ASIAN DEVELOPMENT BANK LAND POOLING IN NEPAL FROM PLANNED URBAN “ISLANDS” TO CITY TRANSFORMATION

viii


TABLES
Although Nepal is one of the 10 least urbanized countries in the world, its urban population is growing very rapidly. Of its total population of about 29.6 million in 2018, 5.5 million (slightly below 20%) was urban-3.8% more than its urban population in 2017, and 28% higher than the corresponding figure only a decade earlier. 1 More than 50% of this urban population is concentrated in the Kathmandu Valley.

2.
Urbanization in Nepal tends to be rapid, sprawling, and fragmented. This trend is driven by the shift from a rural to an urbanizing economy, the associated drive for economic development and competition for land, and either lack of land regulations or weak enforcement capacity, among other factors. The current trajectory is quickly locking Nepal's municipalities into an urban form that is hampering land-use efficiency, mobility, and infrastructure delivery, and making municipal services less accessible and poorer in quality. The failure of the public sector to guide spatial growth in Nepal has generally lowered the living standards of most people in the cities. Environmental degradation is high, and there are few public open spaces such as parks. The inability to provide adequate serviced land for residential, commercial, and public use is stifling the economic potential of Nepal's growing cities.

3.
Nepal's geography is split into three distinct regions: (i) the upper Himalayan mountains, bordering the People's Republic of China and having the lowest concentration of habitation (7% of the population); (ii) the middle hills, with several valleys (43%); and (iii) the southern plains, or Terai, which borders India (50%). 2 Nepal's urban population is concentrated in the Terai and the valleys, including Kathmandu and Pokhara. Besides the expansion of existing municipalities, there are a few major developments such as the Mid-Hill Highway and the East-West Highway, which transect the country and are transforming Nepal's landscape on an unprecedented scale. Urbanization rates are lower in other geographic regions outside the Kathmandu Valley, especially in the hill and mountain regions.

4.
There is a natural tendency across Nepal to build houses, shops, and factories near a main road. Without active land management policies and initiatives to accompany growth, development is rapidly taking place in ribbon or linear forms along highways and peri-urban roads ( Figure 1). This form of narrow sprawl along roads makes it difficult for government to provide basic services because of the long distances to be covered at low densities. Ribbon development is not conducive to building social ties and strong community, and people generally face nuisances associated with heavy traffic such as dust, noise, and compromised safety, at least on major roads like the East-West Highway. In this growth pattern, land located away from the road is left underdeveloped or tends to develop on an ad hoc basis, resulting in inefficient land use. Ribbon development also negatively affects the efficiency of travel on the main road, by building highways through neighborhoods. 3 1 Data derived from the most recent National Population and Housing Census (2011). Figures account for additional population after village development committees were merged with urban municipalities, which occurred after the census was taken. 2 National Population and Housing Census 2011. 3 The United Kingdom introduced the Restriction of Ribbon Development Act in 1935 to limit the number of access points to trunk roads so that these roads can function properly as channels for longer-distance traffic.

5.
The pattern of urban expansion in the Kathmandu Valley shows how a lack of planning and of legal precedent for zoning regulations has resulted in nonsystematic growth. An analysis of population growth and land conversion using satellite data has revealed that the urban area expanded by about 400% between 1989 and 2016, and that this expansion occurred mostly with the 31% of agricultural land converted into built area ( Figure 2). Most of this growth took place over the 20 years between 1989 and 2009. The capital region continues to develop rapidly along the major roads as in the ribbon development shown above, significantly altering the urban form of the valley (see photo on page 3).

6.
In 2017, mayors were elected in Nepal after 15 years without elected representatives at the local level. The 2015 Constitution shifted the country from a unitary to a federal system of government, and strengthened local governments with new judicial, legislative, and executive powers. This shift to federalism has brought renewed focus on urban planning, service delivery, and public investments. Elected representatives are eager to move ahead with development projects to garner public support, but the investments are not always framed within a broader development plan. Devoid of larger plans, investments still tend to be ad hoc and sector-based, lacking integration with any strategic development framework. Municipalities, then, urgently require tools and regulations to actively manage land and population growth, and to acquire land for public purposes.

7.
Land pooling is one of several tools that have been implemented by government across Nepal in the past several decades to secure land for planned residential development, infrastructure, and other public purposes-an especially critical objective for a country like Nepal, where government does not own a significant amount of land. Alternatives to land pooling generally involve land acquisition, or the use of government's power of eminent domain to acquire land for which the owner is paid compensation. Developments using land acquisition have primarily been either guided land development schemes (where only the land needed for infrastructure is acquired from the landowner) or sites and services schemes (where government acquires the entire land area, develops plots, and sells or leases them). Land pooling is seen as an attractive alternative because once land is developed, it is transferred back to the original landowner, and instead of cash compensation, those landowners benefit from infrastructure and higher land values. Government in turn benefits from the self-financing nature of land pooling, since no compensation is paid and infrastructure can be financed from the sale of a portion of the land.

8.
However, the use of land pooling in Nepal has stagnated in recent years on account of bottlenecks that have hampered implementation, and the resistance of landowners to schemes that may take years or well over a decade longer to complete than originally planned. As local governments look to rein in uncontrolled and unsustainable urbanization, some questions beg for answers: Is land pooling a viable planning tool in Nepal? If so, how can the challenges of past projects be overcome, and successes scaled up? What useful lessons can be learned from other countries with long experience using land pooling to shape their cities? Can and should Nepal transition from using land pooling for small, neighborhoodlevel projects to using it as a key strategy to guide urban transformation? If so, what needs to happen before that occurs?

9.
This working paper seeks to answer these questions by examining Nepal's policy and legal framework, planning and implementation processes, and outcomes of land pooling projects, with a geographic focus on the Kathmandu Valley (Box 1). The objective is to draw lessons from Nepal as well as experiences throughout Asia to provide policy and technical recommendations to policy makers, local leaders, donors, and practitioners who are interested in reviving and scaling up the use of land pooling in Nepal as a cost-effective, equitable, and sustainable tool for guiding urban growth. Now is an opportune time, in view of (i) the recent establishment of a federalist government structure; (ii) the recent merger of villages with municipalities across Nepal, resulting in the availability of large tracts of agricultural land within municipal boundaries outside the Kathmandu Valley; (iii) the government's proposals to develop satellite cities and build an outer ring road in the Kathmandu Valley through land pooling; and (iv) proposed changes in land and urban development policy frameworks.

10.
Key recommendations to address implementation challenges of land pooling in Nepal are summarized from page 43.

A.
What Is Land Pooling? 11. Land pooling, more commonly known as land readjustment, is an urban land development technique that involves acquiring fragmented plots belonging to many different landholders, consolidating the various plots into one area, subdividing this area in an orderly and planned manner, providing infrastructure, and redistributing serviced plots back to the original owners ( Figure 3). A portion of the land is typically set aside for hard infrastructure but can also be allocated for social infrastructure (schools, government buildings, etc.) and public amenities such as parks and other open spaces. This urban planning tool is widely seen as a win-win solution whereby government can avoid expensive or legally challenging land acquisition through eminent domain, which can displace landowners. Landowners, in turn, must contribute a portion of their land for a public purpose, but they are "compensated" through better access to infrastructure and higher land values. The process varies widely by country and by project, but the common objective is to maximize the development potential of urban land while minimizing the displacement of landowners.

Box 1: Methodology Used in Preparing This Working Paper
This working paper is focused more on the actual practices and implementation experiences of land pooling in Nepal rather than on an in-depth description of the systems as they exist on paper, as this is already well captured in the literature and other ongoing technical assistance projects of the Asian Development Bank.
To this end, input for this working paper was informed by a wide range of research and stakeholders involved in past and present land pooling projects, including the following: (i) site visits to the Dhulikhel, Gongabu, Ichangu Narayan, Naya Bazar, and Sainbu Bhainsepati land pooling projects; (ii) focus group discussions with landowners in the Sainbu Bhainsepati (completed) and Kamerotar (ongoing) land pooling projects; (iii) discussions with government officials, including the mayors of Bhaktapur and Dhulikhel municipalities, the Kathmandu Valley Development Authority, the Department of Urban Development and Building Construction, and user committees in the Gongabu, Ichangu Narayan, and Sainbu Bhainsepati land pooling areas; (iv) interviews with former project managers, land pooling practitioners, consultants, and academic experts in urban planning, urban design, and engineering; (v) a review of the literature, including research and project documents in Nepal and other Asian countries; and (vi) mapping and spatial analysis, to assess the urban development outcomes of land pooling in the Kathmandu Valley.

12.
The potential benefits cited by planners and government practitioners in Nepal and globally for using land pooling and readjustment in lieu of land acquisition or other planning tools (such as guided land development or sites and services, mentioned earlier) are extensive: 4 (i) Self-financing schemes. A major benefit is that schemes are intended to be self-financing: higher-value plots are sold by government to recover the costs of project administration and infrastructure.
(ii) Equitable projects. Projects are considered equitable since landowners and government share in the distribution costs and benefits of urban development. Government obtains land at a lower cost, landowners share in the gain through higher land values, and displacement is reduced.
(iii) Inclusive process. Integral to the land pooling process is the participation of landowners, who voluntarily contribute land and have a say in how the project develops.
(iv) Well-planned growth. Government is allowed to control growth patterns and provide more efficient service delivery in this process than in one where growth happens organically. Projects can guide future development, and provide a unified process of planning, servicing, and subdivision.
(v) Livability. 5 The city as a whole benefits from less sprawl, infrastructure can be provided more effectively and efficiently in a planned area with rights-of-way, and public amenities such as parks and other public goods can be provided.

13.
While the benefits can be far-reaching, however, land pooling can face challenges as well. Projects can take a long time to prepare and implement, given the need for the continued participation of landowners and for consensus building. The long time lags can pose problems for financial viability, since 4 For example, see Lipman and Rajak (2011);De Souza Ochi, and Hosono (2018); Mathews et al. (2018). 5 The main elements that determine the "livability" of cities, as defined in ADB's draft Urban Operational Plan, are competitiveness, greenness, inclusivity, and resilience. the government's recovery of infrastructure costs depends on a portion of land reserved for sale by the government. Additionally, projects involve considerable up-front costs that local governments may have difficulty financing with their own resources. Without government intervention, land pooling schemes may not end up as socially inclusive: for example, because land is returned to the original owners, they will either remain or sell their newly serviced plots at a higher price that low-income residents cannot afford without a subsidy.

14.
Land pooling has a long history globally, with early concepts of urban land consolidation emerging in Europe as early as the 1600s. But the first tangible example of a major land readjustment project for urban planning was from developing Washington, DC, in the United States in the 1790s. Barcelona, Spain and Kobe, Japan were also early examples of the use of land readjustment practices. However, the concept was not widely applied until Germany developed a legal framework in the late 1800s that enabled more extensive use of land readjustment. Japan later modeled its land laws on this concept. In the 1930s, it was introduced in the Republic of Korea, a Japanese colony at the time, thus cementing the use of land pooling in the Asia region. Policies were updated after World War II, not only to adjust the legislation to the reconstruction context but also in response to rapid postwar urbanization (De Souza, Ochi, and Hosono, 2018).

15.
Since the post-World War II era, several countries throughout Asia have adopted land pooling and land readjustment policies for urban development or reconstruction. Following a literature review on regional examples (Box 2), this paper presents a new typology for roughly classifying land pooling as a mechanism for planned urban development into three stages. The typology is based on the experience of countries that have long used land pooling and readjustment (approximately since the pre-1960s) such as Japan and the Republic of Korea, the state of Gujarat in India, and countries like Bhutan and Nepal where land pooling has been applied more recently. Several themes were employed in establishing general criteria for countries at the early, intermediate, and advanced stages of using land pooling as a tool for sustainable urban development.

16.
None of the countries fit neatly into a single category. These characteristics form a general framework for assessing the state of Nepal's enabling environment, implementation process, and ability to ensure the successful outcomes of land pooling projects with transformational impact on cities and towns as they grow and develop. Figure 4 presents the general characteristics of each phase.

