ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
-A A +A

Examining Electoral Data

An Enquiry into People’s Preferences of Communist Parties in West Bengal and Kerala

Arun Kumar Kaushik ( and Yugank Goyal ( teach economics at O P Jindal Global University, Sonipat, Haryana.

West Bengal and Kerala are often juxtaposed under a common communist identity in most scholarly and policy discourse. We deconstruct these linkages by looking at election result data in these states in the past five decades. Our observations indicate that (i) the assumed supremacy of communist preference in the two states must be diluted, and (ii) the tendency to put both states in the same political basket must be revisited. Since election data are a direct reflection of the people’s preferences, this paper adds an important contribution to the literature by looking at the demand side of the political market.

The authors thank Paavani Pegatraju for her excellent research assistance.

Discourse on the politics of West Bengal and Kerala is varied and rich, attracting interdisciplinary scholars from multiple disciplines to opine on issues of social importance. A vast portion of this literature, however, engages with both the states together, either in a comparative fashion or in a discursive manner. Seminal work has been done on West Bengal (or Kerala) alone, yet the likelihood of finding the two states lumped up in the same research question, under the umbrella of the same title, is predictably high. In fact, it is rare to witness a scholar present her findings on Kerala in a conference and not be asked a question about West Bengal (or the other way around). Scholars across the world in the past two to three decades have often viewed the two states with some sort of intellectual homogeneity. In many ways, West Bengal and Kerala are two sides of the same theoretical coin in social science research: if you bring about one, the other will follow, implicitly or explicitly.

This does not mean that the two states are regarded similar in all important sociopolitical aspects. Social scientists are mindful of the underlying differences between the two states, and the literature is persuasive enough to point out the significant differences between the two states. Yet the policy obsession to view the two states together remains.

In this paper, we try to show why the approach of painting the two states with the same “communist” brush needs re-examination and propose two ideas: first, that the obsession of communist regimes as the most preferred political mandate of the people needs revision, and second, that we recognise that the two states have markedly different political preferences and processes and therefore must be studied with as much
mutual consideration as there is for any two other states in the country. To substantiate these two ideas, we closely examine constituency- and party-level data collected from every assembly election that has taken place in the two states since 1967. This allows us to observe the preference of the people in isolation of whatever their ideology may have been. In other words, since studying voting preferences of people is a neat way of insulating the mandate from speculative predictions and policy analysis, our ex post analysis has lessons for ex ante imagination of what the political preferences of the people are rather than how the government of the states is constructed. We do not engage with political processes or elements of political decision-making in the two states (even though we recognise the
importance); instead, we encourage scholars to re-examine their labels of communism in the state. Our results should merit the attention of those who are interested in how the electorate decides its rulers.

The motivation for the research is to recognise the importance of understanding people’s political preferences, reflected in election mandates. Election results and the micro-constituents of the election data contain a rich source of knowledge about political imagination that shapes the state. There is a considerable value in examining what people want(ed) so as to construct a more robust understanding of political participation, preference, and the value system embedded in the social fabric of the region. In fact, one of the contributions of the paper
is that it knocks out the surprise factor on the defeat of the Communist Party of India (Marxist)—CPI(M)—in 2011 (and later in 2016); if one observes the electoral indices, these results will not look so drastic.

Yet, there is practically no study that focuses on West Bengal and Kerala from the perspective of analysing election data. We study assembly elections data since 1967 and investigate
several indicators. Some of them are standard indices in the literature and reveal crucial information about the quality of the verdicts. For others, we construct new variables to examine the information about the people’s mandate they capture. Our study of this historical data reveals patterns that nudge us to revisit some of our fundamental imaginings concerning the two states.

