Introduction

1This article analyses the structure and dynamics of public opinion in Nepal with regards the radical transformation of the principal features of the country’s political system in the recent past, with a special focus on the similarities and/or differences between the positions taken up by Nepal’s major political parties, and the concerns and priorities of the general public. Furthermore, the article aims to contribute to both the question about the sustainability of Nepal’s new political system and the more general question about elite-mass relations in political science. As this article discusses political events up until 2015 (until the new constitution was promulgated by the Constituent Assembly in September 2015), it does not include the political and social aspects of the Gen-Z movement that recently took place in Nepal in September 2025.

3A study by Karatnycky and Ackerman (2005) reveals that non-violent civic resistance and changes in public opinion are major factors that contribute to the success and failure of transitions to democracy. These scholars find that non-violent civic resistance during the period of transition has a significant effect on yielding sustainable freedom. Referring to a Freedom House study, Shin (2007) argues that the success or failure of the process of democratisation largely depends on the role the general public plays during the transition. The study shows that the likelihood of a transition from authoritarian rule to liberal democracy proving successful is over four times higher if transitions are defended by strong non-violent civic coalitions than in cases where they are not. Doorenspleet and Shrestha (2013) pay particular attention to the interactions between the mass level and various elite levels, arguing that Nepal’s repeated failures to consolidate democracy stem from different degrees of support for democracy at different levels. Though their analysis focuses on support for the idea of democracy rather than on specific constitutional arrangements, as in my study, both studies share the same conviction that the interplay between elite actions and public opinion is vital for ensuring the legitimacy of the political system. Similarly, in the context of Uganda, Moehler (2006) contends that, although large-scale public participation in the constitution-making process is one of the most recommended methods for strengthening constitutional legitimacy, genuine consensus among political elites is essential for ensuring lasting constitutional support.

Why these four reforms?

5Nepali identity has been shaped through these four key components. It was constructed on the basis of the dominant culture, religion and language of the Hindu high-caste hill group. But at the same time Nepali society has been ethnically, culturally, religiously and linguistically diverse since its existence. Though the promotion of the four pillars was unsuccessful in limiting this diversity, it had protected the Hindu high-caste hill group’s domination of the culture, religion and language and had marginalised and even excluded other cultures, religions and languages from the state’s structure. This is the main reason that this article focuses on these four components.

Democratic values and political tolerance: political elites vs general public

6The fact that the structure of the Nepali state was inextricably linked to the cultural, religious and linguistic domination by the Hindu high-caste hill group makes it all the more important to examine and compare the positions taken up by the political elites (ie top-level political leaders of major political parties, legislators and prominent political decision makers) and the decisions taken by the political parties on the one hand and the opinions of the general public on the other hand.

7From a normative democratic point of view, it may seem preferable that elites and political parties listen to public opinion. In particular, when it comes to establishing or reforming the rules and institutions of democratic decision-making, it would seem desirable that political parties seek to translate the concerns of ordinary citizens into a new constitution. But in some respects, this line of argument may be too simplistic. Empirical studies of elite attitudes in a variety of countries have shown that political elites have higher democratic values than the general public. Elites are more tolerant with regard to minority rights and civil liberties than the general public, and they show more support for democratic institutions and principles (e.g. McClosky 1964; Barnum and Sullivan 1989; Sullivan et al 1993; Farag 2020). Samuel Stouffer was the first to demonstrate that community leaders in the USA were generally more supportive of democratic principles and civil liberties than ordinary citizens (Sniderman et al 1991: 349–351). The common argument in these studies is that political elites develop these values due to a process of selective recruitment and due to their socialisation, leading to different degrees of tolerance between political elites and their public (Jackman 1972: 753; Shamir 1991: 1020; Sullivan et al 1993: 52-53). More recently, public opinion surveys conducted in Tunisia between 2010 and 2016 showed that Tunisian elites were more democratic than the masses, and that the masses had developed over time more authoritarian positions than the elites (Farag 2020: 553-559). This became the basis for ‘the theory of democratic elitism’: the principles of democracy are safer in the hands of political elites than in those of the general public.

