The Relationship of Land Assembly and Prices of Cost Equivalent Land in the Equalization of Land Rights Act

The Case Study of Taiwan

Abstract

Since the first revision of the Equalization of Land Rights Act in 1958, urban land readjustment has become a dominant strategy and tool for urban land development. A crucial factor affecting financial self-liquidation of urban land readjustment is the distribution of cost-equivalent land, which is the land used to pay for public facilities within urban land readjustment regions. Cost-equivalent land can be used to fund land readjustment through market bidding. Studies on land prices have revealed that related factors, such as land-use intensity, individuals, and regions, affect the price of land. Studies have begun to focus on whether land transactions in urban land readjustment are similar to normal land transactions. The present study investigated the influence of land assembly on the land prices of cost-equivalent land and identified other factors by estimating hedonic price functions. The results indicate that general characteristics and land assembly are land price determinants. General characteristics, including the floor area ratio, width of the adjacent road, width of the base, depth of the base, street corner conditions, and bidding time, demonstrated a significant effect on the bidding price of cost-equivalent land. Fragmentation of the ownership of adjacent land had little effect on bidding price. Governments handling land allocation operations should consider the location of land. In addition, the aforementioned influencing factors are crucial and can lead to land having higher value.

Introduction

Urban land readjustment is an essential tool for developing new urban areas and revitalizing old cities. The tool has been used to achieve urbanization in Germany, Japan, South Korea, Taiwan, India, and Australia (Larsson, 1997). Taiwan’s urban land readjustment was originally implemented with the sole focus of improving the organization of its cadastral system. The system was later adjusted, however, with the focus of meeting urban development needs and obtaining land for public facilities (Hsieh, 2003). Urban land readjustment is based on urban development trends and urban planning. In accordance with the law, a specified range of deformed land in an area is reintegrated; roads and necessary public facilities are planned to improve the convenience of the urban development area. Apart from the land designated for public facilities, the buildable land in the readjustment areas is redistributed to the original landowners after a cost-equivalent amount of land is deducted as payment for the cost of readjustment.

Urban land readjustment is implemented by either the government or the private sector; few differences exist in the processing procedures of each. Urban land readjustment is a self-liquidating land development enterprise; the land for public facilities, construction and project costs, readjustment operation costs, and loan interest are funded by cash from the landowners in the target areas or from a section of the landowners’ unbuilt land that is equal in value to these costs (i.e., the cost-equivalent land). In cases involving government-led readjustment, the cost-equivalent land is auctioned publicly to obtain funding for construction and project costs. Any surplus in the land readjustment funding is subsumed into the government’s development fund for future readjustment. Overall, urban land readjustment can considerably assist with government financing and the development of urban areas.

Land development is an activity in which land is invested and operated to enhance economic value. For the public sector, land development can provide public facilities, commercial facilities, or housing that enrich urban functions and create a comfortable living environment. The private sector can also more easily obtain expected profits because of land development. Brounen, Eichholtz, et al. (2000) argue that development has the most appeal in real estate business because it leads to considerable profit opportunities. However, it is often accompanied by high risks. After readjustment is completed, the public sector often sets development schedules and allocates bulk rewards for large-scale overall development for land developers to motivate them to more quickly complete urban development and create a distinctive urban landscape through overall planning. Such rewards also often encourage developers to obtain large-scale land through land integration.

Readjustment areas generally have hundreds of plots of land that can be used for immediate construction, and the location or individual conditions vary. To reduce transaction costs, land developers prefer to purchase land at a particular location with a specific set of conditions and few landowners (Lin and Hsu, 2003). The number of landowners is a crucial factor affecting land transactions. The smaller the number of landowners is, the easier it is to negotiate. Simple land ownership reduces the transaction cost of land development and facilitates the promotion of development (Strange, 1995).

In the case of government-led readjustment, municipalities can guide the planning of land allocation. When the complexity of private land sovereignty is considered, the following questions arise: How can the difficulty of land integration be reduced and the transaction cost of land development be lowered in the readjustment process through location planning or scale planning of cost-equivalent land? How can large-scale overall development be achieved and the development process be accelerated through location planning or scale planning of cost-equivalent land? The present study was conducted to answer these questions.

This study analyzed the relationship between the complexity of land ownership and the bidding price of cost-equivalent land from the perspective of land integration. We used the third phase of Linkou New Town’s urban land readjustment in New Taipei City as a quantitative data source. To analyze land price, we referenced Rosen (1974), who proposed hedonic price theory, to analyze the compositional factors for real estate pricing, including location factors, general factors, and individual factors. The research results were expected to elucidate the effect of land integration on land prices and guide future land allocation in land readjustment. Additionally, the results could serve as a reference for other countries or regions that have adopted urban land readjustment as an urban development method.

With consideration of these research motives, this manuscript is organized as follows. In Section 2, urban land readjustment and cost-equivalent land institutions in Taiwan are introduced. In Section 3, we explain and present an analysis of cost-equivalent land’s price factors, including fragmentation of land ownership, location, general factors, and individual factors. In Section 4, we present our research design and hedonic price model. In Section 5, we present an analysis of the empirical results. Finally, we present a conclusion and recommendations for urban land developers and future studies.

Literature Review

Urban land readjustment and the cost equivalent land

Urban land readjustment is based on urban development trends and development policies and is used to reorganize messy, irregular, fragmented, or uneconomic usable land within a specific urban planning area or the periphery of a city (Home, 2007; Turk, 2008). After the public facilities are constructed, the plots of land are adjusted to ensure they have a suitable size, complete shape, and clear boundaries for construction. The land is then reallocated to the original landowners. Urban land readjustment is a well-known policy tool that is used to promote the economic and rational use of urban land and orderly urban development, with funding for planning and construction costs obtained from landowners (Archer, 1989; Liebmann, 2000).

