A Model for Political Restructuring and a new Electoral System for a Federal Nepal

Ram Acharya, CFFN, Ottawa, Canada

Proceedings of Unfolding Futures: Nepalese Economy, Society, and Politics
Friday-Sunday, Octobet 5-7, 2007, Ottawa, Canada


In this paper, I develop a framework for political restructuring of a Federal Nepal, and provide a model for a new electoral system. I identify the natural homelands of 11 ethnic, linguistic, and caste (ELC) groups, called ELC focus regions. I argue that these regions must be an essential component of federation, but that it is not economically desirable to base federation upon them. The desired objective of making Nepal a federation of such ELC groups (i.e. inclusiveness in political power sharing) could be achieved by making these regions electoral constituencies. I further argue that the political constituencies of a federal Nepal should be provinces that extend from north to south as a result of combining ELC focus regions. This north-south corridor would generate immense benefits from the complementarities in natural endowment and in the comparative advantage between northern and southern regions which is not possible if federation is made up of all ELC regions. Moreover, this arrangement will allow all provinces to be bordered with the rapidly growing economies of China and India. Hence, I propose that ELC regions be electoral constituencies and Nepal be a federation of four provinces (each province with three ELC regions), and one territory in the most north-western part. I propose bicameral parliaments at the national and the provincial levels whereby all citizens are equal in the lower house of both levels of parliament, all provinces are equal in the upper house of the national parliament, and all ELC regions are equal in the upper house of the provincial parliament. Finally, I devise a proportional representation system and electoral formula where ethnic, linguistic, caste, gender, and regional issues are addressed to foster an inclusive democracy.
There has been recent and wide coverage in the Nepali media addressing how Nepal’s sub-national entities should form the Federation of Nepal in a way that would support and sustain a democracy that is inclusive of ethnicity, language, gender, region, culture, etc. However, there are only a few studies that tackle the issue of how the various regions of federalism should be formed. Among them, Gurung (2000) has proposed 25 districts (instead of the 75 districts that Nepal currently has), Sharma (2006) has proposed six provinces based on 19 districts, Neupane (2005) has proposed a federal state of eight provinces, and Sharma (2007) has proposed 15 provincial constituencies. Maoists, who have advocated against the exclusion of marginalized groups from political, economic, and decision-making processes, have proposed that Nepal be a federal state of nine autonomous regions; seven based on ethnicity and two based on region. Other political parties have no specific model except for a proposed province to integrate the northern and southern parts of Nepal.In spite of the research that has been conducted in this area, there are several oft-neglected facts that can help us to understand the population distribution of Nepal, which will in turn help us develop more realistic sub-national jurisdictions. As well, most of the previous studies are based on limited data and anecdotal evidence, and are quick to recommend regional formations that do not contain any logical underpinnings. In Nepal’s context, the immediate questions that must be answered are as follows: How should the nation be restructured, how should different groups (especially marginalized groups) be represented in sub-national regions, and what electoral process will achieve inclusiveness. Currently, there are no studies that have dealt with these issues in a comprehensive manner that I am aware of. Thus, the purpose of this paper is to fill in the gaps by providing a model for Nepal’s political restructuring and a new electoral process. The paper provides a complete framework for dealing with Nepal’s needs. The paper’s specific objectives are as follows:

  1. To identify natural homelands for the main ethnic, language, and caste (ELC) groups in Nepal;
  2. To identify the electoral constituencies best suited for inclusive democracy;
  3. To identify the political sub-national entities for the federation of Nepal that can best sustain inclusiveness and promote economic development;
  4. To provide a model of a proportional electoral system for both the national and sub-national parliaments;
  5. To provide a model of representation for women and all ELC groups in the political decision making process.

This paper is based on the premise that the main objective of Nepal’s restructuring is to form a just, prosperous, democratic, and dynamic society. If this is true then Nepal’s challenge is two-fold. First, we must make a consolidated effort to create a political and economic environment where women and ELC groups (both of which have been so far excluded) are empowered as symmetric partners with the affluent. Second, the political structure should be compatible with economic development. The latter will bring broader participation of all groups, including the majority, and will promote mutual trust and harmony. The good news is that these two challenges are not mutually exclusive: If political restructuring is properly balanced, both can be carried out as complements to the other.

In general, the population of Nepal can be divided into Indo-Aryan caste groups and Tibeto-Burmese ethnic groups. “Caste” is defined as social group within the Hindu caste system. “Ethnic” or nationality (Janajati) is defined as a social group within its mother tongue, native area, and religious tradition. In Nepal, the mainstream groups are the so-called high and medium Brahamin, Chhetri, Thakuri, and Sanyasi castes, hereafter known as BC, from the hill area. Although they may live in Terai (i.e. (the plain southern part of Nepal) they are called hill-BC in the population census. Nepali, the only official language in Nepal, is their mother tongue. Despite enormous intra-group vertical disparity, these groups are dominant in Nepal. And account for approximately one-third of the population.

Aside from the BC group, there are eleven groups that each comprise more than 1% of the population and are located in geographic proximity to what could be defined as their natural homelands. Six (i.e. Limbu, Rai, Tamang, Gurung, Newar, and Magar) are mountain and hill ethnic groups with their own mother tongues. They are considered hill janajati, hill ethnic groups or hill nationalities. Three other groups are defined by language and live in Terai (i.e. Maithali-speaking, Bhojpuri-speaking and Awadhi-speaking). A common name for the aforementioned Terai-living people, as well as for other smaller groups whose mother tongues are different, is Madhesi. These people can be separated in three different groups along the line of caste and ethnicity; the so-called upper and medium caste, so-called lower caste, and janajati. There exists another Terai group knwon as Tharu, whose culture and language are not Nepali, and who are different from the Madhesi people. This last hill group is the so-called lower caste who are also known as Dalit. For the sake of brevity, all 11 groups mentioned above will be known as ELC (i.e. ethnicity, language, and caste) groups in this paper. They account for approximately 61% of Nepal’s population, or double the size of BC.

The paper will identify the main settlement of the BC group along with homelands of all ELC groups that contribute to more than 1% of Nepal’s population. Once this is done, I propose that these homelands be focus regions for their respective groups by making them electoral constituencies for both national and provincial parliaments. An existing district is considered a natural homeland of one of these ELC groups if the ELC group is at least the second largest fraction after BC in the district. Hence, if a group is the majority (i.e. more than 50% of the population) among all groups, or in plurality (i.e. having less than 50% of the population but having the highest fraction) in the district, or the largest after BC population then the district will be the natural homeland of that group. In other words, a district is a natural homeland of a group if that group is the largest among other ELC groups whether more or less than BC. Indeed that is the case for 46 out of 75 districts. All districts that qualify as a natural homeland for a given group are combined to make a focus region for that group.

Based on the aforementioned criterion, I was able to distribute 68 districts among eleven regions as the natural homelands for these 11 groups. These groups and their respective homeland regions are: Limbu (Kanchenjunga), Rai (Sagarmatha), Tamang (Gaurishanker), Newar (Kathamandu), Maithali speaking (Mithila), Gurung (Annapurna), Magar (Ridi), Bhojpuri speaking (central Terai), Dalit (Khaptad), Awadhi speaking (Lumbini), and Tharu (western Terai). Among those 68, three districts are placed using geographic proximity rather than natural homeland criterion because there are no adjoining districts that have the same natural homeland characteristics. Of the seven districts that could not be identified as natural regions for an ELC group, three districts (Jhapa, Morang, and Sunsari) contain no group that could be defined as natural dwellers, and the majority/plurality of people have Nepali as their mother tongue, mostly BC (note that Nepali is the mother tongue of Dalit as well). These three districts compose a region without any focus group, thereby giving 12 regions. The remaining four districts (Jumla, Humla, Dolpa, and Mugu) are located in the most remote area of Nepal. Most of the people in these districts are BC, and there is no other distinct group that could be considered a focus for that region. I propose that it be a territory.

All districts that compose an ELC region are geographically adjoined except the Awadhi speaking region, where the Banke district is separated from its two partner districts, Rupandehi and Kapilbastu, by the district of Dang. The same is true for Dang, since it has been separated by Banke from its partner district of Tharu. The regional size differs; the population is as low as 2% in the Kanchenjunga region and as high as 14% in the Ridi region. In terms of area, the Kathmandu region is the smallest with 0.6%, and the Khaptad region is the largest with 15.4% of Nepal’s area.

Dividing the country based on natural homelands shows that the current ideas for a federal Nepal (i.e. that Nepal should be a federal state of autonomous regions based on ethnicity, or that it should be a federal state based on region only) are insufficient. The ELC focussed regions must be taken as building blocks in order for federation to be effective. However, a federation of ethnic regions is not desirable if one looks at population distribution. Population distribution is so mixed that despite my efforts to maximise the share of a group in its focus region, only two groups are in the majority in their regions; Maithali speaking and Bhojpuri speaking. Another three groups represent the plurality of population in their respective regions (Limbu, Awadhi speaking, and Tharu). The shares of the remaining six groups (Rai, Tamang, Newar, Gurung, Magar, and Dalit) are less than those of the BC group in their own focus regions.

The view, which has been expressed in the Nepali press, that an ECL group in its historical region should be guaranteed majority political seats as a pre-right to govern in the region, even if the group is not the majority, goes against democratic principles. This line of reasoning is a result of the historical exclusion of ELC groups from politics. However, one mistake cannot be corrected by another. A more prudent and farsighted approach should be developed for allocation of political representation among the BC and ELC groups. It is not even desirable to have a federation of ELC focus regions if one takes into account the economic challenges, opportunities, and resource complementarities of these regions. An alternative way of federation would be more conducive to economic development. Since the resources and economic activities of the northern and southern regions are quite different, the more practical way of making an economic unit would be to combine the ELC focus regions (i.e. the electoral constituencies) from north to south.

Hence, I propose that Nepal be a federal state of four provinces (Mechi, Koshi, Gandak, and Karnali) and one territory as sub-national political units. Each province will have three ELC regions, one each from the mountain, hill, and Terai. The most eastern province, Mechi, will combine with the Kanchenjunga region, Sagarmatha region (focus of Limbu, and Rai groups respectively), and the eastern Terai region (a region without any focus group). Next, Koshi province combines Gaurishanker, Kathmandu, and Mithila regions (focus of Tamang, Newar, and Maithali speaking groups respectively). The regions of Annapurna, Ridi, and central Terai (focus regions for Gurung, Magar, and Bhojpuri speaking respectively) are combined to form the Gandak province. The far west province, Karnali, includes the regions of Khaptad, Limbini, and western Terai regions. These are focus regions for Dalit, Awadhi speaking, and Tharu groups respectively. Finally, Rara territory is formed of Dolpa, Jumla, Humla, and Mugu districts (see the regional and provincial maps in Appendix B).

The prosperity of Nepal and inclusiveness of all groups requires that we allow equal opportunity so that one group cannot dominate another. Prosperity depends on crafting an inclusive society for both the majority and the minority, for dwellers of all regions (mountain, hill, and Terai), and for both female and male. To this end, I have proposed electoral constituencies based on ethnic, language, and caste groups, and political constituencies based on resource endowment and economic development potential. In doing so, we will be providing an optimal solution for both inclusiveness in the political process and advancement in economic activities.

With this structure in mind, I propose a three-tier government; local, provincial/territorial, and national. At the provincial and the national level I propose bicameral parliaments. The bicameral system at the national level will combine the principle of democratic equality with the principle of federalism – all citizens are equal in the lower house, while all provinces are equal in the upper house. Similarly, at the provincial level, the lower house will maintain the equlity of individuals at the lower house and equality of regions at the upper house. Each of three tiers will have excutive, legislative, and judicial arms. This paper will only discuss the legislative part of the national and provincial levels. Legislative at the local level and issues concerning the executive and judicial in all three levels are not dealt with here.