8However, later studies have questioned whether political elites have a more tolerant attitude. Jackman (1972), for example, reanalysed Stouffer’s data and concluded that the elite-mass difference in tolerance disappeared once demographic differences and the level of education in particular had been taken into account. Similarly, Shamir (1991: 1020) argued that political elites seemed to be more tolerant than the general public due to the selective recruitment whereby people belonging to certain social strata are overrepresented in the elites: the more educated, those of higher social and economic status. Sniderman et al (1991: 369–370) accept that political elites have higher democratic values than the general public, but they stress that such attitudes should be studied within political elites because political parties differ significantly in their attitudes. In addition, evidence of democratic elitism dwindles when the focus shifts from civil liberties to other democratic values, and in particular when elites are studied in systems other than well-established Western democracies (Peffley and Rohrschneider 2007). This finding is especially relevant in a study of democratic reform in Nepal.

9From a Burkean perspective on representative democracy, the discrepancy between the public’s preferences and the decisions taken by their elected representatives is arguably much less problematical. The eighteenth-century political philosopher Edmund Burke famously argued that representatives, once elected, should not act as mere delegates of their voters, but as ‘trustees’ who are elected to use their good judgement to deliberate about the common good. Trustees take decisions based on what they think to be right or just; they do not just follow citizens’ views (Burke 1774: 81). From that perspective, the empirical findings of these studies do not contradict this particular version of normative democratic theory.

10By using data from the 1956–1960 American National Election Study, Converse (1964) argues that relatively few members of the general public understand ideological politics and hold ideologically consistent attitudes. In his influential book chapter titled ‘The Nature of Belief Systems in Mass Publics’, he argues that in the general public, political attitudes are highly unstable and that ordinary citizens essentially lack a structured belief system, unlike political leaders. Hence, he argues that the general public is incapable of guiding policy in a coherent manner. Finally, he concludes that belief systems are constructed and used by political parties and their representatives, not by ordinary citizens. In a similar vein, Zaller (1992) argues that public opinion is largely dependent on the most recent information obtained through exposure to the mass media, which is itself driven by political discourse. He contends that the general public does not have meaningful and stable political opinions on many issues and concludes that public opinion is consequently unreliable; public preferences are no doubt shaped by ‘others’ (Zaller 1992: 6), in practice often by political leaders and other opinion leaders. He affirms that political communications from party leaders that are conveyed in the mass media shape public opinion, but he also argues that the level of political awareness of people is relevant to the impact of political discourse on them. People who are more politically aware are more likely to receive cues from political leaders than those who are politically less aware. By examining the welfare policy of the United States based on findings from the 1992, 1996 and 2000 CPS National Election Studies, Schneider and Jacoby (2005) argue that political leaders’ positions guide and shape public opinion, albeit in the short term. They suggest that political discourse is the prime factor for bringing about changes in public opinion (Schneider and Jacoby 2005: 377). After investigating the opinions of both the general public and local political elites regarding Uganda’s constitution-making process, Moehler (2006: 293–302) argues that public participation had a weak and uneven influence on constitutional support, and that the opinions of political leaders shaped public assessment of the constitution-making process. This was because the general public lacked sufficient knowledge and information to evaluate the constitution on their own and that is why they turned to local elites for cues. Moehler ultimately concluded that public participation during the constitution-writing process alone could not substitute the role of political elites. Consensus among political elites, including opposing elites, was essential to secure an outcome.

11The comparison of the positions of elites and parties with those of the general public in the context of the recent transformation of the Nepali democratic system may thus contribute to the ongoing discussion of the elitist theory of democracy.

