The Authoritarian Origins of Dominant Parties in Democracies: Opposition Fragmentation and Asymmetric Competition in India

Abstract

What explains the electoral dominance of a single party over a prolonged period of time in a democracy? Focusing on the case of India’s former dominant party, the Indian National Congress, this article argues that authoritarian-era politics can influence the likelihood of single-party dominance after democratization. More specifically, the axes of political contestation in the authoritarian era interact with the process of democratization to shape democratic-era party systems. When the authoritarian era’s primary socio-political division becomes irrelevant because the democratization process roundly discredits one side of the division, the resulting party system in the democratic period is likely to feature a single major party and a host of small, disorganized, and inexperienced parties. Such asymmetric party competition is particularly likely to produce a dominant party. This explanation not only accounts for the fragmented nature of the opposition in early post-independence India, which was central to sustaining single-party dominance, but it can also potentially shed light on single-party dominance (or its absence) elsewhere in the world.

Appendices

Appendix 1: Congress performance in national elections

Figure 1 depicts Congress’ vote shares (solid gray line) and seat shares (dashed gray line) in all elections to the Lok Sabha, India’s lower legislative house. The black line represents the vote share won by the largest non-Congress party. The dashed vertical line marks 1989—when single-party dominance unambiguously ended. Three features of Fig. 1 stand out. First, Congress always fell well short of a popular majority, never winning more than 45% of the vote. Second, before 1989, Congress’ seat shares dramatically exceeded its vote shares (except in 1977, the lone election during this period that it lost). Third, again with the exception of 1977, Congress’ largest opponent before 1989 was typically far smaller than Congress, signaling the extreme fragmentation of the non-Congress vote.

Fig. 1

Congress performance in national elections (1951–2019)

Appendix 2: Simulated shares for Congress with varying levels of opposition consolidation

Figure 2 underscores the importance of the opposition by presenting the hypothetical seat shares that Congress would have won had opposition support been more consolidated. The figure considers only single-member district seats. Starting with the 1962 election, all seats in the Lok Sabha were elected in single-member districts. Prior to that, in the 1951–1952 and 1957 elections, some legislators were not. In the 1951–1952 election, 314 of 489 legislators were elected in single-member districts. Of the remaining 175, 172 were elected from 86 double-member districts, and three were elected from a single triple-member district. In 1957, of the 494 legislators, 312 were elected in single-member districts and the remainder in double-member districts. The multi-member districts were used to elect members from historically marginalized groups—Scheduled Castes and Scheduled Tribes. Because one of the legislators from these districts had to be a member of a specified ethnic group, the two highest vote winners did not necessarily win election. Therefore, identifying the hypothetical winners from a more consolidated opposition is not at all straightforward—hence, the focus on single-member districts.

The black circles in Fig. 2 represent the share of seats that Congress would have won if Congress’ vote had remained the same in all constituencies but all of the non-Congress vote was unified behind a single candidate—that is, if each race were a two-way fight between Congress and another candidate. In 1957 and 1971, Congress still would have won a majority of single-member district seats, indicating years when there was almost nothing that the opposition could have done to defeat Congress.

Of course, under almost no circumstances could the opposition unite behind a single candidate in each constituency. The dark gray diamonds therefore indicate the share of seats Congress would have won if the Congress vote remained the same, 90% of the opposition vote was consolidated behind a single candidate, and the other 10% of the opposition vote went to a third candidate. The light gray squares do the same, but with only 80% of the opposition vote consolidated behind one candidate and the remaining 20% going to a third candidate. Finally, to take a slightly different approach, the white triangles present the share of seats Congress would have won if its vote shares remained the same but the votes for the largest and second largest opposition parties were consolidated behind one candidate.

Across all four hypothetical scenarios, Congress would have failed to win a legislative majority in at least one of the first five elections, demonstrating that the opposition’s fragmentation was crucial for explaining how Congress successfully won five successive national elections and therefore governed for nearly 30 continuous years. Congress was particularly vulnerable in 1962 and 1967 (as well as 1980, which is not included in Fig. 2). Congress lost the 1977 election. In the 1980 election, a less fragmented opposition could have certainly defeated Congress. In 1984, Congress turned in its best ever electoral showing, meaning that even a highly consolidated opposition would have been hard pressed to defeat Congress in that election. In short, an opposition that was far less fragmented could very well have produced a party system that produced alternation between Congress and either another large party or, more likely, a coalition of opposition parties.