Ipropose that the lower house of the national parliament (i.e. the National Assembly) have 305 members; 300 from 12 regions and five from the territory. The constituencies for elections are ELC focus regions. Half of the legislaturers are proposed to be elected based on a regional list of candidates and the other half based on national list of candidates that parties will submit during election time. The regional candidates should come from the region, whereas the national candidates can be drawn nation wide. Voters cast votes for parties not for candidates. Voters in each region cast two votes – on the left side of the ballot box for regional representation and on the right-side for national representation. Votes can be split between two parties. The upper house of the national parliament (which I call the National Council of Provinces) will have a total of 70 members, 15 from each province, five from the territory, and five nominated from the government from each of four provinces and territory and from groups other than BC and the eleven ELC groups. One of them must be a Muslim, the group that accounts for slightly more than 4% of Nepal’ population but has no focus region. There will be no direct election for the upper house. The representation of 15 from each province and five from the territory will be decided in proportion to the parties’ involvement at the provincial parliament, for which I have provided a formula. Hence with 305 members in the lower house and 70 members in the upper house, there will be a total of 375 national level law makers.

At the lower level of the provincal parliament (i.e. Provincial Assembly), the number of representatives for each region will be the same as the number of representatives for the National Assembly. However, for the territory I propose 10 members instead of the five I proposed for National Assembly. The election constituency and process for the Provincial Assembly will be the same as that for National Assembly. I propose 15 members for each province (five from each region) at the upper house of provincial parliament (Provincical Council of Regions). Voters in each region will elect five members for the upper house. Each provincial government will nominate one additional member, and if there is no representation from Muslims then that nomination should come from them. With a total of 310 members in the lower house and 64 in the upper house there will be total of 374 members engaged in provincial and territorial parliaments. That will make the total number of legislatures at the national and the provincial level equal 749.

In both levels of govenrment, I propose a proportional representation system to guarantee that every individual and ELC focus region in Nepal has equal say in political matters. To achieve this goal, I propose a list system of proportional representation with Imperiali quota for the regional list and Droop quota for the national list (see Appendix C for definitions) at both national and provincial lower houses. I propose proportional representation with Droop quota for both levels of upper house.

Electoral constituencies are the ELC focus region so this is a built-in mechanism for inclusiveness among different groups. I have also ensured inclusiveness of all groups by developing three criteria for all the political parties to fulfill when choosing candidates at the lower houses of both parliaments (there are no criterion for the upper house and the territory). The criteria at the lower house of national parliament are that (1) the share of candidates from the regional focus group in total regional list of candidates put forward by a party in that region should not be less 80% of this group’s share in regional population, (2) the share of national candidates from a group in total national candidates put forward by a party should not be less than 70% of the share of that group in Nepal’s population, and (3) half of the national (total) candidates should be women. Criterion (1) means that if a group has 10% of its focus region’s population then each party will provide at least 8% of the regional candidates in this region from this group, and if a group has 10% of national population each party will provide at least 7% of its total candidates from that group.

Before proceeding further, I should make a few points regarding the data. First, the data used in this paper are Nepal’s 2001 population Census. This data may be far from perfect, but lacking alternative sources, I have used them extensively. Second, since the data are only at the district level, I have counted each district as a member of a single region when identifying regions. But the population distribution of a district could be such that it would be more appropriate to have some parts of that district in one region and other parts in another depending on how each ELC group is settled in the district. Once one looks at the more detail local (village) level data, it is quite possible that some parts of a district may fall in region other than where it is assigned in the study. However, I do not think that practice will bring a major change in population distribution by regions and provinces. In any case, the basic structure and the layout of the paper will still be valid even if some statistics change.
The rest of the paper is organized as follows: In the next section, we provide a synopsis of population distribution for Nepal based on the 2001 Census, with a focus on the main ethnic, linguistic, and caste groups. The following sections identify regions and then provinces (for the latter, by combining regions that are natural partners). Next, I devise a mechanism for electing representatives for the legislative bodies of a federal Nepal at the national and provincial levels. I then develop a model of group participation in the political process before concluding the paper.

Population Distribution: Ethnicities and Languages
In terms of ethnicity- and linguistic-composition of the total population, Nepal is like a garden with varieties of flowers. According to population Census 2001, there are 23 groups, which constitute at least one percent of Nepal’s total population (see Appendix A, Table A1). Among them, the group of Chhettri Thakuri and Sanyasi constitutes the largest fraction (18.1%) of the population followed by hill Brahamin (12.7%). With so many groups in the list, it is hard to carry a meaningful analysis. Besides there are some common threads (language, ethnicity, socio-economic conditions etc.) that some of these groups could be combined without compromising the effectiveness of the study. The results are presented in Table 1. The largest two groups (Brahman and Chhetri, Thakuri and Sanyasi) are combined into one and called Brahman and Chhetri (BC). This group makes 30.9% (column 2) of Nepal’s population. The other groups in descending order of their shares are Maithali speaking, Bhojpuri speaking, Hill Dalit, Magar, Tharu, Tamang, Newar, Muslim, Rai, Awadhi speaking, Gurung and Limbu. Other (mountain and hill)—referred as OMHJ henceforth—are 2.3% of the total population. The remaining 1% constitute “Others”.

BC constitutes the majority (more than 50%) in 19 districts (first entry in column 3) and the plurality (largest fraction but less than 50%) in another 27 districts with a combined districts of 46 where they are either in majority or in plurality (entry inside the parentheses in column 3). The distribution of ELC groups is rather spread thin across nation. Out of 75 districts in Nepal, the eleven ELC groups are in majority only in 14 districts, and they are in plurality in additional 15 districts, with a total of 29 districts where they are either in majority or in plurality (instead of 46 districts for BC). Among them, Maithali speaking people are in majority in five districts; Bhojpuri speaking people are in majority in three districts, whereas five ethnic groups Magar, Tharu, Tamang, Newar and Gurung and Awadhi speaking people are in majority in only one district each. Therefore only in 33 districts there is either BC or ELC group in majority; in the remaining 44 districts there is no single group which is in majority.

Table 1. Nepal’s population composition by broader caste/ethnicity in 2001

Ethnic, language and caste groups Share in total population
Number of districts with Share in its own population where the group is
(& plurality)
2nd place at least in plurality in 2nd place Total
Brahman, Chhetri, Thakuri, Sanyasi 30.9 19 (46) 19 75.2 18.0 93.2
Total Madhesi (excluding Tharu) 22.5 9 (11) na 65.6 na na
Maithali speaking 12.3 5 (5) 1 82.1 na na
Bhojpuri speaking 7.5 3 (4) 83.6 na na
Awadhi speaking 2.5 1 (2) 91.7 na na
Hill Dalit 7.2 0 (0) 15 0 21.9 21.9
Magar (hill janajati) 7.1 1 (3) 9 17.0 29.5 46.5
Tharu (Terai janajati) 6.7 1 (2) 5 13.1 25.3 38.4
Tamang (hill janajati) 5.6 1 (4) 5 30.9 28.8 59.7
Newar (hill janajati) 5.5 1 (2) 1 21.1 25.7 46.8
Muslim 4.3 0 (0) 6 0 42.5 42.5
Rai (hill janajati) 2.8 0 (3) 4 30.4 22.5 52.9
Gurung (hill janajati) 2.4 1 (2) 4 2.6 34.8 37.4
Limbu (hill janajati) 1.6 0 (2) 1 38.3 11.1 49.4
Other (mountain & hill janajati) 2.3 0 (0) 3 0 1.1 1.1
Others 1.0 0 (0) 0 0 0 0
Total 100 33 (75) 72 38.7 21.1 59.8

In column 3, we have total of 72 districts at the last row because due to lack of data on language, I am not able to identify which group is in second position in the district of Jhapa, Morang and Siraha. The three districts where other (mountain and hill janajatis) are in second place are districts of Mustang, Humla and Mugu. Although the Madhesi group is 22.5% Nepal’s population, the three language groups add only to 22.3 indicating that other remaining Madhesi people speak other than these three languages. na: not available
Among the 29 districts that the ELC group is in plurality, Maithali speaking are in five district; Bhojpuri speaking and Tamang are in four districts each; Magar, and Rai are three districts each; Awadhi speaking, Tharu, Newar, Gurung and Limbu are in two districts each. In column 4, we examine in how many districts each group constitutes the second largest fraction of the population. The data shows that the BC group constitutes the second largest fraction in 19 districts. Hence this group is either the largest fraction or the second largest fraction in 65 out of 75 districts. The other groups wihich constitute the second largest fraction in descending number of districts are Hill Dalit (in 15 districts) followed by Magar (in nine districts) and Muslim (in six districts). Tharu and Tamang are second largest groups in five districts each.

The fifth column shows the population count of each group inhabited only in those districts where it is at least in plurality or more makes only 38.7% of Nepal’s population. The remaining 61% of Nepal’s population is distributed across districts in such a way that no group is even in plurality. BC and Maithali speaking are the only groups where majority of them are living in those districts where they are in plurality (75% and 85% respectively). For other ELC groups, a lot less than half of their population live in the districts where they are at least in plurality (the largest share of 38.3% is for Limbu); others live in districts where they are in minority as a single group.

The absence of many districts with a single ELC group as majority is due to the population distribution of each group being scattered across districts. To capture approximately 85-90% of each ethnic population in a cluster, depending on the group, 10 to 30 districts must be combined (Table 2). For example, 97% of Limbus are settled in 10 districts (71% of them in five districts). Among the groups, the most scattered are the BC and hill Dalit, for which 30 districts account for only 67% of their population. Among the ethnic groups, the most scattered is Magar, comprising 85% of its population in 30 districts.

Table 2. Shares of ethnic/caste population by districts of largest settlement (perecent)

5 Districts 10 Districts 15 Districts 20 Districts 25 Districts 30 Districts
Brahman, Chhetri, Thakuri 18.7 31.2 41.9 51.4 59.5 66.7
Terai 41.9 72.0 86.2 92.1 94.3 95.9
Terai caste 44.9 77.6 90.0 95.7 96.7 97.9
Terai janajati 43.2 69.0 81.5 89.0 91.7 94.3
Terai Dalit 25.8 49.5 75.0 79.2 86.8 88.9
Dalit (hill) 15.3 28.7 40.2 50.1 58.7 66.6
Magar (hill janajati) 29.9 48.9 62.9 72.4 79.3 84.7
Tharu (Terai janajati) 52.1 75.7 92.2 98.0 99.4 99.6
Tamang (hill janajati) 47.9 68.3 78.4 85.2 90.5 94.2
Newar (hill janajati) 53.5 65.6 75.1 82.0 86.8 90.8
Muslim 44.4 76.4 93.7 97.3 98.6 99.2
Rai (hill janajati) 50.3 78.8 92.3 96.1 97.6 98.6
Gurung (hill janajati) 47.9 70.4 77.9 83.8 88.1 91.0
Limbu (hill janajati) 71.2 96.5 98.7 99.2 99.5 99.7
Other (mountain and hill janajati) 41.8 64.7 75.9 82.3 87.3 98.1

In the previous table, we have presented Madhesi population in terms of three languages. However, since we do not have data on districts by mother tongue, we provide data on Madhesi in three other categories: Terai upper and middle caste, Janajati and Terai Dalit.
Identifying Regions
Nepal’s population distribution pattern raises many questions: How should Nepal’s constituencies of federalism be formed‌ How should the balance be maintained between the majority BC and the minority ELC groups‌ How should the ELC groups be recognized in the federation‌ Should the federation be a union of them‌ To answer these questions, we need to identify the focus region for each ELC groups. A focus region for a group includes all those districts in which it constitutes at least the second largest fraction after BC in terms of population share. Hence, for each group, one has to identify in which districts this group is either in majority or in plurality or is only second to BC. The districts where a particular group is either in majority, or in plurality or second to BC, are called the natural homelands for that group. Such districts are then combined to form a region with a focus for that group. These regions will be combined to form provinces, which, in turn, will make up the federation. The analysis in this section will use information provided in Tables A2 through A11 (Appendix A). The tables provide the 15 districts with the largest number of inhabitants for each of the main ELC groups except for Limbu, in which case we have given only ten districts, as they account for 97% of Limbu population).

