Abstract
The relationship between the local and the national in Indian politics has taken a variety of forms, from secessionist tendencies to agitational politics around specific issues. The course of this relationship is typically explored through electoral performance, primarily whether a party wins sufficient seats to form the government. There is much less attention paid to the relationship between votes and seats. This has led to some questions not getting the attention they deserve, particularly whether the first-past-the-post electoral system that India uses is entirely neutral in the dynamic between the local and the national. This article addresses this question by developing a model that captures the effects of the share of the votes of national parties, as well as the concentration of national and local votes, on the performance of national and local parties. The empirical evaluation of this system points to an overall national bias, which is eroded over time by the emergence of regionally dominant local parties.
The relationship between the local and the national has been a recurring theme in the study of post-independence Indian politics (Brass, 1997). Over the decades, the dynamics of this relationship have been viewed through the prism of the dominant political considerations of the time, ranging from the absorption of local politics into national parties (Weiner, 1967) to the rise of regional parties (Pai, 1990). The changing relationship between the local and the national is reflected in the course of India’s electoral politics and in the fluctuating fortunes of local and national parties. There is usually considerable popular and academic discourse on the seats won by these parties, particularly when they cross the threshold to form the government. There is, however, relatively less attention paid to the vote shares of these parties, and much less attention paid to the relationship between votes and seats in India’s first-past-the-post (FPTP) electoral system. This has contributed to at least one critical question tending to be ignored: has the functioning of the FPTP electoral system in India favoured either the national or local parties? And does this bias remain consistent across the country and over time? This article seeks to answer these questions through an analysis of state-wise voting patterns in Indian parliamentary general elections over a period of 35 years.
The Local and the National in Indian Parliamentary Politics
As is to be expected in a large and diverse country, the local has always had a presence in Indian politics. Even as much of the discourse—especially the popular discourse—has been centred around the larger parties, electoral politics has not entirely been around the large parties alone. In the years soon after Independence, the Congress, as the main national party, did try to absorb more local concerns, especially on issues relating to the territories of linguistically defined states (Isaka, 2015). That this effort did not address all local concerns ensured that there was a noticeable place for independents. In India’s first general elections in 1951–1952, 533 independents contested, gaining 15.9 per cent of the vote. Over the decades since then, there has been a decline in the role of independents in parliamentary elections, with independents accounting for a much lower 2.71 per cent of the vote in the parliamentary elections of 2019. A significant portion of the decline in the vote share of independents was accounted for by the rise in the presence of smaller parties. As many as 673 parties contested the 2019 parliamentary election, compared to just 53 in the first election of 1951–1952. The rise of smaller parties contributed to the growing significance of state politics. In the 1990s, in particular, parties with state-level influence were able to gain a significant presence in national coalition governments (Ziegfeld, 2012).
The rise of strong regional parties has sometimes led to state politics being seen as an autonomous domain (Yadav & Palshikar, 2008). This is consistent with the discourse that tends to focus on parties that present very specific regional interests, such as the Dravida Munetra Kazhagam (DMK) in Tamil Nadu (Harris & Wyatt, 2013) or the Akali Dal in Punjab (Hardgrave, 1983). But the autonomy of regional parties can be challenged when state-level coalitions find a place for national parties. It has been argued that the autonomous state domains thesis needs to be more sensitive to the federal framework (Kailash, 2017). The national can also affect the nature of local parties, as when regional parties are no more than breakaways from national parties whose influence is confined to a region. The tendency for some regional parties to be defined by electoral circumstances rather than ideology led Fickett to distinguish, in 1971, between what he called ‘classical regional chauvinist parties’ and breakaway Congress groups that survived (Fickett, 1971). This distinction did continue in later decades, if in a somewhat broader form. The regional parties that emerged in the north-east often demonstrated a strong ethnic character. Elsewhere, other parties breaking away from the Congress, like the All-India Trinamool Congress, not only survived but also went on to become major political forces. And breakaways from other national parties were not unknown either, especially the splinter groups of the Janata Dal. The Janata Dal (United), the Janata Dal (Secular), the Biju Janata Dal and the Rashtriya Janata Dal are all examples of this tendency.