This planning method originated in the Building Law of Zurich in 1893, which stipulated, “after the necessary space of the street is demarcated, all the lots are merged and subdivided, if the problem cannot be solved by only correcting or changing the cadastre, that is, to implement land readjustment” (Lee and Hsu, 2016). The Prussian Housing Law passed by the Prussian Parliament in 1918 also included land readjustment regulations, which prevented private land ownership from being too fragmented due to the strict inheritance system that was not conducive to land integration and construction. Therefore, the country’s housing law added land rezoning norms to prevent urban developable land from being insufficient and residential construction from being affected (Huang, 2013). Several European countries have also implemented land readjustment systems. For example, the Netherlands adopted land readjustment to increase agricultural land production. In East Asia, land readjustment was first implemented in Japan in 1926; after a strong earthquake, Tokyo adopted a land readjustment method for urban renewal.

Taiwan adopted land readjustment during the period of Japanese occupation. Japan implemented land readjustment in Taichung, Kaohsiung City, and other regions in 1937, with 20 readjustment regions spanning 4,397 hectares designated. After the Kuomintang recovered Taiwan in 1945, regulations regarding the readjustment of the city and land were implemented. In the initial stage of readjustment, because of a lack of a workforce, experience, and incomplete development of laws and regulations, the local governments did not implement comprehensive land readjustment practices. When the Urban Land Readjustment Implementation Act was first announced in 1979, the regulations related to land readjustment were relatively complete; thus, land readjustment businesses launched rapidly in urban areas. By 2014, Taiwan had processed 950 readjustment districts with a total area of 15,765 hectares and acquired more than 5,285 hectares of land for public facilities. Readjustment has provided more than 10,164 hectares of land for construction and saved more than NT$837 billion in government construction funds.

As previously mentioned, urban land readjustment is similar to a self-liquidating enterprise; landowners in the district must jointly pay for the provision of public facilities, construction and project costs, readjustment operation costs, and loan interest. The content of the costs include the cost of provision of land for public facilities and other costs. The cost of provision of land involves 10 types of land used for public facilities in readjustment areas, including roads, ditches, playgrounds, neighborhood parks, squares, green spaces, elementary schools, high schools, parking lots, and retail markets. In addition to those of public roads, the costs of ditches, rivers, and unregistered land are offset; landowners within the readjustment region pay the remainder of the cost on the basis of the proportion of the construction benefit.

Other costs include construction and project costs, readjustment operation costs, and loan interest. Construction and project costs include those related to the planning, design, construction, materials, project management fees, and taxes of the 10 aforementioned public facilities. In addition, such costs include signs, pipeline systems, planting, and other projects. Readjustment operation costs may include demolition compensation, fees related to cadastral arrangement, and the necessary business expenses for readjustment operations. Loan interest is considered the interest required to apply for a loan for the aforementioned expenses at a financial institution and must be repaid by all landowners.

According to Taiwan’s Equalization of Land Rights Act, if landowners lack sufficient funds to cover these costs, the landowners can fund readjustment by using unbuilt land whose value is equal to these costs, that is, by using cost-equivalent land.

Factors of cost equivalent land

Factors affecting real estate prices can be roughly divided into general factors, locational factors, and individual (structural) factors (Colwell and Munneke, 1999; Gloudemans, 2002; Goodman, 1978). Additionally, we investigate the effect of the integration of land adjacent to cost-equivalent land on the acquisition price when developers are incentivized by bulk rewards for large-scale development building. The more landowners involved, the higher the transaction costs are in the integration process, and the more difficult integration becomes. Therefore, the number of landowners and transaction costs of a land adjacent to cost-equivalent land are essential factors that affect land price. The following section presents an analysis of how transaction costs and ownership, general factors, regional factors, and individual factors affect the price of cost-equivalent land and our research hypotheses.

Transaction cost and ownership

Coase (1960) first proposed the concept of transaction cost, which is the additional cost that must be incurred in various production and transaction activities and includes search and information costs, communication and negotiation costs, execution costs, supervision costs, and dispute resolution costs (Dahlman, 1979; Coase, 2008). When the transaction cost exceeds the reasonable range of the economic activity, the economic activity cannot occur (North, 1990).

Buchanan and Tullock (1962) discussed the relationship between transaction costs and public choices in The Calculus of Consent. They argued that public concerns that are incorporated into laws involve external decision costs and decision-time costs; thus, a consensus must stem from the “calculus” that balances external and time costs. The concept of transaction costs has gradually expanded from economic and public policy discussions to the study of urban planning and urban readjustment (Alexander, 1992). Lin (2006) discussed the transaction costs that may be incurred in the process of land readjustment. Such costs include information collection, transactional object search costs, negotiation and bargaining costs, decision-making costs, execution costs, and contract supervision costs.

The goal of urban land readjustment is land ownership adjustment and land development and use. If a land developer wants to obtain a building bulk reward through large-scale development, land surrounding the cost-equivalent land must be integrated in addition to cost-equivalent land bidding. The transaction costs incurred during the integration process must be considered. This is particularly true for involuntary development planning, which landowners do not often welcome (Von Haaren and Vollheyde, 2019). Land developers must first consider the degree of fragmentation of land ownership; the more tenuous the ownership of the surrounding land is, the greater the number of co-owners is, which inevitably increases future negotiation costs (Strange, 1995). Land ownership fragmentation can also be caused by the landlord holding out to hinder the integration, being unable to obtain the bulk reward for building, and failing to achieve the original profit target (Cohen, 1991).