1. Mechi river basin
The most eastern and northern part of Nepal, is the homeland of Limbus. The largest number of Limbus reside in Panchthar, Taplejung, Ilam and Terhathum districts (column 2, districts are sorted by the district share of total Limbu population in descending order, Appendix A, Table A2). Panchthar and Taplejung districts are home to 22.7% and 15.6% of total Limbu population, respectively. Relative to other population groups, Limbus are in plurality in Panchthar and Taplejung districts where they make up 40 and 42% of districts population respectively (column 3). They are not in majority in any district (the share of Limbu in total district population in column 4 is less than 50% in each district). In both Panchthar and Taplejung, BC group accounts for the second largest population (column 5) comprising 23.7% and 23% of district population (entries in column 6, which are smaller than Limbu’s share in column 4). Besides these two districts, the other district where Limbus are in plurality among ethnic groups but are less than BC is in Terhathum, where 11% of the total Limbu population reside.

Aside from the aforementioned districts, there are no districts where Limbus comprise the largest ethnic group (or the second largest group after BC). For example, in Ilam, Jhapa and Dhankuta Limbus account for the third largest population. In Ilam and Dhankutta, Rais are in larger fraction than Limbus, and in Jhapa, Terai caste (T-CT) group is larger than Limbu (compare the shares in last column with column 4). Hence, based on the criterion defined above, the natural homelands of Limbus are only Panchthar, Taplejung and Terhathum. All other districts would be natural homelands of other groups. Thus we combine these three districts to form a region called “Kanchenjunga region”, name after the mountain peak that is located in Taplejung, in which the focus group is Limbu. Note that in order to construct this region, we have also included the districts where Limbus are in second position in terms of the size of the population after BC. This strategy is adopted throughout the paper, but that construct should not be used to take away the democratic rights of BC groups.

Table 3. Population composition in Kanchenjunga region (share in percent)

Districts Limbu BC OHJ Hill Dalit Madhesi Tharu Muslim Total
Panchthar 40.3 23.7 27.7 6.6 1.0 0.1 0.0 99.4
Taplejung 41.7 23.0 26.2 8.2 0.5 0.1 0.0 99.7
Terhathum 35.4 36.0 16.8 9.6 0.7 0.0 0.1 98.7
Share in group’s total 49.5 1.7 1.9 0.1 0.0 0 2.0
Share in region’s total 39.5 26.6 24.5 7.9 0.8 0.1 0.0 99.3

Note: BC refers to Brahman, Chhetri, Thakuri and Sanyasi, and OHJ refers to other hill janajati which means hill janajatis other than Limbu. It may also include OMHJ. The districts are ordered in descending order of their shares in Limbu population.
As evident from the second last row of the table 3, these three districts are home to half of the Limbus in Nepal (49.5% of total Limbu population live in this region). And as reported to the last row of Table 3, Limbus are 39.5% of this region’s total population. The remaining groups in the region are BC (26.6%), other Hill Janajati-OHJ (24.5%) and Hill Dalit (7.9%). Among OHJ, the main groups are Rai and Tamang in Panchthar with 14% and 7% of district population respectively. In Taplejung, Sherpas constitute 10% of the district population while Rai, Gurung and Tamang constitute about 5% each. In Terhathum, after Limbu the third largest ethnicity is Tamang with 6% of the district population, and other groups such as Newar, Gurung, Magar and Rai constitutes about 2 to 3% each.

A group residing to the west of Limbu is Rai. Rais are not in majority in any district but are in plurality in three districts (Khotang, Bhojpur and Solukhumbu) (Table A3). They are in second position in four districts (Ilam, Udayapur, Dhankuta and Sankhuwasabha). By combining eight districts presented in Table 4 (those seven districts which are natural homeland of Rais and the district of Okhaldhunga, where Rais make up the third largest group after BC and Hill Dalit) we propose a region, named “Sagarmatha region”—Sagarmatha means Mount Everest in Nepali which lie in Solukhumbu district. We have to include Okhaldhunga district in this region even though this appears to be the homeland of Hill Dalit as there are no other districts nearby with this characteristic. In the three districts that Rais are the largest group, BC is the second largest group, and in all five districts where Rais are either the second or the third largest group, BC group is the largest group in terms of district population. Even though 63.2% of total Rai population is living in this region, they account for only 25.2% of the region’s population. Actually, the share of BC population is higher than that of Rai, making the latter a minority group in its own focus region.

Table 4. Population composition in Sagarmatha region (share in percent)

Districts Rai BC OHJ Hill Dalit Terai Total Tharu Muslim Total
Khotang 38.7 32.5 15.8 11.2 0.7 0.1 0.0 99.1
Bhojpur 34.1 29.5 23.7 11.4 0.7 0.1 0.0 99.5
Ilam 24.4 29.6 36.9 7.1 1.2 0.1 0.0 99.3
Udayapur 16.4 29.2 24.8 11.2 6.8 7.8 0.6 96.9
Dhankuta 23.0 28.1 38.7 7.6 1.1 0.1 0.1 98.7
Sankhuwasabha 22.4 26.8 40.7 8.2 1.0 0.1 0.0 99.2
Solukhumbu 31.5 21.1 36.5 9.1 0.5 0.1 0.0 99.0
Okhaldhunga 11.9 37.3 34.0 14.8 0.9 0.1 0.0 99.0
Share in group’s total 63.2 6.7 8.7 0.6 1.5 0.2 7.0
Share in region’s total 25.2 29.7 30.2 10.1 2.0 1.5 0.1 98.7

Note: BC refers to Brahman, Chhetri, Thakuri and Sanyasi and OHJ refers to other hill janajati other than Rai and may also include OMHJ. The districts are ordered in descending order of their shares in Rai population.
On the southern plain part of the natural homeland of Limbus and Rais, we combine the districts of Jhapa, Morang and Sunsari to form another region. Because there is no particular ELC group that is either in majority or in plurality in these districts and most of them have Nepali as their mother tongue, this region will not be a focus area for any ELC group. Even though Madhesi (people of Terai region) are in plurality in Morang and Sunsari, there is no single language which is used by more people relative to the mother tongue Nepali in these districts. Ideally, we would have preferred to provide a table based on mother tongue, however, these data are not available.

Table 5. Population composition in eastern Terai region (share in percent)

Districts Nepali speaking BC Hill Janajati Hill Dalit Madhesi Tharu Muslim Total
Jhapa 47.8 41.5 20.1 6.3 25.1 1.5 3.1 97.6
Morang 30.2 25.2 19.8 5.0 36.0 7.6 4.4 97.9
Sunsari 25 17.1 19.2 4.3 32.0 14.0 11.1 97.7
Share in group’s total 7.8 8.3 5.9 12.9 6.9 10.5 12.9
Share in region’s total 32.9 27.7 19.7 5.2 31.5 7.6 6.0 97.8

Hill Janajatis and hill Dalit constitute approximately 20% and 5% of this region’s population, respectively. Among the hill janajatis, Rai, Limbu and Newar are the largest group, each contributing at most 7% of the district population. The other remaining groups in the region are Tharu and Muslim. With 7.6% of the region’s population, Tharu is the third largest group.

2. Koshi-Bagmati river basin
Congruent to the west of Rais’ area, is the homeland of another hill ethnic group, Tamang. Based on Table A5, the only district that Tamangs are in majority is Rasuwa (63.7%), and they are the largest fraction in other three districts (Makawanpur, Newakot and Sindhuli). Additionally, they are the second largest group (the first largest ethnic group) after BC in five other districts (Kavrepalanchok, Sindhupalchok, Dhading, Ramechap and Dolakha). These nine districts which are the natural homeland of Tamang (given in Table 6) form “Gaurishanker region”, name given after the Gaurishanker himal in the district of Khotang.

Table 6. Population composition in Gaurishanker region (share in percent)

Districts Tamang BC OHJ Hill Dalit Terai Total Tharu Muslim Total
Makwanpur 47.3 26.5 18.4 4.3 2.6 0.2 0.3 99.8
Kavrepalanchok 33.8 38.2 17.7 7.2 2.7 0.1 0.1 99.8
Nuwakot 38.5 35.2 16.8 6.0 2.5 0.2 0.1 99.4
Sindhupalchok 32.2 34.4 21.6 7.4 3.8 0.2 0.0 99.5
Dhading 21.5 34.5 30.0 10.8 2.9 0.1 0.2 99.8
Sindhuli 25.6 25.1 24.0 14.6 9.7 0.6 0.0 99.7
Ramechhap 20.6 32.9 28.6 13.6 3.7 0.1 0.0 99.5
Rasuwa 63.7 19.4 12.1 2.9 0.9 0.3 0.0 99.3
Dolakha 15.7 45.1 27.9 9.4 1.2 0.0 0.0 99.4
Share in group’s total 59.7 11.4 11.2 1.7 0.3 0.3 10.6
Share in region’s total 31.8 33.1 22.2 8.6 3.6 0.2 0.1 99.6

Note: BC refers to Brahman, Chhetri, Thakuri and Sanyasi and OHJ refers to other hill janajati (other than Tamang) and may also include OMHJ. The districts are ordered in descending order of their shares in Tamang population.
In the four districts that Tamangs are the majority and the largest fraction and in the five districts that Tamangs are either the majority or the plurality and the five districts in which they are in the second largest group, BC group is the second largest and first largest group respectively. About 60% of all Tamang population live in these nine districts, and the remaining 40% are distributed quite thin across several other districts. Even if we take the 15 districts with the highest number of Tamang settlement as shown in Table A5, the total share of Tamang population is about 78%. In terms of the region’s population, with its share of 31.8%, Tamangs are second only to BC which make up 33.1% of the region’s population. The other main group in this region is OHJ, mainly Newar and Magar which make up 10% and 6% of region’s population, respectively.

Moving from east to west and from north to south, we reach the Kathmandu valley. The data on the distribution of Newar population by district (Table A6) show that the top 15 districts encompass only 75% of the total Newar population. They are in majority in Bhaktapur (55.9%) and in plurality in Lalitpur (40.3%) and second only to BC in Kathmandu (29.6%). There are no other districts where Newar is the largest fraction of ethnic population. These three districts (Kathmandu, Lalitpur and Bhaktapur) make “Kathmandu region” where Newar constitutes only 35.4% of region’s population, and less than 50% of Newar reside in Kathmandu region (Table 7). In the two districts, where Newars are the largest population fraction, BC is the second largest group. In terms of population share in the region, BC is the largest group with 37.8%, about 2.5 percentage points higher than Newars’ share. The other large group, other hill janajatis, is mainly the Tamang.