The distinction between regional parties that rely on identity and those that are forced by circumstances to have no more than a regional presence is thus useful, but it is by no means comprehensive. It has been pointed out that state politics also has an element of regions within regions (Kumar, 2009). There are parties like the Gorkha National Liberation Front (GNLF) in West Bengal that represent very specific sub-regional identities (Datta, 1991). There are others that represent castes that happen to be concentrated in particular areas of a state, like the Pattali Makkal Katchi (PMK), seen to be representing the Vanniyars in Tamil Nadu (Harris & Wyatt, 2013). There are yet others that seek to represent larger religious groups but have their influence concentrated in particular areas. The All India Majlis-e-Ittehadul Muslimeen (AIMIM) has a stronghold in the old city of Hyderabad, even as it makes its national ambitions known (Suri, 2013).
The local thus presents a very broad spectrum of political entities, from independents to caste parties to sub-regional parties and regional parties. A comprehensive view of the local would demand that we take all these parties and independents together when defining the local. These parties can have very different origins and ideologies. What they do share is the fact that their effective electoral presence is confined to a particular region. That is, even if they contest elections in more than one region, they are able to significantly influence results in only one region. The term local is thus used in this article as a contrast to parties that have a national presence.
The distinction between parties with local presence and those with national presence gains strength from the empirical reality prevailing in India in the second decade of the twenty-first century. There is a huge gap between the national presence of the Congress and the BJP on the one hand and the remaining parties on the other. If we take a cut-off point of 20 per cent of the valid votes in a state, then the difference in the performance of the two sets of parties is very striking. As shown in Table 1, in the 2019 election to the Lok Sabha, two parties—the BJP and the Congress—each met the 20 per cent of valid votes condition in as many as 25 states and union territories (UTs). No other political party could do so in more than one state or UT. This makes it quite clear that while the BJP and the Congress have a presence across the country, the presence of all other parties is primarily local. This article then has reason to treat the BJP and the Congress as national parties and all other parties and independents as local.
Number of States and Union Territories in Which a Party Has More Than 20% Vote Share in the Lok Sabha Election 2019.
The focus on electoral influence rather than ideology also helps address the divergence between the way parties see themselves and their presence on the ground. There are parties across India that perceive themselves as national parties, though they have no more than a local presence. Several parties with no more than a regional presence prefix their names with the words ‘All India’. This tendency ranges from regional parties in relatively large states, as in the case of the All-India Trinamool Congress in the state of West Bengal, to parties that have a presence in no more than a small UT, as in the case of the All-India N.R. Congress in the Union Territory of Puducherry. Other regional parties may not have the prefix ‘All India’ but are not averse to putting up candidates in states other than the areas of their influence. The Shiv Sena, in its undivided form, may have had its support concentrated in Maharashtra, but it did occasionally put up candidates in other states. A focus on vote shares thus prioritizes actual electoral influence over what parties would like to claim.
The Inherent Bias of FPTP
The sharpness of the distinction in the voting patterns of the two national parties on the one hand and the remaining parties and independents on the other brings to the fore the inherent bias in the functioning of FPTP. As is widely recognized, ‘In a single-member district system of elections, we would never expect to find complete proportionality between a party’s vote share and its seat share. In general, we would expect that the graph of the seats-votes relationship will be an S-shaped curve’ (Grofman, 1983, p. 297). As can be seen in Figure 1, a party with a low share of the votes—if these votes are evenly spread out—would get very few seats, if any at all. Its votes in each constituency would be nowhere near the threshold where it can expect to be the FPTP. When the party’s vote share nears the threshold, it will begin to be able to cross the threshold in an increasing number of constituencies. A further growth in its vote share would take it up to the middle range when its share of seats increases more than its share of votes. By the time the party’s vote share goes past the thresholds in most constituencies, there will be few remaining seats to get, ensuring that the share of seats tapers off at higher voting percentages.
The Relationship Between Vote Share and Seat Share.