Buitelaar (2004) indicated that among the various transaction costs that affect land development, the number of participants or stakeholders is an essential factor affecting transaction costs. Urban land readjustment is a large-scale land integration process, and the number of participants in this process is far higher than that in single-land development. When the number of landowners or stakeholders involved in urban land readjustment increases, the negotiation process becomes more complicated, and the contracting and bargaining costs increase. Lin and Hsu (2003) argued that the number of people involved in land ownership is also an essential factor affecting land prices.

As the number of people involved in land ownership increases, the amount of time and size of the workforce required to integrate increases; moreover, if uncertainty increases, land transaction costs also increase. Therefore, we proposed the following hypothesis (H1):

H1: The number of landowners is negatively related to cost equivalent land prices.

General factors

General factors refer to the factors that affect the aggregate market condition, which consequently affects real estate prices. Although general factors have less influence on individual real estate prices than regional and individual factors do, they are the basis of real estate prices. General factors can be subdivided into four categories: environmental, social, economic, and administrative factors.

Environmental factors include natural conditions, such as those related to geology, soil, topography, and climate. Environmental factors also include artificial conditions, such as those related to noise, pollution, public security, and culture, and can either increase or reduce the potential for land development. Social factors include the population status, public construction, real estate transactions, habits, architectural styles, and lifestyles. For example, if the population increases, the demand for real estate, particularly in densely populated areas, also increases, which affects the local prices. This concept has been incorporated into theoretical models; for example, Karashimaet, Ohgai, et al. (2015) considered population to be a factor influencing land price in their model.

Economic factors include public savings, fiscal and financial status, interest rate levels, changes in industrial structure, and tax burdens. The effect of economic factors on real estate prices is relatively slow and occurs over a long period. Administrative factors include land-use plans, land control regulations, real estate transaction control, taxation, and housing policies. These factors can considerably affect the price of real estate.

Locational factors

Regional factors are the combination of natural conditions and social, economic, and administrative characteristics of the region in which the real estate is located that form the region’s characteristics and that affect the price level of the real estate in the region. For example, a cultural facility or school in a specific area can be zoned as a cultural and educational district with more favorable living functions, which increases the price of real estate in the area. However, if the area often floods, the real estate price decreases. Regions can readjust after natural disasters to improve their amenities, as Eom and Suzuki (2019) demonstrated by analyzing a region in Tokyo with higher real estate value, which could be traced back to land readjustment after the Great Kanto Earthquake of 1923. In addition to regional environmental differences, the focus of the conditions involved in regional factors may differ with the intended use of real estate. For example, the focus for residential areas is generally the distance from the city center, the quality of transportation facilities, and the comfort of the living environment, whereas the focus for commercial areas is generally the profitability of the real estate. For commercial land use, the economic value, types of businesses surrounding the real estate, and the state of competition are crucial. For industrial real estate, providing industrial infrastructure and convenient transportation, attracting a labor force, and procuring markets and raw materials are the focus.

Lin (2000) proposed that the factors influencing urban land prices include the distance between the land and central business area, the distance between public facilities, the density of buildings, and residential density in addition to general, regional, and individual factors. In addition, factors such as road width, street corner conditions, building strength, and the distance from the main road are the factors that affect the price of land (Lin, 2010; Hsu, 2012). Therefore, locational factors considerably influence cost-equivalent land prices, and the second hypothesis was formulated as follows:

H2: Locational factors of a parcel are related to cost equivalent land prices.

Individual (structural) factors

Individual (structural) factors refer to the elements that contribute to the individuality of real estate and individual prices. Individual factors can be land and building factors and include elements such as location, area, topography, geology, width, depth, shape, road conditions, and distance from public facilities. Several researchers (Gloudemans, 2002; Goodman, 1978; Kalra and Chan, 1994) have explored the relationship between price and real estate sales time by using sales time as a dependent variable. The empirical results revealed that location, area, the age of a house, and other variables affect the price of real estate. Sirmans, Stacy, et al., (2005) used a hedonic price model and determined that a house’s location, area, and age significantly affect real estate prices. Thus, we proposed the following hypothesis:

H3: Individual factors of a parcel (form, size, width of abutting roads, a street lot) are related to industrial land prices.

The literature review and research concept can refer to the following Figure 1.

Figure 1. The literature review and research concept

Methodology

Data collection

The present study investigated the case of the third phase of Linkou New Town’s readjustment area. The readjustment area is in the central area of Linkou District, New Taipei City and includes 631 cost-equivalent properties with an area of 48.47 hectares. The distribution is presented in Figure 2; the yellow area represents the distribution of the cost-equivalent land, the red area represents the individual building area, and the purple area represents the land for public facilities. We removed the plots land that were unsuccessfully marked and that were acquired through the policy project. We included 558 properties in our analyses; 338 were for separate construction, and applications to merge the construction assets with neighboring land had been received for 189. Because this study analyzed whether the fragmentation of land ownership in neighboring land affects the price of cost-equivalent land, the 189 properties were used as the sample for our empirical analysis.

Figure 2. Distribution of the cost equivalent land in the third phase readjustment of the Linkou New Town

To meet the urban development needs of Linkou New Town and accelerate its development, Linkou New Town is planned to be divided into residential and commercial zones in the third phase of the readjustment. On the basis of density and function of use, the residential and commercial zones are planned to have buildings with different coverage ratios, floor area ratios, and widths and depths.

Large-scale developers of building bases can receive bulk rewards. A base developed at a scale of 3000–5000 m² can receive extra 5% of the floor area; a base with a scale of 5000 m² or more can receive extra 5% of the floor area. In the present case, sites were more than 5000 m², and whole-block development was implemented; thus, 15% of the floor area of the floor area standard could be obtained, which served as the strongest incentive for land development through land integration.

