The value of sub-national data: The dynamics of contentious politics in Nepal

Abstract

Over the last decade, a renewed interest in the empirical analysis of contentious politics has led to significant improvements in the quality and quantity of data. The related wave of research has thus turned to analysis of the dynamics of contentious politics from a comparative perspective. Unfortunately, these studies use country level datasets for their analysis, creating a mismatch between the original actor-based theoretical framework and the empirical conceptualization of contentious collective action. We discuss the conceptual challenges this theoretical-empirical gap generates and highlight how contentious dynamics can be studied with a bottom-up sub-national approach. To do so, we present a new fine-grained dataset of contentious collective action in post-conflict Nepal (2007 to 2010). We descriptively show how this type of data can help us improve our understanding of repertoires of contentious politics. This paper highlights the relevance of gathering information on active organizations to study the inherent dynamics of contentious politics.

Keywords

Contentious politics actors goals tactics dynamics

Introduction

In recent years, there has been renewed interest in the empirical analysis of contentious politics based on the finding that nonviolent actions can be effective at generating significant political change (Chenoweth and Stephan, 2011). This research agenda aims to explain why and when activist groups choose between violent and nonviolent forms of dissent. Parallel political developments have nourished this research. Protest movements in autocratic regimes during the Arab Spring and recently in Hong Kong or the spike in right-wing terrorism in Western democracies demonstrate the political relevance of better understanding the tactical repertoire of activist groups.

Research on social movements and contentious politics now include diverse forms of organized resistance with varied political goals and tactics (e.g. Cunningham et al., 2017; Muñoz and Anduiza, 2019). Combining information on contentious campaigns and events, recent scholarship has turned to analyze previously neglected dimensions in the study of contentious collective actions (CCA), including groups’ characteristics, composition, and tactical preferences—whether violent or not (Horowitz et al., 2018; Cunningham et al., 2017).

Despite these advancements and the growing interest in better understanding contentious dynamics, there is a mismatch between the actor-oriented theoretical frameworks and the prevalence of rather aggregated empirical analyses. A survey of six representative political science journals between 2012 and 2017 shows that 71% of all the CCA quantitative publications use information at the country or year levels.¹ These empirical studies are useful to compare general patterns across countries, however newer research has turned to the same data to investigate the dynamics of CCA. Unfortunately this path risks overstretching theoretical concepts, as aggregated information on CCA is inadequate to investigate a rapidly evolving actor-center phenomenon.

We contribute to this discussion and provide ways forward in the field in two distinctive ways. First, we discuss the challenges cross-country studies face regarding the conceptualization of CCA. We focus on comparing three conceptual dimensions, namely actors, goals and tactics, and demonstrate that existing data efforts and empirical designs tend to focus on one dimension at the expense of the other two. This has generated a gap between concepts and empirical design that hampers our understanding of CCA dynamics. As measurement becomes increasingly distant from theory, we cannot use cross-country data to examine, for instance, how ideology affects tactics or how competition with other social groups shapes goals over time.

Second, we demonstrate how sub-national data can help us overcome this gap. With a descriptive overview of new data on CCA in post-conflict Nepal (2007–2010), we study all three CCA dimensions and show the strength of sub-national datasets in investigating contentious dynamics. In doing this, we also explain limitations related to sub-national data that researchers should be aware of, including reporting problems and issues associated with external validity. We conclude with a discussion on the utility of sub-national data to closely match theoretical concepts and empirical reality in CCA. This level of data collection offers a fertile ground to improve our understanding on mobilization processes, changes in political actors’ tactical repertoires and interactive responses between activist groups, civil society and state institutions.

Empirical research on contentious collective action

The dynamics of CCA are related to the social and political context where activist groups operate. This contextual nature requires us to take into account a plurality of goals and tactics that conform with the reality activist groups experience on the ground. However, it is precisely this characteristic of CCA that has generated a tension between theoretical concepts and their empirical measurement. Traditionally, we have collected information across countries to look at general patterns of actions or to compare the conditions under which instances CCA take place. These empirical efforts have sidestepped activities of individual activist groups, who are at the centre of the theoretical frameworks to explain contentious politics (Tilly and Tarrow, 2015). We contend that the collection of sub-national data can alleviate this theoretical-empirical tension. By focusing on a specific country, researchers can collect temporally fine-grained information on single activist groups to examine the dynamics of contentious politics (Adcock and Collier, 2001: 530).