The increase in the seats a party gets for a given increase in its share of votes is, however, not dependent on the magnitude of its vote share alone. It also depends on the number and strength of other contestants. Calvo (2009) has pointed out that if two parties are competing under the FPTP electoral rules, a new entrant who draws votes more heavily from one of the old parties would have a dramatic effect on the expected seat shares of all parties. The precise threshold on the S curve, where a party moves from being underrepresented in terms of seats to getting a greater share of the seats than its increase in vote share, is influenced by the votes of other parties. In a situation where the votes of other parties are widely dispersed, it would be possible for a party to win seats with a relatively smaller share of the votes. In an electoral battle characterized by two parties alone, it would be necessary for one party to get more than 50 per cent of the vote in order to win the seat. When the votes of parties are widely dispersed across multiple candidates, the FPTP system would ensure that a party could win the seat with a much smaller proportion of the vote.
The effect of this varying threshold on the shape of the S curve is depicted in Figure 2. We can begin with the curve marked S1, which is the same as that in Figure 1. The 45° line in the figure represents the path of perfect equality between support (measured in terms of votes) and power (measured in terms of seats), with the share of votes on this line always being equal to the share of seats. When the seat–vote ratio follows the path presented in curve S1, the equality will only be achieved at the point marked E1, corresponding to curve S1. If we now have a situation where the number of parties increases, then the threshold to win a seat will occur at a lower level of the share of the vote. The lower end of the S curve will then move closer to the line of equality, as presented in S2, with the point of equality now being E2. If the increase in the number of contestants with a significant share of the vote occurs across all constituencies, then the threshold in the share of votes to win these constituencies will remain low even after the S curve crosses the line of equality. That is, S2—both below and above the line of equality—would move closer to the Y axis, losing much of its S shape.
S Curves and the Distribution of Votes of the Losing Parties.
In addition to the size of the share of votes of a party and the number of opponents in the fray in particular constituencies, there is a third element that influences the ability of a party to convert its share of votes into seats: the distribution of its vote across the country. A party whose votes are concentrated in a particular constituency alone could have a good chance of winning that seat, though its share of the national vote would be miniscule. Again, a party whose votes are concentrated in a particular region could be close to the threshold that would give it seats in that region, though its vote share would not be significant at the national level. The ability of a party to convert its votes into seats would then depend on the magnitude of its vote share, the number of significant contestants in individual constituencies, and the spatial concentration of its votes.
The dynamics of two of these three factors would tend to be different among national and local parties. National parties, which would have their votes spread out across the country, would have the potential to cross the threshold in a larger number of constituencies, even as the spread of their votes would bring with it the possibility of missing the threshold in many constituencies. The local parties, which would have their votes concentrated in particular constituencies, would have the ability to cross the threshold in those constituencies, even though they have little, or no, presence in most constituencies in the country. Both national and local parties thus have the potential to benefit from the inherent bias of the FPTP system. The specific benefits to national and local parties would then have to be evaluated on a case-by-case basis. It is to this task, in the case of India, that the article now turns.
The Indian Experience
India’s association with the FPTP system goes back to debates in the Constituent Assembly that framed the constitution the country adopted in 1950 (Patidar, 2005; Sridharan, 2002). The success of India’s democracy, especially in the early years of independence, depended to a great extent on the stability of its governments. India’s diversity made it very unlikely for a single party to get a large majority of the votes in its favour. The FPTP system overcame this danger. The overall effect of the FPTP system is to increase the majority, in terms of seats, of the parties with the largest share of the vote. The ability of the dominant party to get compelling majorities in terms of seats, despite falling short of a majority of the votes, enhanced the sense of stability of Indian governments—a sense of stability that they would not have had in a system where the seats were allocated in direct proportion to a party’s share of the votes.
The stability provided by the FPTP system in India did have its costs. There were times when exaggerated majorities in seats made governments appear to have more support than they actually did. This forced the opposition onto the streets, which had its consequences. Mrs Indira Gandhi’s overwhelming majority in parliament in 1971 did little to prevent mass protests by 1973, leading to the imposition of the internal Emergency in 1975. While the Emergency has not been repeated since then, governments have not hesitated to use their exaggerated majorities in seats to bring in unpopular laws. A recent example is that of the three farm laws passed by the BJP-led Union government. The passing of the laws prompted a year-long protest by farmers before the laws were repealed.
As we have argued earlier, the exaggeration of majorities in terms of seats is particularly great when there are a number of electorally significant candidates in the fray. As the number of electorally effective parties increases, the threshold drops even further. The anomaly of candidates winning with low shares of the vote in the FPTP system was expected to diminish over time. Duverger argued that voters were unlikely to vote for a third party with a lesser chance of winning. This tendency was expected to confirm Duverger’s law that a FPTP single-ballot system would lead to the emergence of a two-party system (Duverger, 1954).
