Since the milestone victory of the Bharatiya Janata Party (BJP) in the 2019 general elections, political observers have been closely watching the Muslim minority’s response in the electoral arena. Available data suggest that in subsequent state elections, Muslims have largely voted against the BJP and in favor of the leading challenger party. There are several caveats, however. Alongside this consolidation of the Muslim vote, salient social and political differences within the Muslim population have simultaneously surfaced. For example, Muslim caste hierarchies and other sub-identities such as sect have become important topics of mainstream political conversations. This fact raises pressing questions for electoral dynamics as the 2024 general election approaches. Should the idea of a coherent Muslim voting bloc be retired? To what extent is caste relevant to the understanding of Muslim voting behavior? Do group-based solidarities around religion and caste exist for Muslims?
A Monolith or a Mosaic?
In popular election analyses, Muslims are often treated as a homogeneous voting bloc. Yet, as Hilal Ahmed’s research points out, Muslims hardly make up a monolithic community that mechanically acts in a coordinated fashion; just like Hindus, Muslims are divided on class, sect, caste, and regional lines—complicating claims that they behave like a uniform vote bank. Prior to 2019, Muslim voting behavior was significantly fragmented, and there were no clear signs of constituency-level coordination behind a single political party. Notably, however, voting behavior has become more complex since the BJP’s 2019 victory.
Political scientists typically use a measure known as the Herfindahl-Hirschman Index (HHI) to examine the level of fractionalization or consolidation that exists in the distribution of votes. An index score close to zero implies that all Muslims voted for different parties, while a score of 1 indicates their support for the same party. An application of the HHI for the 2019 general elections reveals some fractionalization of Muslim votes, despite rising Hindu majoritarianism. In addition, there is considerable variation across states (see figure 1). This fractionalization is marked when looking at Muslim votes both for individual parties and for alliances.
By comparison, studies of Muslim voting in state assembly elections since the 2019 general elections and the ushering in of a new dominant party system distinctly show some consolidation. Surveys by the Lokniti Programme for Comparative Democracy at the Centre for the Study of Developing Societies show that in the 2020 Bihar elections, about 77 percent of Muslims voted for the Mahagathbandhan (an alliance of parties opposed to the BJP); in the 2021 West Bengal elections, 75 percent of Muslims voted for the incumbent Trinamool Congress; and in the 2022 Uttar Pradesh elections, 79 percent of Muslims voted for the opposition Samajwadi Party.
A range of factors appear to shape Muslim voting in state and national elections. Some scholars suggest that Muslims vote according to a state’s electoral context (based on how many parties are contesting and who these parties are) or individual candidate prospects (with Muslims’ votes divided in multipolar contests). My own research, which focuses on state assembly elections, argues that Muslim voting behavior is influenced by the presence (or absence) of a coethnic representative. I show that the existence of a Muslim representative can activate internal divisions among Muslims that shape their subsequent voting behavior. These divisions can manifest along sub-identities within the broader Muslim identity.
Caste Within Religion
Despite scholarly inattention, caste remains a relevant sub-identity in the Muslim community. While observers largely agree that caste among Muslims may not have a religious or ideological underpinning, it manifests socially and occupationally in a hierarchical form similar to caste in Hindu communities. Muslim caste identity is divided into three categories: Ashraf, Ajlaf, and Arzal. The Ashraf are self-proclaimed descendants of Muslim immigrants who emigrated to the Indian subcontinent from the Middle East and Central Asia; the group includes Sayyids, Shaikhs, Mughals, and Pathans, who all have different hierarchical positions within the category. The Ajlaf and Arzal are primarily Hindus who converted to Islam and correspond to Backward Class (OBC) and Dalit subcastes, respectively. Dalit Muslims, however, do not benefit from affirmative action protections that exist for Scheduled Castes under the Constitution of India. Since the 1990s, activists among non-Ashraf Muslims—Ajlaf (OBC) and Arzal (Dalit) Muslims—have sought to mobilize against Ashraf dominance in politics and have begun to refer to themselves collectively as Pasmanda (a Persian term meaning “those left behind”). Today, “Pasmanda Muslims” is used as an umbrella term for OBC and Dalit Muslims.