To better understand the ongoing mismatch between the theories and the empirical conceptualization of CCA, we turn to the most used definition of contentious politics, originally coined by Tilly and Tarrow (2015: 7): “[the] interactions in which actors make claims bearing on other actors’ interests, leading to coordinated efforts on behalf of shared interests or programs, in which governments are involved as targets, initiators of claims, or third parties.”

This definition highlights the relevance of groups that collectively act in relation to (broadly defined) political goals. By becoming active, groups use a particular tactic from an available set of contentious options (Tilly and Tarrow, 2015: 15). For this reason, empirical data efforts to study CCA must include three constituent dimensions: (a) unitary actors; (b) political goals, and (c) tactics.

In the following, we introduce these theoretical concepts and explain how six prominent datasets in the field have tried to empirically capture them. Figure 1 provides an overview of the coverage of these data sources, differentiating between their geographical or issue outreach (global, regional, and single issues) across dimensions.² We show that these data collection efforts tend to focus on one CCA dimension at the expense of the other two, generating a gap between concepts and empirical design. An ideal CCA dataset based on the definition by Tilly and Tarrow (2015) should have all boxes in Figure 1 ticked.

Figure 1.

Overview of contentious collective action conceptualization across country level datasets.

Actor dimension

Who conducts the contentious activities? Actors³ comprise groups of individuals or organizations who carry out dissent activities. For instance, actors are civil society groups, political parties or state agents. Explanations for contentious dynamics are therefore interested in explaining how these actors behave. To do so, this agent-centred understanding assumes that these actors are unitary (Cunningham et al., 2017; Horowitz et al., 2018).

Figure 1 shows that the coverage of actor types is limited and most often does not identify unitary groups, particularly in global and regional datasets. This shortcoming prevents us from testing theories with a focus on agency. For instance, datasets with campaign level information do not allow us to analyze who conducts the activities and why they do so (e.g. the Nonviolent and Violent Campaigns and Outcomes (NAVCO) data project). The agency problem is not so relevant for the cross-country datasets with a single issue focus such as the Strategies of Resistance (SoR) (Cunningham et al., 2017) and the Minority at Risks Organizational Behavior (MAROB) datasets. Both datasets include information about ethnic organizations and their activities which allow us to take into account the agency dimension of CCA concepts.

Goal dimension

What goals should be included? Goals are objectives that a given unitary actor wants to achieve. CCA includes activities with broadly defined political goals and points towards the involvement of the government or specific policies (Tilly and Tarrow, 2015). This distinction allows us to register actors with multiple goals or those that change goals over time. Efforts to systematically collect cross-country data on goals are constrained to a reductionist operationalization to guarantee comparability across cases. For example, NAVCO looks only at campaigns with maximalist goals—antiregime, antioccupation, and secession campaigns—to investigate the effectiveness of nonviolence compared with violence (Chenoweth and Stephan 2011: 13). Most of the other datasets share this maximalist focus, or a subset of such goals. As Figure 1 shows, exceptions are the Mass Mobilization Protest Data (MMPD), Social Conflict Analysis Data (SCAD) and MAROB, which include a broader socio-political spectrum of goals.

This strategy is reasonable as there is a limit to the amount of information that it is possible to gather in a cross-country design, but has consequences for the theoretical-empirical gap. First, this form of data collection narrows the possible theoretical assumptions that can be used. For example, opportunities and motives to engage in a campaign with maximalist goals are different to those for activities with other goals, such as higher wages. Holding everything else constant, activists will need to have stronger motives and favourable opportunity structures to mobilize people. Second, the selection criteria affect what we can study empirically, as these data collection efforts tend to assume that goals are constant over time. In many cases, groups or campaigns are even defined by their goals (e.g. SoR). Cross-country datasets therefore lack the required broad range of political goals to investigate dynamics related to the pursuit of political or policy change. We cannot then answer questions such as how do larger protests emerge or whether contentious actions start with small, achievable goals but then extend the objectives as the movement grows stronger (or weaker for that matter).