17.
The most advanced use of land pooling attains citywide and even regional development goals that can transform entire cities. Projects are large in scale and flexible in their use-for example, for rebuilding cities after a time of war or natural disasters, as in the case of Japan, or for developing citywide transport infrastructure like highways and public railroads. Those in the "intermediate" phase also use land pooling on a strategic level as a key tool for realizing citywide or local-level plans. Land pooling projects are well integrated into land use and infrastructure plans, as in the case of Ahmedabad in the state of Gujarat in India, which was largely planned using land pooling under a town planning scheme (TPS) structure.

18.
The early phase of land pooling is characterized by a more limited use of the tool for developing residential land and ensuring vehicular access to plots to restrain sprawl on the urban fringe and ease housing pressures. Importantly, land pooling projects in this early phase are generally implemented ad hoc and are not well aligned with any larger urban or infrastructure plans. While the city government does benefit from the self-financing nature of infrastructure provision, positive impact is much more focused on higher property values for individual landowners than for the overall sustainable development of the city.

C. Existing and Potential Uses of Land Pooling in Nepal
19. The land pooling process was first used informally in Nepal in 1976, before the current legal framework was in place, for a Chipledhunga, Pokhara project covering 13.5 hectares (ha). Over a decade later, the Town Development Act (TDA) of 1988 provided supporting legislation for land pooling, spurring wider and more systematic use starting with the Gongabu project in the Kathmandu Valley. Land pooling quickly became the preferred method of planned urban development in place of the sites and services approach, which typically relied on land acquisition with cash compensation or development of public land rather than voluntary landowner contributions. In the Kathmandu Valley, 13 projects have been completed and 8 others are in various stages of implementation (Table 1, Figure 5). Ten additional projects are in the planning stage but have not yet been approved. Taken together, the completed projects have developed over 350 ha of land and nearly 16,000 serviced plots across the Valley in the past 30 years. 6 Smaller cities outside the Kathmandu Valley have also implemented land pooling projects; 3 have been completed and 15 others are at various stages of planning and implementation. 6 Calculated by the authors using data from IOE-TU (2017) and information provided by the Kathmandu Valley Development Authority.

Box 2: Land Pooling in Asia
This working paper drew from the experience of several Asian countries that have used land pooling or land readjustment to varying degrees.
Bhutan. In Thimphu, the local government has used land pooling as a key tool for implementing the Thimphu Structure Plan (2002-2027) through local area plans. While still in progress, 1,210 ha and 2,558 plots have been developed, largely under projects financed by the Asian Development Bank (ADB 2006). and the World Bank (UN-Habitat, 2018).
India. The state of Gujarat has taken a town planning scheme (TPS) approach to overall city planning since 1915. From 2000 to 2010, the Ahmedabad Urban Development Authority prepared and implemented 103 TPS schemes covering 154 square kilometers (km 2 ) of newly zoned areas for development. Taken together, the schemes provide a well-defined road network that also enables the efficient provision of basic infrastructure-water supply, sewerage, storm water drainage, and street lighting (UN-Habitat 2013).
Japan. Land readjustment has been used extensively in Japan for over a century. Projects across the country, covering over 3,000 km 2 , about one-third of its urban area (De Souza, Ochi, and Hosono 2018), have had many different purposes, including urban rebuilding after World War II, more recent transitoriented development linked with the expansion of commuter rail lines through private investment, and the development of urban greenbelts. In Nagoya City, about 70% of the city's land was developed through land readjustment (Murayama 2017).

Republic of Korea.
The Republic of Korea has used land readjustment since the 1930s. In the 50 years between 1934 and 1984, it had 397 land readjustment projects with a total land area of 436 km 2 . The projects have been used for urban expansion, arterial roads, railway stations, and housing. During the postwar rapid urbanization process, large-scale projects totaling between 300 and 400 hectares (ha) were implemented in the suburbs of urban centers, including Seoul (De Souza, Ochi, and Hosono 2018).

Source: ADB.
Land Pooling in Nepal 9 Land readjustment relies on poor-quality cadastral data and land market is largely informal; land contributions are standard regardless of plot. Grievance redress procedures are not specified in legislation. Disputes are frequently taken up in court. Land-pooling is a reaction to haphazard growth, used on a project basis to regularize residential plots and provide road access.
Land-pooling is used as a mechanism for carrying out urban plans, for residential, commercial, and public development.
Land-pooling is used to transform entire cities, as a means for reaching citywide (or regional) infrastructure, social, and planning objectives.
Legal and institutional framework is in place but regulations are vague and some aspects not legally binding; institutional mandates are conflicting and lack resources.
Urban plans are done but implementation is weak. Zoning and development controls are not in place or not enforced. Financial resources are insu cient to finance infrastructure, and projects rely on early sale of reserve plots to finance implementation.

Phase
Legal and institutional framework includes clear and specific land-pooling laws, regulations, and institutional roles backed by resources.
Urban plans are in place and framework exists for land-use zoning and development controls, which are moderately enforced. Financial resources are available for financing project implementation and serviced plots are sold to recover costs.
Legal and institutional framework is strong, with legally binding rules and procedures for all project development steps and interinstitutional coordination.
Urban planning is institutionalized, planning departments have strong authority, and land-use regulations are in place and enforced. Financial resources include significant public contribution to reduce reliance on land contribution, and sophisticated financial incentives.

20.
As in other countries, the use of land pooling in Nepal has been an ongoing iterative process. Nearly 20 years passed between its introduction in the mid-1970s and the time when land pooling gained wider use across the country (in the early 1990s). Amendments to the TDA first had to be made to give more clarity to the process on the basis of lessons learned from early projects. Most land pooling projects in the Kathmandu Valley were begun between 1991 and the early 2000s (Table 1), and projects also proliferated outside the valley during this time. However, land pooling in Nepal has since stagnated, with eight projects still ongoing more than a decade after their approval. Among these projects are the Manohara Phant project (a 13-year project) and the Chamati project (a 16-year project) (Box 3).

21.
Plots developed so far in the Kathmandu Valley number almost 16,000, but this falls far short of the total needed for planned, serviced plots in a rapidly growing capital region. Land pooling has mostly been limited to residential projects, and while there are several proposals such as the Outer Ring Road and Smart City initiatives in the Kathmandu Valley, Nepal has struggled to use land pooling successfully for large-scale projects with citywide impact (Box 4). Yet recent institutional reforms could help revive and scale up land pooling. Since 2017, Nepal's municipalities have democratically elected mayors, with the authority to carry out city planning. The Department of Urban Development and Building Construction (DUDBC) is supporting 185 municipalities in integrated urban development planning to provide municipalities with land-use and investment plans. With technical assistance from the Asian Development Bank (ADB), DUDBC is also promoting the adoption of urban economic corridor development strategies in three provinces. 7 These plans take note of the critical importance of land availability for economic development, and emphasize the use of land pooling as a tool for obtaining land for infrastructure, industry, and urban planning. 7 Final drafts of the strategies for Provinces 1 and 5 were completed in 2019. The strategy for Province 2 is expected to be completed in mid-2020.

22.
Cities where the use of land pooling is more advanced are better able to define the characteristics of a strong enabling environment, including a clear and detailed regulatory framework. Usually part of this regulatory framework is legislation specific to land pooling with legally binding rules and procedures, implementing agencies staffed with professional urban planners and having the financial and human resources needed to plan and implement projects, a transparent land market, and strong controls over land use and development that apply equally to land pooling and private developments. In cases where

Box 3: What Causes Project Delays?
In the Kathmandu Valley, few projects have been completed on time. Projects are expected to be completed 5 years from approval, but only 4 of 21 projects have reached this goal. All other completed projects took between 7 and 12 years to reach completion. Ongoing projects have been under way between 12 and 17 years, and some are still far from being completed. The Chamati project, for example, has been ongoing for 16 years and was still only at 60% completion in 2015 (Thapa 2015). Why do such delays occur?
Interviews with key stakeholders with firsthand experience in implementing land pooling projects yielded several reasons for the delays. The most commonly cited reasons were: (i) Grievance resolution and lawsuits. For example, the Kamerotar project took a full year to resolve grievances with landowners, a court case with Nepal Telecom stalled the project for 2 years, and another lawsuit with a factory owner who refuses to make the required land contribution has been ongoing for over 7 years. a The project is still ongoing, over 13 years after approval. (v) Financial shortfalls. Implementation depends on the sale of plots to finance infrastructure. At times when the real estate market declines or when external financing is not available, for example, government often has insufficient resources to finance the infrastructure and projects undergo implementation delays.
While stakeholders were candid about the challenges, many were also optimistic that the recent reestablishment of local government and elected mayors will help in resolving grievances and provide the stability, resources, and leadership needed to overcome implementation issues faced in the past.
a Meeting with the Kamerotar Users Committee, 28 June 2019.

Box 4: Land Pooling for Citywide Infrastructure-The Kathmandu Outer Ring Road Project
The Kathmandu Outer Ring Road Project was launched by the MOUD in FY2004/5 to improve the transportation network in the Kathmandu Valley. The project would involve the construction of a 72-kilometer-long, 50-meter-wide trunk road with eight lanes, and the development of a 250-meter buffer on either side of the road through land pooling (photo below, left).
The project's strategic objectives include containing population growth over 15-20 years, creating a new bypass for the Kathmandu metropolitan area, defining a rural-urban boundary, and creating an integrated infrastructure development corridor including water, electricity, telecommunications, and sewerage facilities (DUDBC 2008). The road design itself includes highway lanes, a service road, bicycle lanes, parking, a greenery belt, and a pedestrian path, and would be the first "complete streets"-style design in Nepal (photo above, right).
Using land pooling for a project of this size is without precedent in Nepal. The project would require about 42-52 km 2 of largely agricultural land, which, the MOUD recognizes, would be socially complex, given that farmers depend on the land for their livelihoods and the project has created a sense of "unease" among the affected population despite potential financial gains from higher land values (MOUD 2017). The first 6.6 km phase (Chobhar-Satungal) would cover about 420 ha belonging to over 14,000 landowners, making this first phase alone about 10 times larger than the next-largest completed project in Nepal, the 42.8 ha Naya Bazar Project.
Gaining the consensus of landowners needed for land pooling has been a challenge so far, but the alternative of acquiring land by cash compensation is not financially feasible, considering the scale and number of projects that are now required to cope with rapid urbanization in Nepal. The Outer Ring Road project has already stalled for over a decade, largely because of issues related to obtaining sufficient land, but the detailed project report was approved by MOUD in 2017. JICA reports that in view of the difficulties in obtaining the land through land pooling, the KVDA has shifted to a strategy that involves both land acquisition and land pooling, depending on the current land use in different segments (JICA 2017). land pooling is used on a large scale, schemes are tools for carrying out strategic plans for land use, infrastructure, and social purposes such as affordable housing, transit-oriented development, and open space networks. Lastly, critical to the successful and timely implementation of large-scale projects is having substantial up-front funding for infrastructure, with a high contribution from government to reduce the private land contribution. Together, these factors ensure that projects can be fully implemented at once, rather than incrementally as funds become available, and increase acceptability and confidence among landowners.

23.
Land pooling in Nepal is allowed by law as part of the authority and procedures for local government under the current legal framework, and not under specific land pooling legislation, as is true of countries like Japan, which passed its Land Readjustment Act in 1954. The legal basis for land pooling by Nepal's central or local government derives largely from the TDA and the Kathmandu Valley Development Authority (KVDA) Act, both from 1988. Amendments to the TDA in 1997 provided a general outline of the implementation process, such as procedures and conditions for securing landowners' approval and government approval, and for drawing up the block plan and having it legalized in preparation for the drafting of the certificate of title. 8 A further amendment in 2007 reduced the landowner consensus needed to formally initiate a scheme from 75% to 51%, to ease the implementation process. The consensus requirement will be discussed in more detail in later sections of this paper.