Research reveals that some Muslims still experience practices of untouchability that are also observed in Hindu communities. For example, during a survey on caste among Muslims in Uttar Pradesh, some Dalits reported regularly encountering instances of untouchability, such as not being permitted to bury their dead in the same graveyard as upper-caste neighbors or their children being seated separately during meals at school. They also reported facing untouchability in interactions with both upper-caste Muslims and Hindus. Moreover, the survey responses of non-Dalit Muslims indicated that they engage in practices of untouchability to various extents; for example, 21 percent of non-Dalit Muslims reported that they do not visit the homes of Dalit Muslims.
The 2006 Sachar Committee Report published by India’s Ministry of Minority Affairs is often cited to highlight the educational and economic differences between Muslims and Hindus, but it also examines differences within the Muslim population, specifically along caste lines.1 Grouping non-upper-caste Muslims into the OBC Muslim category, the report compares OBC Muslims to General Muslims and OBC Hindus. In the domains of education, employment, and economic status, Muslim OBCs fared poorly relative to both General Muslims and Hindu OBCs. (In the report’s groupings, “General Muslims” are those who fall in the Ashraf category, whereas OBC Muslims include the Ajlaf and Arzal, who constitute the Pasmanda category.)
Politicization of Muslim Caste Identity
In addition to being relevant to the socioeconomic lives of Muslims, caste is increasingly becoming part of conversations on politics and group relations. This is evidenced by, for instance, the sharp rise in Google searches for the term “Pasmanda” in India over the past four years (see figure 2). This rise is most prominent in North India, which has also witnessed the most public mobilization of Pasmanda identity.
In terms of civic life, there is growing evidence of Muslim caste-based mobilization. For instance, one mobilizing force at the center of the Pasmanda movement is the All-India Pasmanda Muslim Mahaz (AIPMM) organization. The AIPMM, based in Bihar and led by former Rajya Sabha parliamentary member Ali Anwar, has been at the forefront of civic efforts to advocate for Dalit Muslims (as well as Dalit Christians) to have Scheduled Caste status, which allows for affirmative action in government jobs, education, and politics and legal protection against atrocities.
Over the past several years, political elites have also been expanding their efforts to mobilize Muslims along caste lines. While most of this mobilization is based on the umbrella “Pasmanda” grouping, there is some evidence of mobilization of the Ansari subcaste in Bihar. In the lead-up to the 2020 Bihar state elections, the Ansari Mahapanchayat (AMP) advocacy group was formed and engaged in grassroots organizing in the central and western parts of the state, arguing specifically for elected representation for Ansaris. However, eventually, the group’s leadership fractured, with top officials throwing their support behind different alliances at the state level.
Perhaps the most prominent attention given to the Pasmanda grouping surprisingly came from the BJP during its 2022 national executive meeting in Telangana’s capital city Hyderabad. In fact, the aforementioned Google Trends for the term Pasmanda reached its ten-year peak in the same month as the event. Despite his party’s Hindu nationalist bona fides, Prime Minister Narendra Modi urged BJP workers to reach out to Pasmanda Muslims. Modi subsequently held several events with notable Pasmanda Muslims later in the year. While these meetings were the party’s first large-scale public outreach effort, it has engaged with Pasmanda Muslims in the past. For instance, sociologist Khalid Anis Ansari noted that in 2013 in Uttar Pradesh, the BJP formed a “weavers’ cell” to engage some Pasmanda Muslims, and in 2017 in Odisha, the party highlighted the need for Pasmanda Muslims to receive the same benefits as OBCs and even called out Sayyids and Pathans for “usurping welfare measures.”