Data on public opinion and elite opinion

12Longitudinal opinion polls based on random (probability) sampling, titled ‘Nepal Contemporary Political Situation’ or NCPS, were conducted with support from The Asia Foundation Nepal between 2004 and 2012, and the author of this article was one of the principal researchers. Furthermore, other longitudinal opinion polls (also based on random sampling), called ‘People’s Perception of Safety and Security’ or PPSS, were conducted with support from The Saferworld UK between 2007 and 2010, in which this author was also involved. The availability of these longitudinal public opinion polls at such a crucial period in Nepal’s history means that it is possible to measure and to track people’s perceptions, choices and priorities during the momentous historical transition and to detect the continuities and changes in people’s opinions of the same topics and issues over the period. It would not be possible to do so if such longitudinal public opinion polls had not been conducted. Such longitudinal polls are very rare outside Western established democracies. Conducting public opinion polls is even more challenging in a third-world country like Nepal due to the small number or poor quality of roads. In addition, an armed conflict between state security forces and Maoist rebels increased the risk of collecting data in the field. Despite these challenges, we were able to conduct these opinion polls without compromising their quality. So, the public opinion data used in this study is of great value in itself.

13Ideally, we should be able to compare the opinions of citizens with those of members of the political elite. Fortunately, there was an Elites Opinion Poll conducted in July 2007 (i.e. about a year before Constituent Assembly elections) as part of the Nepal Democracy Survey undertaken by the International Institute for Democracy and Electoral Assistance (International IDEA) (see International IDEA 2008). The respondents of the poll were members of the Interim Parliament (MPs). Interviewers questioned as many MPs as possible during a one-month period. They conducted interviews with 300 MPs out of a total of 330 (International IDEA 2008: 11). The poll contained questions about various political issues including republicanism, secularism, multilingualism and federalism. Unfortunately, it was undertaken only once, which means that the poll does not allow us to map any continuities or changes in the perceptions of political elites. However, the first decisions in the transition process, taken by the elected Constitutional Assembly in 2008, were in line with the views expressed by the political elites in this poll.

Methodology of public opinion polls

14Before the new constitution was promulgated on 20 September 2015, Nepal was administratively divided into five development regions along an east-west axis, namely the Eastern Development Region (EDR), Central Development Region (CDR), Western Development Region (WDR), Mid-Western Development Region (MWDR) and Far-Western Development Region (FWDR). Since then, three ecological regions have been identified along a north–south axis: the mountains, the hills and the Tarai. With regards the two following parameters, ecological region and development region, Nepal has been divided into 15 distinct eco-development regions (i.e. 3 ecological regions x 5 development regions), namely the eastern mountains, eastern hills, eastern Tarai, etc. The Kathmandu Valley is treated as a separate 16th region given the fact that it contains the capital city, not to mention its large population. These 16 eco-development regions constitute the different ‘strata’ of the sample design. Each stratum has distinct features (i.e. physical, cultural-linguistic and social) and within a given stratum there is a relatively large degree of homogeneity, whereas there is generally a large degree of heterogeneity across strata.

16In this way, the surveys followed random (probability) sampling techniques at every stage so that the findings of these surveys could be generalised to the entire population under study (see e.g. Patton 1990, Sayer 1992, Weisberg 2008). Random (probability) sampling is arguably the only scientific basis that allows inferences to be drawn from a sample to a population, although there is inevitably still a (small) margin of error between sample statistics and population parameters (O’Muircheartaigh 2008).

17The samples of these surveys truly represented the national population. The sample makeup in terms of ethnicity, sex, age group, region, religion etc was very consistent with the population’s makeup, as per Nepal’s 2001 national census (See The Asia Foundation 2005; 2006a; 2006b and 2008; Interdisciplinary Analysts 2011a for comparisons between the sample’s makeup and the population’s makeup). Therefore, the findings of these surveys are likely to closely mirror the opinions of Nepal’s entire adult citizenry (with the usual margin of error, of course), not only of sample respondents. In other words, the findings of these surveys can be generalised to the entire population.

Recent political development and transformation

19The CPN (Maoist) signed the Comprehensive Peace Accord (CPA) with the government in November 2006, ending the decade-long armed insurgency. The signing of the CPA paved the way for holding Constituent Assembly elections, as well as many other actions such as promulgating an interim constitution, forming an interim parliament and an interim government. The CPA also stated that the first session of the Constituent Assembly would decide the fate of the monarchy by a simple majority vote.