24.
The Local Government Operation Act (LGOA) of 2017, reflecting the recently adopted three-tiered federalism model of government, outlines the roles and responsibilities of federal, provincial, and local government. But while the LGOA does empower local governments to implement land pooling projects, it does not specify the steps or procedures for doing so. In view of the fact that the TDA and the KVDA Act preceded the federalism structure, the Ministry of Urban Development (MOUD) is drafting a new Urban Development Act (UDA) that would supersede both. 9

25.
Private property rights in Nepal are strong, but the government does have power to acquire land through eminent domain. The Constitution guarantees the right to property, and states that the government "shall not, except in the public interest, acquire, requisition, or create any encumbrance on the property of any person." 10 The Constitution goes on to state that "provisions of clauses (2) and (3) shall not obstruct the state in carrying out land reforms, management and regulation by law in order to increase the production and productivity of land…and managed housing and urban development." 11 This is backed by the Land Acquisition Act (1977), which allows the government to acquire land for public needs.

26.
The legal framework for urban planning and land use is less clear. The National Land Use Policy (2015) calls for the development of land-use plans for human settlements (Policy 4, Strategy 1) and for drafting of "basic guidelines" for human settlements, urbanization plans, and building construction (Policy 6, Strategy 4). However, the policy has not yet been implemented effectively, although a 2015 amendment made to the Land Act of 1964 (BS 2021) provides a legislative basis for implementing the policy. This complicates urban planning: governments are technically allowed to guide spatial development under the law but have found it difficult to do so, given the protection afforded under the Constitution for private control over land use, and the lack of precedent for zoning or other land-use restrictions in Nepal. Building bylaws are effectively the only legal tool at the disposal of municipalities 8 Town Development Act, Section 12. 9 As of August 2019, the UDA was in draft form awaiting Cabinet approval. for urban growth management (Shrestha 2013b;Karki 2004). For example, in Kathmandu, the building bylaws are based on the land-use map of 1976 and do not place any restrictions on how the land is used but do regulate building setbacks and height.

27.
The primary institutional actors involved in the planning, approval, and implementation of land pooling projects in Nepal are as follows: (i) MOUD. This apex agency approves detailed project reports, including any changes in the cadastral map. Additional changes in the project design (e.g., project boundaries) after the initial approval must be approved by the MOUD before they can be implemented.
(ii) Implementing agency. The implementing agency varies, depending on the project. Both the KVDA (for projects in the Kathmandu Valley) and the DUDBC (for projects outside the Valley) are national agencies that fall under the MOUD. At the local level, the town development committee (TDC) 12 or the municipality itself can also implement projects. The implementing agency is responsible for doing the feasibility study and financing the project, including securing seed money for preparation activities through grants or loans and financing the infrastructure.
(iii) Project management committee (PMC). The PMC is formed after the project is approved and sets up a site office to handle project operations. The PMC manages consultants, develops the block plan and determines landowner contributions, closely coordinates with the users committee, handles landowners' grievances, and is responsible for overall project administration.
(iv) Users committee. This committee is essentially the "bridge between the people and the PMC." 13 It is generally made up of about 15-20 landowners involved in the project, and acts as liaison advising the PMC on each step of the process, from the block plan to landowner contributions to plot readjustment and distribution. The committee is critical in convincing landowners to participate and negotiates directly with the PMC.
(v) Utilities. Agencies such as the water utility (Kathmandu Upatyaka Khanepani Limited, or KUKL), Nepal Telecom (NTC), and the Nepal Electricity Authority (NEA) also have a role in providing services to land pooling areas. The PMC should coordinate service provision in terms of design, financing, timing, and construction.

28.
The institutional structure for land pooling is fairly well defined under the law but has constrained project planning and implementation. The 15-year gap in local leadership was a major hindrance to project implementation and, in many cases, was one of the primary causes of project delays and stagnation. During this time, the TDCs (appointed by the central government only for projects outside the Kathmandu Valley) and the users committees were strongly involved in keeping projects on track and engaging with landowners. All users committees interviewed for this paper pointed to the lack of local leadership as a challenge that, it is hoped, will be overcome with the new municipal leadership in place. 12 The formation of TDCs was authorized under the Town Development Act at a time when there were no municipal governments. TDCs have the authority to carry out town planning and implement land pooling projects. Members are appointed for a set term by the central government. With the recent federalism reforms, TDC authority is viewed as a duplication of municipal government authority, although some TDCs are chaired by elected mayors. 13 Interview with Bhaktapur municipal official on 4 July 2019.

29.
The PMC is composed of civil servants and an appointed project manager. Project management is widely cited in the literature and highlighted in focus group discussions as a major issue in the timely completion of projects: PMC staff experience high turnover and a lack of consistency in the team over the life of the project. Karki (2004) points out that there was a 10-month gap in project management in the Sinchitar project, and that the Sainbu project had at least six project managers during implementation. Project staff tend to work only part-time and have other responsibilities, are often transferred to other jobs by the MOUD, and generally lack project management and negotiation skills. Both the Kamerotar and Sainbu focus groups saw this as a challenge in their projects. According to both groups, management was ineffective, with the project manager rarely devoting full-time attention to the scheme.

30.
The users committee holds much of the power in the design and implementation of projects. Stakeholders from all levels agreed that the strength of the committee benefits projects. Because the committees are composed of the landowners themselves, they can negotiate most effectively with other project landowners and convince them to participate, and also negotiate with government on the land contribution and plot readjustment. Because implementing agencies tend to lack financial resources, projects are financed solely through the sale of reserve plots, and there are few statutory requirements for minimum infrastructure standards or adherence to city or local area plans, the government is in a relatively weak position to negotiate with the users committee. Landowners' demands to reduce their contributions are generally met by the PMC, with unintended consequences such as narrower streets and less open space. The outcomes of this power dynamic between users committees and the government are discussed in more detail in Sections IV and V of this paper.

31.
Now that mayors have been elected, there is renewed interest among them in using land pooling for urban development and in leading project implementation, and this is a promising development favoring smoother project implementation. 14 For example, mayors, instead of government appointees, now chair TDCs, so there may be greater incentive to complete projects on time and greater likelihood of prioritizing budget allocations for project preparation and implementation. Some municipal governments in the Kathmandu Valley have already requested the transfer of implementation authority for ongoing land pooling projects from the KVDA to the municipality, as in the case of the Kamerotar project in Madhyapur Thimi municipality. 15 There is some evidence that projects implemented by municipalities take less time to implement. The Liwali Land Pooling Project, implemented by the Bhaktapur municipality, took 3 years to complete and was one of the few projects in the Kathmandu Valley to be completed on time. The Naya Bazar Land Pooling Project, implemented by the Kathmandu Metropolitan City, took 4 years and was the largest land pooling project implemented thus far (Karki 2004). Landowners themselves indicated a preference for municipal leadership, with focus groups in both Sainbu and Kamerotar stating that they would rather have their elected officials implement projects, with the KVDA providing support as needed.

32.
On the other hand, while recent changes in institutional structures are promising, the capacity to prepare and implement land pooling projects is still weak. Partly because of staffing gaps and a shortage of professional planners, designers, and engineers in the civil service, the implementing agencies and the PMCs rely heavily on private consultants for most activities including feasibility studies, site plans, engineering designs, consultations, and negotiations with landowners. The focus groups called attention to project management problems within the PMC, with project managers frequently transferred to other jobs or involved only part-time in the project when project management should be a full-time job with accountability for results. Within the municipalities, planning and engineering departments are under-resourced and formal record-keeping systems are often lacking. The Dhulikhel mayor reported that even simply obtaining a copy of the detailed project report (DPR) on the municipality's land pooling project from the municipal engineer took several days. 14 Discussions with Bhaktapur and Dhulikhel mayors, 1 July and 4 July 2019, respectively. 15 Meeting with the Kamerotar Users Committee and the PMC, 28 June 2019.

33.
As mentioned previously, there are no legal requirements for urban planning or zoning in Nepal, although local governments have the authority to adopt plans. Traditional towns in Nepal did have a clear form, with compact settlements surrounded by agricultural land, but rapid urban growth has distorted this pattern without accompanying land-use regulations. Today, where plans have been developed, implementation is weak. For example, master plans were prepared for the Kathmandu Valley in 1963, 1969, 1976, 1984, and 1991, but none have been fully implemented so far (Shrestha 2013a). The National Urban Development Strategy (NUDS) of 2017 highlights the need for more urban planning without giving prescriptive guidance.

34.
In the absence of master plans and zoning regulations, the tools most commonly used by local governments in Nepal to guide urban development are (i) periodic plans, which define budget priorities over a 3-5-year horizon; (ii) integrated development plans, which lay out a vision, multiyear investments, and special initiatives; and (iii) bylaws to regulate the development of plots and buildings. Of these, the only legally enforceable planning tool at present is building bylaws, although enforcement is inconsistent. In Nepal's new federal structure, land management is now within the purview of local governments, and not TDCs or central government. But clear and legally binding tools that will enable local government to guide land use proactively (e.g., zoning) have yet to be developed.

35.
The TDA states that land pooling schemes should be part of a "town planning area" and services should also be consistent with town planning. 16 However, given the lack of spatial planning, land pooling projects are not used as a means of implementing investment plans or spatial growth plans for municipalities. This situation is unique. In other countries, cities typically plan first and use land pooling as a means of implementing those plans ( Figure 4); in Nepal, on the other hand, land pooling is a proactive measure used to rein in haphazard urban development, but also a reactive measure in the absence of planning. Throughout the Kathmandu Valley, projects are scattered wherever topographic and sociopolitical conditions are favorable, rather than being guided or prioritized with the help of a land-use plan. Some early projects such as the Bagmati Corridor were even sited on floodplains (Karki 2004). In Figure 6, the scattered nature of projects in the Kathmandu Valley can be seen in comparison to the more contiguous nature of the town planning scheme (TPS) approach in Ahmedabad, India, where land readjustment was used in a phased manner over many years to achieve the city's 10-year development plan objectives.

36.
The state of Nepal's land market further complicates the planning and development of land using land pooling, without some of the prerequisites noted in other country examples. The characteristics of the land market include the following: (i) Informal land transactions. The land market in Nepal is predominantly informal and operates without a land information system, so that buyers and sellers usually rely on unlicensed brokers to facilitate land transactions. A lack of transparent information on prices and ownership is a common reason given for the propensity of new land buyers to leapfrog to unplanned peripheral greenfield areas-thus worsening residential and industrial sprawl (ADB 2019).
(ii) Highly fragmented land. Land fragmentation is a major factor behind haphazard urbanization. Individuals can subdivide their land as they please, with the only restriction being that plots must maintain a minimum size of 80 square meters (m 2 ) in the Kathmandu Valley and urban areas of adjacent hill districts. Fragmentation occurs as agricultural land on the periphery is sold in small lots to maximize returns, and on account of inheritance entitlement structures, which often lead to the division of property among family members. 16 Town Development Act, Section 12.1.2a.
(iii) Non-transparent values. The government sets land values every year through each district's land revenue office, but actual transaction values are usually much higher. According to the Nepal Land and Housing Developers' Association, land prices in Kathmandu have risen by 300% since 2003 (Shrestha 2015). Land acquisition by the government is a long, tedious, and contentious process, given that what the government will offer as compensation at the official values is much lower than prevailing values on the informal market, leading to legal challenges by landowners.
(iv) Rampant speculation. Land speculation results from and contributes to the three preceding points and keeps land pooling projects from maximizing their impact. By encouraging landowners to hold onto their land without investing, speculation creates an artificial shortage in the market and drives land values even higher. Individuals who own land in land pooling areas but live elsewhere will often leave their plots undeveloped and wait for the land to appreciate, leaving large areas in land pooling projects idle for years (Karki 2004).
(v) Easily serviced unplanned land. The few restrictions on utilities in terms of providing connections to basic services on private land allow the land market to flourish even in unplanned agricultural areas on the urban periphery.