Fig. 2

Hypothetical Congress seat shares with a less fragmented opposition

Appendix 3: Comparison of Congress and the British Conservative and Labour parties

Figure 3 presents the vote shares for Congress in Indian elections from 1951 through 1972 and the vote shares for the Conservative and Labour Parties in British elections from 1950 through 1970. During this period, Congress was a dominant party; it won all five elections, securing single-party majorities in each election and governing continuously from 1951 through 1977. Meanwhile, the Conservative and Labour Parties alternated in power. Labour formed the government after elections in 1950, 1964, and 1966 elections, while the Conservatives came to power after the 1951, 1955, 1959, and 1970 elections. What is particularly noteworthy is that during its time as a dominant party, Congress’ vote shares were not consistently higher than those of the two British parties. Indeed, when comparing temporally proximate elections in India and the UK, Congress sometimes won smaller vote shares than both Labour and the Conservatives. Congress’ vote shares in the 1951 and 1967 Indian elections were lower than both the British parties’ vote shares in the 1951 and 1966 British elections. In other words, Congress’ vote shares alone were insufficient to make it a dominant party as both the Labour and Conservative Parties won similar vote shares but were not dominant. Had Congress faced a single large rival, as Labour and the Conservatives did, it would not necessarily have been a dominant party.

Fig. 3

Vote shares for Congress, Conservatives, and Labour (1950–1971)

Appendix 4: Congress’s competition in electoral districts

Riker’s (1976) argument suggests that Congress dominance rested on the party’s centrist position. Faced by parties to both its left and right, the opposition parties could not coordinate against it because voters favoring a party to Congress’ right (left) would prefer Congress to a left-wing (right-wing) alternative. Riker’s assumption about the nature of political competition does not hold in most constituencies. Figure 4 breaks down constituencies Congress contested by the type of opposition candidates Congress faced. I code all party-based candidates winning more than 5% of the vote as “Left,” “Right,” or “Other.” Left and Right parties are those that clearly and consistently differentiated themselves from Congress in ideological terms, whether on a social or economic dimension.

The parties coded as “Left” are the Bolshevik Party of India, Communist Party of India, Communist Party of India (Marxist), Dravida Munnetra Kazhagam, Forward Bloc (and its various factions), Indian Socialist Party, Kerala Socialist Party, Kisan Mazdoor Praja Party, People’s Democratic Front, Praja Socialist Party, Peasants and Workers Party, Republican Party of India (and its various factions), Revolutionary Communist Party, Revolutionary Socialist Party, Scheduled Caste Federation, Socialist Party, Samyukta Socialist Party, and Socialist Unity Centre. The “Right” parties are the Bharatiya Jana Sangh, Ganatantra Party, Hindu Mahasabha, Congress (O), Ram Rajya Parishad, Shiv Sena, and Swatantra Party. All other parties are counted as “Other.” Most of the “Other” parties are splinters from Congress, regional parties with no clear economic or social appeals that could constitute a plausible “left-wing” or “right-wing” appeal or highly personalistic parties with vague ideological positions.

This coding is designed to maximize the number of constituencies that could be coded as having both left-wing and right-wing opposition candidates. For example, the Kisan Mazdoor Praja Party (KMPP) and Congress (O) were both composed mainly of Congress dissidents who broke with the parent party for largely non-ideological reasons. Nevertheless, because the KMPP eventually merged with other left-wing parties and Congress (O) was eventually thought of as comparatively conservative relative to Congress, I have coded these parties as left and right, respectively. Despite the liberal coding of parties as “Right” or “Left,” Fig. 4 shows that most constituencies did not feature candidates both to the left and right of Congress.

Figure 4 only includes elections through 1971 because the coding of parties becomes especially difficult in 1977 with the formation of the Janata Party, which was the result of a merger of parties previously to the left and right of Congress. Inclusion of subsequent elections would, if anything, further reinforce the point made by Fig. 4, since the Janata Party’s ideological heterogeneity means that it and its offshoots could hardly be classified as parties clearly differentiating themselves from Congress on ideological grounds.

From the perspective of Riker’s argument, the lavender slices in each pie are the most important slices since those constitute evidence in support of Riker’s argument that Congress competed against parties to both its left and right. But, in no election did these constituencies account for even half of all races. In the overwhelming majority of constituencies, Congress faced a left-wing opponent but no right-wing opponent (red slices), a right-wing opponent but no left-wing opponent (blue slices), or opponents that were neither left-wing nor right-wing (gray slices). Constituencies coded as “Left opposition only” and “Right opposition only” sometimes include candidates from parties coded as “Other.” The broad point underscored by Fig. 4 is that in most constituencies, Congress did not face multiple opponents who were implacably opposed to one another on ideological grounds.

Fig. 4

Types of opposition candidates to Congress (1951–1971)