20The historic and long-awaited Constituent Assembly elections took place on 10 April 2008. The Maoists participated in these elections and emerged as the winning party with 238 seats out of a total of 601, followed by the Nepali Congress with 115 seats and the CPN (UML) with 109 seats. The CPN (Maoist) failed to secure a majority of the seats in the Constituent Assembly. However, this was less of a problem as there was by now a general consensus of opinion among the major political parties on federalism, republicanism, secularism and multilingualism. Nepal was declared a federal, democratic, secular and republican state by the first sitting of the Constituent Assembly on 28 May 2008. This declaration was a momentous event in Nepal, which led to the start of a transformation of the state’s political structure. King Gyanendra’s putsch and his refusal to work with the political parties, and the subsequent alliance between the parliamentary parties and the CPN (Maoist) consequently proved to be fatal, first for the autocratic King-chaired government and eventually for the core features of the Nepali state, including the centuries-old institution of the Gorkhali Shah dynasty.

21The Constituent Assembly recognised the need for efficient participatory constitution-making and, on 4 November 2008, approved the Constituent Assembly Rules to assist the Constituent Assembly in drawing up a constitution and in facilitating people's participation. A constitutional committee and 10 thematic committees were created. Each thematic committee was assigned responsibility to prepare a preliminary draft along with a concept paper on its respective subject. The constitution committee was mandated to prepare a draft constitution by integrating preliminary reports submitted by the thematic committees. The latter, according to the Constituent Assembly Rules, were required to seek proposals from the general public through public hearings, interaction programmes, seminars and workshops (see Articles 65 and 66 of Constituent Assembly Rules 2065 [2008]). The general public were invited to submit their opinions through advertisements in public media. Furthermore, consultations were held with experts and the government officials concerned. The committees were also mandated to study reference materials and political party manifestos, as well as concept notes and proposals related to the constitution. The committees publicly appealed to the general public, political parties, civil society organisations and other institutions to submit their views, opinions and suggestions either directly to committee offices, or to their village development committees (VDCs), municipalities, district development committees (DDCs), or district administration offices (DAOs). The general public were also given the option of sending their suggestions by email, fax, or a Freephone number (Khanal 2014: 15).

22In order to support and facilitate these processes efficiently and effectively, three procedural committees – the Committee on Civic Relations (CCR), the Committee on Collection and Coordination of Public Opinions (CCCPO), and the Committee on Capacity Development and Resources Management (CCDRM) – were set up with specific responsibilities. The first two were responsible for organising interaction programmes for the general public and the last one was responsible for organising interaction programmes for Constituent Assembly members. The CCR was responsible for creating a way for citizens to have easy access to the Constituent Assembly; for disseminating information on the constitution-making process to citizens; for coordinating with other collaborators such as government agencies, NGOs, INGOs, civil society and the media; for monitoring and evaluating the activities of these collaborators with regard to the Constituent Assembly and the constitution-making process; and for informing citizens of how to participate in the constitution-making process (see Article 67(1) of Constituent Assembly Rules 2065 [2008]). The CCCPO was responsible for facilitating the collection of public opinion and suggestions about the draft constitution; for conducting public hearings; for organising interaction programmes, seminars and workshops on the draft constitution; and for preparing a report on these suggestions and submitting it to the Constituent Assembly (see Article 67(2) of Constituent Assembly Rules 2065 [2008]). The CCDRM was responsible for organizing interaction programmes for Constituent Assembly members on the constitution and constitution-making process; for conducting studies and research on the constitutional issues; for making a provision of resource materials for members and their distribution; for developing and managing a resource centre (see Article 67(3) of Constituent Assembly Rules 2065 [2008]).