Tactic dimension

What do the activists do? Cross-country datasets have different strategies to address the challenge of collecting information about tactics. First, various datasets focus on a high level of aggregation and provide general information on the types of tactics. The predominant strategy is to distinguish between violent and nonviolent forms of mobilization (e.g. NAVCO). Despite this advancement, identifying only the most frequent or salient tactic limits our ability to make a fine distinction between different levels of violence or the strategic use of supplementing different tactics across the violent-nonviolent spectrum.

Second, event-based datasets generally focus on protest events and classify those into demonstrations, riots, or strikes depending on them turning violent or not (MMPD, SCAD, Armed Conflict Location & Event data project (ACLED)). The focus on one tactic conceived as a political phenomena in itself, however, limits the study of tactical decision-making or the relevance of substitution and complementary dynamics as information on other tactics is missing. Finally, issue specific datasets (SoR and MAROB) include tactical diversity, but only gather yearly information on ethnic groups. This level of temporal aggregation constraints the study of consecutive tactical choices within a year. Most problematic remains the exclusive focus on ethnic actors, not knowing whether tactical choices are exceptional compared to other non-ethnic actors—something CCA theoretical approaches highlight as relevant (Tilly and Tarrow, 2015; Soule and King, 2008). Whether actors simultaneously engage in multiple tactics or dynamically change them—either as a result of learning or competing with other groups, adapting to different resource availability, or government responses—remain then as open questions.

Cross-national data on sub-national conflict

In recent years, new data collection projects have tried to overcome some of these limitations by offering cross-national data on sub-national conflict. We can distinctively categorize such effort into two groups. On the one hand, there are machine coding efforts such as NAVCO 3.0 that automatically extract event data from news reports published by international news agencies (Lewis et al., 2016). This data effort comprehensively uses the Conflict and Mediation Event Observations (CAMEO) ontology to isolate relevant conflict events (Gerner et al., 2008). In relation to CCA, however, the reliance on news wire reports from Agence France Presse and pre-defined dictionaries (Lewis et al., 2016: 1, 34) prevent detailed sub-national coverage that contextually identifies relevant actors, goals and tactics (cf. Nam, 2006).

On the other hand, xSub offers a repository on sub-national event data on violence (Zhukov et al., 2019). This astute effort relies on existing datasets, harmonizing spatial and temporal measures to facilitate its access for research, educational and policy-making purposes. The reliance on existing datasets, however, imposes some limitations on this endeavor, including an uneven temporal coverage of countries, reliance on the data generation process that guided each original dataset (e.g., threshold for event inclusion), or varying numbers of unique data sources per country. When it comes to the study of contentious politics more broadly, xSub is better suited to investigate violent events than other forms of political contestation (Zhukov et al., 2019).⁴

Overall, these new and promising projects on cross-national data on sub-national conflict represent a significant step forward in research on contentious politics. These advancements, however, still suffer constraints on data coverage and quality, affecting the reliability and validity of measures on actors, goals and tactics. Instead, sub-national data efforts on representative countries offer more adequate coverage of these CCA concepts, narrowing the distance between the original actor-based theoretical framework and the empirical conceptualization of CCA (Gerring, 2006).

The value of sub-national data: Introducing the Nepal Political Contentious Event Dataset (NPCED)

Existing data efforts and empirical designs tend to focus on one dimension of CCA at the expense of the other two. This has generated a gap between concepts and empirical design that hinders our understanding of CCA dynamics. Sub-national data collection can overcome the measurement limitations as it collects reliable information on sub-units within a given country. The focus on sub-units offers the opportunity to provide insightful information on actors, goals and tactics including their spatial and temporal variation. Because the collection of sub-national data is resource intense, such data is generally collected on a single case or selected number of cases.