37.
Landowners seem to support improvements in urban planning to mitigate the issues highlighted above. Focus group participants in both Kamerotar and Sainbu pointed out that having urban plans in place would reduce speculation and provide clarity to landowners regarding future plans for their area, thus dampening rumors and allowing the landowners to invest more wisely.

38.
The self-financing nature of land pooling is one of its most attractive characteristics, giving it a decided advantage over land acquisition, which is tedious, contentious, and expensive, and can cause displacement. However, while self-financing is in theory attractive, ensuring that projects are financially

Project Completion Year
Note: Map on the left shows land pooling areas in the Kathmandu Valley. Map on the right shows town planning schemes in Ahmedabad.
Sources: ADB (left); Shirley Ballaney (right). viable and will maximize returns is complex. Significant resources are needed up-front for administrative costs and infrastructure, which are later recovered through mechanisms to capture increased land values. Self-financing also requires expertise in real estate development and land value capture. The most basic model involves financing project administration and infrastructure up-front and recovering the costs later by selling fully serviced reserve plots that have appreciated with the provision of infrastructure. However, more advanced models with larger-scale projects in the examples from other countries show that government contributes financial resources to projects in order to minimize the amount of private land that is needed to cover project costs.

39.
Funds for administrative costs in the initial preparation and design phases of land pooling projects in Nepal have limited financing options. Projects in the Kathmandu Valley are eligible for seed funding from the KVDA, which secures grants from the Ministry of Finance and provides project loans at 7% interest. According to the KVDA, enough resources are available to fund the start-up costs for five or six projects at one time. For projects outside the Valley, the Land Development Fund (LDF) under the MOUD provides modest resources of up to NRs5 million (about $50,000) per project through a revolving fund. The loans also carry 7% interest and must be repaid within 3 years. The total LDF capitalization is NRs50 million (about $5 million), which is used to finance projects across Nepal (other than those in the Kathmandu Valley).

40.
Donor funding has also been used to cover both infrastructure and administrative costs. This funding includes soft loans through the central government, as in the case of ADB support for the Naya Bazar project implemented by Kathmandu Metropolitan City from 1995 to 2000 (Box 5). These loans can have a positive effect on project implementation where resources are scarce and local governments lack access to credit. In land pooling projects in Nepal as well as in Bhutan, ADB support for both administrative and capital costs has helped the projects mitigate delays stemming from financial shortfalls, give landowners confidence that the projects would be viable and their land would be returned to them within a reasonable amount of time, and generate positive returns. The Naya Bazar project, for example, resulted in a financial internal rate of return of 15.5% (ADB 2001) and was completed in 5 years despite being the largest land pooling project in the country. The implementation of Bhutan's local area plans (LAPs) in Thimphu had ADB and World Bank support, recognized by the government as critical to the planned expansion and financing of infrastructure in the capital city. 17 ADB financing was particularly vital in this case, since value capture mechanisms such as the sale of reserve land were not part of Thimphu's LAPs.

41.
The most common method of financing capital projects is selling some of the projects' reserve land before it is developed, since the government has only enough funds (e.g., from LDF and KVDA loans) to cover administrative costs and municipalities have limited access to credit. While necessary, given the lack of viable alternatives, this financing method was shown to have negative consequences during project implementation. Relying heavily on land sales for cash flow, infrastructure development proceeds in increments. If there is a dip in the real estate market, the project may hold on to plots and bring development activities to a halt until values rise, thus causing implementation delays. In many projects, only a basic road track and drainage can be built with the limited funds up-front before the project is handed over to the landowners. The practice of selling non-serviced plots is contrary to the DUDBC's own land pooling guidelines, which outline a project flow where reserve land is to be sold after it is developed and serviced. Moreover, though the benefits of development are meant to be shared between landowners and government, a total reliance on private land to finance the project skews the power dynamic in favor of landowners since the government does not contribute financially to the project. 17 Thromde, Thimphu. Frequently Asked Questions and Answers. https://tinyurl.com/y4srd7ly.

42.
Financing all infrastructure and utilities solely through landowners' contributions also results in a perceived inequity with unplanned areas. It is assumed that other government agencies will, in time, improve roads and provide utilities and basic services in non land pooling areas, so landowners in those areas will benefit without contributing any of their land-despite the fact that servicing irregular settlement patterns is more costly and complex (Subba, Castro-Wooldridge, and Chitrakar, 2017;Karki 2004). This sentiment was echoed in the focus groups, where landowners from Kamerotar posited that their contributions should be used only for road infrastructure and that other services and utilities should be paid for by government, with its own resources and not with project funds.

43.
The financial viability of land pooling hinges on value capture. The most basic forms include selling off a portion of the project's reserve land after it is developed and making use of tools such as betterment levies and property taxes to capture increases in land value after development. Nepal's model uses only one method of land value capture: selling reserve plots. However, land value capture is not maximized in practice. Some reserve land is often sold before development (and consequently without benefiting from the appreciation of land value post-completion). Because the loan period for seed funds is short (3 years), projects must recover costs such as consulting services and project management early, and this provides an incentive for early sale. Reserve land is also sold early to finance infrastructure rather than recover infrastructure costs, such that less ambitious and less comprehensive projects may result, given the limited up-front funding.

Box 5: ADB Financing and the Naya Bazar Project
The Naya Bazar project in Kathmandu Metropolitan City (KMC) was financed differently from most other land pooling projects in Nepal. While those other projects, as a rule, rely solely on the sale of a portion of private land contributions to finance infrastructure development and recover other costs, in this case infrastructure development was financed by an Asian Development Bank (ADB) loan under a subcomponent of the Kathmandu Urban Development Project. At 25 hectares and over 1,300 plots, Naya Bazar is Nepal's largest completed land pooling project. It was completed in 5 years despite its large size and a private land contribution ratio (56%) well above that for other projects (generally between 12.5% and 46%; see Table 1). Most projects take at least 7 years to complete after approval.
In the course of implementation, the project scope changed dramatically from what was planned at appraisal. The plan was to use a guided land development (GLD) mechanism and cover only 297 plots. However, during project implementation, the government and ADB agreed to change the scope in order to increase benefits to the area. GLD would have improved only 25% of the area through road access, and plot shapes would have remained irregular. Land pooling presented an opportunity to expand road access to 95% of the project area, give the land plots a more standard shape, and allow for open spaces. The change in scope hiked project costs from about $1.7 million at appraisal to $7.7 million, but by increasing the number of serviced plots from the original 297 to 1,316, it also made the project more beneficial and valuable.
The project was implemented in three phases, with ADB providing funding support for the first phase, including the construction of new roads to a sub-base level, the installation of drainage facilities, and river training for sediment control. From the ADB funds, the KMC made a loan to the project management committee, which was repaid with interest after the reserve plots were sold. Sales of surplus land were also used to finance infrastructure development in phases 2 and 3.
Importantly, the reserve plots were sold after they were serviced, contrary to common practice in land pooling projects in Nepal, which, for lack of budget resources or sources of credit, often resort to selling reserve plots before servicing, to finance infrastructure. Because of the ADB loan, the KMC had enough resources to complete the first phase of infrastructure development.
continued on next page 44.
In many cases, the early sale of undeveloped plots involves transactions with other government agencies at below-market rates. In Kamerotar, for example, the project sold about 2.5 hectares (ha) of undeveloped land to Nepal Telecom (NTC) in 2005, at about $300,000/ha to cover initial project costs. Five years later, land prices of new plots with road access had a value of over $4 million/ha at auction. Had the project taken out a loan instead of selling undeveloped land, landowners could have contributed less land or surplus resources could have been used for higher infrastructure standards (Shrestha 2015). The Sainbu project also sold undeveloped land to the NTC, the Nepal Electricity Authority (NEA), and other agencies (Shrestha 2015). The Ichangu project was an exception: undeveloped land was sold to the DUDBC at above the market rate, which DUDBC accepted because of an urgent mandate to construct resettlement housing for evicted squatters (K. C. 2015-see photo in Section V.D.).
The Naya Bazar project demonstrated that a land pooling investment can have a high rate of return. The financial internal rate of return noted in the project completion report (15.5%) was the highest achieved by any of the three components of the overall project. The two other components had returns of 2.9% (for the Kathmandu Core Area Upgrading component) and 0.5% (for the storm-water drainage component) (ADB 2001).
On the other hand, while Naya Bazar exemplified the effective use of donor finance to facilitate land pooling, the project outcomes suggest that, with more ADB support and technical assistance-for designing infrastructure, enforcing building regulations, and ensuring that operation and maintenance systems are in place-the needs and preferences of the intended beneficiaries would have been better served. The photo below (left) shows the poor condition of roads and lack of building setbacks. A municipal engineer interviewed for this study described planning in Naya Bazar as sub-standard: the street layout is confusing, streets are narrow, and there is no proper road hierarchy (see site layout plan below right). Violations of building bylaws in the project area have reportedly gone unchecked, with buildings exceeding plot coverage and height limits. According to a users committee member, such violations exist in 80%-90% of buildings within the project area (Pradhan 2008

45.
Using funds from the sale of plots in one land pooling area to finance urban development outside the project boundary is considered socially unacceptable in Nepal, as pointed out by the Sainbu focus group mentioned earlier. In this case, the attempt by the KVDA to use the funds for that purpose aroused strong opposition from landowners. The increase in land values is viewed as the rightful benefit to the land pooling area's landowners, because they alone have contributed to the project. This is contrary to practices elsewhere, such as in Ahmedabad, where revenues from selling reserve land are used to create a land bank (Mahadevia, Pai, and Mahendra 2018) (Box 6).

46.
Transparency and the people's trust in the way public funds are managed and accountability is exercised is critical. The Kamerotar and Sainbu focus groups expressed their distrust of the ability of the PMC to manage revenues. Sainbu landowners complained that the PMC had used NRs20 million from the sale of plots to provide an interest-free loan for an indefinite term to another KVDA land pooling project (Sainbu Nakhudole) as seed money, on the understanding that the funds would be reimbursed. Kamerotar participants pointed out that both the users committee and the PMC, called on to generate funds from the sale of plots, had little experience in doing so, and landowners had doubts about the financial management capabilities of the implementers. In Sainbu, where an unanticipated surplus of land after the plot survey had produced a financial windfall, the focus group recommended stronger financial monitoring to avoid any potential misuse of funds.

47.
International cases suggest that more advanced land pooling systems have a regulatory framework that clearly defines procedural steps from feasibility assessment to construction, including detailed requirements for public consultation and the resolution of disputes with landowners, as well

Box 6: Ahmedabad's Land Bank
The Ahmedabad Town Planning Scheme (TPS) model shows how reserve land within one scheme can be used to finance citywide infrastructure development through the creation of a land bank. This is in stark contrast to the system in Nepal, where reserve land and use of funds from land sales are strictly limited to the project area and its landowners.
Gujarat state law allows for the creation of a land bank to commercially sell land to raise funds for infrastructure and services. Ahmedabad has raised significant revenue from the sale of reserve land from TPSs. Sixty-five percent of the revenue of the Ahmedabad Urban Development Authority (AUDA) in 2006 came from land sales. In that same year, AUDA auctioned off 20 plots of land to large real estate firms for $38.0 million. AUDA has used this money to finance large infrastructure projects, mostly roads and water, sanitation, and drainage facilities.
Sources: Mahadevia, Pai, and Mahendra (2018); De Souza, Ochi, and Hosono (2018). as legal obligations to complete projects within a specified time frame (Box 7). Having the necessary financial resources gives government greater leverage to implement projects, since eminent domain can more readily be used to resolve bottlenecks when grievances cannot be settled, rather than allow projects to stall indefinitely. Project feasibility is assessed in detail, and while projects often encounter grievances, systems are in place both for consultation and for grievance redress at the project level.