Table 7. Population composition in Kathmandu region (share in percent)

Districts Newar BC OHJ Hill Dalit Terai Total Tharu Muslim Total
Kathmandu 29.6 41.0 20.4 2.2 4.0 0.8 1.1 99.3
Lalitpur 40.3 32.5 19.7 3.3 2.5 0.7 0.3 99.4
Bhaktapur 55.9 30.0 10.0 2.2 0.9 0.3 0.1 99.3
Share in group’s total 46.8 8.8 2.2 1.1 0.8 1.4 7.2
Share in region’s total 35.4 37.8 18.8 2.5 3.3 0.7 0.8 99.3

Note: BC refers to Brahman, Chhetri, Thakuri and Sanyasi and OHJ refers to other hill janajati (other than Newar) and may also include OMHJ. The districts are ordered in descending order of their shares in Newar population.
The next homeland, region of Mithila, consists of five districts (Saptari, Siraha, Dhanusa, Mahottari, and Sarlahi), where majority of inhabitants speak Maithali as their mother tongue. Because we do not have data on distribution of Maithali speaking people in other districts, we have not been able to provide similar table on linguistic as we have done on ethnicity in our appendix. However, for comparison purpose, we have given the table for Madhesi people, Table A4, where Madhesi could be anyone whose mother tongue is Maithali, Bhojpuri, Awadhi, Rajbanshi or any other Terai languages. These five districts are inhabited by about 83% of the Maithali speaking population, and they constitute 77% of the total population in the region. More than 25% of Muslims in Nepal live in this region, and they constitute about 9% of the region’s population. BC group constitutes 5.4% of the population.

Table 8. Population composition in Mithila region (share in percent)

Districts Maithali speaking Madhesi BC Hill Janajati Hill Dalit Tharu Muslim Total
Dhanusa 89.7 75.6 5.6 4.6 4.1 0.6 8.5 99.0
Siraha 85.0 76.1 3.0 4.4 1.7 4.8 7.3 97.3
Mahottari 82.5 71.9 3.8 5.3 2.9 1.6 13.5 99.0
Saptari 75.1 67.1 4.6 3.5 1.3 12.8 8.5 97.8
Sarlahi 54.4 66.7 9.7 9.5 2.3 3.2 7.7 99.0
Share in group’s total 82.8 41.9 2.3 2.8 4.1 8.7 27.7
Share in region’s total 77.2 71.5 5.4 5.5 2.5 4.5 9.0 98.4

The districts are ordered in descending order of their shares in Maithali speaking population.
3. Gandak river basin
This discussion takes us to the Gandak river basin area, a northern region generally inhabited by Gurungs. The Gurugn population is quite scattered. For instance, if we combine the top 15 districts inhabited by Gurungs, we will only account for 77% of their population (Table A7). They make up the majority only in two districts, Manang and Mustang (note that Mustang is not listed in Table A7 as the share of Gurung residing in Mustang is only 1.2%, less than 1.3% in Morang, the 15th district). Gurung constitutes 75% of the Manang population and slightly less than 50% of the Mustang population. In three districts (Kaski, Gorkha and Lamjung), Gurungs are in second largest group after BC. Although these are the three districts with the largest number of Gurungs, they constitute only 35% of the Gurung population. So, five districts (Kaski, Gorkha, Lamjung, Manang and Mustang) make the “Annapurna region”, a focus of Gurung group. Interestingly, in all fifteen districts that Gurungs have resided heavily, the largest group is BC (second in two districts where Gurungs are in majority) except in Rupandehi where Terai-caste is the largest group. Note that this region has only 37% of Gurung population; Gurungs are less than one-quarter (23.4%) of the region’s population. The share of BC is more than 14 percentage point higher than that of Gurung. Other Hill janajati also constitutes substantial share (18.4%) followed by 15.6% of hill Dalit. This region seems to be more diverse than other groups.

Table 9. Population composition in Annapurna region (share in percent)

Districts Gurung BC OHJ Hill Dalit Terai Total Tharu Muslim Total
Kaski 18.1 47.1 14.9 14.7 3.8 0.2 0.7 99.6
Gorkha 22.3 31.5 23.9 16.8 4.0 0.0 0.9 99.4
Lamjung 31.7 32.4 15.8 17.0 2.3 0.1 0.4 99.8
Manang 75.9 5.2 16.1 2.1 0.5 0.1 0.0 99.9
Mustang 45.2 11.3 31.7 9.4 1.7 0.3 0.0 99.6
Share in group’s total 37.4 4.7 7.3 0.6 0.1 0.6 3.8
Share in region’s total 23.4 37.9 18.4 15.6 3.5 0.2 0.7 99.6

Note: BC refers to Brahman, Chhetri, Thakuri and Sanyasi and OHJ refers to other hill janajati (other than Gurung) and may also include OMHJ. The districts are ordered in descending order of their shares in Gurung population.
Next we examine the region of Nepal where Magars are concentrated. They are in majority (50.9%) in Palpa, and occupy plurality shares of population in two districts (Rolpa and Myagdi, 44% and 42% respectively—Table A8). They have the second largest share of population in nine districts (seven as shown in the Table A8, and the eighth is the district of Arghakhanchi, where both Magar and Hill Dalit are 16.4% of district population and ninth is the district of Salyan).

Thus, there are twelve districts where Magar are either the largest or the second largest group (after BC) in district population, therefore being eligible to be called Magar’s natural homeland. In Parbat, Magar is the third largest group (with 10.7% of the population) after BC and Hill Dalit (19.7% of the population). Geographically, Parbat is surrounded by Magar’s natural homeland. As a result, we have include Parbat to Magar’s focus region, increasing the number of districts to 13 in “Ridi region”, name after the place which is famous in Magars’ history and is located in Gulmi district . Note that these 13 districts encompass 51% of Magar population (Table 10). In terms of region population, they are far less relative to BC (26% vs. 41%). Another group which is in large number in this region is Hill Dalit. In fact, this region is home of 14% of Hill Dalit. They are the third largest group in 10 of 15 districts of top Magars inhabitant.

Table 10. Population composition in Ridi region (share in percent)

Districts Magar BC OHJ Hill Dalit Terai Total Tharu Muslim Total
Palpa 50.9 28.9 4.0 11.4 4.2 0.1 0.6 99.8
Nawalparasi 17.2 24.5 5.9 6.7 24.8 5.8 16.5 99.5
Rolpa 43.8 37.6 0.7 14.8 2.6 0.0 0.1 99.9
Tanahu 26.8 28.0 22.0 15.2 6.6 0.0 0.2 99.7
Baglung 27.7 43.4 4.2 21.7 2.6 0.0 0.1 99.9
Syangja 21.2 47.3 13.5 15.5 1.7 0.1 1.0 99.9
Pyuthan 30.6 43.0 3.6 18.1 4.1 0.2 0.1 99.9
Gulmi 19.9 54.4 2.8 16.9 5.6 0.1 0.3 99.8
Myagdi 41.8 26.6 7.7 20.8 2.5 0.0 0.3 99.5
Rukum 23.1 68.6 0.8 5.8 1.5 0.0 0.1 99.9
Parbat 10.7 59.0 8.1 19.7 2.0 0.2 0.3 99.8
Arghakhanchi 16.4 56.4 3.4 16.4 6.3 0.0 0.1 99.7
Salyan 17.2 61.9 3.0 12.6 2.9 0.2 0.8 98.6
Share in group’s total 51.1 18.7 8.8 12.9 6.2 3.4 14.0
Share in region’s total 26.0 41.2 6.8 14.2 7.5 3.0 1.0 99.7

Note: BC refers to Brahman, Chhetri, Thakuri and Sanyasi and OHJ refers to other hill janajati (other than Magar) and may also include OMHJ. The districts are ordered in descending order of their shares in Magar population.
The majority of the people in three districts in central Terai (Bara, Parsa and Rautahat) has Bhojpuri as mother tongue. These three districts combined with district of Chitwan make up the central Terai region. Based on the criterion, Chitwan should have been the natural homeland of Tharu. However, we include it in this region because there are no Tharu homeland districts close to Chitwan. In all three districts where Bhojpuri speaking are in majority, the second largest group is Muslim and the third largest groups are BC in Rautahat and Tharu in Bara and Parsa. This region encompasses 63% of the people whose mother tongue is Bhojpuri and this group constitutes slightly more than 50% of the region’s population. Additionally, Madhesi constitutes 48% of the region’s population followed by BC (15.3%), Muslim (12.6%), and Tharu (9.3%).

Table 11. Population composition in center Terai region (share in percent)

Districts Bhojpuri speaking Madhesi BC Hill Janajati Hill Dalit Tharu Muslim Total*
Bara 76.1 56.9 9.1 6.1 2.4 11.3 13.4
Parsa 83.6 60.9 7.3 5.3 1.0 8.2 15.4
Rautahat 42.8 63.4 6.2 3.0 2.1 5.0 19.5
Chitwan na 6.2 41.6 29.6 8.7 12.7 0.8
Share in group’s total 62.7 19.5 4.5 3.6 3.8 12.5 26.8
Share in region’s total 51.8 48.0 15.3 10.5 3.4 9.3 12.6 99.1

na: not available
*total is not given as there are no data available by mother-tongue. The districts are ordered in descending order of their shares in Bhojpuri speaking population.
4. Karnali river basin
Next we examine an area where Dalits are the focus. In contrast to janajatis, Dalits have no traditional homeland; they are similar to the BC group in the sense that both are distributed across the country. The fifteen districts which are home to the highest number of Dalits are homeland to 50% of Dalit population (Table A9). There are 11 districts (reported in Table 12) that classify as natural homelands of Dalits, as they have the second largest population after BC in these districts. They are far from majority in any of the 11 districts. For instance, their population composition is highest in Kalikkot (27.5%) followed by Jajarkot (25.6%), and Dailekh (22.4). In terms of their own total population, the largest number of Dalits are in Kailali followed by Baglung, Surkhet, Kaski and Achham (each with about 3% share).

These 11 districts make up the “Khaptad regionn”. This region encapsulates only (17.8%) of the Dalit population and the Dalits in this region represents only (18%) of the region’s population. In contrast, the largest fraction of the region, BC constitutes 68% of the region’s population. Among the hill janajatis, the main are Magars in Surkhet, Dailekh, Jajarkot, Salyan and Kalikot and Terai untouchable in Doti and Terai caste in Bajura.

Table 12. Population composition in Khaptad region (share in percent)

Districts Hill Dalit BC Hill Janajatis Terai Total Tharu Muslim Total
Dailekh 22.8 62.4 11.8 2.7 0.0 0.2 99.9
Jajarkot 25.6 62.3 9.2 2.5 0.0 0.1 99.7
Surkhet 21.5 46.5 23.2 6.0 2.1 0.4 99.7
Achham 22.4 68.3 0.9 6.8 0.0 0.1 98.5
Doti 17.8 66.7 3.9 9.5 0.1 0.1 98.0
Bajura 17.0 71.8 1.4 8.2 0.0 0.1 98.5
Kalikot 27.5 65.3 3.9 2.8 0.5 0.1 100.0
Bajhang 13.0 81.6 0.3 3.8 0.0 0.1 98.8
Baitadi 11.1 78.1 0.4 8.3 0.0 0.0 98.1
Dadeldhura 15.9 74.5 3.7 4.5 0.0 0.1 98.7
Darchula 7.8 86.6 1.4 3.8 0.0 0.0 99.7
Share in group’s total 17.8 17.7 2.0 2.1 0.4 0.3 8.0
Share in region’s total 18.0 68.0 6.6 5.8 0.3 0.1 99.0

The districts are ordered in descending order of their shares in Dalit population.
The southern districts Kapilvastu, Banke and Rupandehi, (Table 13) form a region call, “Lumbini region” (name after Lumbini, the birthplace of Buddha) where the majority of people has Awadhi as mother tongue. Unfortunately, in the absence of language data for the district of Rupandehi, we are not sure whether this district is a homeland of Awadhi speaking people. It most likely is. If it is not the case then it will be a homeland of Tharu (10.6% of the district population). However, even in that case, it will not be feasible to allocate this district to the Tharu region since the Tharu region is not congruent to Rupandehi district.