Nevertheless, the BJP’s explicit outreach in 2022 was not entirely well received. Several Muslim groups warned members of the Pasmanda community to be wary of the party’s outreach. For instance, the AIPMM emphasized that no political party should take the support of Pasmanda Muslims for granted, while the organization’s leader, Anwar, personally criticized the party’s Pasmanda outreach by underscoring that Pasmanda Muslims have been disproportionately affected during communal attacks. Asaduddin Owaisi, the leader of the All India Majlis-e-Ittehadul Muslimeen party—which regularly evokes Muslim solidarity—also lambasted the BJP’s outreach, highlighting that Pasmanda Muslims have been victims of vigilante violence under the current BJP government.
Results of the BJP’s outreach have been mixed. The party fielded four Pasmanda Muslims in the Delhi Municipal Corporation elections in December 2022, but none of the candidates won and only one came close as a runner-up. However, the results of the May 2023 Uttar Pradesh by-elections in the city of Suar suggest that the party’s efforts may have gained some traction. After Abdullah Azam Khan, a representative of Suar on the Uttar Pradesh Legislative Assembly, was sentenced to prison, by-elections were held in the city, where Muslims constitute around 60 percent of the population. The Apna Dal (Sonelal) party, a regional partner of the BJP-led National Democratic Alliance, fielded a Pasmanda Muslim, Shafeek Ahmed Ansari, with the backing of the BJP. Ansari ultimately defeated the opposition Samajwadi Party candidate, with key BJP leaders claiming that Ansari’s win was a direct result of the party’s efforts to mobilize the Pasmanda Muslim vote. However, because it was a Muslim candidate from Apna Dal contesting against a Hindu candidate from the Samajwadi Party, it is hard to discern whether caste or religion played the greater role.
Contours of Pasmanda Political Behavior
While Pasmanda politics has figured more prominently in mainstream political mobilization in recent years, the community’s voting behavior is less clear. Understanding the heterogeneity in political behavior within the Muslim population remains difficult for two reasons. First, most surveys seek to capture a representative sample of India or a particular Indian state, which limits the sample of Muslim respondents. Second, they often do not capture details on Muslim castes. To help address these limitations, I conducted a survey of nearly 2,000 Muslims in Uttar Pradesh in 2022, with a particular focus on the contours of political behavior within the Muslim community.2 The survey was also conducted with almost 2,000 Hindus in Uttar Pradesh in 2022 to allow for comparisons across religious groups.
When asked about state elections in Uttar Pradesh in 2012, a small percentage of Muslim voters reported that they voted for the BJP candidate. By 2017, 12.6 percent of General Muslims and 8 percent of Pasmanda Muslims reported support for the BJP. Interestingly, by the 2022 state elections BJP support among General Muslims fell to 9.8 percent while support among Pasmanda Muslims increased slightly to 9.1 percent (see figure 3). While vote choice in state elections does not automatically translate to behavior in national elections, the BJP’s recent inroads with the Pasmanda community—along with targeted 2024 election outreach (for example, through enlisting “Modi Mitrs” or friends of Modi)—suggests that increased Pasmanda support for the BJP is entirely possible. But when this outreach occurs alongside efforts to mobilize Hindus on their collective religious identity, voting for the BJP can become a hard sell for Muslims, regardless of caste.
While the wider conversation about caste hierarchies among Muslims suggests that individuals may think of their caste in terms of the broader categories of Pasmanda and General, the 2022 survey in Uttar Pradesh—and the aforementioned Ansari-specific mobilization in Bihar—suggest that Muslims may think more in terms of their subcaste. When asked in the survey whether a respondent agreed with the statement, “In an election, it is important for members of the same jati [subcaste] to vote for the same party,” 43 percent of General Muslims agreed with the statement, compared to 53 percent of Pasmanda Muslims.3 This latter percentage is also higher than the share of Hindu respondents who agreed with the statement; 34 percent of General Hindus, 37 percent of Scheduled Caste Hindus, and 41 percent of OBC Hindus reported that it was important for members of the same subcaste to vote as a bloc. If the BJP can gain ground with the Pasmanda community, there could be far-reaching consequences given heightened perceptions that caste-based coordination is important. As national elections approach, the extent to which Pasmanda Muslims view the BJP’s outreach as sincere (or not) needs to be tracked.