48.
In Nepal, the process of preparing and implementing land pooling projects is not fully defined in detail in regulations. Section 12 of the TDA does outline some steps, but mostly grants authority to carry out projects to local bodies and defines the roles of the user committee, the local body, and the TDCs. The DUDBC's land pooling guidelines provide more details to guide the process, but are not legally binding, and the process differs between projects that are implemented in the Kathmandu Valley by the KVDA (which has its own Act), those implemented by municipalities, and those implemented by TDCs or the DUDBC. Figure 7 provides a general outline of the land pooling process, which is based on the DUDBC guidelines, legislation (e.g., TDA), and input from stakeholder interviews on how projects are carried out in practice.

Box 7: Legally Defining the Land Pooling Process
While Nepal's land pooling process has legal backing in the Town Development Act and the Department of Urban Development and Building Construction (DUDBC) has a set of detailed guidelines, the country's legally binding rules and regulations provide little detail on how projects are carried out. In other countries, detailed procedures for land pooling are spelled out in the law.
In India, the Gujarat Town Planning and Urban Development Act (GTPUDA) of 1976 and the corresponding Rules (1979) outline over 50 implementation steps and include standard forms for processes such as calculating land values and finances. a In Japan, the Land Readjustment Law describes concepts related to land readjustment, such as replotting, reserve land, development restrictions, and the fair sharing of costs and benefits. It also establishes the rights and duties of landowners, leaseholders, implementation agencies, and other third parties involved from initiation until project completion.
Germany's land readjustment system (Umlegung) in the country's Federal Building Code contains details on the following:

49.
Potential project sites are generally proposed by landowners themselves or selected through site observation by local or national authorities (such as KVDA in the case of the Kathmandu Valley). Information about land ownership and cadastral maps are gathered during this early phase and a feasibility study is undertaken, usually contracted out by government to a private consulting firm, which, at a minimum, assesses the economic, technical, and social viability of the project (Box 8). A preliminary block plan is developed with land requirements for roads, drainage, and open space. Project management and the infrastructure costs per developed plot, as well as land values after the project, are estimated. From this information, the land "contribution ratio" (the ratio of land taken for public use versus land returned to the owner as a developed plot) is determined, and the area and cost of developed plots to be returned to landowners are estimated.

50.
The feasibility study including the preliminary block plan, which usually presents three alternatives, and the contribution ratio are discussed with landowners. At this stage, the proposed contribution ratio is an average percentage across all landowners, which may later go up or down for individual plots when the land redistribution and plot readjustment are done. At this point, according to the TDA, the project must obtain the consent of 51% of landowners for the project to be approved by the MOUD. Landowners, then, have leverage to negotiate for a lesser land contribution, which often results in projects reducing infrastructure and open space allotments to minimize the overall public land take, even if future growth projections indicate a need for higher standards (Shrestha 2015).

51.
Landowners can raise objections at this early stage, and consultations with stakeholders indicate that the most common objections are put forward by (i) landowners who already have road access and refuse to contribute land when they expect to receive no benefit in return; (ii) farmers and others who are more interested in the use or livelihood value of their land than in its development potential; (iii) landowners with small plots, especially where the plot size after contributions would not meet the minimum plot size of 80 m 2 ; and (iv) those demanding more and better services. However, as long as 51% of landowners agree, then it can be assumed that all landowners will proceed with the project.

Box 8: Is Project Feasibility Adequately Assessed?
A case study of the Misato Chuo project in Japan demonstrates how an in-depth analysis of a project's financial viability must be done before a project is even approved. A detailed budget and financial proposal, including all expenditures and revenue sources for the project, were prepared during the feasibility stage and were then submitted to the Minister of Land, Infrastructure, Transport and Tourism for approval. An exercise similar to Nepal's "long sheet," which estimates individual landowner contributions, was also carried out before landowners were even engaged in the consensus-building process and was undertaken by government planners rather than consultants. Then a project plan was made, an implementation ordinance was passed by the municipality, and a land readjustment council was established (De Souza, Ochi, and Hosono 2018).
Similarly in Gujarat, a draft "long sheet" is prepared early as part of the preliminary town planning scheme, and drafted before landowners are even consulted or the project is approved. After individual contributions are estimated, hearings are held during which each landowner is consulted on the plan and allowed to raise objections before the plan is revised. There is also scrutiny of the detailed financial proposal in the feasibility study, which requires approval by the government revenue office.
In contrast, the feasibility stage in Nepal's land pooling process involves less detailed analysis and planning than feasibility assessments in other country examples. In Japan and Gujarat, India, detailed land surveys and land ownership information are completed at the feasibility stage and factored in the financial viability assessment of projects before landowners are engaged. In Nepal, data collection and analysis at this level is done much later, during the detailed design phase and even during construction. Only basic financial viability is assessed during feasibility, on the assumption that reserve land will be sold after servicing rather than before development, as is done in practice. Detailed land ownership information is also collected during detailed design rather than during the feasibility study.

52.
The TDA does include a clause that government can acquire land with compensation for sites and services and guided land development schemes, but importantly this clause does not apply to land pooling schemes. 18 According to an interview with a DUDBC official, "compensation is never paid in cash." 19 One key exception for projects in the Kathmandu Valley is in cases where a single plot would be less than 80 m 2 after the landowner's contribution is made: if the landowner does not have financial capacity or is unwilling to buy enough land to meet this minimum plot size, monetary compensation is paid in lieu of serviced land. Yet the use of cash compensation is a key option for dispute resolution in the global literature on land pooling. According to Hong and Needham (2007): State power to take private property for a public use with just compensation will continue to be important when negotiations between private individuals and the government break down or when land resources are not fully utilized. In these situations, the taking of land is unquestionably for the well-being of the public at large, and eminent domain may be best suited to handle disputes arising from land assembly.

53.
Given that project approval requires a consensus of 51% of landowners, this leaves a large proportion that could potentially object to the project. Yet given that the use of eminent domain is contrary to Nepal's land pooling process, project planners are essentially left with two options for dealing with opposing landowners during implementation: (i) redraw project boundaries to exclude them, as has been done in projects such as Gongabu; or (ii) proceed anyway and continue to negotiate. Both of these options pose problems for implementation. While redrawing boundaries is the easiest way to resolve disputes in the short term, in the longer term, areas that were excluded from the project end up developing anyway, but in a haphazard way, whereas they could have been properly serviced if included in a land pooling area. In the second option, project design might carry on, but disputes with objecting landowners can languish in court for years. In the Kamerotar project, for example, one unwilling factory owner has had a case pending against the project for over 7 years, and it is still unresolved. 20 This project has also had several other disputes, but the case illustrates that a single landowner is able to halt a project if there is no legal mechanism to opt out. Box 9 summarizes lessons on how grievances are handled in Japan and the state of Gujarat, India, both of which have detailed legislative processes that can help to resolve issues before they escalate to the courts.

54.
A robust consultation process is a prerequisite for the successful implementation of land pooling, but perceptions of the quality of the consultation process in Nepal are uneven, and significant time and skill are needed to explain to landowners what land pooling actually is. In the case of the Naya Bazar project, just this initial step took about 18 months since landowners were unfamiliar with land pooling and many were uneducated. The Kamerotar focus group reported that consultations with communities were weak and landowners did not understand the process or the potential benefits, and that a lack of meaningful consultation caused implementation issues that halted the project for a year. However, the same focus group reported that consultation meetings were useful in resolving grievances and the members regarded the consultation process as effective. Similarly, in Sainbu the focus group was split between those who found consultations effective and those who did not, though the group did agree that landowner participation can be efficient if carried out effectively. Importantly, both groups noted that more extensive consultation is needed before the project is designed so that their input can be reflected in the design, rather than concentrating consultation during the implementation phase. The DUDBC also pointed out that spending more time on consensus building early will save time later. 18 Section 16, Acquisition of Land: "The Government of Nepal may acquire the required land as per the prevailing law for the committee in order to implement town planning."

55.
After the 51% consensus is obtained, the executive agency (KVDA, TDC, etc.) makes a formal decision whether or not to proceed with the project, and if a favorable decision is arrived at, an official notice is published and a users committee, usually consisting of landowners and tenants in the project area, is legally formed. 21 In the guidelines, a 7-to 15-member users committee is suggested, 21 While the guidelines suggest that the users committee is formed after the project is approved, it can be formed at a different time, as project experience has shown. In some cases, users committees have been formed before or during the feasibility stage.

Box 9: Grievance Redress: Lessons from Japan and Gujarat
Regardless of the country, disputes between landowners and government frequently arise in land pooling projects. The ability of landowners to paralyze project implementation, however, owes much to how well the legislation in a country defines grievance redress procedures. In most systems, court cases arise as an appeal against a decision made by a lower body such as a committee or a quasi-judicial entity, but in the case of Nepal, disputes can go directly to court. In the projects examined in the Kathmandu Valley, all had gone to court at some point, aside from the Liwali project in Bhaktapur and the Naya Bazar project in Kathmandu Metropolitan City. Private property rights are enshrined in the Constitution, so land cases are often taken up by the Supreme Court.
Who handles grievance redress and how grievances are handled is not defined in Nepal's land pooling process, but informal channels exist at the project level. Stakeholders reported that complaints should first be raised to the users committee, escalate to the PMC, and, if unresolved, then go to court. The user committee plays a vital role in resolving disputes and convincing landowners, but the TDA does not define how land-related grievances are handled, and neither the users committee nor the PMC/TDC includes legal experts, skilled arbitrators, mediators, or community mobilizers. Yet there is no statutory requirement that this process should be followed for a case to have standing in court.
Japan provides an example of how the legal framework helps in avoiding this setback. In the Japanese process, a landowner can disagree entirely or partially with the project-e.g., if he or she is not in favor of the prescribed contribution ratio or the location of the plot after the redistribution-but cannot stop the project or take it immediately to court. The Land Readjustment Law requires the submission of written complaints to the central ministry during a 2-week public inspection period. The ministry examines complaints and orders the implementing agency to make modifications if the complaints are found valid, or notifies the objecting landowners if the claims are found invalid. If legal actions are later taken, the project is still valid and its implementation can continue until the legal action is judged under strict rules and procedures. These strict rules and procedures are considered lengthy processes that weaken the landowner's potential opposition and strengthen his or her disposition to obtain consensus and the best results possible.
In the case of Gujarat, the legal framework also has built-in mechanisms for dispute resolution. Early in the TPS implementation process, a town planning officer, a quasi-judicial officer, is appointed and gives a legal hearing to each landowner during the process of plot redistribution. These types of individual hearings are mandatory at several stages of project implementation. The town planning officer has a clear mandate and accountability for resolving disputes related to physical planning and financial issues. PMC = project management committee, TDA = Town Development Act, TDC = town development committee, TPS = town planning scheme.
Sources: Mahadevia, Pai, and Mahendra (2018); De Souza, Ochi, and Hosono (2018); Karki (2004). but stakeholders reported that the committees for most projects have closer to 20 members. Projects, regardless of their size, have one users committee. The PMC is also formed at this time, to handle dayto-day planning and administration.

56.
At this point, the public is likewise notified of a moratorium on any development or subdivision of land in the project area for 2 years (or 3 years, as in the case of the Kathmandu Valley) to enable the government to carry out a "town planning exercise." 22 Technically, this should mean that the government is required to complete the project up to the stage of an approved block plan and detailed implementation plan including the redistribution of individual plots within that 2-or 3-year period. However, in practice, if the planning is not completed within the specified time frame, the government can extend the "town planning exercise" by publishing a new gazette notice. There is no penalty for government if it does so, nor a limit on the number of extensions, which can go on indefinitely. But the ability to freeze land rights for an uncertain amount of time causes an issue for landowners who rely on land as collateral for bank loans, for example, or who want to invest in developing their land.