Nepal is a good case to demonstrate the added value of such an approach to analyzing contentious dynamics. After emerging from a ten year armed conflict at the end of 2006, Nepal exemplifies the difficulty of post-conflict periods which are often marked by contentious politics. Postwar societies face adverse economic, political, and social conditions that affect political participation and attitudes that reignite the risk of violent conflict (e.g. Collier et al., 2008). At the core of the disputes are socio-political power struggles in response to contextual changes inherited from the war. In this regard, Nepal was no exception. The country experienced a significant number of contentious events with geographical differences in terms of economic centers, population density, settlement areas of ethnic groups and war zones (Holtermann, 2016; Lawoti and Hangen, 2013). Moreover, Nepal dealt with a transition phase to a new constitution despite having low state capacity following the civil war (Vüllers and Schwarz, 2019; Upreti et al., 2013). These features are also representative of diverse post-conflict countries such as Afghanistan, El Salvdor and Ivory Coast.

Nepal is therefore an important case for collecting and analyzing sub-national data on contentious events. By doing so, we can shed light into various research topics including how peace processes affect the contentious dynamics or how different social identities influence tactical decisions in the immediate post-conflict phase. To date, a detailed dataset covering post-conflict nonviolent and violent CCA events in Nepal is missing.⁵ In what follows, we overcome this limitation by introducing NPCED, a sub-national data collection effort to capture the CCA dynamics in Nepal from 2007 to 2010.

Bottom-up data collection process

The NPCED originates from printed news outlets, collecting raw and detailed descriptions of local contentious politics. The coding is based on the Himalayan Times and Rising Nepal, two of the leading private Nepalese daily newspapers. The original sources are daily digitalized newspapers provided by the Nepal Press Council.⁶ The two national newspapers provide a good overview of contentious politics in Nepal.

The main goal of the sub-national data collection was to identify all types of CCA event. Therefore we have not used any key terms to identify protests or clashes. Instead, we looked through each individual article to see if a CCA event was reported or not. This approach makes it possible to recognize specific country conditions that surround the development of contentious politics. Compared to cross-country data efforts, for which keywords and restrictions are required for feasibility and comparability purposes, our sub-national data collection efforts help us better bridge the gap between the original actor-based theoretical framework and the empirical conceptualization of CCA.

Figure 2 provides a graphic overview of the steps in the data collection process. Before the process started, coders were trained to identify CCA events using multiple random samples of articles. In addition, we established a system of reliability checks during the data collection process in which the principal investigators regularly verified the coded information. In the following, we describe the data collection process focusing on the three constituent dimensions of CCA: actors, goals and tactics.

Figure 2.

Bottom-up data collection approach.

In Table 1 we use one example of newspapers reports to illustrate how the coding process worked. The first step is the collection of the raw information from the newspaper reports in very general categories (e.g. actor, event date, tactic). In the second step, we organize the raw information and classify it into the three CCA dimensions. This stage produces a highly disaggregated dataset that still requires further elaboration to clearly define it relative to the context of Nepal. The third step in the bottom-up data collection process focuses on additional research based on secondary literature, official documents (e.g. electoral commission) and intensive web searches for individual organizations (e.g. to identify local organizations). These three steps constitute the building blocks of the sub-national data collection effort.

Table 1.

Coding process example.

	Actors	Goals	Tactics
Step 1	NHWU; UCPN-M	health sector reform	bandh; shut-down of hospital; disruption of traffic; demonstration
Step 2	NHWU; UCPN-M	health sector reform	bandh, shut-down of health institutions; disruption of traffic; protest
Step 3	NHWU is identified as a union based on web search; UCPN-M is a political party (electoral commission)	health sector reform	bandh is a particular type of protest in Nepal resulting in a shut-down of all economic and public institutions in a given place (secondary literature); shut-down of health institutions; disruption of traffic; protest
Step 4	economic interest organization (NHWU); political party (UCPN-M)	social goal (health sector reform)	economic non-cooperation (bandh, shut-down of health institutions); nonviolent intervention (disruption of traffic); protest (protest)

NHWU: Nepal Health Workers Union; UCPN-M: Unified Communist Party of Nepal - Maoist.