Caste Versus Religion
The increased relevance of caste to Indian politics raises questions about how much solidarity Muslims have with members of their superordinate identity (religious group) compared to their subordinate identity (subcaste group). Scholars of race and ethnic politics across contexts often use measures of “linked fate” to understand group-based solidarity and consciousness. These measures seek to understand how much an individual’s well-being is linked to their group’s well-being. Adapting survey questions generally used in the American context, the 2022 survey in Uttar Pradesh asked Muslim respondents what better or worse conditions for their subcaste group and their religious group would mean for themselves as individuals (see figure 4).4
Muslims generally indicated high levels of linked progress and linked hurt based on their caste identity. Around 90 percent of Muslim respondents perceived that if things got better for their subcaste, things would also get better for them individually. And about 72 percent believed that if things got worse for their subcaste, they would get worse for them, too. By comparison, Muslims generally indicated high levels of linked progress but quite low levels of linked hurt based on their religious identity. Among all Muslims, caste-based solidarity appeared stronger relative to religion-based solidarity.
Across statements probing both progress and hurt, Pasmanda Muslims indicated higher levels of overall linked fate than General Muslims did, suggesting greater group solidarity. While this understanding of solidarity among caste and religious groups is limited to one state, it underscores the need to incorporate various Muslim sub-identities into researchers’ mapping of political dynamics. Indeed, sect (another salient sub-identity) has long shaped Muslim political behavior in some parts of India such as Lucknow, the capital of Uttar Pradesh. In my survey, Shia Muslims in Lucknow exhibited higher levels of support for the BJP, a pattern also found in other national-level election surveys. And, at times, members of the Deobandi and Barelvi sects of Sunni Islam have placed their weight behind conflicting parties and candidates in both state and national elections. For example, my survey revealed that in state elections in Uttar Pradesh, Barelvi Muslims have regularly supported the Samajwadi Party at higher rates compared to Deobandi Muslims.
Toward the 2024 Elections
Since the 2019 general elections, Muslims have become more likely to vote in a unified way, but the political salience of caste in the Muslim community appears to have increased at the same time. Thus, in the coming 2024 elections, while it still seems unlikely that Muslims will vote entirely on their caste or sect identity, it is clear that sub-identities will shape both campaigning and voting to some degree. With the BJP’s mobilization of Pasmanda Muslims gathering steam and with opposition parties plotting countermeasures, it would be unwise for political analysts to limit investigations of caste to the Hindu community alone. If the BJP is even moderately successful in consummating its outreach to Pasmanda Muslims, researchers may need to retire the idea of a Muslim voting bloc and pursue a comprehensive, nuanced understanding of how and when sub-identities shape Muslim voting behavior.
In the months ahead, Carnegie scholars and contributors will be analyzing various dimensions of India’s upcoming election battle—including the role foreign policy plays, the impact of welfare schemes, and how technology has reshaped campaigning. Keep up to date with the project here.
Notes
1 The Sachar Committee was a high-level committee established in March 2005 by the government of former prime minister Manmohan Singh of the Indian National Congress (also known as the Congress Party). Headed by former chief justice of the Delhi High Court Rajinder Sachar, it was tasked with studying the social, educational, and economic status of India’s Muslim community. The committee submitted its report to Parliament in November 2006.
2 This survey was carried out by the author across eighteen districts in eastern, central, and western Uttar Pradesh. The sample included 1,468 Muslims who identified as Pasmanda (OBC and Dalit) and 480 who identified as General or Upper Caste. Additionally, the Hindu sample of the survey included 355 Hindus identifying as Upper Caste, 1,180 identifying as OBC, and 389 identifying as from the Scheduled Castes.
3 Agreeing with the statement involved saying that the respondent “strongly agreed” or “agreed.”
4 The questions on linked progress and linked hurt were adapted from Jae Yeon Kim and Alan N. Yan, “Unbundling Linked Fate: How Respondents Interpret Linked Fate Question,” SocArXiv, Working Paper, November 15, 2021, http://dx.doi.org/10.31235/osf.io/k6c9x.