Land value can vary with land zoning and whether a bulk reward for large-scale overall development building can be obtained. Land value can affect the land developers’ profits. Therefore, developers focus on land integration to obtain the highest bulk rewards they can and to maximize development profits. However, land integration is significantly associated with the degree of fragmentation of land ownership, which affects the cost-equivalent land price. Thus, we considered the floor area ratio, fragmentation of land ownership, and depth and width of building bases of land as variables in our empirical model to analyze the effect of cost-equivalent land prices.

The model

The hedonic pricing model

Rosen (1974) combined consumer theory, utility theory, competitive price theory, and advanced hedonic pricing theory to estimate the implicit prices of the characteristics that compound the price of composite goods. Hedonic pricing is implemented under the assumption that characteristics compound the price of a composite good and affect its market. Because lacking theoretical guidelines on which functional form is preferable in the hedonic pricing model, and the bias in estimated results from improper functions, Rosen (1974) proposed that numerous functions can be compared to identify those that are preferable and prevent biased empirical results from being obtained. Transformation was applied to the dependent variable to test the functional forms (Box and Cox, 1964). This method was used to identify the preferable functional form for the study data. The Box-and-Cox transformation model is as follows:

  

$$y^{\left(\theta \right)}=\sum _{i=1}^n{\beta }_i{x}_i^{\left(\lambda \right)}+\sum _{k=1}^m{\alpha }_k{Z}_k+\epsilon$$

$Y^{\theta }\{\begin{array}{c}\hfill \frac{(Y^{\theta }-1)}{\theta },\theta \ne 0\hfill \\ \hfill \mathrm{\text{ln}}Y,\theta =0\hfill \end{array}{X_i}^{\lambda }\{\begin{array}{c}\hfill \frac{({X_i}^{\lambda }-1)}{\lambda },\lambda \ne 0\hfill \\ \hfill \mathrm{\text{ln}}{X}_i,\lambda =0\hfill \end{array}$ …………(1)

where Y is the price of the cost-equivalent land; $X_i$ is an independent variable that is transformed by the Box-and-Cox transformation (e.g., numerical data); $Z_k$ is an independent variable that is not transformed by the Box-and-Cox transformation (e.g., dummy variables); $\epsilon$ is an error term; and ${\beta }_i$ , ${\alpha }_k$ , θ, and λ are the estimated coefficients.

According to the model, when θ = 1 and λ = 1, the function is a linear function; when θ = 0 and λ = 0, the function is a log–log function; and when θ = 1 and λ = 0 or when θ = 0 and λ = 1, the function is a semi-log function. The values of θ and λ were determined using a t test, which was used to identify the optimal function type rather than test each function.

Empirical model

We used the Box–Cox regression model to test Eq. (1) and the maximum likelihood estimation method to determine a suitable function for the hedonic price of the cost-equivalent land. The results did not reject the null hypothesis, in which the values of θ and λ were 0. Therefore, we used the log–log function and rewrote Eq. (1) as follows:

$lnL={\alpha }_0+{\sum }_{i=1}^n{\alpha }_{i}ln{X}_i$ ………………....……(2)

In Eq. (2), L is the unit price of the cost-equivalent land, and X is the independent variable.

Variables

This study was conducted on the basis of the premise that land developers wish to meet large-scale overall development regulations through land integration and obtain building bulk rewards. This study evaluated the factors affecting the bidding price of cost-equivalent land. Additionally, we analyzed whether the fragmentation of land ownership in the areas surrounding cost-equivalent land affected the bidding price. The dependent variable was the unit bidding price of the cost-equivalent land (NT$10,000/m²). The independent variables were separated into two categories: structural and locational variables. The structural variables included individual factors, such as site area, adjacent road width, width and depth of the base, corner land, floor area ratios, time variables, and the degree of fragmentation of adjacent-land ownership. Locational variables included the distances to schools, parks, and town squares, which are associated with a price–distance gradient and are neighborhood attributes that are block-like effects and uniform within a spatially delimited area.

The relationship between individual factors and the price of land indicates that flexibility of the planning, design, and use of a building increases with the size of the area of its building base, thus individual factors positively affect the land price. It results from the higher floor area ratio of the base, the more floor area can be built, which reflects selling profit, eventually increase the value of a base. In addition, the greater the width of the road facing the base is, the greater the volume is of vehicles that can be supported. Consequently, development of ancillary facilities, such as greening, sidewalks, or bicycle paths, becomes more feasible. Similarly, the greater the width of the road is, the more favorable the road conditions are, which positively affects land prices. Furthermore, the larger the width of the base is, the squarer the base is, and the larger the contact area with the street is, the more flexible planning and land use can be, which positively affects land prices. The land price of adjacent street bases that are closer to the street is higher; the land-use value and land price increase with the closeness of the base to the road line. Therefore, a deeper base can negatively affect spatial use and such a base would have a negative sign. Corner areas face more roads and can be more conveniently reached. Moreover, the view, lighting, and ventilation are more favorable in corner areas than they are on land further away from the road, which leads to corner plots having higher commercial value. Therefore, a plot of land being corner land likely has a positive effect on land price. Corner land was considered to be a dummy variable that equaled 1 if the base site was on a street corner and 0 if not.

Regarding the degree of fragmentation of ownership of adjacent land, the land-use control authority of the third phase of the new township of Linkou New Town formulated large-scale development incentives that could be obtained through land ownership integration. The degree of difficulty of integration increases with the complexity of ownership of adjacent land; complex land ownership can lead to higher transaction costs and negatively affect land prices. In the present study, the amount of land adjacent to a base was used as the denominator. After whether each neighboring plot exhibited co-ownership characteristics was determined, the number of co-owned properties was calculated as the numerator. The resulting fraction was considered to indicate the degree of fragmentation of adjacent-land ownership. In other words, when the land adjacent to the cost-equivalent land was not co-owned, its value was zero, which was the minimum value. When the land adjacent to the cost-equivalent land was co-owned, the maximum value was 1. The analyzed bidding period was 2004 to 2006. Because the land price increased during the bidding period, the present study considered 2004 to be the base year. We predicted that the later the bidding period was, the higher the land price would be; that is, we predicted bidding time would positively affect land prices.