Table 13. Population composition in Lumbini region (share in percent)

Districts Awadhi speaking Madhesi BC Hill Janajati Hill Dalit Tharu Muslim Total
Kapilvastu 71.3 47.7 13.4 3.8 2.5 12.6 19.4
Banke 44.2 26.5 22.0 8.1 4.3 16.4 21.2
Rupandehi na 38.0 22.5 14.9 4.0 10.6 8.9
Share in group’s total 91.7 11.7 4.4 2.6 3.1 13.0 24.4
Share in region’s total 32.6 38.1 19.6 9.8 3.6 12.6 15.1 98.9

na: not available
The districts are ordered in descending order of their shares in Awadhi speaking population.
The remaining region on the Karnali river basin formed by the districts of Kailali, Bardiya, Dang, and Kanchanpur, is the Tharu region. As given in Table A10, Tharus are the majority in Bardiya district, are the plurality in Kailali, and are the largest ethnic group in Dang and Kanchanpur. The district of Kailali is home to the largest number of Tharus (about 18%) followed by Bardiya (13%) and Dang (10%). Although Tharu’s settlement stretches to many teria districts, these are the only four districts which are natural homelands for Tharu and make up the western Terai region.

Table 14. Population composition in western Terai region (share in percent)

Districts Tharu BC Hill Janajati Hill Dalit Terai Total Muslim Total
Kailali 43.7 31.6 5.0 9.7 6.5 0.6 97.1
Bardiya 52.6 23.4 4.6 6.0 10.1 3.0 99.8
Dang Deokhuri 31.9 36.9 13.4 10.6 6.0 1.0 99.8
Kanchanpur 23.3 48.5 4.9 8.3 12.0 0.1 97.1
Share in group’s total 46.0 9.1 2.2 8.8 3.0 2.1
Share in region’s total 38.4 34.7 7.0 8.9 8.3 1.1 98.3

The districts are ordered in descending order of their shares in Tharu population.
5. Rara territory
The most underdeveloped part of Nepal is the remote west-north corner of Dolpa, Humla, Jumla and Mugu districts. In 2001, approximately 164 thousand people were residing in this area, with approximately 32 thousand households in 20 thousand square feet of land. These four districts cover 7% of the land in which only 0.7% of the total population (40% of them are in Jumla district alone) reside. Not only is this are underdeveloped, but the terrain is difficult and the settlement is quite scattered. For these reasons, we propose to make this area a territory—called “Rara territory”, the name derives from the Rara lake that is situated in Humla. We believe that combining these four districts into a single territory will bring forth consolidated efforts to raise the living standards of the most neglected area of Nepal. By making it a territory, the federal government will have more responsibility of providing more revenue to this area.

Table 15. Population composition in Rara territory (share in percent)

Districts BC Hill Janajati Hill Dalit Terai Total Tharu Muslim Total
Dolpa 50.4 39.9 7.8 1.6 0.2 0 99.8
Humla 69.7 16.6 8.9 4.7 0.1 0 99.9
Jumla 79.7 2.2 13.8 3.8 0.1 0 99.6
Mugu 66.2 30.1 3.2 0.2 0.2 0 99.9
Share in group’s total 1.6 0.4 0.9 0.2 0 0 0.7
Share in region’s total 70.7 13.1 9.7 6.2 0.1 0 99.8

In the Rara territory, about 71% of the region’s population is BC, 13% is other Hill Janajatis and 10% is Hill Dalits. In the Dolpa district, the Hill janajatis are Gurung (23%) and Magar (13%). In Humla, Sherpas are 14% of the population and Thakalis are 15%. In Jumla, however, the share of Hill Janajatis is small, and the second largest group is Hill Dalit. For Mugu, the second largest group is one of Dhusadh, Chamar, Dhobi and Mushar.

We have identified 11 regions based on ethnicity, language and caste. Additionally, we identified a region on the eastern Terai which has no specific focus group. Lastly we identified a territory on the north-west where the overwhelming majority of people are BC and there is no ELC group that could be identified as natural inhabitant of that region. In terms of population, the largest region is Ridi region (Magar foucs) with 14% of Nepal’s population followed by Mithila region with 13% of Nepal’s population. The smallest is the Kanchenjunga region with 2% of Nepal’s population. In terms of district, the region with highest number of districts is again Ridi region (13), while four regions (Kanchenjunga, eastern Terai, Kathamandu and Lumbini) have the three districts each. They are: regions. Rara territory covers 13.3% of the country’s area but only 0.7% of the population inhabits there.

IV. A Model of Federation: Identifying Sub-National Units
Now the question is: Should Nepal be considered a federation of these 12 regions and one territory or should there be an alternative arrangement‌ From socio-economic development perspective, one would like to have federation as a union of those political sub-national units that can efficiently manage their economic, social and educational policies. Although the identified 12 regions and one territory could potentially be constituencies of the federation, that may not be optimal. In this context, we propose necessary and sufficient conditions for making Nepal a federation of the ELC regions. NECESSARY CONDITION: In order to make a federation as a union of regions based on ethnicity, language and caste groups, two conditions must be fulfilled: (i) the group must constitute at least simple majority of its homeland region’s population and (ii) the region must be home for at least simple majority of the homeland group’s national population.

In forming these regions, we tried to put each ethnic, linguistic and caste group together if they are the largest group after BC. In other words, in forming the regions, our objective has been to maximize the share of each of these groups’ in the region’s population. Despite this attempt, the distribution of population is so mixed that most of the ELC groups are in minority even in their own regions (see Table 16 where all summary statistics are drawn). There are only two ELC groups—Maithali and Bhojpuri speaking groups—that are majority in regions’ population (column 6). Maithali and Bhojpuri speaking people make up approximately 77% and 52% of the total population in their respective regions. Three groups are in plurality (Limbu, Awadhi speaking and Tharu), in their regions, but with less than 40% of own region’s population. The remaining six groups are overshadowed by BC even in their own regions (compare columns 6 and 7). For these six groups, their share of population in their respective regions range from as low as 18% for Dalit to 35% for Newar. Hence, only Maithali speaking and Bhojpuri speaking regions fulfill Necessary Condition (i).

Even in terms of the shares in their own population, these ELC groups are not quite representative; the natural homelands are not clustered at all (column 6). Of the 11 focus groups only six focus group have majority of their groups’ population residing in their regions, thereby fulfilling Necessary Condition (ii). For the three language focus groups the shares are very high (92% for Awadhi speaking, 83% for Maithali speaking and 63% for Bhojpuri speaking). The other three groups with more than 50% of their total population in their regions are Rai (63%), Tamang (60%) and Magar (51%). The remaining five focus regions (for Rai, Newar, Gurung, Dalit and Tharu) house less than 50% of their own population. Thus the regions are weak in terms of having similar ethnic, linguistic and caste groups. Thus, only Maithali speaking and Bhojpuri speaking qualifies for both part of Necessary Condition.

The above condition can be considered a bare necessity. However, the sufficient condition can be as follows. SUFFICIENT CONDICITON: In order to make a federation as union of ethnicity, language and caste groups, the focus group must constitute at least two-thirds of the region’s population and the region should be home for at least two-thirds of focus group’s population.

None of the 11 ELC regions except Maithali speaking area passes the sufficient condition. Hence, given the population distribution, it is fair to say that the a federation based on regions that are solely defined on ethnic, linguistic and caste groups is not feasible.

Is it desirable economically to have a federation based solely on ELC regions‌ To answer this, we have to consider the natural resources that each ELC region has and examine the prospect of each region becoming a viable economic unit. There are three regions why we conclude that the federation based on only ELC regions is not desirable economically. First, considering the small size of some of the regions, one could imagine that the administrative, economic and social costs of forming an union on these regional bases will be very high. In the lack of added benefit in making federation of these regions, this reason itself is sufficient to search for another basis for sub-national entities.

Table 16. Summary by regions and provinces

Regions Focus group Number of existing districts* Population (%) Area (%) Region’s share in focus group’s population (%) Focus group’s share in region’s population (%) BC’s share in region’s population (%)
Mechi province 14 18.2 17.6 25.9 21.7 28.4
Kanchenjunga region Limbu 3 2.0 3.8 49.5 39.5 26.6
Sagarmatha region Rai 8 7.0 10.6 63.2 25.2 29.7
Eastern Terai region 3 9.2 3.2 8.3 27.7
Koshi province 17 31.0 16.4 74.0 58.8 22.4
Gaurishanker region Tamang 9 10.6 11.7 59.7 31.8 33.1
Kathmandu region Newar 3 7.2 0.6 46.8 35.4 37.8
Mithila region Maithali 5 13.2 4.1 82.8 77.2 5.4
Gandak province 22 26.9 27.0 60.3 38.2 32.0
Annapurna region Gurung 5 3.8 8.9 37.4 23.4 37.9
Ridi region Magar 13 14.0 14.1 51.1 26.0 41.2
Center Terai region Bhojpuri 4 9.1 4.0 62.7 51.8 15.3
Karnali province 18 23.1 25.7 50.1 37.7 41.8
Khaptad region Dalit 11 8.0 15.4 17.8 18.0 68.0
Lumbini region Awadhi 3 6.9 3.7 91.7 32.6 19.6
Western Terai region Tharu 4 8.1 6.7 46.0 38.4 34.7
Rara territory 4 0.7 13.3 1.6 70.7
NATIONAL TOTAL 75 100 100 61.4 30.9

Note: The sum of last two columns (92.3%) do not add to 100 because the share of Muslim population (4.3%), and other mountain and hill janajatis (2.3%) and others (1%) is not included in either of the two column.
*The district names for different regions are as follows. Kanchenjunga includes, districts of Taplejung, Terhathum and Panchthar. Sagarmatha region includes eight districts: Ilam, Sankhuwasabha, Solukhumbu, Bhojpur, Khotang, Okhaldhunga, Dharan, Udayapur. The region of eastern Terai includes Jhapa, Morang and Sunsari. Gaurishanker region has nine districts: Dolakha, Sindhuliplachok, Rasuwa, Ramechap, Kaverpalanchok, Nuwakot, Dhading, Sindhuli and Makawanpur. Kathmandu region includes, Kathmandu, Patan and Bhaktapur districts. Mithila region includes Saptari, Siraha, Dhanusa, Mahottari and Sarlahi. Further to the west is Annapurna region with five districts, Gorkha, Lamjung, Manang, Mustang and Kaski. Ridi region has 13 districts: Tanahu, Syanja, Parbat, Myagdi, Palpa, Gulmi, Baglung, Arghakhanchi, Pyuthan, Rolpa, Rukum Salyan and Nawalparasi. Central Terai has districts of Rautahat Bara, Parsa and Chitwan. Khaptad region has 11 districts: Surkhet, Jajarkot, Dailekh, Kalikot, Achham, Bajura, Bajhang, Doti, Darchula, Baitadi and Dadeldhyra. Lumbini region includes Rupandehi, Kapilbastu and Banke. The four districts Dang, Bardia, Kailali, and Kanchanpur are in western Terai region. Finally, Rara territory includes Dolpa, Jumla, Mugu and Humla.
Second, the point of making Nepal an inclusive democracy is to allow all to share the fruits of development and raise the overall standard of living. For this to happen, political restructuring is essential but that in itself cannot bring all the desired results. What is equally important is a comprehensive economic policy by each pillar of federation for which resources that it owns will be important. Nepal, which has a very different ecology from north to south, also has different resources in these regions. One could make a convincing argument that differences in resource should be a part of the equation in forming a federation. In other words, put the regions together that are dissimilar in their endowment and hence complement to economic activities.

