 6) − 1.79 × 10− 6
structure (e.g. open carparks) and in these cases, we treated the car McFadden R squared 0.0883 No. of observations: 1440
parking areas as GFA. Thirdly, we assumed that each application was an Horowitz R squared 0.0855
independent event, but it is possible that the TPB members could be ME1: marginal effect evaluated at the mean. ME2: mean marginal effect.
influenced by the decisions of a similar application, particularly those Standard error in the parentheses. ***, ** and * indicate significant at the 1 %, 5
previously approved or rejected by the same TPB members. Fourthly, % and 10 % respectively.
5
K.W. Chau et al. Cities 137 (2023) 104291
Table 4 Table 6
Result of testing Hypothesis 2. Result of testing Hypotheses 2 and 3 together.
Variable Coefficient ME1 Variable Coefficient ME1
ME2 ME2
6
K.W. Chau et al. Cities 137 (2023) 104291
R2 is 0.1453 (McFadden R2 is 0.1510) which is within Hemmert’s reject private residential applications due to self-interest in sustaining
benchmark range of 0.09–0.17 (McFadden R2 range: 0.11–0.20). This residential prices.
suggests that our empirical data fit the model rather well.
Although our data fit the model well, one should bear in mind that a 8. Conclusion
low pseudo-R2 does not necessarily invalidate the empirical results. This
is especially true for dependent variables that involve a lot of human This study found by a logit analysis of 1440 sets of disaggregate
errors or judgements, since these variables tend to be noisy and a sig planning application and decision data, against intuition, that a policy
nificant portion of their variations cannot be explained by the observed change allowing greater public participation in objecting to planning
variables. In Hemmert et al.’s sample of published papers in top business proposals did not lower but increased the success rate of planning ap
and economics journals, close to 40 % of the papers involving logistic plications made by the private organizations, though the impact was
regression have pseudo R2 lower than their derived benchmark range of mainly on applications for low-value land uses. Consistent with the idea
“good fit”. of a lower perceived Approval Risk, non-profit making organizations
generally stood a higher chance of winning planning approvals and or
7. Discussion dinary developers found it harder to get high-value land uses approved.
It is hoped that this study will stimulate further research on the effect of
This study applied a principal-agent model to predict the TPB’s de public opinion on the planning authority decision making patterns.
cisions on planning applications for change of land zoned GIC use to Some possible policy implications can be identified from this study.
other uses. Central to model is the TPB’s (agent’s) behavior to minimize The shortage of developable land for housing and non-residential use in
Approval Risk under the constraints of the involvement of the principal Hong Kong has been a long term social and political problem in Hong
(citizens). Kong, which has led to a housing affordability issue and the gradual loss
The change of policy in 2005 provides an excellent opportunity for of international competitiveness for the territory’s business sector.
testing the impact of increasing the principal’s involvement in lowering Despite such shortages, the increase in the supply of developable land
the Approval Risk. We used planning application data concerning sites has been slow. An increase in developable land should benefit the so
zoned solely as GIC only so as to minimize the effect of other zonal ciety as a whole, although this may, at the same time, create more
factors that may affect the TPB’s decisions. business opportunities for developers. However, the Approval Risk
All empirical results are consistent with the predictions of the three minimizing behavior of the TPB has resulted in planning decisions that
hypotheses derived from the principal-agent analysis. First, it is gener may not be optimal for the citizens of Hong Kong, taking them as a
ally perceived that public and community use applications do not whole. Although greater involvement of the general public can lower the
involve private economic gain and thus lower perceived Approval Risk. Approval Risk, such involvement is not sufficient to lower the Approval
Second, the higher the perceived economic benefits that could be ach Risk for high-value land uses, which are the most needed for Hong Kong.