When analyzing the regional factors of bases, we considered the distances to schools and parks to be empirical variables. Because primary and high schools are essential public facilities in readjustment areas, the farther the distance from a base to a school is, the lower accessibility is, which would negatively affect land price. This distance was calculated as the shortest straight-line distance from a base to a school on a map. Neighborhood parks are primary public facilities in daily life; thus, the farther a base is from a park, the lower its accessibility is. Thus, the distance from a base to a park would negatively affect land prices. This distance was calculated as the shortest straight-line distance from a base to a park or plaza. The empirical variables of this study and their descriptions are listed in Table 1.

Table 1. Variables in empirical model

Variable	Symbol	Name	Description
Dependent Variable	Price	Unit price	Unit bidding price of the cost equivalent land (NT$10,000 / m2)
Structural Variables	Area	Area	The area of the cost equivalent land (m2)
	Ratio	Floor area ratio	The ratio of buildable floor area to a unit buildable land area (%)
	Road	Width of the adjacent road	The maximization width of the road adjacent to the cost equivalent land (m)
	Width	Width of the base	The cost equivalent land’s base width adjacent to main road (m)
	Depth	Depth of the base	The distance of a line extending from the front street line to the rear lot line (m)
	Corner	Corner land	If the land is on the corner of street is coded 1, and 0 otherwise.
	Owner	Fragmentation of the ownership of adjacent land	The number of lands adjacent to the cost equivalent land is used as the denominator, and the number of adjacent co-owner lands is the numerator. The value of the fraction is between 1 and 0, and with the value closer to 1, it represents the adjacent land with more complex co-ownership structure.
	Time	Bidding time	The bidding time of each cost equivalent land, the base period is 2004, and the initial value of time-variable is 0. Adding 0.5 with half-year, for instance, the case traded in March 2005, the time value is 1.5. (year)
location Variable	School	Distance from the base to the school	The shortest straight-line distance from the cost equivalent land to the school on the map. (m)
	Park	Distance to the neighborhood park	The shortest straight-line distance from the cost equivalent land to the neighborhood park on the map. (m)

Note: NT$10,000/m² represented that unit bidding price of the cost equivalent land is in NTD 10,000 per square meter.

Empirical Results and Discussion

This section first presents an analysis of the basic attributes of the sample data. Thereafter, a description of a correlation test between the variables is provided. This test was used as a reference for the variable selection of the hedonic price model. Finally, a description of an empirical test of the hedonic price model is presented.

Sample data analysis

To determine the variables that affect the hedonic price of cost-equivalent land, we used the actual unit bidding price as the dependent variable. The independent variables are presented in Table 1. The characteristics of the sample data are analyzed as follows.

The dependent variable – the unit bidding price of the cost equivalent land

This study analyzed the actual unit price of 189 samples of cost-equivalent land in 2004 to 2006 as the dependent variable. The highest unit price was NT$71,666/m², and the lowest was only NT$23,243/m². The average unit price was approximately NT$38,156/m². These values indicated considerable differences in the standard error.

Independent variable

Regarding the sample’s individual (structural) factors, the largest land area was 2565.34 m², and the smallest was 133.17 m². The average area was 588.70 m². The largest land area did not reach the 3000 m² that would enable its development to qualify as large-scale overall development, and therefore, building bulk rewards could only be obtained through land integration. The maximum floor area ratio was 500% in the central commercial area, and the minimum was 120% in the first category of residential area. Among the 189 samples, 27 of their bases were on corner land, and 162 were on noncorner land. Regarding the fragmentation of adjacent-land ownership, the minimum value was 0, which indicated that the land adjacent to the cost-equivalent land was not co-owned. Additionally, the maximum value was 1, which indicated that the adjacent land had the highest degree of fragmented ownership.

In terms of time, the minimum value was 0, which indicated that the bidding period occurred in 2004, and the maximum value was 6, which indicated that the bidding period occurred in 2006. The average waiting period from development to public bidding was 1.5 years. In terms of locational factors, the shortest linear distance from a base to a school was 20 m, the longest was 641 m, and the average was 225.71 m. The shortest linear distance to a neighborhood park or plaza was 1 m; the base was adjacent to a park. The longest distance was 437 m, and the average was 124.31 m. The characteristics of the samples are presented in Table 2.

Table 2. Descriptive statistics of characteristic variables of the cost equivalent land’s bid price

Variables	Name	Min	Max	Average	Standard error	Variance	Sign
Dependent Variable	Price	23243	71666	38156.39	13157.34	17311.3
Independent Variable	Area	133.17	2565.34	588.70	441.61	195015.06	positive
	Ratio	120%	500%	238%	127%	161%	positive
	Road	6.00	40.00	12.96	7.25	52.57	positive
	Width	6.02	90.51	19.22	12.20	148.91	positive
	Depth	7.35	73.76	29.24	9.23	85.16	negative
	Corner	-	-	-	-	-	positive
	Owner	0.00	1.00	0.43	0.31	0.10	negative
	Time	0	6	1.5	0.90	0.82	positive
	School	20.00	641.00	225.71	129.40	16743.86	negative
	Park	0.00	437.00	124.31	86.47	7477.64	negative

Empirical result

Before the model analysis, we performed a correlation coefficient analysis between the variables. The results for the independent and dependent variables listed in Table 2 were significant (α = 0.05), which indicated a degree of correlation between the variables.