23The Thematic Committees completed the outreach campaigns (for collecting opinions and suggestions) by April 2009. They then started writing concept notes and preliminary reports on their respective thematic areas. All the committees completed their preliminary drafts by January 2010. The different views of the political parties and other stakeholders, such as civil society, NGOs, human rights activists and the media, began to emerge immediately after preliminary draft reports had been made public. The Constitutional Committee deliberated on the contentious issues via dialogue and workshops. However, the actual form of federalism remained one of the most contentious issues throughout the process (Khanal 2014: 21, Breen 2018: 417). The other contentious issues identified were the form of government and electoral system, the composition of judiciary and judicial appointments; the right to self-determination, citizenship, pluralism, and special political prerogatives for indigenous people in the leadership of provincial governments (Khanal 2014: 21).

29Among its other main features, the constitution adopted a mixed electoral system for the first time in the country’s history: the House of Representatives was to consist of 275 members of whom 165 (60 per cent of the total number of MPs) were to be elected through a first-past-the-post electoral system in 165 electoral constituencies, while another 110 MPs (40 per cent) were to be elected according to a proportional electoral system where citizens vote for party lists, considering the whole country as a single electoral constituency. A provision was to be made to ensure the representation of women, marginalised groups and backward regions. The constitution was to establish the right to social justice for women, marginalised groups and backward regions in employment by state institutions on the basis of the principle of inclusion.

Elite-mass comparison of opinions on the four reforms

31This section compares the preferences of public opinion regarding the four reforms with the preferences and decisions of the political elites. The main reason for exploring the relationship between recent political events, the major parties’ decisions and public opinion is scholarly curiosity: there is an ongoing debate on the nature and direction of the relationship between political elites and public opinion, and the prime concern here is: how can the case of Nepal contribute to this debate; how can this general relationship work out in this specific case; and what does this mean for the scholarly debate. As discussed in the previous section, most studies of the relationship between public opinion and elite decisions have so far dealt with policy issues that are essential to the general public such as civil rights, social welfare, defence spending and economic issues. This article adds to this literature by comparing the views of the general public and the major political parties’ decisions on the design and fundamental reform of the political system itself.

32In the Elites Opinion Poll (conducted in July 2007), MPs were asked whether the monarchy should be upheld or abolished. The exact phrasing of the question was ‘What do you think about the monarchy; should it be upheld or abolished?’. Although the only possible answers were ‘upheld’ or ‘abolished’, some MPs (i.e. 9 per cent) insisted on answering that the people or the Constituent Assembly should take the decision. The poll showed that 84 per cent of political elites favoured abolishing the monarchy while only seven per cent wanted to retain it (International IDEA 2008: 114). It was clear in July 2007 that there was majority support among the elites for the abolition of monarchism and the transformation of Nepal into a republic. In fact, the Elites Opinion Poll of July 2007 already indicated what the elected Constitutional Assembly was to declare in May 2008, and it showed that the preferences of the political elites were different in more ways than one from those of the general public.

33An overwhelming majority of Nepali citizens still supported the monarchy in December 2004 (81 per cent) and January 2006 (74 per cent), showing only a marginal decline. This seems to contradict the plausible diagnosis that the royal palace massacre of 2001 had already curbed support for the monarchy, but the massacre may have had an indirect effect: a decline in the general public’s faith in the divine aura of the monarchy and the weakening of the monarchy’s emotional relationship with the general public, thus ‘preparing’ public opinion for real change to come. The real change became visible in the September 2006 survey (see Table 1): public support for the monarchy dropped from 74 to 53 per cent in the space of only a few months. During these months the Jan Andolan II took place. It seems likely that, once the mobilisation of a mass movement succeeded and its anticipated outcomes had been achieved, this began to affect ordinary citizens’ positions towards the monarchy. The successful cooperation between the SPA and the Maoists in the Jan Andolan II contrasted with the failure of King Gyanendra’s attempt to bring an end to the ten-year long conflict by usurping absolute power the year before.