ieved by private applicants, the less likelihood of obtaining approval by Since the involvement of the principal occurs after the appointment of
TPB. Third, an increase in the opportunity for public participation in the the TPB members (ex post involvement), the impact on the alignment of
planning approval process has decreased the Approval Risk of the TPB the interest of the agent and principal is limited. One possible way to
members. However, the reduction in Approval Risk does not affect the further align the interest of the agent and principal is to involve the
approval of applications for high-value for-profit private land uses, agent in the selection of the principal by a more democratic process (ex
which usually result in potentially higher economic benefits for the ante involvement).
applicant. Moreover, regarding the principle-agent relationship, it is essential
The name of the GIC zone apparently has a systematic psychological to ensure that the TPB members are agents acting collectively on behalf
effect on the pattern of approvals against profit making applicants of their principal (all citizens). Systematic and effective communication
especially those land uses that can realise higher land values. Thus, it should be developed on a long-term basis, for example, better use of
was easier, as the results above show, for the TPB to feel more social media and platforms should be adopted rather than primarily
comfortable in accommodating public and community uses rather than using formal channels in the statutory planning system. Such means can
the for-profit private uses, as the perceived Approval Risk for the former enhance the decisions made by TPB increasing the probability that they
is lower. This interpretation of the findings is in line with the popular are in line with the citizens’ will.
sentiment against developer “profiteering”, which has the background Public participation can also be further enhanced by empowerment
that the original uses were leasehold land grants that were often, but not of the citizens’ capabilities and knowledge in the urban planning system,
always, charged a nominal premium for charitable or public utility and by a more collaborative planning approach where interactions be
purposes. tween different stakeholders in development projects can be adopted in
In any event, any use higher in value than the original use, unless both public and private sectors. The recent government directive to
otherwise permitted by the Crown (Government) Lease, attracted a constrain public participation may be unwise.
premium to be paid by the applicant and hence profiteering was a
perception rather than a real issue. In any case, compared with other CRediT authorship contribution statement
large-scale development areas such as Comprehensive Development
Area (CDA) zone development, the time required for GIC zone devel Chau, K. W: Conceptualization, Methodology, Software, Investiga
opment is still shorter, and hence GIC zones are attractive to developers tion, Formal analysis, Writing - Original Draft, Writing - Review &
as vacuoles for quicker development and return. Editing. Choy, Lennon H. T.: Conceptualization, Methodology, Writing -
Since residential use is one of the major for-profit high-value private Original draft preparation. Chua, Mark H.: Data Curation. Lai, Lawrence
uses and that many TPB members own one or more private residential W.C.: Conceptualization, Methodology, Data Curation, Writing - Orig
units, it is possible that some TPB members reject applications for inal Draft, Writing - Review & Editing, Supervision. Yung, Esther H.K.:
change of GIC zone to private residential zone so as to limit supply and Conceptualization.
sustain residential prices. To test this possibility, we replaced HIGH-
VALUE with a dummy variable (RESIDENTIAL) which equals 1 if the Declaration of competing interest
application is for change to private residential land use and zero
otherwise and re-estimates the models in Tables 4 and 6. The results are The authors declare that they have no known competing financial
similar, except that the coefficients of RESIDENTIAL are less significant interests or personal relationships that could have appeared to influence
than those of HIGH-VALUE. Therefore, it is unlikely that TPB members the work reported in this paper.
7
K.W. Chau et al. Cities 137 (2023) 104291
Data availability Lai, L. W. C., Ho, D. C. W., Chau, K. W., Chua, M. H., & Yu, R. K. K. (2015). Repeated
planning applications by developers under statutory zoning: A technical note on
delays in private residential development. Surveying and Built Environment, 24(1),
The authors do not have permission to share data. 8–36.
Lai, L. W. C., Ho, D. C., Chau, K. W., & Chua, M. H. (2016). Repeated planning
applications by developers under statutory zoning: A Hong Kong case study of delays
Acknowledgements and design improvements in private residential development. Land Use Policy, 57,
709–718.