57.
A DPR, involving a preliminary design level with the block layout, infrastructure rights-of-way, open space, and reserve plots, is prepared by consultants, and importantly a "land pooling policy" that outlines how much land would be contributed is proposed. The contribution is based on the location and attributes of existing plots; while some past projects have used standard land contributions, projects in Nepal generally follow an area-based model. In the area-based model, contributions are determined on a plot-by-plot basis-for example, a plot that already has road access might contribute less land than an interior plot with no road access. The DPR is presented to the users committee and stakeholders for comments and revised as needed, and the final report is submitted for ministerial-level approval by the MOUD. Since projects are multi-sectoral, projects are also approved by the Council of Ministers so that all sectors are informed. Once the project is approved, a public gazette notice is published for 35 days.

58.
Engineering surveys and the gathering of plot ownership details are carried out after the DPR is approved by the central government and the 35-day public notice period is completed. Joshi (2014) points out that this phase may mark the first time that landowners who live abroad or outside the project area become aware of the project. In Ichangu, for example, the users committee described a scenario where community opinion about the project fractured after survey equipment was deployed to start field measurements. Though the project design, including the land contribution ratio, had already been approved earlier with the required 51% landowner consensus, 40% of landowners later opposed the project because of objections over the land contribution (originally 40%-50%), and raised a court case that took 2 years to resolve. As a result, the KVDA had to reduce the contribution ratio to 20% for those landowners having existing road access and 38% for those without. To accommodate this change, street widths in the project area were reduced from the planned 12-meter roads with sidewalks to 8-meter roads without sidewalks, and the total open space allocation was reduced from 5% to 3% of the project area. Despite consensus and consultation requirements, there was still a lack of public awareness of the project, little understanding of benefits to landowners in terms of infrastructure improvements, and strong objections to the project design. 23 In time, the grievances were resolved through consultations and negotiation, and the project was completed after 8 years.

59.
Because the feasibility study and the DPR are based on older paper cadastral maps rather than field surveys and geographic information system (GIS) maps, implementation issues also arise as a consequence of discrepancies between surveys and field verification, as well as delays due to the length 22 Town Development Act (Sections 8 and 12). 23 Meeting with the vice president of the Ichangu Narayan Users Committee, 30 June 2019 . of time it takes to acquire land records from district land offices. Focus groups in Kamerotar and Sainbu both described discrepancies between the DPR and survey data. In Kamerotar, a major discrepancy between the DPR and the survey findings nearly caused the project to fail. The project agreed to sell 5.3 ha of the reserve land to the NTC, but this deal was made before the detailed survey was carried out. The DPR was later found to be inaccurate-only 2.6 ha was available-and this resulted in a financial shortfall that was resolved only when agreement was reached with the project to reimburse the NTC later. In comparison, the Sainbu Bhainsepati project had a financial windfall when the survey revealed that landowners had larger parcels than what was recorded in the cadastral maps from the 1960s, because the landowners had reported smaller plots to minimize their property taxes. The project ended up with about 2.7 ha of surplus land, which provided it with financial security, but the focus groups raised questions about how the implementing agency planned to use the surplus land.

60.
As discussed earlier, in international examples, reserve land is generally sold after the land is fully serviced, to maximize land value capture. Nepal's land pooling guidelines are consistent with this practice, indicating that the plots are sold for cost recovery after the project is handed over. In practice, projects do not typically follow the guidelines and at least a portion of the reserve land is sold during the design phase to mobilize funds for construction. 24 Once funds are secured, the contractor is engaged.

61.
During procurement and after contractor mobilization, the most complex step in the land pooling process-land redistribution and replotting-happens in parallel. This is also a trigger point for landowner grievances, and requires intensive negotiation between the users committee, the PMC, and landowners, who are interact constantly regarding where their plot will be and how much land they will contribute. The project-specific redistribution policy agreed in the DPR is applied here through the preparation of a "long sheet," which is a detailed and complex exercise that tabulates the percentage of land deducted from each plot as well as the location of each returned plot. Because there is room for negotiation, this is mostly done on a trial-and-error basis until all landowners and the PMC reach an agreement they view as fair. 25 This is a third point in the process where landowners negotiate to reduce their land contribution yet at the same time might advocate higher infrastructure standards. 26 Once the plot layout is agreed on, a land readjustment map is prepared and submitted for approval by local authorities.

62.
The transfer of developed plots back to the original landowners can occur at various levels of completion of infrastructure works but is generally done once there is at least a gravel road track and after the local body approves the land readjustment map. Landowners are then provided with a temporary land certificate. A major benefit to land pooling projects is that the old cadastral maps are updated as part of the project, and include new cadastral plot numbers and land records. Once all records in the project area are updated with the Land Revenue Office, landowners receive a permanent ownership certificate.

63.
When the project is deemed completed, the management of assets (such as open spaces) is handed over either to the users committee or to the municipality, and any remaining reserve plots are sold at auction. Depending on the project, additional infrastructure may also be constructed later or infrastructure might be improved once funds are available, as many projects are deemed "complete" even with only a gravel road track.

64.
Land pooling projects are intended to improve the overall quality of the urban environment, and international examples show great promise in terms of connecting newly developed areas to existing neighborhoods, providing comprehensive infrastructure and basic services, generating new open spaces and parks, and using reserve land for social benefits such as affordable housing. This section summarizes findings on how land pooling has affected urban development through observations from site visits, interviews, and a geographic information system (GIS) analysis for projects in the Kathmandu Valley.

65.
Land pooling areas tend to develop later than surrounding areas. In land pooling areas in the Kathmandu Valley, 45% of buildings were constructed after 2013, compared with 28% within a 500-meter buffer zone (Figure 8). This shows that the unplanned surrounding buffer areas develop during the time that land pooling areas are being prepared and implemented. Figure 9 shows the Bagmati Nagar land pooling area, a 63 ha project in Kathmandu Metropolitan City municipality, which took 11 years to implement (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014). The map shows less dense development within the land pooling area, with most buildings constructed since 2013 (71%). In contrast, 77% of buildings in the surrounding area were constructed before 2013. Given the haphazard development of surrounding areas, providing access, utilities, and basic services to the land pooling area can be complicated. Source: ADB analysis of historical building footprints.

66.
In terms of population density, Shrestha (2013a) calculated that in land pooling areas, the figure is lower than the recommended gross density of 300 persons per hectare in the KVDA's Development Plan 2020 of the Kathmandu Valley. Density in Kuleswore, for example, is 159 persons per hectare; in Gongabu, it is 143. However, a spatial analysis for this study showed that for completed land pooling projects, average building density is slightly higher (by about 12%) within land pooling areas than in surrounding buffer areas.

67.
Regularized plots also have higher land values in completed land pooling areas than in adjoining areas. Focus group participants in Kamerotar and Sainbu broadly agreed that they have benefited from having regular plots, and that land values rose significantly after the projects. Discussions with officials in Ichangu reflected this finding as well: from about 1 lakh/anna (about $900/32 m 2 ) when the project started 14 years ago, the land price rose to 30 lakh/anna ($27,000/32 m 2 ) in 2019. Land prices have skyrocketed across the Kathmandu Valley, but data show faster appreciation in land pooling areas than in adjoining areas. In Kamerotar, the market price of land outside the planning area is 45% to 50% less; in Sainbu, it is 50% to 60% less (Shrestha 2015).

68.
However, the benefits of planning are confined within project boundaries. Because land pooling projects in Nepal are planned on an ad hoc basis, they tend to have few direct linkages with surrounding areas in terms of street grid or block patterns. The experience of the Ichangu project demonstrates the importance of embedding land pooling in a wider city plan for transportation and land use (Figure 10). The Ichangu project lies on the periphery of the city, near the Nagarjun forest. A 1.5 km swath of dense unplanned area with narrow and irregular streets separates the planned area from the ring-road highway. How the project area would connect with the ring road was not considered in the design phase, though landowners had said this was a necessary consideration in any decision to proceed with the project. When the landowners' demands were not met, the issue escalated to a lawsuit that paralyzed the project for years. Eventually, the KVDA and the MOUD committed funds to the widening of an access road to the land pooling area. This subproject is mostly complete, aside from a 300-meter stretch where the high compensation costs for five households in the right-of-way have stalled progress. If the access road had been included in the project from the start, 14 years ago, the court case would have been avoided and the compensation costs would have been significantly lower.

69.
Land pooling areas tend to have better infrastructure than surrounding areas. Projects are generally designed to include roads, water supply, sewerage, drainage, electricity, and in some cases streetlights. An analysis showed that road density was higher within land pooling areas in the Kathmandu Valley than within a 500-meter radius of the project area-on average, about double the figure for areas outside the project (32% vs. 18%), but with wide variance between projects (Figure 11). Deadend roads are also about three times more prevalent in the surrounding area, as seen in the examples of Ichangu and Sainbu projects in Figure 12, indicating that planned areas have better traffic circulation than surrounding areas but benefits are limited to the project area as connectivity and a contiguous street network are lacking.

70.
While the design phase may anticipate a comprehensive infrastructure package all at one time, implementation tends to be incremental over a number of years. There are two main reasons for this. First, construction can start only once the project raises funds from selling reserve plots. Even then, the amount recovered may not be enough for all infrastructure included in the designs. The review of completed projects showed cases where the project is handed over after having completed a gravel road track, and the users committee or municipality finances infrastructure such as sewers and streetlights with their own resources years later. For example, in Gongabu, the project financed gravel roads with drains and water pipes. Later, the users committee raised funds from community contributions for road blacktopping and solar streetlight installation. The users committee is now installing sewer lines; this involves digging up the tarmac road and re-blacktopping.

71.
The other main reason for the incremental implementation is that projects rely on other line agencies for utility service provision. Yet, those agencies are still struggling to provide services in the city core and have not yet turned their attention to the urban fringe, where land pooling projects are usually located. Because the projects do not have up-front financing for infrastructure, there is less financial incentive for utilities to service those areas that lack paying customers in the immediate term. The Kamerotar focus group reported that "line agencies (e.g., those in water and electricity) were not supportive of expanding the urban services in the project area during the implementation process," and the respondents' view was that integrating service provision into the original project execution would have been more cost-effective and viable.

72.
Reducing street widths is a common way for projects to negotiate with landowners who object to the proposed contribution ratio. This tactic, which may facilitate project implementation but neglects to consider the city's future growth and growing car ownership, was described earlier in the case of the Ichangu Narayan project. In Kamerotar, a survey of landowners indicated that 60% of respondents would contribute land for a 6-meter road, 30% for an 8-meter road, and only 10% for a 12-meter road (Shrestha 2015). In the 1980s, Gongabu landowners successfully negotiated this contribution down to 4-meter roads, but current residents now complain that the roads are too small and the government should have made a stronger case at the time. 27 As can be seen in the photo above (left), there is no demarcated space for pedestrians and streetlight poles had to be constructed in the road right-ofway. At the time many land pooling projects were planned, there were significantly fewer vehicles and landowners (many of whom were farmers) could not conceive of automobile congestion or the need for public transport becoming an issue.

73.
Stakeholders interviewed in completed land pooling projects reported infrastructure design issues that stemmed from poor planning and engineering studies (Box 10). Both the Sainbu and Ichangu projects included the construction of piped water systems, but without an assessment of the reliability of water sources. The pipes constructed in the Sainbu area deteriorated from lack of use, and focus group participants pointed out that the area currently faces a water shortage and most residents rely on water delivered from tanker trucks. In Gongabu, the assessment made for the DPR did not cover the potential power load or sewerage needs. Services were designed for 300 households when now there are 700, partly because of further plot fragmentation after project completion. Because a sewerage system was not constructed during project implementation, individual landowners built their houses with septic tanks, which they then connected to the project area's roadside stormwater drains. The drains were not designed with adequate capacity or gradient for sewerage, resulting in overflow with contaminated effluent during the rains. The Ichangu Users Committee reported similar issues from under-designed sewer pipes. Gongabu's streets (pictured in the right photo above) were designed without consideration of a safe turning radius. The resulting blind corners caused accidents and bottlenecks for emergency vehicles. However, street design in land pooling projects has improved over time, as can be seen in the above photos, which compares typical streets in Gongabu (approved in 1988) andIchangu (approved in 2006). 27 Interview with former Director General of the DUDBC.
Comparing street widths in land pooling project areas. Photo on the left shows a 4-meter and 6-meter street in the Gongabu land pooling area, designed with narrow widths and no consideration of a safe turning radius for vehicles. Poles for solar streetlights had to be constructed on the street since residents did not provide enough land to allow for footpaths. Photo on the right shows a street junction in the Ichangu Narayan land pooling area, where eight-meter-wide streets have a safe turning radius and treatments such as zebra crossings (photos by Amy Faust, ADB site visits July 2019).