We additionally employed the log–log model and ordinary least squares test. The results revealed that when noninfluential variables were eliminated and the regression test was repeated, fragmentation of adjacent-land ownership was significant (α = 0.1), and other variables also reached significance (α = 0.05), with the exception of base area, distance to a school, and distance to a park. Thus, fragmentation of adjacent-land ownership significantly affected the price of cost-equivalent land. Overall, the result of the F test of the model was significant (α = 0.01), and the adjusted R² reached 0.97, which indicated that the model estimation had a high degree of explanatory power. Because collinearity between independent variables may cause model estimations to be biased, the variable inflation factor (VIF) was used to determine whether collinearity was present between the independent variables. The VIF of the independent variable results revealed all independent variables to have values of <10 (Table 3), which indicated no collinearity was present among variables.

Table 3. The result of the bid price of cost equivalent land from the log-log model of the hedonic price model

Variables	Original			Revised
	Coefficient	T test	P value	Coefficient	T test	P value	VIF
Constant	-72419.10	-3.63	0.00***	-84725.32	-4.51	0.00***
Area	-0.20	-0.26	0.80	-	-	-
Ratio	10452.13	23.93	0.00***	9903.78	67.26	0.00***	1.36
Road	259.02	9.89	0.00***	265.89	10.40	0.00***	3.75
Width	58.51	1.46	0.15	50.07	1.97	0.05**	3.75
Depth	-84.49	-3.04	0.00***	-82.22	-3.13	0.00***	2.30
Corner	2740.49	3.28	0.00***	2704.27	3.99	0.00***	2.21
Owner	-0.64	-1.83	0.09*	-18.90	-1.92	0.09*	1.03
Time	939.08	4.57	0.00***	1043.57	5.26	0.00***	1.26
School	0.89	0.58	0.56	-	-	-	-
Park	3.12	1.53	0.13	-	-	-	-
F-test	505.77***			821.42***
R2	0.97			0.97
Adj R2	0.97			0.97

Note: ***represent the significant is α=0.01; ** represent the significant is α=0.05; * represent the significant is α=0.1.

Analysis and discussion

Transaction cost and ownership

The proximity variable of land integration, namely fragmentation of adjacent-land ownership, was only significant at α = 0.1; which may support H1. However, we also observed t-test values that revealed a minor effect on land price in the results of both the original and revised model. The development process of Taiwan may partially explain why the variable had only a minor effect. In most land readjustment cases, landowners are inclined to develop large areas; thus, integration is essential after readjustment. Through the integration process, the land price reflects the effect of fragmentation of adjacent-land ownership over time. Fragmentation was correlated with parcel size; Ritter, Hüttel, et al. (2020) reported similar results in which plot size positively affected land price. Therefore, the aforementioned results weakly support H1. Developers often wish to own land outright, which creates a price gap in land ownership.

When land is put on the market, the land can be developed independently to maximize profits. Some individual development is completed in relatively small lots (Figure 1). Therefore, the fragmentation of adjacent-land ownership is not an absolute determiner of land price, and we can infer that the land with more complex ownership and adjacent to the samples may be lower in value. If buildings cannot be built on neighboring land, a developer can achieve their development goals through separate developments.

Locational factors

Two variables, distance to a school and distance to a park, did not have significant results; therefore, H2 was not supported. This may be because the land that was intended for schools was not indicated on the development schedule and the school site remained under construction when the development of the third phase of the readjustment area of Linkou New Town was completed. Thus, proximity to a school did not cause the bid price to increase, which was different from the findings of most land price studies, such as Lee (2015). Nevertheless, 43 parks and plazas were evenly distributed in the readjustment area, which satisfied the needs of each region. Therefore, little difference was observed in the conditions of each base, and the bid price of the cost-equivalent land was not notably affected.

Individual (structural) factors

According to the empirical results, six individual (structural) factors (i.e., floor area ratio, maximum road width, base width, base depth, corner land, and bidding time) were significant at α = 0.05. These individual factors affected the price of cost-equivalent land. This finding indicates that because the amount of floor area that could be built increased with the floor area ratio of a base, land with larger bases had higher bid prices. These results were consistent with those of other studies (Gao, Asami, et al., 2006; Kim, Park et al., 2007). In addition, the wider the adjacent road was, the more accessible the plot was, and the higher the economic activity and bid price of the land was. Similarly, the greater the width of the base was, the more flexible the land allocation and land use was; base width demonstrated a positive relationship with the bidding price. Corner bases had higher land use and were more conveniently reached through transportation, which contributed to an increase in their bid prices. This finding coupled with the rising trend of land prices in the Linkou area indicated that the later the bidding period was, the higher the bidding prices were, and the greater the base depth was, the farther away the adjacent road was, which resulted in unfavorable bid prices. The empirical results were in line with the previous inferences of the present study.

The flexibility of building construction or expansion was predicted to increase with the size of the land area, which was predicted to have a positive relationship with the bid price. However, land area was negatively associated with bid price, which indicated that land with a larger area does not have a higher price. This finding may be explained by land transaction costs causing a nonlinear relationship to develop between land price and land area. The land transaction costs included in estimations of land price (Colwell and Munneke, 1999) can cause convexity in the price. The empirical data did not fully support H3.

Conclusion and Recommendation

Urban land readjustment is a self-liquidating enterprise. Government-led urban land readjustment involves reorganizing land within an urban planning area and adjusting roads and necessary public facilities in accordance with a set of regulations. Consequently, the urban development area becomes an efficient street block. With the exception of the land zoned for public facilities, the buildable land in the readjustment area is redistributed to the original landowners after cost-equivalent land is deducted as payment for the cost of readjustment. The government then balances the costs of urban land readjustment by auctioning the cost-equivalent land. Cost-equivalent land with a higher bidding price can be used to pay for development costs. The development surplus can then be reinvested in the readjustment area and used to support the government.