The Indian experience, however, runs counter to this expectation, with the country remaining an exception to Duverger’s Law (Chhibber & Murali, 2006; Diwakar, 2007). Contrary to any tendency towards two major parties, the Election Commission of India, in 2021, recognized eight national parties and 54 state parties, even as it listed 2,796 unrecognized parties. This meant that a vast majority of the parties were at the lower end of the S curve, where the seat share of a party could be expected to be lower than its vote share. Elsewhere in the world, this has often led to movements for reforms of the FPTP system, if not its abandonment altogether (Blais, 2008). Yet this trend has not extended to India to any significant degree. Together with Britain, the United States and Canada, India remains among the group of four countries that are widely seen as primary examples of the practice of the FPTP system in its pure form (Reynolds et al., 2005). There has also been no major political demand for a shift from the FPTP system to one based on proportional representation. It would appear that even the smaller parties that are nowhere near the effective portion of the S curve are comfortable with the FPTP system.
This continuing commitment to the FPTP system, despite its inherent bias against small parties working alone, is a commentary on the role of the local in Indian politics. India’s diversity ensures that there is a demand for politics that takes into account the specifics of the local. Such diversity has been known to work against Duverger’s law. It has been noted that a strong relationship between social diversity and the number of parties exists even under FPTP electoral rules (Milazzo et al., 2018; Ordeshook & Shvetsova, 1994). In India, as in other diverse societies, politics is also about protecting the interests of specific social groups. This could lead to voters being loyal to the party most closely associated with their social group. In situations where the vote of the social group is concentrated in a particular constituency and the votes are sufficient to win that constituency, there would be no electoral cost in remaining loyal to the party. If, however, that core vote is widely spread out, then the party’s share in each constituency will be towards the lower end of the S curve and hence too low for it to win seats on its own. It would then be in the interest of that party to seek alliances that would help it move closer to the threshold that would enable the alliance to win seats. Such an alliance would also suit a large party in constituencies where that party is close to the votes needed to win a seat but is likely to fall a little short. When such alliances work, the small party can find itself in a much more rewarding part of the S curve, that is, it would win a greater proportion of seats than its share of votes. In the 2017 election to the assembly of India’s largest state, Uttar Pradesh, a small party—Apna Dal—could get only 0.98 per cent of the votes, but through an alliance with the much larger BJP, it won 2.33 per cent of the seats. The nine seats the Apna Dal won must be seen in the context of the Indian National Congress, which received 6.25 per cent of the vote in that election winning only seven seats.
The ability of smaller parties to counter the effects of the larger national parties is also enhanced in a large country like India by the concentration of votes in particular regions. A local party whose votes are concentrated in a particular region can become the largest party in the constituencies of that region. In such cases, it is the local party that would become beneficiary of FPTP in that region. Indeed, in states like Tamil Nadu, local parties benefit from FPTP to the extent that both the ruling and the main opposition parties at the state level are local parties. The fact that local parties, despite having a smaller vote share at the national level than national parties, can benefit from FPTP in parliamentary elections in constituencies in their spheres of influence would explain the acceptance of this system across the political spectrum.
The acceptance of the system does not, however, mean that both national and local parties are equally affected by the working of FPTP. The fact that the vote of national parties is spread across the country may give them more constituencies where they can cross the threshold to win seats. But the concentration of votes of the local parties in particular regions gives them an advantage in these regions. The concentration of local party support in particular regions may prevent individual local parties from competing with national parties in terms of overall seats in parliament, but taken as a whole, local parties could, as they did in the 1990s, pose a serious challenge to the dominance of the national parties. It is thus important to estimate the distribution of the benefits of FPTP between the national and local parties in the Indian experience.
When seen in terms of the overall national picture, the bias in the functioning of FPTP in India is in favour of national parties. Simple graphs of the seat share and vote share of the two sets of parties reflect the benefit to the national parties. The timeframe for this analysis is determined by the emergence of the BJP and Congress as the two national parties. As the BJP was formed in 1980, our analysis begins from the first parliamentary election the party contested, in 1984. Since there are often different factors at play in national and state elections, there is a need to treat the two separately. In this article, we focus on parliamentary elections alone, covering the 12 elections to the lower house of Indian parliament, the Lok Sabha, that have taken place between 1984 and 2019. In each of the 543 parliamentary constituencies, the votes of the BJP and the Congress have been taken together in determining the vote share of the national parties. The remaining votes in each constituency have been treated as local votes.