The mountain, hill and Terai regions are very different in their natural endowment due to their ecology and geography. Mountains can be used for herbs, husbandry and horticulture, and hills can be used for farming, and horticulture, while the Terai can be used for agricultural and industrial purpose. The natural resources, water, minerals and forests are located in the mountain and hill areas, whereas the plain land “grain basket of Nepal” is located in the Terai. Besides other things, mountains and hills can be developed into popular tourist spots, whereas the Terai could be developed for other services such as safari. Hence, considering the topology, vegetation and natural endowments, the three regions spanning from north to south in every river basin that we have defined complement each other. Because all three groups are distinct in terms of ethnicity or language, the combination of three groups will make a political unit that encapsulates a wealth of diversity.

Third, Nepal is bordered with two giants, India on the south and China on the north. These are one of the fastest growing economies in the world and are the leading economic powers in the developing world. If Nepal could put its policies together and have a long term plan on human capital building (a radical reform in education and skill promotion), the opportunities it can tap from these highly growing economies are enormous. Nepal could turn from “a yam between two giants” to “a diamond between two economic powerhouses”. Hence it is imperative that we give due weight on making constituencies of federalism that bordered both with India and China.

Based on the above background, we propose the formation of a province by combining three regions. The eastern three regions, Kanchenjunga, Sagarmatha (focus regions for Limbu and Rai) and eastern Terai make “province of Mechi”. The next three regions, Gaurishnaker, Kathmandu and Mithila (the respective focus regions for Tamang, Newar and Maithali speaking) make “province of Koshi”. Annapurna, Ridi and center Terai (focus regions for Gurung, Magar and Bhojpuri speaking) make “province of Gandak” consists of). And, the remaining three regions, Khaptad, Lumbini and western Terai (focus regions for Dalit, Awadhi speaking and Tharu) form the “province of Karnali”. All these provinces border India to the South and China to the North. These four provinces and Rara territory are the five pillars of Nepal’s federation (map with regions and provinces is provided in Appendix B).

The provinces are balanced in terms of area and population. In the four provinces, the population shares range between 18-31%; the area share range between 18-27%, and the number of existing districts range from 14-22. In terms of population, Koshi province will be the largest with 31% of the population covering 17 existing districts followed by Gandak province with 26.9% of the population and 22 districts (Table 16). Other remaining provinces will have 23.1% (Mahakali province) and 18.2% (Mechi province) of the population.

The formation of provinces in the north-south manner may sound like the same old story of regional development practiced by the previous regimes for the last few decades that kept the country stagnated and backward. But there is fundamental difference. Even though we have defined province on the north-south manner, there are well-founded regional ingredients within each province. Furthermore, the additional problem in the past was due to political, economic, social and cultural policies. With massive inclusiveness in politics, radical reform in economic endowments, broad-based high quality universal education and independent provincial economic policies, the act of integrating three regions stretching north to south to form a province with these provinces forming an union is the best strategy for Nepal.

No doubt, we need to provide special attention to the minority ELC groups and women who have been excluded from Nepalese politics and economic fortunes. There is a huge disparity among different groups in terms of wealth, access to education and access to health (Table 17). Poverty rates are highest among hill and Terai Dalits (57.8%) and hill janajatis (48.7%). While 45% of Newars completed at least grade eight, only 12% of Dalits, 15% of Hill Janajatis, and 16% of Terai middle casts did so. More than half of Nepal’s population is illiterate and the illiteracy rate for Terai and Hill Janajatis is even greater. The average time required to reach school is highest for Hill Janajatis, indicating that households are far away from the educational institute. The case is even worse in terms of access to health centers.

Table 17. Different indicators by ethnicity and caste in Nepal, 2003-04

Poverty headcount (%) Illiterate (%) Completed at least grade 8 (%) Population living in urban area (%) Time to reach school (hours) Time to reach health centers (hours)
BC 34.1 36.9 33.9 15.5 0.35 0.78
Terai middle class 28.7 64 15.6 6.6 0.20 0.47
Dalits (hill-Terai) 57.8 59 12.3 9.5 0.30 1.10
Newar 19.3 31.2 45.1 53.1 0.25 0.72
Hill janajatis 48.7 52.3 15.0 10.5 0.43 1.36
Terai janajati (tharu) 53.4 54.2 17.7 1.5 0.22 0.54
Others 63.1 63.1 14.5 15.6 0.25 0.50
Muslim 46.1 7.3 0.21 0.45
Nepal 41.8 50.1 22.7 15.0

Source: World Bank (2006), Table 1.14, page 16; Table 2.3, page 23; Table 2.8, page 27; Table 2.9, page 28. This table is based on “Nepal Living Standard Survey 2003/04,” Volumes I and II, Central Bureau of Statistics, Nepal.
This is one side of the coin; the other side is that there are disparities among regions (Table 18). When comparing the skilled nonagriculture sector, the wage differences among regions are huge. Despite vast hydroelectric resources, access to electricity is low in rural areas, whereas all in Kathmandu valley reported have electricity. Poverty rates differ greatly by regions, ranging from 3.3% in Kathmandu to 43% in rural eastern hills and 38% in rural western Terai.

Table 18. Different indicators of regional disparity in Nepal, 2003-04

Poverty headcount rate Distribution of the poor Distribution of population Average daily wage in skilled nonagriculture in 1995-96 rupee Share of households with an electricity connection
Kathmandu 3.3 0.6 5.4 672 99
Other urban 13.0 4.1 9.7 170 81
Rural Western Hill 37.4 23.6 19.4 111 26
Rural Eastern Hill 42.9 29.4 21.1 137 25
Rural Western Terai 38.1 18.9 15.3 126 32
Rural Eastern Terai 24.9 23.5 29.1 159 27
Nepal 30.9 100 100 295

Source: World Bank (2006), Table 1.3, page 6; Table 3.6, page 48, and Table 6.10, page 81. This table is based on “Nepal Living Standard Survey 2003/04,” Volumes I and II, Central Bureau of Statistics, Nepal.
Thus there are disparities based on ethnic and caste groups, and there are alarming disparities based on regions. As Nobel Laureate Amatya Sen (2006) describes the divisive factors region, gender, caste, religion and class are complement and they add together. In his view, class, in particular can make the influence of the other sources of disparity much sharper. In his words, “class is not only important on its own, it can also magnify the impact of other contributions to inequality, enlarging the penalties imposed by them (Sen, 2006, p. 206). That is what has happened in Nepal. Although the proportion of people living under abject poverty are higher for other ELC groups than those for BC and Newar groups, there are alarming class differences even inside those groups, a vertical inequality. Not all “upper” castes are wealthy, not all “lower” castes are poor (on this issue see Pradhan and Shrestha, 2005).

Hence the thinking that the ethnic autonomy would necessarily entail economic prosperity could turn out to be a myth. What is needed is a broader policy perspective providing opportunities for all to participate in nation building, whether in education, in politics or in economic activities. For this to happen, the major condition is to empower people of all groups, irrespective of ethnicity, language and gender. We claim that the political framework we have proposed in this paper will be a precondition for addressing both vertical and horizontal disparities across groups and regions.

We provide a summary population distribution of ethnic, language and caste groups for the provinces and territory in Table 19. In the Mechi province, the largest group is Hill BC with 28.4% of the population; the second largest group is Rai with 13.4% of the population, and the third largest group is Limbu (note that this province has two natural homelands: for Rai and Limbu). For all provinces, the entries with bold regular font and bold italics indicate the largest group and the second largest group respectively. In Koshi, the largest, Maithali speaking group, with 32.8% of population is ten percentage points greater than the second largest group BC. The third and the fourth largest groups are Tamang and Newar respectively (the other two focus groups for the province). In Gandak and Karnali provinces, BC is the largest group with others a distance second. In Gandak, after BC the three largest groups are Bhojpuri speaking, Magar and Gurung (the only three groups with homelands in the province). And for Karnali, the three largest groups after BC are Tharu, Hill Dalit and Awadhi speaking.

Table 19. Distribution of provincial population by ethnicity, language and caste

Mechi Province Koshi Province Gandak Province Karnali Province Rara Territory NEPAL TOTAL
Hill BC 28.4 22.4 32.0 41.8 70.7 30.9
Hill Dalit 7.4 4.6 10.7 10.5 9.7 8.1
Hill & mountain janajatis 38.3 33.4 26.5 7.7 13.1 26.4
Magar 4.1 3.3 15.0 5.6 1.9 7.1
Tamang 4.6 13.7 1.8 0.2 1.2 5.6
Newar 4.1 12.2 2.9 0.7 0.3 5.5
Rai 13.4 0.9 0.2 0.1 0.0 2.8
Gurung 1.5 1.2 5.6 0.8 3.2 2.4
Limbu 8.3 0.2 0 0.1 0.0 1.6
Other mountain & hill janajatis 2.3 1.9 0.9 0.2 6.4 1.4
Maithali speking 0 32.8 0.0 0 0 12.3
Bhojpuri speaking 0 0 17.5 0 0 7.5
Awadhi speaking 0 0 0 9.8 0 2.5
Other Terai groups 16.8 0 3.1 6.6 6.2 0.2
Tharu 4.5 2.1 4.7 17.4 0.1 6.7
Muslim 3.1 4.1 4.9 5.0 0 4.3
Others 1.7 0.5 0.5 1.3 0.2 1.0
Total 100 100 100 100 100 100

An Electoral Model for the Federal and Provincial Legislatures
With region and provinces defined, we examine how the national and provincial legislations should be formed. We propose a three-tier government: central, provincial and local. Both at the national and provincial levels, there will be a bicameral parliament; whereas the territorial government will be unicameral. At the national level, the lower house (the National Assembly) is elected to represent the people, and the upper house (National Council of Provinces) is made up of representatives from the provinces. The bicameral system at the national level will be a method of combining the principle of democratic equality with the principle of federalism — all citizens are equal in the lower houses, while all provinces are equal in the upper house. Similarly, at the provincial levels, all citizens are taken as equal at the lower house, and all regions are equal in the upper house. In what follows, we disucss how these two levels of parliments should be formed.

1. National parliament
We start with the National Assembly, where we propose to have total of 305 legislatures. Of the 305, 300 will be eleceted from 12 regions based on the population proportion in each region. The remaining five will be elected from the territory. For the election of the national assembly, each region will be a constitutency. Each party will submit two lists of candidates: regional/territorial (for each of 12 regions and one territory) and national. The candidates in the regional list should be from the region, however, the national list can be drawn from all over the country. For each party, the total number of candidates for each regional list should not exceed the number of legislatures designated for that region. The national list should not exceed more than 305 candidates, and the candidates who are in regional lists can also be included in the national list.

Based on these lists, the votes will be translated to seats based on “proportional list system”. In each region, voters will vote to the party and not for the candidates. The ballet box will have two columns of party names, one for regional and the other for national. The names of parties in the ballot for the regional list could vary if all parties are not putting their candidates in all regions, but the national list of parties will remain the same throughout the country. Voters will have the right to split the vote by choosing different parties for regional and natioanal representations. About half of the seats, 150, will be elected based on regional list and the remaining seats will be filled using the national list.