The fourth and final step is the aggregation of the cleaned information to measure specific theoretical concepts. This final step links our data to the theoretical background concepts in the contentious politics literature. The researcher can decide how to aggregate the data based on the theoretical concepts of interest. This enhances the comparability of the NPCED with other studies on contentious politics. For instance, by aggregating the actors, goals or tactics from a very specific level (e.g. bandhs) to a more general concept (e.g. economic non-cooperation) the results of the NPCED can be compared to other research (e.g. SoR data). Figure 3 illustrates how the disaggregated tactics are aggregated to the concepts later used in this research note.

Figure 3.

Examples of aggregating tactics in NPCED.

Overall, this sub-national data collection approach offers the possibility of linking the data with different established concepts and cases discussed in the literature. It provides the possibility of making use of very disaggregated information to test particular theoretical concepts or of using more aggregated information to compare results from cross-sectional datasets. The NPCED offers researchers the opportunity to adapt the data to their specific research interests. Based on the disaggregated and aggregated information on actors, goals and tactics, researchers can create their own categories to investigate, for example, the use of tactics by a subset of actors or to identify a cluster belonging to the Maoist movement and examine whether they differ in their tactical choices.

Before we turn to show how the NPCED can help us better understand the dynamics of CCA in post-conflict Nepal, we want to discuss some potential limitations. Despite the efforts and resources invested in the collection of this comprehensive sub-national dataset, the NPCED is not immune to reporting biases similar to those in cross-sectional data efforts (e.g. Herkenrath and Knoll, 2011; Earl et al., 2004). We detail the two most prominent sources of bias and the steps taken to alleviate these concerns.

First, newspapers may focus on developments in the economic and political centers, for example in the Kathmandu valley. Such geographic targeting would systematically underestimate CCA events in the rest of the country. To assess this type of bias, we conducted personal interviews with former editors and journalists from various Nepalese newspapers, who emphasized that these national newspapers do report on events in all regions of the country (April to May 2014). Still, when using the data, geographical differences should be controlled for in quantitative models, for example by including fixed effects at the district or regional level.

Second, newspapers tend to report events with a high news value—such as violent events and events with a high number of supporters (Herkenrath and Knoll, 2011:166). Given that the NPCED originates from national news sources, it may not be exempt from this source of bias. However, national newspapers are more likely to report low turnout and nonviolent events than international news agencies because readers are particularly interested in such national events (Herkenrath and Knoll, 2011; Vüllers and Schwarz, 2019). The NPCED thus offers a comparative advantage relative to other data collection efforts which are only based on international media. Moreover, the NPCED provides an opportunity to control some of these reporting biases by using subsamples of regions (e.g. districts with more than one municipality) or events (e.g. protests, bandhs) that should be newsworthy enough for the newspapers to include reports of these events (Weidmann, 2016).

Insights from the NPCED: Contentious dynamics

In this section, we turn to the NPCED to demonstrate the advantages of sub-national data efforts in advancing our understanding of contentious politics. We focus on showing the wealth of the NPCED by descriptively examining variation across actors, goals and tactics. The uniqueness of this dataset allows us to go one step further relative to cross-country data efforts. The disagregated nature of sub-national events facilitates the study of goals and tactical portfolios across actors. In addition, we can examine how actors change these two dimensions over time, hinting at learning or adaptation strategies to (potentially) respond to their immediate environments. In the following, we offer descriptive insights into these two additional extensions of CCA. We start by identifying actors, goals and tactics in the NPCED.

Actor dimension

The open data collection approach allows the inclusion of all groups who have at least one CCA event. There is a total of 1545 unique actors who were active in Nepal between 2007 and 2010. We follow the systematized concepts in the literature to distinguish between unitary actors according to their characteristics. NPCED thus identifies five mutually exclusive actor types: political parties, economic interest groups, ethnic organizations, local organizations, and societal groups (e.g. villagers, teachers, women).