34Near the time of the Elite Opinion Poll, in May 2007, public support for the monarchy had further declined to a minority (46 per cent), but there was still quite a large difference with the mere seven per cent of political elites who continued to support the monarchy. Nearly a year later, shortly before the Constituent Assembly would formally end the monarchy, public support for the monarchy was still just below 50 per cent; however, after the Constituent Assembly’s the declaration, the gap with the political elite eventually narrowed further. The formal abolition of the monarchy may have accelerated the erosion of support for the monarchy in public opinion. Nevertheless, four years after the Constituent Assembly’s declaration, a substantial minority of more than one third of citizens still preferred a monarchy over a republic. Here, political elites had not heeded the preferences of citizens when they decided to abolish the monarchy. Furthermore, there had been no convergence of elite and mass over time: the political elites did not compromise on their decision to radically end all forms of monarchy. Over time, I observed that public opinion gradually, yet far from completely, shifted to the position taken up almost unanimously by political elites early on.

35It was the House of Representatives, reinstated after the successful Jan Andolan II, that not only stripped the King of his executive powers in May 2006, but also proclaimed Nepal a secular state rather than a Hindu state. The series of surveys I used contained no earlier data on public opinion with regard to the religious identity of the state. International IDEA (2008: 86) cited data from 2004 showing 66 per cent public support for a Hindu state at that time. A few months after the declaration by the House of Representatives, the September 2006 survey showed that 59 per cent supported a Hindu state, which is not much lower (see Table 2). Later confirmation of the transition to a secular state in the interim constitution and in the declaration by the first Constitutional Assembly did not seem to have affected public opinion in this respect. There was remarkable continuity in public opinion with regard to the relationship between the state and Hinduism. At mass level, a clear majority favoured Nepal being a Hindu state. This proportion was stable throughout the series of surveys at national level: in the latest survey (April 2012) the proportion favouring a Hindu state was only marginally lower than it had been in 2006 (55 per cent).

36This robust and stable public opinion stood in stark contrast to the early and nearly unanimous decision by the political elites to abolish the Hindu state. In the Elite Opinion Poll of July 2007, only six per cent continued to support a Hindu state (International IDEA 2008: 128). However, it might be argued that the political elites’ stance had shifted to take into account citizens’ concerns. The 2015 constitution not only reconfirms that Nepal is now a secular state, but it adds an ‘explanation’ of the crucial term secular: religious freedom, but also protection of religion as practised since ancient times. I argue that this rather ambiguous ‘explanation’ can be interpreted as a compromise, and in that sense, the political elites seem to have acknowledged some of the concerns of the majority of the citizens.

Table 2: Comparison of support for a Hindu state and a secular state between the general public and political elites (all in %)

	May 2006	Sept 2006	Jan 2007	May 2007	Jul 2007	Jan 2008	May 2008	Aug 2010	Feb 2011	Jun 2011	Apr 2012	Sept 2015
Public opinion
Support for Hindu state	-	59	-	59	-	59	-	55	57	56	55	-
Support for secular state	-	31	-	32	-	31	-	39	34	37	37	-
Do not know	-	10	-	9	-	10	-	6	9	7	8	-
Political elites’ opinion
Support for Hindu state	-	-	-	-	6	-	-	-	-	-	-	-
Support for secular state	-	-	-	-	94	-	-	-	-	-	-	-
Important political events
	Parliament reinstated: Nepal is a secular state	-	Interim constitution: Nepal is a secular state	-	-	-	CA: Nepal is a secular state	-	-	-	-	Constitution adds ‘explanation’ to secularism

Source: International IDEA 2008 and Sen 2025

Table 3: Comparison of support for Nepali as the only official language and multilingualism between the general public and the political elites (all in %)

	Sept 2006	May 2007	Jul 2007	Dec 2007	Jan 2008	May 2008	Aug 2008	July 2009	Sept 2015
Public opinion
Support for Nepali only	55	48	-	-	41	-	55	55	-
Support for multilingualism	39	47	-	-	53	-	41	40	-
Do not know	6	5	-	-	6	-	4	5	-
Political elites’ opinion
Support for Nepali only	-	-	18	-	-	-	-	-	-
Support for multilingualism	-	-	82	-	-	-	-	-	-
Important political events
	-	-	-	Fourth amendment to interim constitution: official status of other languages	-	CA declaration to adopt multilingualism	-	-	Constitution: Nepali at national level; provinces may add other languages at their level

Source: International IDEA 2008 and Sen 2025

38In other words, the apparent gap narrowed between elite opinion and majority public opinion until January 2008, then widened again in the later surveys. The final text of the constitution, maintaining Nepali as the only administrative language at national level, but allowing multilingual policies at subnational level, can perhaps be regarded as a compromise, with political elites taking into account some of the concerns of majority public opinion.