Taken together, the vote share of national and local parties, their seat share and the seat–vote ratio tell us an interesting story. In the seven decades since India became a republic in 1950, Indian prime ministers have always belonged to either the BJP or the Congress, with the exception of brief stints of less than a year each by V. P. Singh, Chandrasekhar, H. D. Deve Gowda and I. K. Gujral. Yet the vote shares of the national and local parties, especially in the twenty-first century, do not reflect this dominance. As Figure 3 tells us, if we take vote share to reflect the support parties have, the support of local parties is not consistently below that of the national parties. The local vote, in fact, exceeds the national vote not only in 1996 but also in two twenty-first-century elections to the Lok Sabha.
This pattern is, however, not seen in the share of seats of national and local parties in the Lok Sabha over the same period. As can be seen in Figure 4, the share of seats of national parties is consistently higher than that of local parties, sometimes by a very substantial proportion.
The dominance of the national parties over local parties in terms of their share of seats, despite a more mixed pattern in the share of votes, is consistent with the trends in the seat–vote ratios of the two sets of parties. As can be seen in Figure 5, the seat–vote ratio of the national parties (share of seats won by national parties to the share of their votes) is higher in every Lok Sabha election than the seat–vote ratio of local parties (share of seats won by local parties, including independents, to their share of votes) for the entire period considered in this article.
The consistent bias in favour of national parties over local parties and independents taken together tells us something about the working of FPTP in India. It suggests that the benefit to national parties of a presence in a larger number of constituencies is greater than the benefit local parties gain in terms of the concentration of their vote. Exploring the precise balance between the spread of the vote of the national parties and the concentration of that of local parties would, however, demand more detailed estimates of these factors.
The Model
In order to estimate the working of the relative benefits of FPTP to national and local parties, we first generate panel data covering all the general elections to the Lok Sabha from 1984 to 2019. Taking the units of analysis to be states and UTs, we calculate the seat share and the vote share, as well as the concentration of votes, in each state and UT. The N in the estimate is thus the number of states and UTs in each election times the number of elections.
In keeping with the argument outlined in this article, the seats won by national and local parties are a function of their share of the votes, as well as the concentration of the votes. The vote percentage of national parties and that of local parties would add up to 100. We can then take only the vote share of national parties into consideration, expecting a positive relationship with the seats of national parties and a negative one with the seats of local parties. We take the concentration of national and local votes separately, that is, the share of the main national party in the total national party vote and the share of the main local party in the total local party vote. In other words, the seat share of national parties in each state would be dependent not just on the vote share of national parties but also on the ability of the main national party to concentrate the national party votes in its favour, as well as the extent to which the main local party is able to concentrate local party votes in its favour. Correspondingly, the seat share of local parties in each state would be dependent not just negatively on the vote share of national parties but also positively on the ability of the main local party to concentrate the local party votes in its favour as well as the concentration of national party votes in favour of the main national party.
In formal terms,
where Sn = seat share of the national parties, Vn = vote share of the national parties, Cn = proportion of main national party in total vote of national parties, Cl = proportion of main local party in total vote local parties and εn is the iid residual errors.
And the model for the seats of local parties would be
where S Ɩ = 1 – S n = seat share of local parties, β0Ɩ = 1 – β0n and ∈Ɩ = 1 – ∈n .
As the results of these OLS equations, presented in Table 2, tell us, the seats won by the national parties are dependent not only on the share of national party votes, but also on the concentration of both the national and local votes. The national parties’ share of the votes and the share of the main national party in the votes of national parties have a positive effect on the seats won by national parties. What is equally important to note is that the concentration of the main local party’s votes also affects the seats won by the national parties, though this effect is understandably negative. That is, the greater the concentration of local votes in the main local party in a constituency, the fewer the seats won by the national parties. The p values in all three cases would be considered significant, being most significant for the share of the national party vote, followed by the concentration of local party vote and then the concentration of the national party votes. As is to be expected from the fact that the independent variables are the same in the second model, the estimate of the seats won by the local parties is a mirror image of the estimate of the seats won by national parties. The seats won by local parties are negatively related to the share of the votes of national parties and the concentration of national party votes, while being positively related to the concentration of local party votes.