C.
Open Space

74.
Cities in Nepal are starved for open space. The National Urban Development Strategy points out that open spaces are critical for environmental quality, social interaction, beauty, livability, and, critically, safe evacuation during natural disasters such as earthquakes. Table 2 shows open space coverage in several municipalities across Nepal. In all but one, open space accounts for less than 0.5% of the total land area.

Box 10: Consequences of Feasibility Study and Design Shortcomings
The case studies examined for this paper demonstrate how inadequate feasibility studies and engineering designs can lead to financial, social, and technical issues during implementation.
During interviews, Kamerotar Project Management Committee members reported that the engineering consultant made a critical error in neglecting to factor in a 20-meter right-of-way required by the Department of Roads for the construction of a new highway bordering the project area. Because no compensation is provided for land taken for highways, the project faced a major financial shortfall in losing public land that was intended to be sold later to recover the costs of the land pooling project. The feasibility study also left out details regarding the high number of plots that were smaller than the minimum plot size of 80 m 2 , which the project would have to "compensate" with additional land from the reserve plots to ensure that the minimum standard is met. This represented an unexpected loss of 1.7 hectares and another reason for the project's financial shortfall.

Kamerotar Land Pooling Project
Hig hw ay

75.
While there is no standard rule of thumb for the recommended percentage of an urban area that should be publicly accessible open green space, the World Health Organization (WHO) recommends 9 m 2 per person, a metric that takes into account population density while ensuring that accessibility to open space keeps up with population growth. Current bylaws in the Kathmandu Valley require planned residential zones to have 2.5%-5% of total land area as open space depending on the size of the development. Similarly, the National Urban Development Strategy proposes that 2.5% of existing urban areas and 5% of new urban areas is the "desirable condition" for open space at the ward level (MOUD, 2017).

76.
For more recent land pooling areas, a 5% contribution for open space is now the norm. Table 3

77.
Land pooling projects tend to have more open space than the city as a whole, but open space provision varies widely. In the Kathmandu Valley, open space in land pooling projects ranges from 0% (Bagmati Corridor) to 12.9% (Sainbu). In earlier land pooling projects, reducing open space was noted as a common bargaining tool for minimizing land contributions, much like road widths. For example, the focus group in Kamerotar reported that the originally planned amount of open space was perceived to be too much, and was thus successfully reduced to 4% of the project area. However, this practice is less prevalent in more recent projects since the establishment of the bylaw requiring open spaces for new developments.

78.
The actual use and quality of open spaces varies among the land pooling areas visited. The design and construction cost of open spaces is not generally included in the land pooling project. Once a project is complete and handed over to the municipality or users committee, the latter can use funds from the sale of reserve plots or direct contributions from the community to design and construct parks, playgrounds, and other facilities. The users committee is also generally tasked with the maintenance of such public areas using their own funds. The users committee for the Ichangu Narayan project is in the process of constructing a park in two phases. The first phase is nearly complete (see photo below, left). In Sainbu, designated open spaces are used for badminton and other sports. Other users committees use the demarcated open spaces for community amenities that also generate revenues, such as a public pool in the Gongabu project area that charges an entry fee (see photo below, right). Open spaces in land pooling areas. Photo on the left shows a park under construction in Ichangu Narayan using users committee funding. Photo on the right shows a public pool in the Gongabu land pooling area used to generate revenues for the users committee (photos by Amy Faust, ADB site visits July 2019).

79.
While land pooling areas contribute to open space, more could be done to ensure that parks, recreation areas, and green spaces are used more effectively. In block planning, open spaces are sometimes relegated to marginal plots, such as in Gongabu, where open spaces were placed beneath existing high-voltage power lines. Instead, spaces strategically located within walking distance from most plots serve an equity function as well as contribute to public safety in case of evacuation in an emergency. On account of the need for the users committee or municipality to raise additional resources for constructing parks in the open space plots, some lie vacant for years, reportedly attracting crime and drug users. Others have been encroached on by buildings, as in the case of the Naya Bazar and Gongabu projects (Figure 14), though this is a common issue across the Kathmandu Valley and not unique to land pooling projects. Source: ADB spatial analysis.

80.
Undersupply of housing is a major issue faced by Nepal's cities. The National Shelter Plan (2013-2014) estimates the number of housing units required in urban areas by 2023 at about 1.4 million (MOUD 2016). The 2015 Constitution guarantees housing for all, and key policies and guidelines include stipulations for affordable housing. For example, the Town Development Guidelines require the allocation of 10% of developed land parcels to economically weaker sections (EWSs) (Subba, Castro-Wooldridge, and Chitrakar, 2017). Both land pooling and private sector development are contributing to meeting the demand for new residential housing stock. In the Kathmandu Valley, the private sector has constructed over 6,000 plots since 2012 (or about 850 plots per year), while completed and ongoing land pooling projects together will have contributed just over 22,000 plots in the Valley over about 30 years (or about 733 plots per year), less than the private sector's contribution.

81.
While playing a role in meeting housing demand, land pooling has had little use as a means of housing the urban poor while there is precedent for doing so in other countries (Box 11). Developed plots in land pooling areas are affordable only to wealthier individuals, as prices are far beyond what even middle-class salary earners could afford (Karki 2004). Projects have attempted to include housing for the urban poor, but without success so far.

Box 11: Land Pooling and Affordable Housing
Land pooling has been used as an effective means of providing public housing for the poor, the landless, and evicted squatters. In the 1960s, Seoul in the Republic of Korea started to face a major housing shortage and surging demand for land in a small area that caused a sharp spike in land prices. In 1980, the country's government mandated that 30% of funding for low-income housing was to come from land readjustment.
In Ahmedabad, India, lands made available through the town planning scheme (TSP) mechanism allowed for the construction of about 80,000 dwelling units under various social housing schemes, including grants from the national government, for which Ahmedabad was eligible because it already had available land. Of these units, 33,000 were constructed under the Basic Services to the Urban Poor (BSUP) program between 2007 and 2012, and social housing was used to resettle people who were affected by city-level infrastructure and beautification projects. The BSUP housing scheme ensured that resettled households were not relocated to the city's periphery (Mathews et al. 2018). The distribution of housing projects across the city can be seen in the figure below, which maps Ahmedabad's social housing schemes (Mahadevia, Pai, and Mahendra, 2018).

82.
In Kamerotar, the idea was considered and rejected after landowners objected to the proposed land contribution even without public housing included ). In the Ichangu Narayan scheme, the DUDBC purchased reserve land from the users committee at 20% above the market rate (NRs110 million) for the construction of housing for the nearly 1,000 slum dwellers who had been evicted in 2012 from a settlement along the Bagmati River. Multilevel apartment buildings were constructed in the land pooling area by the DUDBC (see photo), but in view of objections from both Ichangu residents and the evicted slum dwellers, the resettlement process has not occurred, and the households have returned to the site where their structures were demolished. Evicted squatters claim they were consulted only after the buildings were constructed, and raised several reasons for rejecting the resettlement site: the long distance from their work in the city center and the lack of public transportation, the lack of schools, the additional expenses of apartment living such as utilities, and the inadequate size of the housing units, which were about half the size of their demolished dwellings. Ichangu landowners also objected to the resettlement scheme, with one resident stating to the press: "This is a VIP area … we can't have low-level people coming into the neighborhood and disrupting the environment" (Gurung and Ojha 2019). The buildings, completed in 2014, are still largely vacant and deteriorating, apart from some units used for municipal and other public offices.
Ichangu Narayan low-income housing development. Low-income housing development in the Ichangu Narayan land pooling project, intended to provide apartment housing for evicted slum dwellers but still largely vacant to this day, aside from a few public offices (photo by Amy Faust, ADB site visit 30 June 2019).

A. Has 30 Years of Land Pooling Delivered Results for Nepal's Cities?
83. This study has found that since the formal adoption of land pooling in the legislation in 1988, cities and towns across Nepal have completed land pooling projects, successfully developing some planned areas that otherwise would have developed haphazardly, like surrounding non land pooling areas. The observed land pooling practice in Nepal exhibits a high emphasis on equity and on the need to ensure that no one is displaced, and landowners have benefited from higher land values and better infrastructure services. Government, in turn, has benefited from regular plots that are easier to service and has not had to pay compensation to acquire land for new roads, open space, and utilities. The legal and regulatory framework has been an iterative process but has adapted to implementation lessons over the years. In the Kathmandu Valley, housing development through land pooling has not quite kept up with the pace of the private sector, though the average annual contribution to the housing stock has been comparable. Land values in land pooling areas are higher than those in nearby unplanned areas, infrastructure standards are better, and land pooling areas have considerably more open space than the overall city. These successes demonstrate that land pooling is possible and can deliver positive results compared with the status quo of unplanned and individually driven urban development.

84.
While there have been some positive results, the impact thus far on guiding rapid population growth at a citywide scale has been limited. Land pooling tends to be done on a project-by-project basis, where land pooling areas are essentially islands in a sea of haphazard development. No cities or towns in Nepal have managed to use land pooling for transformative results, unlike cities such as Ahmedabad, Seoul, or Tokyo. Why is this so?

85.
Many factors that hinder the use of land pooling on a larger, more ambitious scale are external to the land pooling process (Figure 15). Among the external factors that have been found to complicate land pooling are the long gap in local leadership in times of political instability, institutional fragmentation, urban land-use plans that are outdated or do not exist at all, and a lack of zoning regulations combined with strong private property rights and an informal, non-transparent land market-all of which create an environment that constrains large-scale development efforts, whether government-led land pooling or private development.

86.
Other challenges are specific to Nepal's land pooling system itself. While there is a legal and institutional framework in place, both the system as written and the system in practice have gaps as compared with systems in countries that use land pooling for larger-scale initiatives. For example, the legal framework lacks specifications for grievance redress and project timelines, and overemphasizes the need for projects to be self-financing and not displace any landowners even when they wish to opt out of projects. When the external factors combine with weaknesses in the land pooling system, they lead to consequences such as delayed project implementation and land-related legal battles, and more broadly keep projects from reaching their potential to maximize land value capture and provide a high-quality urban environment that benefits all residents.

87.
Without the basic prerequisites found in external factors, such as urban land-use plans, land market regulations, and the unlocking of sufficient municipal finance for infrastructure, land pooling projects in Nepal are likely to continue to suffer implementation challenges. Systemic change and strong political will are needed to draft and implement land regulations at a national scale. Unless substantial public resources are made available for sharing the costs of infrastructure and amenities, as well as compensating non-consenting landowners, more ambitious large-scale projects are likely to be difficult. That said, these challenges apply equally to land pooling, private sector land development, and land acquisition. In view of the high and rising land values in Nepal, land acquisition is and will continue to be prohibitively expensive and legally complex, so land pooling is still an attractive option for planned urban development now and will continue to be so in years to come, especially in provincial capitals and secondary cities outside the Kathmandu Valley that still have large tracts of undeveloped peri-urban areas. While the external factors will be slow to change, now is an opportune time to make changes in the land pooling policy framework, build capacity, and improve processes, given recent institutional reforms and ongoing policy development.

88.
The following seven recommendations focus on short-to medium-term measures for addressing land pooling-specific factors posing more immediate implementation challenges. Addressing the external factors is beyond the scope of this paper. Each recommendation includes a set of specific actions to address key issues that were found through this review. Recommendation 1: Update and strengthen the legal and regulatory framework for land pooling

89.
The current framework as defined in the Town Development Act (TDA) must be revisited, either through a stand-alone Land Pooling Act and Regulations or through the proposed Urban Development Act (UDA) which is currently in draft form and which will replace the TDA and the KVDA Act. The MOUD should carefully weigh these two options, noting that countries with robust frameworks almost always have stand-alone legal and regulatory frameworks for land pooling. The following actions apply regardless of the legal and regulatory modality chosen.