In contrast to the discussed datasets at the country level, this conceptualization and empirical identification of actors has several benefits. Using this sub-national data, we are able to observe a more complete picture of the actors active in CCA. This information also distinguishes between multiple actors involved in the same event from actors who singly handle dissident activities. Further, this differentiation between actors allows us to investigate why they became active or used a particular tactic.

Goal dimension

For each event registered in the NPCED we included a description of the goals. To enhance comparability with existent studies, we distinguish in this paper between political goals (e.g. claims regarding political structure, peace process), social goals (e.g. claims related to education or the health system) and economic goals (e.g. claims about salaries or prices regulated by the state).

Figure 4 (left graph) maps the distribution of goals different actors have in Nepal. Political parties, ethnic groups, and local organization tend to predominantly stand for political goals. Economic interest groups have a tendency to economic goals, as they operate in predominantly economic sectors. Societal groups share this economic inclination; teachers or women groups are likely to mobilize for the improvement of local living conditions. To a lesser extent, economic interest groups, local organizations and societal groups also seem to have a drive to social goals. This visualization already demonstrates that goals do not determine what an actor is. Different to country level datasets that define actors based on their goals, we see that in Nepal actors tend to address different objectives beyond their most relatable concern.

Figure 4.

Actors’ goals and tactics.

Tactic dimension

NPCED includes information about the tactics used in a given event. We were able to identify tactics which would have been neglected with broader search terms, for example the waving of black flags as sign of resistance or the use of bandhs.⁷ Neglecting these case specific forms of contentious activity would have called into question theoretical models on CCA as the action and tactical choices of various actors in Nepal would have been inaccurately recorded.

Following the existing literature and to enhance comparability, we group tactics into protests (e.g. demonstrations, protests), economic non-cooperation (e.g. strikes, halt of work, mass resignation, bandhs), nonviolent intervention (e.g. sit-ins, occupations) and violent actions (e.g. clashes between groups, vandalism, bombing). With this distinction we demonstrate how valuable it is to capture the country’s peculiarities regarding CCA while ensuring comparison with other cases.

Figure 4 (right graph) shows the distribution of tactics. In general, Nepalese groups tend to avoid violent tactics. Only political parties seem to be more prone to violence, representing over 25% of their events. This finding is expected given that Nepal ended a ten year long armed conflict in November 2006 and all conflict actors remained active as political parties (Ishiyama and Marshall, 2015). With the exception of ethnic organizations, other groups seem to have a slight preference for nonviolent forms of intervention. This difference is consistent with the singularity of the case. Ethnic groups systematically use bandhs as a form of economic disruption. This tactic allows ethnic groups to publicly display their level of popular support and it effectively forces less inclined citizens to side with these groups—see on bandhs in Nepal Lakier (2014); on the theoretical argument Tilly (2003).

The sub-national data allows us to go one step further and study whether groups choose a specific tactical portfolio given their goal in Figure 5. We find that actors tend to differ on their tactical choices based on the goal they pursue. For groups with a political aim (left graph) there is variation on their tactical portfolios. Economic interest and societal groups show a preference for protests and economic non-cooperation. This is different for ethnic and local organizations who share their frequent choice of nonviolent forms of intervention. In contrast, groups with social (middle) or economic (right) goals seem to share a preference for a nonviolent portfolio of tactics.

Figure 5.

Actors’ tactical portfolio by goals.

Actors and the dynamics of goals

To understand contentious politics we need to study the changing dynamics actors experience in their struggle to alleviate their grievances. With the sub-national data on Nepal, we can trace organizations that are active beyond a single instance and evaluate whether they persist in their main goal overtime or they divert from it.⁸ Table 2, column “Goal ∆”, shows that activist groups tend to be consistent with their objectives, and rarely shift to other goals at subsequent events.

Table 2.

Percentage of actors that change their goals and tactics.

Group type	Goal ∆	Tactic ∆
Political parties	20.0%	40.0%
Economic interest groups	25.5%	35.5%
Ethnic groups	12.5%	30.0%
Local organizations	19.5%	39.0%
Societal groups	35.0%	49.5 %

Note: The delta symbol stands for change.