Model Based Estimates of Seat Share of National and Local Parties.
The fact that the concentration of votes in local parties has a negative impact on seats won by national parties points to the effect of the growth of regionalism and other forms of localism on the performance of national parties. As local votes get consolidated around specific local players, the ability of the national parties to benefit from dispersed local votes diminishes. It is, therefore, important to take on board the impact of localism on the ability of national parties to convert their votes into seats. One broad indicator of the extent of localism is provided by whether a local party has been in power in a state. We can then classify the states into three broad categories: those where a local party has not been in power, those where a local party has been in power at least once, but such parties have not regularly occupied the positions of both the ruling and main opposition parties at the same time and those where local parties have occupied the space of both the ruling party and that of the main opposition.
In identifying these categories in empirical terms, further specifications are necessary. We must remember that our period of analysis is between 1984 and 2019. Thus, the categories are only relevant for this period. For instance, a state that is classified as one in which a local party has not been in power, only meets this criterion for the years 1984–2019. In order to ensure consistency over the period of the study, states in this exercise have been taken in terms of their composition in 1984. That is, states that have been divided since 1984 have been treated in terms of their old undivided status. This means that for a state to be declared as not having had a local party in power, this condition must hold for both states after the original state has been split. The condition that a local party has been in power at least once would be met if the state in its unified form or any of its descendants had a government of a local party at least once. Similarly, the condition that a state had local parties occupy both its ruling and opposition spaces would hold even if some of the elections were for the United States and others for states that emerged after the original state was divided. It must also be remembered that not all UTs have legislative assemblies. If they have not voted for a local party, it could be only because they have not had the option to do so, but their votes in national elections cannot be ignored. As a local party is not a part of their electoral calculations, they have been included as a part of the first category of states and UTs that have not had a local party in power. A further empirical qualification is needed for the distinction between the second and third categories. The states and UTs that have had local parties as the ruling party as well as the opposition would necessarily also meet the condition that they have had such a party in power at least once. The empirical dividing line used in this article to distinguish between these two categories is the frequency of states having local parties as both ruling and opposition parties. States and UTs that have had local parties in both ruling party and opposition spaces after at least three state elections between 1984 and 2019 have been taken to belong to the third category. The remaining states and UTs that had local parties in power after at least one election are taken to belong to the second category.
As can be seen in Table 3, there were 10 states and UTs that did not have a local party in government between 1984 and 2019. Not all of them, however, had the option of electing a local party. Andaman and Nicobar Islands, Chandigarh, Dadra & Nagar Haveli, Daman and Diu and Lakshadweep do not have legislative assemblies. Four other states that have legislative assemblies—Goa, Himachal Pradesh, Madhya Pradesh and Rajasthan—have not elected a local party government since 1984. Meghalaya presents a less clear case. It had its first local party government in 1998, with a representative of the United Democratic Party (UDP) as the Chief Minister. The party was, however, a junior partner in an alliance with the Congress. Since the UDP had fewer seats than the Congress, and our focus in this article is on the relationship between seats and votes rather than governments, we have placed Meghalaya too in the first category of states that have not voted to make a local party lead the government during the period of the study. The category that is most populated is the one in which there has been at least one local party in government during this period, with 17 states and UTs meeting this condition. The phenomenon of local parties playing the role of both ruling and main opposition parties is confined to four states, though three of them—Uttar Pradesh, Tamil Nadu and West Bengal—are large and have had a prominent place in Indian political discourse.
States and Union Territories Classified by Presence of Local Parties in Government Since 1984.
Estimates of Seat Share of National and Local Parties in States and Union Territories Where No Local Party Has Been in Government Since 1984.
In the states where there has been no local party in power between 1984 and 2019, there is understandably little role for how widely the vote is distributed across either local or national parties. As Table 4 tells us, the concentration of both the national party vote and the local party vote has little impact on the seats won by the national parties, with the coefficients of both these variables being quite some distance from being significant. The seats of the national parties are determined primarily by their share of the total valid vote, with the corresponding coefficient being significant. This has a direct effect on the seat–vote ratios. During the period of the study, as Figure 6 tells us, the seat–vote ratio of the national parties was consistently higher than the seat–vote ratios of local parties.