The authors are thankful to our Research Assistant Mr. Ronald K.K. Lalli, M., & Thomas, C. (1989). Public opinion and decision making in the community.
Evaluation of residents’ attitudes towards town planning measures. Urban Studies, 26
Yu for data collection and processing.
(4), 435–447.
Ley, D., & Teo, S. Y. (2014). “Gentrification in Hong Kong? Epistemology vs. Ontology.
References International Journal of Urban and Regional Research, 38(4), 1286–1303.
Longarela-Ares, Á., Calvo-Silvosa, A., & Pérez-López, J.-B. (2020). The influence of
economic barriers and drivers on energy efficiency investments in maritime
Anderson, S., & Newell, R. G. (2003). Simplified marginal effects in discrete choice
shipping, from the perspective of the principal-agent problem. Sustainability, 12(19),
models. Economics Letters, 81(3), 321–326.
7943.
Besley, T. (2006). Principled agents? The political economy of good government: Oxford
Manville, M., & Osman, T. (2017). Motivations for growth revolts: Discretion and pretext
University Press on Demand.
as sources of development conflict. City & Community, 16(1), 66–85.
Besley, T., & Burgess, R. (2001). Political agency, government responsiveness and the
Marks-Bielska, R. (2021). Conditions underlying agricultural land lease in Poland, in the
role of the media. European Economic Review, 45(4–6), 629–640.
context of the agency theory. Land Use Policy, 102, Article 105251.
Besley, T., & Case, A. (2003). “Political institutions and policy choices: Evidence from the
McGuire, R. A., & Ohsfeldt, R. L. (1989). Self-interest, agency theory, and political voting
United States. Journal of Economic Literature, 41, 7–73.
behavior: The ratification of the United States constitution. The American Economic
Besley, T., & Kudamatsu, M. (2007). Making autocracy work. In LSE STICERD Research
Review, 79(1), 219–234.
Paper No. DEDPS48.
McFadden, D. (1974). Conditional logit analysis of qualitative choice behaviour. In
Bhat, C. R. (1997). Covariance heterogeneity in nested logit models: Econometric
P. Zarembka (Ed.), Economic theory and mathematical economics. Frontiers in
structure and application to intercity travel. Transportation Research Part B:
econometrics (pp. 105–142). New York: Academic Press.
Methodological, 31(1), 11–21.
McVittie, E., & Swales, J. K. (2007). The information requirements for an effective
Bonsón, E., Royo, S., & Ratkai, M. (2017). Facebook practices in Western european
regional policy: A critique of the allsopp review. Urban Studies, 44(3), 425–438.
municipalities: An empirical analysis of activity and citizens’ engagement’.
Miller, G. (2005). The political evolution of principal-agent models. Annual Review of
Administration & Society, 49(3), 320–347.
Political Science, 8, 203–225.
Butsic, V., Lewis, D. J., & Ludwig, L. (2011). An econometric analysis of land
Mitnick, B. M. (1975). The theory of agency. Public Choice, 24(1), 27–42.
development with endogenous zoning. Land Economics, 87(3), 412–432.
Monkkonen, P., & Manville, M. (2019). Opposition to development or opposition to
Carlevaro, F., & Sénégas, M. A. (2006). Simplified marginal effects in discrete choice
developers? Experimental evidence on attitudes toward new housing. Journal of
models: A correction. Economics Letters, 92(1), 44–46.
Urban Affairs, 41(8), 1123–1141.
Campbell, R. J., & Turnbull, G. K. (2003). On government structure and spending: The
Mustafa, A., Cools, M., Saadi, I., & Teller, J. (2017). Coupling agent-based, cellular
effects of management form and separation of powers. Urban Studies, 40(1), 23–34.
automata and logistic regression into a hybrid urban expansion model (HUEM). Land
Chau, K. W., & Lai, L. W. C. (2004). Planned conversion of rural land: A case study of
Use Policy, 69, 529–540.
planning applications for housing and open storage uses in agriculture zones.