As the vote will be for parties and not for individuals, and the rule of proportionality has to be invoked, we need to define what electoral formula to be used to translate votes into party seats. We propose for “largest remainder” system with Imperiali quota for the regional list and Droop quota for the national list. For an excellent discussion on major electoral systems that are practiced in the world, see Farrell (2001). A brief summary of relevant concepts, derived from Farell, is given in Appendix C.
Let us first discuss the mechanism for assigning seats based on the regional list by defining quota votes per seat in region r ([image]), as
(1) [image],
where [image] = total number of valid votes cast in region r, [image] = total number of seats in region r. And
(2) [image],
where[image] = total number of seats to be awarded to a party p in region r; and [image] = total number of votes cast to a party p in region r. Any fraction (remainder) will be discarded in the first round and will be used in the second round (if the seats are still available after assigning according to equation 2), awarding seats from the highest remainder party towards the lowest ones until the seats are exhausted. Hence this system is called, the “largest remainder” system. Identifying the number of seats party p wins in the territory by [image], the total seats for party p based on regional list, [image], is given by
(3) [image]
For the other half candidates which have to be fulfilled through national list, we propose a similar process. As in the previous case, a quota of votes per seat at the national level, [image], shall be determined by dividing the total number of valid votes cast throughout Nepal, [image], by total number of legislative members, [image], plus 1 and the result plus one (Hare quota), disregarding fractions. As a balance, we propose for Droop quota at the national level. In notation term,
(4) [image], and
(5) [image],
where [image] is the total number of valid votes cast nationally in favor of party p, and [image] is the total number of seats that the party p should receive based on national votes. However since the party has already won [image] number of seats as given in equation (3), the additional seats the party p wins based on national list, [image], is given by
(6) [image]

By construct, [image]. If there is a fraction in equation (5), and the seats are not exhausted then the remaining seats should be awarded to the party or parties concerned in sequence of the highest remainder (fraction), up to a maximum of three seats so awarded. Provided that subsequent awards of seats still remaining unawarded seats shall be made in sequence of those parties having the highest average number of votes per seat already awarded. Whichever party has the largest number of seats [image] will form the national government.

This discussion completes the process of the lower house. At the upper house, the National Council of Provinces (NCP), there will be 15 seats each from each province and 5 seats from Rara territory. The parties involved in the provincial parliaments will nominate legislatures for NCP. The number of seats of a party from a province to the NCP will be determined as follows:
(7) [image];
where [image]is the number of seats party p nominates in NCP from province I; [image] is the number of seats the party holds in the provincial legislature (both lower and upper houses), and [image] is the total number of seats in the legislature (both lower and upper houses) in province i. In case of territory (given by sub-script t), the multiplication and the division will be by 5 as given by
(8) [image],
where [image] and [image] are the number of seats a party holds in the territorial legislature and total number of seats in the legislature, respectively. With this provision, the total number of seats of a party p in the NCP, Bp, will be given by:
(9) [image];

Besides those elected 65 representatives, there will be five nominated by the federal government, one from each province and one from the territory, thus making the total number of seats in the upper house 70. The nomination must be out of the eleven ELC groups that has been recognized as the basis for regional formation. And at least one of them must be Muslim. The spirit of this nomination is to bring small ethnic groups who may not be present in the upper house.

Based on this discussion and the most recent population Census data of 2001, the total number of National Assembly members that have to come through regional and national lists for each region and territory are given in column 3 and column 4, respectively, in Table 20. The distribution of NCP is given in column 5.

Table 20. Distribution of legislatures at the national parliament

Number in lower houses Number in upper houses
Population share (%) Total number of legislatures
Of which to be elected from regional list
National Council of Provinces Provincial Council of Regions

Total number of legislatures
Mechi Province 18.2 55 27 15 15 140
Kanchenjunga region 2.0 6 3 5 17
Sagarmatha region 7.0 21 10 5 47
Eastern Terai region 9.2 28 14 5 61
Koshi Province 31.0 94 47 15 15 218
Gaurishanker region 10.6 32 16 5 69
Kathmandu region 7.2 22 11 5 49
Mithila region 13.2 40 20 5 85
Gandak Province 26.9 81 40 15 15 192
Annapurna region 3.8 12 6 5 29
Ridi region 14.0 42 21 5 89
Center Terai region 9.1 27 13 5 59
Karnali Province 23.1 70 34 15 15 170
Khaptad region 8.0 25 12 5 55
Lumbini region 6.9 21 10 5 47
Western Terai region 8.1 24 12 5 53
Rara territory 0.7 5 (10) 2 5 20
Nominated 5 4 9
NATIONAL (PROV) 100 305 (310) 150 (155) 70 64 749

Note: The column “total legislatures” is about the 3 times the first column (the share of population), except for the row “Rara territory”.
2. Provincial/Territorial parliament
We propose bicameral parliaments for four provinces and unicameral for the territory. The number of seats in the lower house of the provincial parliament will be equal to the number of seats in the lower house of the national parliament, the National Assembly. Each region will have the same number of representatives in the lower house of the provincial parliament as it does in the lower house of the national parliament. Each party will provide regional and provincial lists of candidates. The total number of seats in the lower house of the provincial parliament divided by 2 (fraction discarded) will be elected based on regional list, and the remaining will be elected by provincial list. Thus the voters in the provincial election also will vote in a ballet where they choose the parties for the regional and for the provincial seats.

The mechanism of translating votes to seat will be exactly the same as in the National Assembly. For the regional list, the procedure for assigning winning seats will be as given by equations (1) through (6), with the difference that in equations (4) through (6) it is province not nation that is relevant for the formulas. So the total number of seats at the province and their composition by region and the representatives to be elected based on the regional list will be as given in columns 3 and 4, respectively, of Table 20.

We are now left with only the upper house of provinces “the Provincial Council of Regions (PCR)”. We propose that each province will have 15 seats, five seats from each region and 10 for the territory. The voters in a region will choose a party for the representation. The seats will be assigned based on the quota votes per seat in region r ([image]), which is defined as
(10) [image],
where [image] = total number of valid votes cast in region r. And
(11) [image],
where[image] = total number of seats to be awarded to a party p in region r; and [image] = total number of votes cast to a party p in region r. Any fraction will be discarded. The distribution of PCR is given in column 6. The total number of legislatures (from both levels and houses) for each region and province is given in last column, where the total of 749 legislatures has been distributed.

A Model for Inclusiveness of Ethnic, Language and Caste Groups
The above procedure takes care of how the seats will be allocated to parties. In this section, we discuss how different groups should be represented both at the national and provincial parliaments of lower houses. Note that the criteria proposed here do not apply to the upper houses (both national and provincial) and Rara territory. One of the reasons for forming the region with focus on particular ethnic, language and caste groups is to make an inclusive democracy so that all groups are represented in the parliament. The fact that the ELC regions are the electoral constituencies, by construct, the chances for ELC groups to be elected representatives are higher. In addition, we provide a framework that ensures candidacy from each ELC group. The framework is developed based on the following four considerations. It should (1) foster political competition among parties, (2) accommodate regional candidates from the focus group for each region, (3) accommodate representation of groups other than those which are focus groups using national list, and (4) provide half of the total candidates to women.

1. National Assembly
First we start from national level and in the next sub-section we will discuss the criteria for provincial parliaments. For each of the eleven groups (Limbu, Rai, Tamang, Newar, Maithali speaking, Gurung, Magar, Bhojpuri speaking, Dalit, Awadhi speaking and Tharu) that are the basis of regional formation, a party must fulfill the following criterion so that the number of candidates for any of these groups from their own designated region is not less than the number given by:

(12) [image], r = 1, .., 11; g = 1,…, 11
where [image] is the number of candidates that a party p has to include from a group g in a region r (note that group g is the focus group of region r, such as if g is for Limbu, then r is for Kanchenjunga region); [image] is the share of a group g in its own focus region r’s population; and [image] is the total number of candidates that party p has filed from region r. Any fraction will be discarded. Note that if we make the regional list very restrictive in a sense that the focus group at the region should have at least the same share of candidates as its population share in the region (that is, weight of unity instead of 0.8 applied in equation 12), then it could happen that not many parties can fulfill this criterion thereby limiting political competition.

Similarly, at the national list (out of total 300 candidates), the number of candidates of a party p from each group should not be less than the number given by the following criterion:
(13) [image], g = 11 ELC groups, BC, Muslim and Others
where [image] is the number of candidates that a party p has to include from each group g in its national list; [image] is the share of that group in Nepal’s total population; and [image] is the total number of candidates that the party has contested nation-wide excluding those from the territory (if the party gives all the candidacy, then it will take the value of 300, not 305). Note that the scale factor is 0.7 instead of 0.8 that was applied for regional list of candidates (equation 12). The results based on population Census 2001 calculated using equations (12) and (13) are given in Table 21.

Table 21. Minimum candidates at the National Assembly by groups

Focus Group

Share in total population

Share in own focus region’s population
Total members at the National Assembly from the corresponding region Members of National Assembly for focus group
Of which Regional list
National list

Of which Regional list
BC 30.9 64
Limbu 1.6 39.5 6 3 3 0
Rai 2.8 25.2 21 10 5 2
Tamang 5.6 31.8 32 16 11 4
Newar 5.5 35.4 22 11 11 3
Maithali 12.3 77.2 40 20 25 12
Gurung 2.4 23.4 12 6 5 1
Magar 7.1 26.2 42 21 14 4
Bhojpuri 7.5 51.8 27 13 15 5
Dalit 7.2 17.9 25 12 15 1
Awadhi 2.5 32.6 21 10 5 2
Tharu 6.7 38.4 24 12 14 3
Muslim 4.3 9 0
OMHJ 2.3 4 0
Others 1.0 2 0
NATIONAL TOTAL 69 38.6 272 134 202 37

Note: “OMHJ” means other other mountain and hill janajatis. For both OHMJ and “Others” see footnote 16.

The second column in this table provides the main groups’ shares in Nepal’s total population. From the last row “national total” it is clear that 11 groups that form the regions constitute 61.4% (= 100 – 30.9 – 4.3 – 2.3 -1) of Nepal’s population. Note that the first group (BC) and the last three groups “Muslim”, OMHJ and “Others” are not focus groups. The third column is the share of each group in its own focused region’s population. For example, the entry 39.5 is the share of Limbu in its corresponding (Kanchenjunga) region’s total population. In average, the 11 groups make up 38.6 percent of these 11 region’s population (the sum of 11 focus group population, one group from each region divided by sum of 11 regions’ total population). The fourth and fifth columns are the total number of seats that have to be represented in the National Assembly from the 11 regions that corresponds to these groups (repeated from columns 3 and 4 in Table 20). For example, the total number of seats from Rai focus region (Sagarmatha) is 21 (the second entry, column 45). The total number of candidates for National Assembly that has to be represented from 11 regions is 272, remaining 28 come from eastern Terai, and five from territory. Out of this 272, 134 will be represented from regional list (column 5).

A party which provides full candidates in all regions and in the nation, should have the minimum number of candidates for each ELC groups as given in columns 6 and 7. Column 6 is the outcome of column 2 multiplied by 300 (the total number of non-territorial legislatures) multiplied by 0.7, the outcome of equation (13). So the number of candidates from all groups combined will not be less than 202. Out to them from the 11 groups, in total, there will at least 123 (202 minus 64 from BC minus 9 from Muslim groups minus 4 from OMHJ and 2 from “others” groups) which is 45% of the total candidates of 272 that are proposed from these 11 regions. And the outcome of equation (12) is given in column (7). Out of those minimum 123 candidates from these 11 groups, at least 37 must be from the regional list. Note that since the Muslim and the “other group” have no defined region, they do not have regional representation but they would get the national share as all other groups.