To promote the development of Linkou New Town and create a distinct urban landscape, the Linkou District plan offered a bulk reward for large-scale overall development building or reward at the time of development. The results of the present study reveal that in the third phase of readjustment in Linkou New Town, floor area ratios, widths of adjacent roads, widths and depths of bases, plots being corner land, and bidding time significantly affected the bid price of cost-equivalent land. Fragmentation of the ownership of adjacent land had only a minor effect on bid prices. Therefore, governments should consider such factors in addition to a plot’s location to increase land value when allocating land.

The number of completed public facilities in a readjustment area can affect the bid price of cost-equivalent land. We discovered that the distance from a base to a school and to a park did not significantly affect the bid price of the cost-equivalent land, which may have been because of public facilities being underdeveloped and public maintenance being lacking. After land for public facilities is allocated, appropriate maintenance or management protocols should be enacted if immediate construction is not required. For example, small sports venues, green spaces, or temporary parking lots can be constructed. In addition to improving the environmental aesthetics, such projects can reduce maintenance and managerial costs.

Although the present study analyzed a considerable amount of primary data, we could not determine the auction times and number of bidders for each plot of land because the data were out-of-date. If these missing data could be collected, we could explore the effect of ownership on the bidding price of cost-equivalent land in greater depth. Future studies can analyze the number of auctions and the price of cost-equivalent land. Because cost-equivalent land is normally not auctioned off the first time it is put up for auction, future studies could investigate the effects of the number of times a plot is put up for auction and the final transaction costs on the bidding price. Additionally, studies could analyze the number of bids and the bid-off price of cost-equivalent land to understand which characteristics of cost-equivalent land attract more bidders.

Author Contributions

Conceptualization, L.H.R.; methodology, L.S.W. and S.H.J.; writing—original draft preparation, L.S.W., S.H.J. and L.H.R.; writing—review and editing, L.S.W. and S.H.J. All authors have read and agreed to the published version of the manuscript.

Ethics Declaration

The authors declare that they have no conflicts of interest regarding the publication of the paper.

Acknowledgments

We would like to thank the editors and anonymous reviewers for their valuable and constructive suggestions for improving this paper.