Nguyen, H. (2011). The principal-agent problems in health care: Evidence from
Environment and Planning B: Planning and Design, 31(6), 863–878.
prescribing patterns of private providers in Vietnam. Health Policy and Planning, 26,
Chen, W. (2007). Analysis of rail transit project selection bias with an incentive
no. suppl_1, i53–i62.
approach. Planning Theory, 6(1), 69–94.
Nissim, R. (2016). Land administration and practice in Hong Kong Hong (4th Edition). Kong
Coats, J. C. (2002). Applications of principal-agent models to government contracting
University Press.
and accountability decision making. International Journal of Public Administration, 25
Permadi, D. B., Burton, M., Pandit, R., Walker, I., & Race, D. (2017). Which smallholders
(4), 441–461.
are willing to adopt Acacia mangium under long-term contracts? Evidence from a
Connolly, J. M. (2017). The impact of local politics on the principal-agent relationship
choice experiment study in Indonesia. Land Use Policy, 65, 211–223.
between council and manager in municipal government. Journal of Public
Poon, A. (2005). Land and the ruling class. Hong Kong: Alice Poon.
Administration Research and Theory, 27(2), 253–268.
Ross, S. A. (1973). The economic theory of agency: The principal’s problem. The
Daneshvary, N., & Clauretie, T. M. (2013). Agent change and seller bargaining power: A
American Economic Review, 63(2), 134–139.
case of principal agent problem in the housing market. Journal of Real Estate Finance
Sager, T., & Ravlum, I.-A. (2005). From projects to strategies: A transaction cost
and Economics, 47(3), 416–433.
approach to politicians’ problems with strategic transport planning. Planning Theory
Fearon, J. (1999). Electoral accountability and the control of politicians. In
& Practice, 6(2), 213–232.
A. Przeworski, B. Manin, & S. Stokes (Eds.), Democracy, accountability, and
Scott, A., & Vick, S. (1999). Patients, doctors and contracts: An application of principal-
representation. New York: Cambridge University Press.
agent theory to the doctor-patient relationship. Scottish Journal of Political Economy,
Ferejohn, J. (1986). Incumbent performance and electoral control. Public Choice, 5–25.
46(2), 111–134.
Ferris, J. M. (1992). School-based decision making: A principal-agent perspective.
Sheng, J., Zhang, S., & Li, Y. (2017). Heterogeneous governance capabilities, reference
Educational Evaluation and Policy Analysis, 14(4), 333–346.
emission levels and emissions from deforestation and degradation: A signaling model
Gailmard, S. (2010). Politics, principal–agent problems, and public service motivation.
approach. Land Use Policy, 64, 124–132.
International Public Management Journal, 13(1), 35–45.
Shi, X., Chen, S., Ma, X., & Lan, J. (2018). Heterogeneity in interventions in village
Gailmard, S. (2012). Accountability and principal-agent models. In Chapter prepared for
committee and farmland circulation: Intermediary versus regulatory effects. Land
the Oxford handbook of public accountability.
Use Policy, 74, 291–300.
Haila, A. (2017). Institutionalization of ‘the property mind’. International Journal of
Sng, T. H. (2014). Size and dynastic decline: The principal-agent problem in late imperial
Urban and Regional Research, 41(3), 500–507.
China, 1700–1850. Explorations in Economic History, 54, 107–127.
Hamman, J. R., Loewenstein, G., & Weber, R. A. (2010). Self-interest through delegation:
Stiglitz, J. E. (1989). Principal and Agent (pp. 241–253). UK: Palgrave Macmillan.
An additional rationale for the principal-agent relationship. American Economic
Strand, A., & Næss, P. (2017). Local self-determination, process-focus and subordination
Review, 100(4), 1826–1846.
of environmental concerns. Journal of Environmental Policy & Planning, 19(2),
Hayek, F. A. (1944). The road to serfdom. Chicago: University of Chicago.
156–167.