Table 4: Comparison of support for a unitary state and a federal state between the general public and the political elites (all in %)

	Sept 2006	Jan 2007	Jul 2007	Dec 2007	Jan 2008	May 2008	Aug 2008	Jul 2009	Aug 2010	Feb 2011	Jun 2011	Apr 2012
Public opinion
Support for a unitary state	1	-	-	-	5	-	2	8	18	14	16	14
Support for a federal state	5	-	-	-	16	-	16	16	27	22	24	28
NH/NU/DK	94	-	-	-	79	-	82	76	55	64	60	58
Political elites’ opinion
Support for a unitary state	-	-	7	-	-	-	-	-	-	-	-	-
Support for a federal state	-	-	93	-	-	-	-	-	-	-	-	-
Important political events
	-	Interim constitution did not mention federalism	-	Interim constitution amended to include federalism		CA: Nepal is a federal state	-	-	-	-	-	-

Note: NH/NU/DK = Not heard/Not understood/Do not know

Source: International IDEA 2008 and Sen 2025

41Overall, there was a remarkable shift from monarchism to republicanism in public opinion, considerably reducing the gap with the political elites. On the question of the Hindu state versus a secular state, public opinion remained more or less stable, with majority support for maintaining a Hindu state. The political elites, however, did not waver in their support for a secular state; however, by inscribing the term secularism in the 2015 constitution, it can be argued that the gap with the general public narrowed somewhat. With regard to the question of unilingualism versus multilingualism, public opinion seemed to have shifted towards the position of political elites by January 2008. But when there was a reverse trend after the Constitutional Assembly’s declaration in May 2008, the subsequent article in the 2015 constitution, which maintained unilingualism at national level, could be interpreted as a concession to majority public opinion. In that sense, public opinion had (initially) reacted to elite decisions and the elites had (eventually) reacted to public opinion: a reciprocal, not simultaneous relationship. Regarding the issue of federalism, although a gradual reduction of item non-response could be observed, the issue remained the great unknown at mass level years after the proposal to federalise the Nepali state had been introduced. This shows that the decisions by political elites on this question had not reached the ears of the common masses.

42The Elite Opinion Poll 2007 also showed that there was broad consensus among Nepali political elites on the transformation of the political system that Nepal was about to undertake. Only a small proportion of the elite was in support of the old state structure. By July 2007, they had agreed, at least in terms of their individual opinions or preferences as stated in the survey, to rebuild the Nepali state as proposed by the CPN (Maoist). The overwhelming majority in favour of the new state structure acknowledged that Nepal’s political parties and elites were very unlikely to retain the old state structure. It was therefore not surprising that a year later (in May 2008) the elected Constitutional Assembly formally abolished the monarchy, the unitary form of governance, the Hindu state identity and unilingualism. The exact form that federalism should take was the only aspect that was not part of this consensus; the consensus on the other issues may help to explain why the elites did nothing to narrow the gap with public opinion. The recent political reforms formalised in the new constitution were ultimately guided by the elite or political parties’ decisions, not by public opinion.

Conclusion

43The promulgation of the new constitution of Nepal in September 2015 formally concluded one of the most radical transformations of a political system in recent history: from a monarchy to a republic, from a Hindu state to a secular state, from a unitary state to a federal state, and from a unilingual state (Nepali-only) to a multilingual state. This also marked a dramatic change in the country’s political development. Nepal has always been characterised by great cultural diversity but, since the foundation of the Gorkha Empire in 1768, Nepal’s history can be summarised as a long-standing attempt at political and cultural unification modelled on the culture of the dominant Hindu high-caste hill group. The new constitution has abandoned that project and seeks to embrace and formally acknowledges the country’s diversity.