References

•  Alexander , E. R. (1992). “A Transaction Cost Theory of Planning”, Journal of the American Planning Association, 58: 190–200. doi: https://doi.org/10.1080/01944369208975793.
•  Archer , R. W. (1989). “Transferring the urban land pooling/readjustment technique to the developing countries of Asia.”, Third World Planning Review, 11 (3): 308–30. Retrieved from https://www.proquest.com/openview/d59d3b83bb591b2167ea6c20f8559fc1/1?pq-origsite=gscholar&cbl=1820928.
• Box, G. E. P., and Cox, D. R. (1964). “An analysis of transformations”, Journal of Royal Statistic Society, Series B, 26, 211-243. doi: https://doi.org/10.1111/j.2517-6161.1964.tb00553.x.
•  Brounen,  D.,  Eichholtz , P. M. A., et al. (2000). “The effect of property development on the performance of REITs”, Real Estate Finance, 16(4), 17-22. Retrieved from https://hdl.handle.net/11245/1.182277.
• Buchanan, J. M. and Tullock, G. (1962). The Calculus of Consent. Ann Arbor, MI: The University of Michigan Press. doi: http://doi.org/10.3998/mpub.7687.
•  Buitelaar,  E. (2004). “A transaction-cost analysis of the land development process”, Urban Studies, 41(13), 2539-2553. doi: https://doi.org/10.1080/0042098042000294556.
•  Coase , R. H. (1960). “The problem of social cost”, Journal of Law and Economics, 3, 1-44. doi: https://doi.org/10.1086/674872.
•  Coase , R. H. (2008). “The institutional structure of production” in Ménard, C. and Shirley, M. (eds.), Handbook of New Institutional Economics, Berlin: Springer, 31-40. doi: https://doi.org/10.1007/978-3-540-69305-5_12.
•  Cohen,  L. (1991).” Holdouts and free riders”, Journal of Legal Studies, 20,351-362. doi: https://doi.org/10.1086/467890.
•  Colwell , P. F. and  Munneke , H. J., (1999). “Land prices and land assembly in the CBD”, Journal of Real Estate Finance and Economics, 18(2), 163-180. doi: https://doi.org/10.1023/A:1007714624700.
•  Dahlman , C. J. (1979). “The problem of externality”, The Journal of Law and Economics, 22(1), 141-162. doi: https://doi.org/10.1086/466936.
•  Eom,  s. and  Suzuki,  T. (2019). “Spatial distribution of pedestrian space in central Tokyo”, International review for spatial planning and sustainable development, 7(2), 108-124. doi: https://doi.org/10.14246/irspsda.7.2_108.
•  Gao,  X.,  Asami,  Y., et al. (2006). “Evaluating land-use restrictions concerning the floor area ratio of lots”, Environment and Planning C: Government and Policy, 24, 515-532. doi: https://doi.org/10.1068/c0531.
•  Gloudemans , R.J. (2002). “An Empirical Analysis of the Incidence of Location on Land and Building Values”, Lincoln Institute of Land Policy Working Paper WP02RG1, 113 Brattle St., Cambridge, MA. Retrieved from https://www.lincolninst.edu/sites/default/files/pubfiles/122_gloudemans02.pdf.
•  Goodman , A. C. (1978). “Hedonic prices, price indices and housing markets”, Journal of Urban Economics, 5(4), 471-484. doi: https://doi.org/10.1016/0094-1190(78)90004-9.
•  Home,  R. (2007). "Land Readjustment as a Method of Development Land Assembly: A Comparative Overview", The Town Planning Review, 78(4), 459-483. doi: https://doi.org/10.3828/tpr.78.4.4.
•  Hsu , H. J. (2012). Applying Hedonic Price Model to Evaluate the Price of the Land Rezoning in the Previous Tainan City (master’s dissertation), University of Kang Ning, Tainan, Taiwan. Retrieved from https://hdl.handle.net/11296/ecnku3.
•  Hsieh , J. C. (2003). “The discussion of urban land consolidation”, Journal of the Land Bank of Taiwan, 40(3), 15-31. Retrieved from https://tpl.ncl.edu.tw/NclService/JournalContentDetail?SysId=A03028211
•  Huang , Z. L. (2013). “Urban land readjustment policy and regional impacts: readjustment by private sector effect in Taichung city”, Bulletin of the Geographical Society of China , 51, 35-51. doi: http://doi.org/10.29972/BGSC.201312_(51).0003.
•  Kalra,  R., and  Chan, K. (1994). “Censored sample bias, macroeconomic factors, and time on market of residential housing”, Journal of Real Estate Research, 9(2), 253-262. doi: https://doi.org/10.1080/10835547.1994.12090750.
•  Karashima,  K.,  Ohgai,  A., et al. (2015). “A Spatial Simulation Model to Explore Agglutination of Residential Areas and Public Service Facilities”, International review for spatial planning and sustainable development, 3(4), 57-74. doi: http://dx.doi.org/10.14246/irspsd.3.4_57.
•  Kim , K. S.,  Park , S. J., et al. (2007). “Highway traffic noise effects on land price in an urban area”, Transportation Research Part D: Transport and Environment, 12(4), 275-280. doi: https://doi.org/10.1016/j.trd.2007.03.002.
•  Larsson,  G. (1997). “Land readjustment: A tool for urban development”, Habitat International, 21(2), 141-152. doi: https://doi.org/10.1016/S0197-3975(96)00059-8.
•  Lee,  T. C and  Hsu , Y. C. (2016). “City Rezoning in Urban Planning”, Taiwan Architect Magazine, 42(1), 84-89. Retrieved from https://tpl.ncl.edu.tw/NclService/JournalContentDetail?SysId=A16003157
•  Lee , Y. K. (2015). “School districting and the origins of residential land price inequality”, Journal of Housing Economics, 28, 1-17. doi: https://doi.org/10.1016/j.jhe.2014.12.002.
•  Liebmann , G. W. (2000). “Land readjustment for America: A proposal for a statute”, Urban Lawyer, 32 (1), 1–20. Retrieved from https://www.jstor.org/stable/27895225.
•  Lin, C. M. (2010). The Effects of bidding price of the cost equivalent land in reorganization area -Take Kaohsiung county for example (master’s dissertation), National Sun Yat-sen University, Tainan, Taiwan. Retrieved from https://hdl.handle.net/11296/8vd3sg
•  Lin , T. A. (2000). Land Prices and Floor Area Ratios (master’s dissertation), National Cheng Kung University, Tainan, Taiwan. Retrieved from https://hdl.handle.net/11296/3jacn6
•  Lin , T. C. and  Hsu , M. F. (2003). “Adjustments to ownership on individual sites prior to land development -A case study of land readjustment area”, Journal of Taiwan Land Research, 6(2), 1-16. Retrieved from https://tpl.ncl.edu.tw/NclService/JournalContentDetail?SysId=A03050164
•  Lin , Y. W. (2006). The Study of Applying Transaction Cost Theory to Analyze the Land Development Mechanism in Taiwan-A Comparison between Old Downtown and New Development Areas (master’s dissertation), National Cheng Kung University, Kaohsiung, Taiwan. Retrieved from https://hdl.handle.net/11296/8s6w3x
• North, D. C. (1990). Institutions, Institutional Change and Economic Performance, Cambridge: Cambridge University Press. doi: https://doi.org/10.1017/CBO9780511808678.
•  Rosen,  S. (1974). “Hedonic prices and implicit markets: product differentiation in pure competition”, The Journal of Political Economy, 82(1), 34-55. doi: https://doi.org/10.1086/260169.
•  Ritter,  M.,  Hüttel,  S., et al. (2020). “Revisiting the relationship between land price and parcel size in agriculture”, Land Use Policy, 97, 104771. doi: https://doi.org/10.1016/j.landusepol.2020.104771.
•  Sirmans,  G.  Stacy ,  David  A. M. and  Emily  N. Z. (2005). “The composition of hedonic pricing models”, Journal of Real Estate Literature, 13(1), 3-43. doi: https://doi.org/10.1080/10835547.2005.12090154.
•  Strange , W. C. (1995). “Information, Holdouts, and Land Assembly”, Journal of Urban Economics, 38, 317-332. doi: https://doi.org/10.1006/juec.1995.1036.
•  Turk , S. S. (2008). “An Examination for Efficient Applicability of the Land Readjustment Method at the International Context”, Journal of Planning Literature, 22(3), 229-242. doi: https://doi.org/10.1177/0885412207310283.
•  Von Haaren,  C. and  Vollheyde,  A. (2019). “Landscape planning in Germany: Not loved by all, but badly needed”, International Review for Spatial Planning and Sustainable Development, 7(4), 148-166. doi: https://doi.org/10.14246/irspsda.7.4_148.
•  Williamson , O. E. (1987). “Transaction cost economics: The comparative contracting perspective”, Journal of economic behavior & organization, 8(4), 617-625. doi: https://doi.org/10.1016/0167-2681(87)90038-2.