Hemmert, G. A., Schons, L. M., Wieseke, J., & Schimmelpfennig, H. (2018). Log-
Tang, B. S., & Choy, L. H. T. (2000). Modelling planning control decisions: A logistic
likelihood-based pseudo-R 2 in logistic regression: Deriving sample-sensitive
regression analysis on office development applications in urban Kowloon, Hong
benchmarks. Sociological Methods & Research, 47(3), 507–531.
Kong. Cities, 17(3), 219–225.
Horiuchi, A., & Shimizu, K. (2001). Did amakudari undermine the effectiveness of
Tang, B. S., Wong, S. W., & Lee, A. K. W. (2007). Green Belt in a compact City: A zone for
regulator monitoring in Japan? Journal of Banking & Finance, 25(3), 573–596.
conservation or transition? Landscape and Urban Planning, 79(3/4), 358–373.
Horowitz, J. (1982). An evaluation of the usefulness of two standard goodness-of-fit
Tang, B. T., Wong, S. W., & Lee, A. K. W. (2005). Green belt, countryside conservation
indicators for comparing non-nested random utility models. Transportation Research
and local politics: A Hong Kong case study. Review of Urban and Regional Development
Record, 874, 19–25.
Studies, 17(3), 230–247.
Huang, Y., Zhang, C., & Liu, W. (2019). Who drives the formation and adoption of the
Tang, B. S., & Tang, R. M. (1999). Development control, planning incentive and urban
‘increasing versus decreasing balance policy’?—Evidence from a policy process
redevelopment: Evaluation of a two-tier plot ratio system in Hong Kong. Land Use
analysis. Land Use Policy, 80, 175–184.
Policy, 16(1), 33–43.
Jensen, M., & Meckling, W. (1976). Theory of the firm: Managerial behavour, agency
Van Slyke, D. M., & Hammonds, C. A. (2003). The privatization decision: Do public
costs and ownership structure. Journal of Financial Economics., 3(4), 305–360.
managers make a difference? American Review of Public Administration, 33(2),
Lai, L. W. C., & Fong, K. (2000). Town planning practice: Context, procedures and statistics
146–163.
for Hong Kong, 1. Hong: Kong University Press.
Waterman, R. W., & Meier, K. J. (1998). Principal-agent models: An expansion? Journal
Lai, L. W. C., & Ho, W. K. O. (2001). Small is beautiful: A probit analysis of planning
of Public Administration Research and Theory, 8(2), 173–202.
applications for small houses in Hong Kong. Environment and Planning B: Planning and
Weber, A. (2015). Implementing EU co-financed Agri-environmental schemes: Effects on
Design, 28(4), 611–622.
administrative transaction costs in a regional grassland extensification scheme. Land
Lai, L. W. C., & Ho, W. K. O. (2017). Change in Use of Land: A Practical Guide to
Use Policy, 42, 183–193.
Development in Hong Kong (3rd ed.). Hong Kong: Hong Kong University Press.
8
K.W. Chau et al. Cities 137 (2023) 104291
Wong, S. H. W. (2015). Real estate elite, economic development, and political conflicts in Yu, K. H., & Hui, E. C. M. (2017). An empirical analysis of Hong Kong’s planning control
postcolonial Hong Kong. The China Review, 1–38. decisions for residential development. Habitat International, 63, 89–102.
Wong, S. K., Li, L., & Monkkonen, P. (2019). How do developers price new housing in a Zhang, H., Mu, J. E., & McCarl, B. A. (2018). Adaptation to climate change via
highly oligopolistic city? International Real Estate Review, 22(3), 307–331. adjustment in land leasing: Evidence from dryland wheat farms in the US Pacific
Yip, M. K. C., Lee, J. W. Y., & Tang, W. S. (2020). From God’s home to people’s house: northwest. Land Use Policy, 79, 424–432.
Property struggles of church redevelopment’. Geoforum, 110, 14–24.