For the fewer organizations that do change goals there is a very revealing pattern, displayed in Figure 6. Organizations with prior claims on societal or economic conditions (middle and right graphs) tend to later promote their goal as political. As they advance their grievances away from sector-specific issues, these groups seem to consolidate their agenda towards maximalist goals. In contrast, groups with a formerly political agenda (left) re-define their objectives as economic or social—the reverse trend. This change towards specific claims refines the groups’ political grievances perhaps in pursue of concrete government concessions or to mobilize support from aggrieved sectors of the population. Figure 6 also shows that actors’ substitution of goals is predominately to a single issue objective, instead of holding multiple claims in followup events.

Figure 6.

Substitution of goals by actor groups.

Actors and the dynamics of tactics

While actors seem to be rather stable with their goals, their choice of tactics fluctuates at a higher frequency, as shown in Table 2. Focusing on groups that do change their tactics across events, we find that these groups follow a substitution logic: 87% of these groups substitute former tactics for another in subsequent events, while a fewer 13% opt for adding new tactics to their existing portfolio. We explore this substitution logic in Figure 7, where we see which tactic groups tends to switch to based on their previous contentious portfolio. For groups who previously used protests (top left) or violence (bottom right), nonviolent intervention becomes the natural substitution in subsequent events.

Figure 7.

Substitution of tactics by actor groups.

There is no clear dominant substitution tactic for groups who previously used economic non-cooperation (top right) and nonviolent intervention (bottom left). For the latter set of groups, they either turn to protests or to violent tactics. This bi-directional change seems to reflect groups’ incentives to (de-)escalate tensions previous tactics may have generated in the immediate political environment. When it comes to groups with an economic non-cooperation past (top right graph) violence becomes relatively salient, though not across all group types. In here, local organizations seem to behave differently by completely discarding protest as a substitution tactic. Instead, local organization sustain their political engagement either with violent or nonviolent forms of contestation.

Conclusion

How can we analyze contentious dynamics? Comparative research has greatly contributed to our understanding of patterns of contentious politics across cases. However, most studies use country level datasets. This is problematic because efforts focus on gathering information on only one of the three CCA dimensions (actors, goals, tactics). This limitation has generated a gap between concepts and empirical design that has prevented us from better understanding contentious dynamics.

The sub-national approach overcomes such limitations by comprehensively including all three CCA dimensions. The descriptive evidence from a new dataset on CCA in Nepal shows three main takeaway points. First, using sub-national data we capture organizations which are active at the local level and manifest dissent with a low number of participants. Second, while active groups are generally persistent in their goals over time, they are prone to vary their tactical trajectories at different events. Finally, when it comes to different groups’ tactical repertoires, we observe that mobilized groups follow a substitution logic to change tactics rather than a complementary one.

We descriptively illustrated the utility of sub-national data for the study of contentious dynamics. There are multiple avenues for research to further shed light on CCA. Scholars can use sub-national data to follow the development of organizations over time (e.g. organizational structure, size, or support base) and whether coexisting with other groups enhances collaboration or competition among them. When it comes to the goals dimension, future research can systematically explore under what conditions groups decide to broaden or lessen their claims for policy change, relative to their CCA trajectory. While groups’ mobilization potential largely depends on this CCA dimension, changes may indicate their ability to learn and adapt in order to sustain their resistance.

Finally, by analyzing sub-national data such as the NPCED, future studies will also be able to examine why actors diversify their tactical portfolio beyond their initial choice. In Nepal, activist groups show a preference for substituting rather than complementing tactics, with remarkable differences across actor types. Pursuing this line of research, the field of CCA will be able to account for the selection process guiding strategic choices and eventually why and when actors are successful in achieving large scale mobilization.

Supplemental Material

Appendix – Supplemental material for The value of sub-national data: The dynamics of contentious politics in Nepal

Supplemental material, Appendix for The value of sub-national data: The dynamics of contentious politics in Nepal by Belén González and Johannes Vüllers in International Area Studies Review

Footnotes

Acknowledgements

We would like to thank participants at a workshop in Trondheim for their feedback, in particular Charles Butcher. We thank Anselm Fliethmann for his assistance during this research. We also thank the two anonymous reviewers and the editors for insightful comments and suggestions. Johannes Vüllers is solely responsible for the Nepal Political Contentious Event Dataset (NPCED).