This picture changes with the emergence of local parties. As local votes consolidate around one major local party in a state, the local parties as a whole become efficient in the conversion of their votes into seats. As Table 5 tells us, the votes of the national parties become more sensitive to the number of opponents, whether the opponent is another national party or a local party. Thus, the number of seats won by national parties is dependent not only on their share of vote but also on the concentration of their vote. The concentration of the local party vote does have a statistically significant negative effect on the seats won by national parties. But, as Figure 7 shows us, the national parties are still, with the exception of 1 year, the more efficient set of parties in converting their votes into seats.
Estimates of Seat Share of National and Local Parties in States and Union Territories Where a Local Party Has Been in Government At Least Once, but No Local Parties in Both Government and Main Opposition Positions after Three or More Elections Since 1984.
As the local vote increases further, it has a transformative effect on the ability of the national parties to convert their votes into seats. In the states where there are two major local parties—so that it is possible to have both the ruling party and the main opposition from among local parties—the national parties have to depend on their vote share alone to win seats. This is evident in Table 6, where the coefficient of the vote share of the national parties has a significant effect on their seats, and the concentration of the winning local party votes has a significant negative impact on the ability of national parties to convert their votes into seats. The concentration of national parties’ votes is not a significant player in this equation. Consequently, as Figure 8 tells us, the local parties get more efficient in converting votes into seats, being more efficient than national parties in 2 years of the period under study, 1989 and 2004.
Estimates of Seat Share of National and Local Parties in States and Union Territories Where Local Parties Have Been in Government as well as the Main Opposition after Three or More Elections Since 1984.
Conclusion
The working of the FPTP electoral system in India has a clear bias towards national parties in the country’s parliamentary elections. The spread of their vote across the country ensures that they are higher up the S curve in more constituencies than parties that have a more local influence. Their share of the vote in specific constituencies is also aided by a frequent tendency in Indian politics for national votes to gravitate towards a single national party. In some years, as in the election when Rajiv Gandhi first became prime minister, the move was in favour of the Congress; in more recent years, it has been the BJP that has benefitted. This has contributed to the concentration of the vote of national parties, which has a statistically significant effect on the seat share of national parties. The bias of the FPTP system in favour of national parties is further compounded by the existence of a very large number of local parties. This division of the votes against the national parties lowers the threshold for a party to win.
Conversely, as a local party increases its share of local votes, it can also begin to defeat national parties. The concentration of local party votes thus has a statistically significant negative impact on the seat share of national parties. As a result, as local votes consolidate in each state, the bias of the FPTP system in favour of national parties will diminish. We can then see the bias of the FPTP system as a changing phenomenon. It is at its peak when there are no politically significant local parties around. As the local parties increase, they first add to the bias in favour of the national parties by dividing the local votes. As the local vote consolidates around a single local party, this division reduces, thereby eating into the bias in favour of the national party. As local parties get a greater share of the local vote, they begin to become the main beneficiaries of the FPTP system.
In a large and diverse country, it is to be expected that different states go through these stages at different times. Consequently, the extent of the bias the FPTP system offers national parties would vary from state to state, depending on the extent and nature of the influence of local parties. In states and UTs dominated by national parties, all that matters is the vote share of national parties and the concentration of their votes. In states and UTs where the consolidation around a local party has grown to a level where a local party has been in power at least once since 1984, the gap in the ability of the national and local parties to convert votes into seats is substantially reduced. As the consolidation around local parties grows to a point where these parties can occupy both the ruling party and main opposition party spaces, the vote share of national parties is so low that it does not really matter how these votes are distributed between the national parties. The seat share of national parties is entirely dependent on their share of the votes as well as the extent of the concentration of local party votes. In the process, there is a further erosion of the bias of the FPTP system in favour national parties. The possibility of local parties growing to the point where they become the beneficiaries of the FPTP system ensures that there is no significant movement in the country against the FPTP electoral system, though, on the whole, the system still favours national parties.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
This paper is the result of work on a larger project funded by Tata Consultancy Services.