In developing these criteria, we have tried to balance between the minimum representation and the degree of political competition among parties that could appear as trade off in some instances. If we put higher weight at the national level (more than 0.7 given here), we are putting stringent condition that each party should have the same type of political influence in each of 11 groups. This is neither possible nor necessary. Note that with the criteria we have developed, if a party has no national candidates of one group, it has to lower the total number of candidates respectively. Hence we do not want the political competition to be compromised by putting higher weight.

So far we have discussed about the number of candidacies but what matters is how the party assigns its candidates to the winning seats. Mere representation of different groups in candidacies does not guarantee inclusion. So we propose that in seat distribution, all parties must fulfill two criteria. The first one is given by equation (14) as follows:
(14) [image],
Each group should obtain at least the number of seats that would have been if we multiply the share of this group in national population by additional number of seats that a party wins at national level (given by [image] in equation 6). In other words if a party wins 50 additional seats from national list, it should provide 3 candidates out of these 50 from Bhojpuri speaking people, which are 7.5% of the population (3 = 0.075 * 50). The second criterion is that half of the total seats should be awarded to women.

1. Provincial Assembly
We provide similar criteria for representation of different groups in provincial candidates in this sub-section. The symbols used are the same as used for the national level, but they are subscripted by i to denote that they are for provinces. For any party that contest for election in a province i should fulfill the following criterion on its regional list of candidates:

(15) [image],
where [image] is the share of focus group g in the region r’s population in province i; and [image] is the total number of regional list of candidates put forward by party p in region r. Any fraction will be discarded. This regional formula applies only to the focus group of the region; there is no such rule in the candidacies of other groups which are not focus for that particular region. Similarly, in total provincial candidates, the number of candidates for each group that a party proposes must not be less than the number given by following criteria:
(16) [image], g = 11 ELC groups, BC, Muslim and Others
where [image] is the lower bound of the number of candidates a party should put in its provincial list from group g; [image] is the share of that group in province i’s population; and [image] is the total number of candidates put forward by part p in the provincial list.

Using the shares of different groups in total population in a province given in Table 19, and multiplying them by the total number of seats in each province reported at the bottom of the Table 22 (entries outside the parentheses) and scaling it by 0.8, we have provincial candidates lower bound for the groups which are given by column (open entries) in Table 22. Thus this calculation is based on equation (16) assuming that a party contests for all seats. The entry in the parentheses is the lower bound of regional candidates for the focus groups in its own region, obtained using equation 15. For that we multiply the number of regional seats for each province given in column 5 by entries in column 3 in Table 21 and scale it by 0.8. Note that the entries will be only for those which are focus groups for the province. For example, in Mechi, the minimum criterion applies only to Limbu and Rai (the two focus groups of the province).

Table 22. Minimum candidates at the Provincial Assembly by groups

Group Mechi Province Koshi Province Gandak Province Karnali Province Rara Territory TOTAL
BC 12 16 20 23 71
Limbu 3 (0) 0 0 0 0 3
Rai 5 (2) 0 0 0 0 5
Tamang 2 10 (4) 1 0 0 13
Newar 1 9 (3) 1 0 0 12
Maithali 0 24 (12) 0 0 0 24
Gurung 0 0 3 (1) 0 0 4
Magar 1 2 9 (4) 3 0 17
Bhojpuri 0 0 11 (5) 0 0 12
Hill Dalit 3 3 7 5 (1) 0 18
Awadhi 0 0 0 5 (2) 0 5
Tharu 2 1 3 9 (3) 0 15
Other Mountain & Hill Janajati 1 1 0 0 0 1
Muslim 1 3 3 2 0 9
Others 0 0 0 0 0 0
Other Madhesi 6 0 0 0 0 6
Total seats without BC 25 (2) 53 (19) 38 (10) 24 (6) 0 140 (37)
Total seats with BC 37 (2) 69 (19) 58 (10) 47 (6) 0 211 (37)
Total seats in the lower house 55 (27) 94 (47) 81 (40) 70 (34) 10 (5) 310 (153)

Note: the numbers inside the parentheses are the members to be elected based on regional list.
The third last row “total seats without BC” shows the total number of minimum candidates that have to be filed from different groups without counting candidates from BC. It shows that the number of candidates combined from all these 11 groups and Muslim and Other hill and mountain janajatis should not be less than 25 in the Mechi region, 53 in the Koshi region and so forth. The second last row also includes BC, which means that the total minimum candidates for Mechi region are 37 (25 + 12 from BC). The last row shows the total members in the provincial lower house. So any difference between the last row and the second last row will be the room for party to have candidates from any groups. For example, in Mechi a party can have 18 (55 – 37) candidates from any groups, which is 32% of the total 55 seats in the province.

The last column shows the total number of minimum candidates from all four provinces combined. For example, without BC, all other groups (11 ELC plus Muslim plus OMHJ) will have total of 140 candidates, with BC that number will be 211. So, out of total of 300 (not counting those for Rara Territory) candidates for all provinces, remaining 89 (30%) may come from any groups.

The summary of minimum candidates from each group, for lower houses of both national and provincial parliament, is given in Table 23. The second column shows the share of population (taken from column 2 of Table 2). The third column is the sixth column of Table 21 and the fourth column is the last column of Table 22. The sum of columns 3 and 4 is provided in column 5, which shows the total minimum number of candidates that a party should put if it is contesting in all seats in the provinces and the national lower house (out of total 600). For example, a party should have at least 64 candidates for National Assembly and 71 candidates for Provincial Assembly with total of 135 from BC group and so on. In total, 413 out of 600 (68.8%) candidates are pre-assigned for these groups, whereas the remaining 31% may come from any of these groups. The last column shows the share of candidates from each group (given in column 5) in total of 413. This share closely matches population share of the group (column 2). Some discrepancy between these shares for a group in two columns is due to rounding.

Table 23. Minimum candidates at National and Provincial Assemblies by groups

Share in total population (%) National Assembly (number) Provincial Assembly (number) Total (number) Share of candidates (%)
BC 30.9 64 71 135 32.7
Limbu 1.6 3 3 6 1.5
Rai 2.8 5 5 10 2.4
Tamang 5.6 11 13 24 5.8
Newar 5.5 11 11 22 5.3
Maithali 12.3 25 24 49 11.9
Gurung 2.4 5 3 8 1.9
Magar 7.1 14 15 29 7.0
Bhojpuri 7.5 15 11 26 6.3
Dalit 7.2 15 18 33 8.0
Awadhi 2.5 5 5 10 2.4
Tharu 6.7 14 15 29 7.0
Muslim 4.3 9 9 18 4.4
OMHJ 2.3 4 2 6 1.5
Others 1 2 0 2 0.5
Other Madhesi 6 6 1.5
Sum of minimum candidates 100% 202 211 413 68.8%
TOTAL Candidates 100% 300 300 600

Note that since the three language groups in Terai (Maithali speaking, Bhojpuri speaking and Awadhi speaking) do not cover all Madhesi population, we have also provided the number of candidates from other Madhesi groups these in the last row.
A nation becomes stronger when its citizens are empowered and have opportunities to advance. If previously marginalized people who were excluded from political, economic, and social aspects focus on their vision of a prosperous Nepal, a changed political situation can be the beginning of the empowerment process. While the end of the monarchy is a necessary condition for Nepal’s prosperity, is not sufficient in itself. In order for Nepal to thrive and its people advance, there must be an appropriate political structure, access to physical resources, and access to education.

This paper has outlined a political restructuring – one of the pre-conditions for increasing Nepal’s prosperity – that offers the best compromise for power-sharing among different ethnic, linguistic, and caste (ELC) groups, including the dominant Brahamin and Chhetri groups. This political structure will be able to develop political institutions that are inclusive for all.

However, political structure itself cannot address the striking horizontal (i.e. resource disparity between groups) and vertical class differences (i.e. disparity of resource ownership within groups) in Nepal. These differences must be addressed by a complete package of economic policies – the second pre-condition for prosperity – that are designed to provide an opportunity for everyone to contribute to the process of wealth generation. This is an important issue that needs research and rethinking.

The only way for Nepal to be a prosperous and dynamic nation is by educating and improving all of its citizens. That is also the only way to take advantage of our rapidly growing neighboring giants India and China. We must have an educated population if we are to compete in the international market or make use of technology. The issue of access to education – a third pre-condition for prosperity – is the single most important issue in Nepal. A thorough discussion on educational policy must occur if Nepal is to benefit from its changed political situation.

Besides the issues of resource ownership and access to education, there are several other important issues raised in this paper. For example, what should the language policies be in a restructured Nepal‌ What should the language be for media delivering education to the different ELC focus regions and provinces‌ What should the jurisdictions be between federal (national) and provincial (sub-national) governments‌ What should the formula for federal cash transfers to provincial and territorial government be‌ Additionally, Nepal is endowed with vast water resources. Thus, there should be a focus in this area. A well-formulated plan on the use of water resources is an important issue for both research and policy. A serious debate and research on all of these issues could be useful instruments to help achieve what all Nepalese are dreaming and aspiring for – a just, prosperous, democratic, and dynamic society.

Devkota, S.R. and Gautam, S. July 2006. Model of Federal System, Kantipur daily newspaper, 25, http://www.kantipuronline.com/kolnews.php‌&nid=80799.

Devkota S.R. and Gautam, S. September 2006. Federal Model – II, Kantipur daily newspaper, 14, http://www.kantipuronline.com/kolnews.php‌&nid=86167.

Farrell, D.M. 2001. Electoral Systems: A Comparative Introduction, New York: Palgrave.

Finance Canada. May 2005. Expert Panel on Equalization and Territorial Formula Financing (TFF), Canada, Ottawa, http://www.eqtff-pfft.ca/english/issuespaper0.asp

Gurung, H. February 2000. Idea of Decentralization: District Strengthening, New Nepal with Twenty Districts, 23 Himal Newspaper.

Hachhethu, K. April 2003. The Question of Inclusion and Exclusion in Nepal: Interface between State and Ethnicity, Paper presented to the conference on The Agenda of Transformation: Inclusion in Nepali Democracy.

Lawoti, M. July 2006. Self-Rule: An Efficient Weapon of Conflict Management (in Nepali, Himal Newspaper, 16 Bhadau – 31 Bhadau 2063

Neupane, G. 2005. The Nationalities’ Question of Nepal: Social Convergence and Partnership Building through Multiculturalism and Federation, Revised Internet Edition (English), Kathmandu: Center for Development Studies, Nepal.

Newar, N. January 2007. Nepalis Want to Stand and be Counted, 332 Nepali Times, 19 January – 25.

Pradhan, R. and S. Ava. June 2005. Ethnic and Caste Diversity: Implications for Development, Working Paper Series No. 4, Asian Development Bank.

Pokharel, B. 2005. Restructuring the Nation-State: Nepal, Mimeo.

Sen, A. 2005. The Argumentative Indian: Writings on Indian History, Culture and Identity. New York: Picador, Farrar, Strus and Giroux.

Sharma, P. July 2006. Regional Structure of Republican Nepal (in Nepali), Mulyankan 10, Saun 2063.

Sharma, V. 30 January 2007. Autonomous Ethnic-Federal System (in Nepali),” Kantipur daily newspaper, http://www.kantipuronline.com/kolnews.php‌&nid=99190.

Tamang, P. July 2006. Why is ethnic-state necessary (in Nepali)‌ Himal Newspaper, 16 Bhadau – 31 Bhadau 2063.

The World Bank. 2006. Unequal Citizens: Gender, Caste and Ethnic Exclusion in Nepal.

The World Bank. June 2006. Nepal Resilience Amidst Conflict: An Assessment of Poverty in Nepal, 1995-1996 and 2003-04.

Leave a Reply