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Johannes Vüllers

Supplemental material

Supplemental material for this article is available online.

Notes

References

Adcock

Collier

(2001) Measurement validity: A shared standard for qualitative and quantitative research. American Political Science Review 95(3): 529–546.

Chenoweth

Stephan

(2011) Why Civil Resistance Works: The Strategic Logic of Nonviolent Conflict. Columbia: Columbia University Press.

Collier

Hoeffler

Söderbom

(2008)Post-conflict risks. Journal of Peace Research 45(4): 461–478.

Cunningham

Dahl

Frugé

(2017) Strategies of resistance: Diversification and diffusion. American Journal of Political Science 61(3): 591–605.

Earl

Andrew

McCarthy

et al. (2004). The use of newspaper data in the study of collective action. Annual Review of Sociology: 65–80.

Gerner

Abu-Jabr

Schrodt

et al. (2008). Conflict and mediation event observations (CAMEO): A new event data framework for the analysis of foreign policy interactions. In: International Studies Association, New Orleans. Paper presented at the International Studies Association.

Gerring

(2006) Case Study Research: Principles and Practices. Cambridge: Cambridge University Press.

Herkenrath

Knoll

(2011) Protest events in international press coverage: An empirical critique of cross-national conflict databases. International Journal of Comparative Sociology 52(3): 163–180.

Holtermann

(2016) Relative Capacity and the Spread of Rebellion: Insights from Nepal. Journal of Conflict Resolution 60(3): 501–529.

10.

Horowitz

Perkoski

Potter

PBK

(2018) Tactical diversity in militant violence. International Organization 72(1): 139–171.

11.

Ishiyama

Marshall

(2015) Candidate recruitment and former rebel parties. Party Politics 21(4): 591–602.

12.

Johari

(1982) Comparative Politics. New Delhi: Sterling Publishers Pvt. Ltd.

13.

Lakier

(2014) The spectacle of power: Coercive protest and the problem of democracy in Nepal. PhD thesis, University of Chicago, USA.

14.

Lawoti

Hangen

(eds) (2013) Nationalism and Ethnic Conflict in Nepal: Identities and Mobilization after 1990. Abingdon: Routledge.

15.

Lewis

Chenoweth

Pinckney

(2016) Nonviolent and Violent Campaigns and Outcomes 3.0: Effects of Tactical Choices on Strategic Outcomes Codebook. University of Denver. Available at: https://www.du.edu/korbel/sie/research/chenow_navco_data.html

16.

Muñoz

Anduiza

(2019) ‘If a fight starts, watch the crowd’: The effect of violence on popular support for social movements. Journal of Peace Research

17.

Nam

(2006). What you use matters: Coding protest data. PS: Political Science & Politics 39(2): 281–287.

18.

Soule

King

(2008) Competition and resource partitioning in three social movement industries. American Journal of Sociology 113(6): 1568–1610.

19.

Tilly

(2003) The Politics of Collective Violence. Cambridge: Cambridge University Press.

20.

Tilly

Tarrow

(2015) Contentious Politics. Oxford: Oxford University Press.

21.

Upreti

Poudel

Ghimire

(2013) Ignored or Ill-represented? The Grievance of Terai-Madhes Conflict in Nepal. New Delhi: Adroit Publisher.

22.

Vüllers

Schwarz

(2019) The power of words: State reactions to protest announcements. Comparative Political Studies 52(3): 347–381.

23.

Weidmann

(2016) A closer look at reporting bias in conflict event data. American Journal of Political Science 60(1): 206–218.

24.

Zhukov

Davenport

Kostyuk

(2019) Introducing xSub: A New Portal for Cross-National Data on Subnational Violence. xSub Portal. Journal of Peace Research 56(4): 604–614. Available at: http://cross-sub.org/

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.16 MB