44The citizens of Nepal were not directly or intensively involved in this major transformation. There had not been, for instance, a referendum on any of the political reforms. Formal decisions had been taken by popularly elected bodies such as the reinstated House of Representatives in 2006, the Interim Parliament in 2007, and the two Constituent Assemblies (elected in 2008 and in 2013). But the actual decisions had been taken even earlier, in negotiations between the main political parties such as the 12-point agreement between the SPA and the CPN (Maoist) in 2005.

45The strong emphasis on and dominant role of political elites in the transformation is not surprising. However, the legitimacy and consolidation of the new political system ultimately depend not only on support from the elites but also on that from ordinary citizens. If a significant part of the general public remained vehemently opposed to one or several of the reforms, there would always be the risk that this dissatisfaction would grow and be readily mobilised by an anti-system party or political entrepreneur, threatening the stability of the new system. This makes the question of public opinion on the main reform issues of great social relevance.

46The disparity between elite and public opinion that emerged from the surveys was perhaps too marked. If we look more closely at elite decisions, it might be argued that, on some of the issues, the parties moderated their positions to some extent. This was most evident regarding the abolition of the Hindu state. The elites maintained their decision to move over to a secular state but added to the constitution a definition of ‘secularism’ that included the protection of traditional religious practices. In addition, a law against proselytisation was adopted in 2017. Article 158 of the National Panel Code (2017) prohibited proselytisation to allay the fears of the Hindu majority. With regard to the language policy, the eventual implementation of multilingualism, with Nepali remaining the administrative language at national level and other languages being recognised as administrative languages in each province, can also be interpreted as a concession to public opinion. However, even such a more nuanced assessment does not deny the discrepancy between elite and public opinion, and there is no evidence that such concessions have led to greater public support.

47Leaving aside for the moment the reorganisation of the unitary state into a federal state, the transformation of Nepal’s political system resulted in one reform on which political elites and public opinion ultimately converged, and in two reforms on which elite and public preferences remained largely opposed. From a normative point of view, the situation of congruence seems preferable both with regard to the legitimacy and the stability of the political system. However, Nepal is a very diverse country, deeply divided along ethnic, religious, linguistic and cultural lines. In a society of this type, congruence between political elites and a popular majority (or even a plurality) may allow the preferences of this majority to obscure the rights of minorities. This would no doubt undermine the minorities’ allegiance to the state, with the long-term risk of creating political instability. The question of monarchy versus republic was not closely related to the diversity of the Nepali people. With regard to this issue, the risk of a ‘tyranny of the majority’ seems remote. But the religious identity of the state, the official language and the choice between unitarism and federalism are of immediate relevance to cultural diversity. That the major political parties opted for a secular state and multilingualism served to protect the interests of religious and linguistic minorities despite opposition to both issues on the part of most of the general public (also see Sen 2020; Sen 2021; Sen 2022; Sen 2023). According to the most favourable interpretation, this lends support to the view in the literature that political elites are the benevolent custodians of political tolerance and minority rights.

The analysis and interpretation presented in this article are part of my doctoral research entitled ‘Transforming Nepal’s Political Systems: Party Positions and Public Opinion (2004–2012)’ and conducted at the Political Science Department of the Faculty of Social and Behavioural Sciences, Leiden University, the Netherlands. I am indebted to the late Prof Dr Rudy Andeweg and Prof Dr Joop van Holsteyn for their critical comments and regular guidance that helped me to develop insights into interpreting survey data. I am grateful to the Asia Foundation Nepal and the Saferworld UK for allowing me to use raw data from public opinion surveys for my study. My heartfelt thanks go to Mr Caspar ten Dam, a Netherlands-based political scientist and conflict analyst, for always encouraging me to publish this article. I am much obliged to the two anonymous peer reviewers for their valuable comments and suggestions that allowed me to improve on this article.