1.1
Same procedures for all projects, regardless of size or objective Differentiate between procedures for localized, residential land pooling projects, and those for larger projects of strategic importance (including clear classification criteria).

Indefinite freezing of land rights
For a specified time, allow for the freezing of the subdivision of land in land pooling areas and the implementation of physical changes in such land, but include a provision that prohibits indefinite extensions without land reevaluation or some form of compensation to landowners, in order to hold government accountable and incentivize project implementers to complete projects on time.

Institutional overlaps and conflicting mandates
Be consistent with the federalism structure, and clarify the institutional roles of the Kathmandu Valley Development Authority, municipalities, provinces and federal government for projects sited within one jurisdiction, as well as larger projects that cross municipal boundaries.

Cumbersome approval process without binding timelines
Streamline approvals, and for each step of the process specify mandatory review and approval timelines.

Private sector involvement not specified in law
Analyze suitable options for private sector participation in land pooling projects, and include provisions for private sector participation in the regulatory framework.

1.7
Informal and project-based grievance redress procedures Define in detail grievance redress procedures (see Recommendation 6), including a legally binding quasi-judicial mechanism to avoid immediate court cases.
1.8 Weak land value capture Allow municipalities to use land pooling projects as a land bank, and capture revenues from projects as a means of financing land acquisition elsewhere and future infrastructure development.

90.
There has been some discussion within government and in the literature of further reducing the current 51% landowner consensus required to initiate a land pooling project. The findings reported in this paper suggest that while lower consensus might ease project approval at the start, grievances may be more likely later. In fact, there is no evidence that the first reduction in the consensus requirement from 75% to 51% resulted in more effective project implementation. For projects geared toward larger-scale trunk infrastructure or public amenities with a wider public benefit, a lower consensus requirement (or none at all) may be appropriate. However, for residential projects, lowering the consensus requirement could be detrimental since community buy-in is more crucial to implementation. If government decides to lower or eliminate the consensus requirement, projects should be prepared to commit to cost sharing with funds that are roughly proportionate to the overall public benefit of the project, and also budget for compensation to landowners who do not wish to participate.

2.1
Potential effects of reducing landowner consensus requirement not analyzed Carry out a policy analysis of the effects of reducing consensus requirement. This could include a legal review of other country frameworks, a cost-benefit analysis of such provision, and the development and recommendation of cost-sharing strategies. The analysis should include evidence-based recommendations for different types of projects (e.g., residential, strategic, economic).

Costs for non-consenting landowners not included in project budgets
Factor an estimated compensation budget in the financial feasibility and project financial proposal as a contingency measure for non-consenting landowners.
Recommendation 3: Integrate land pooling in urban and regional planning

91.
While strengthening urban planning processes is an issue that extends beyond land pooling in Nepal, it is a critical underpinning for the successful use of land pooling to guide growth and efficiently provide basic services. The current patchwork of projects does reduce haphazard growth in select areas, but local and national governments have not yet effectively managed to link land pooling with planning for a municipality's future growth. The following actions would contribute to better integration of land pooling and urban planning.

Issue Actions
3.1 Ad hoc land pooling projects Make central government approval of land pooling projects contingent on a municipality having an approved master plan or integrated urban development plan indicating the proposed land pooling areas. This type of provision could be included in land pooling regulations as well, thus helping to reduce speculation and rumors among landowners about which areas might be slated for land pooling.

3.2
Infrastructure and open space standards not based on urban growth projections Determine minimum infrastructure and open space standards on the basis of growth projections for the larger area, include these standards in municipal bylaws, and factor them in the project's contribution ratio and financial feasibility as a nonnegotiable minimum from the start. These minimum standards should apply equally to land pooling and non-land pooling areas to ensure the consistency and equity of all development.

Minimal connection with overall urban form
From the start, design land pooling projects such that they go beyond the immediate project area and consider how the area will connect with surrounding areas, trunk infrastructure (e.g., water supply and sewerage), and transport networks. Additional resources needed to make these connections should be included in the financial plan.

92.
Under the current modality, the economic feasibility of projects is determined on the basis of insufficient information (e.g., without a detailed analysis of landowners and land values), and cost recovery estimates are based on the value of fully serviced sales plots. In reality, plots are often sold before development to finance road construction and other infrastructure, thus minimizing the potential for land value capture. Implementing agencies must be prepared to share the cost of strategic projects as well. The larger the land pooling project, and the more it contributes to city or regional development objectives, the less likely it is to be fully self-financing. Contributing sufficient public resources to finance infrastructure and amenities will help give implementing agencies leverage over project design elements and assure landowners that projects will be implemented to a high standard in a timely way. Donor finance can also play a role, as land pooling projects in countries where local governments face capacity and resource shortfalls, including ADB projects in Nepal, showed that those projects were implemented faster and with fewer delays caused by landowner grievances. The following actions are intended both to ensure that the project finance is more robustly assessed and to maximize land value capture in the longer term.

Inadequate project budgeting process
Ensure that the financial feasibility assessment, financial plan, and budget can cover infrastructure construction, open space, and administrative costs up-front, with minimal reliance on selling undeveloped reserve plots to finance the project.

Insufficient scrutiny of project budget
Subject the financial feasibility of projects to closer government scrutiny, including a review and approval process before the projects advance to the detailed design stage.

4.3
Few project financing options, especially for implementation Carry out a study to determine a suite of financing options for land pooling projects. These options could include reviving and modernizing the Land Development Fund in the Department of Urban Development and Building Construction and making use of opportunities to tap the Town Development Fund, donor finance, subsidies, grants, public-private partnerships, and the economic potential of land pooling to generate revenues for a land bank that can be used to finance infrastructure in future projects and can be continuously replenished.

4.4
No mechanism for private sector involvement Establish a legal framework and regulations for involving private sector developers in land pooling projects.

Financial benefits accruing only at the project level
Provide more flexibility and ensure financial transparency to allow the sale of reserve plots in one scheme to finance the costs of other schemes.

4.6
Limited use of donor finance Consider donor finance to help fill the resource gap and provide technical assistance. There is potential for development partners to support projects and set a good precedent that can increase public confidence in land pooling for urban development.

93.
Scaling up land pooling projects, especially with new local and provincial governments that lack human and financial resources, will require capacity building at the executive and technical levels, as well as structural means of avoiding project delays due to gaps in project management. The following actions are aimed at improving capacity at all levels and at all stages of projects.

Limited institutional and technical capacity
Develop a national capacity-building program to train municipal government, project management teams, users committees, consultants, and other stakeholders in regulations, finance, planning, design, negotiation, consensus building, and other skills needed for effective land pooling projects. A training center and curriculum could be established at the central level in the DUDBC, the Urban Planning and Development Center (UPDC), or a university.

5.2
Weak project management Ensure that PMCs have a dedicated organizational structure; enough resources based on project size, type, and complexity; and a dedicated project manager with a support team for the duration of the project and a mechanism for ensuring that project managers are not often transferred. Introducing mechanisms for involving the private sector in implementation could also help bridge capacity gaps.

Unavailability of project information and data
Build a library database of past projects, including a repository for project documents, spatial and financial data, plans, and lessons learned, as well as a resource management system for ongoing and future projects.
Recommendation 6: Build consensus, resolve disputes early, and develop mechanisms for non-consenting landowners

94.
Implementing agencies can better capitalize on the strengths of communities to resolve projectrelated grievances early in the process before they end up in legal battles that can tie up projects for years. At the same time, a major weakness of the land pooling framework is the lack of a mechanism that would allow non-consenting landowners to opt out of projects. Moreover, resolution is needed to end the paralysis of projects that have been ongoing for a decade or more: indefinite timelines reduce landowner confidence in participating in land pooling projects, and place an undue burden on landowners and project implementers alike. The following actions are intended to improve stakeholder engagement and reduce the likelihood that projects will face delays due to landowner grievances.

Issue Actions
6.1 Informal and project-based grievance redress procedures Define a systematic approach to grievance redress in the regulatory framework (or alternatively at least in the project-specific land pooling policy), which lays out step-by-step how complaints are handled and must be followed by the implementing agency. This quasi-judicial system should be binding, such that landowners cannot raise a case in court unless these procedures fail to resolve the grievance. See also Recommendation 1.7.
6.2 More strategic consultation process needed Improve consultation processes and techniques so that landowners better understand what a project entails in order to avoid conflicts based on misunderstanding. The Project Management Committee and its consultants should be capable of using maps, renderings, and other technical tools to meaningfully engage with landowners and generate strong buy-in for projects.

6.3
No mechanism for non-consenting landowners Include mechanisms in land pooling reforms so landowners are allowed to opt out of projects with fair compensation, and ensure that resources are available for compensation. The current paradigm where all landowners must participate is not realistic for large-scale projects, and will contribute to delays.
6.4 Paralysis of ongoing projects Develop a time-bound strategy of either resolving the impasse in ongoing projects or canceling projects where resolution is not possible.
6.5 Lack of resources for stakeholder engagement Early in the project, dedicate sufficient budget resources for consultation and dispute resolution.
Recommendation 7: Use land pooling as a means of achieving high-quality infrastructure and contributing to livability

95.
Thus far, land pooling projects have been used largely to provide residential housing plots with basic road access and services, but such projects hold promise to do much more. The inability of government to restrict land use complicates the realization of larger-scale land pooling. Minimizing landowner contributions often comes at the expense of higher-standard infrastructure and open space. Open space exists, but often is not used to its greatest potential for public recreation and safety in urban environments starved of greenery. In cities struggling to provide affordable housing, land pooling could be used to revive stalled plans and create new strategies both for affordable housing programs and for the relocation of informal settlements. With more thoughtful design and with additional financial resources, land pooling, especially larger projects, can set an example for sustainable urbanization through the following actions.

Lack of government contribution for infrastructure and public amenities
Ensure that governments contribute financially for infrastructure, services, and amenities to help mitigate the tendency of landowners to negotiate for minimized standards so that their land contribution is reduced, and to give government leverage in providing higher-standard services and amenities that contribute to the greater public good.
7.2 Full package of infrastructure and amenities often not financed by project Offering a full package of infrastructure and basic services up-front can help to avoid the tendency for users committees to finance infrastructure with their own resources as they become available-which can result in piecemeal and inefficient implementation. A full package should include at least tarmac roads, water supply, sanitation, drainage, streetlighting, and utility service ducts. Having a detailed financial proposal and resources secured up-front will incentivize line agencies and utilities to provide services in greenfield areas during project implementation rather than according to their own time frames.

7.3
Open space requirement not scientifically determined Consider setting the minimum standard for open spaces on a per-capita basis rather than as a blanket percentage, at a minimum meeting the World Health Organization guideline of 9 m 2 per person. Costs for the design and construction of parks, playgrounds, and other recreation facilities should be built into the overall project cost. Communities/users committees should be involved in the vision and design of these spaces to generate buy-in.

Structures built in designated open spaces
Prohibit building in open space areas, aside from structures needed for the functioning of the space, and ensure that open spaces are designed to contribute to public service provision such as storm-water management or offer safe spaces for emergencies.

7.5
Land pooling areas not successfully used for affordable housing Consider how land pooling can be used for affordable housing, including resettlement from informal areas. Work together with housing nongovernment organizations and other experts to develop consultation and negotiation methods for engaging communities in both informal settlements and land pooling areas in reaching mutually agreeable strategies for resettlement plans, locations, and housing design and finance mechanisms.

Minimal use of urban design methods
Use an urban design approach, which should be built into the scope of work for design consultants, and include urban design expertise in the design consultant team. This will help to ensure that social and environmental aspects are captured in designs, as well as improve linkages with the overall urban fabric and arrangement of plots.