Backdrop on Elections in West Bengal and Kerala

The rise of communist parties in the political landscape of West Bengal and Kerala had very different beginnings and motivations (Harriss and Törnquist 2015). The emergence of a socialist party in Kerala (in the form of the Congress Socialist Party) as an alternative to the nationalistic Congress—and an unsympathetic view towards the British—took place much earlier (Desai 2001; Heller 2005; Menon 1994). In West Bengal, communist tendencies erupted not only under the umbrella of the Congress through the elite bhadralok (prosperous, well-educated people) but also in tacit acceptance of British rule (Basu and Majumdar 2013; Kohli 1990). Where Kerala’s first elected communist leadership created headlines around the world as early as in 1957 (Devika 2010), in West Bengal,
this was to happen only in 1977 (in the form of the CPI[M]).
The social movement in Kerala therefore not only predated that in West Bengal but also superseded it in scope and size (Basu and Majumdar 2013). In fact, even during independence, West Bengal’s Communist Party of India (CPI) had limited political power as compared to its Kerala counterpart
(Desai 2001).

West Bengal

Even though West Bengal had ousted the Congress from political dominance as far back as during the assembly elections of 1967 (the Congress was the largest party in terms of both seat share and vote share, but a United Front coalition proved its majority in the assembly),1 the CPI(M) emerged victorious in 1977 and retained its power until 2011. Securing more than 60% of seats and one-third of the votes, the CPI(M) became a powerful force in West Bengal politics in 1977. Jyoti Basu emerged as the longest serving chief minister of any state in India, with the CPI(M) as the leading party in each assembly election (1977, 1982, 1987, 1991, and 1996). In all these five elections, the CPI(M)’s seat share remained within an impressive range of 54%–64%. The party again came to power in 2001 and 2006, but this time, Buddhadeb Bhattacharya was
appointed chief minister.

During 1997–98, the Congress had witnessed a split within its members in West Bengal, led by Mamata Banerjee, who formed her own party, the All India Trinamool Congress (AITC). The AITC had appeared as a non-trivial contender to the CPI(M) in 2001 and cost the Congress and the CPI(M) 60 seats in 2001 and 30 in 2006 (of the total 294 constituencies).2

Assembly elections in 2011 were groundbreaking. The AITC emerged as the largest party in terms of both seat share
and vote share in West Bengal, dealing the CPI(M) a crushing defeat and marking the end of what is claimed to be the world’s longest-elected communist government (BBC News 2011; Bhaumik 2011). The United Progressive Alliance with the Congress and the AITC secured a rise of 196 seats from 2006, while the seat share of the Left Front (led by the CPI[M]) fell by a massive 171. Individually, the seat share count of the AITC and the CPI(M) were practically reversed in 2011, as compared to 2006 (the AITC’s numbers rose from 30 to 184, while the CPI[M]’s fell from 176 to 40). This was, in many ways, a rehashing of the 2009 Lok Sabha elections in West Bengal. Political commentaries blamed the Singur protest of 2008 (against the land acquisition by the CPI[M] state government through eminent domain for Tata Nano’s industrial project) as the progenitor of the fall of the CPI(M) and the rise of the AITC, whose leader Mamata Banerjee was the chief supporter of the protests (Bardhan et al 2014; Chakrabarty 2011a, 2011b; Chatterjee and Basu 2009), which also attracted widespread media attention.

In the 2016 assembly elections, the people of West Bengal overwhelmingly voted in favour of the AITC again. The CPI(M) garnered a mere 26 seats, while the AITC secured 211 seats. If the 2014 Lok Sabha elections were of any predictive utility,
the results of the 2016 West Bengal assembly elections could be understood.


Built on anti-feudal principles, peasant consolidation, and welfare state models, the CPI won the first elections held in 1957 in the then three-year-old state of Kerala (Radhakrishnan 1989). E M S Namboodiripad became its first non-Congress chief minister. Due to anti-communist protests, however, President’s Rule was imposed in the state in 1959. In 1960, the Congress offered support to the Praja Socialist Party (which had merged with the Samyukta Socialist Party, led by George Fernandes, in 1972 to re-form the Socialist Party, which in turn became a part of the Janata Party following the Emergency in 1977). The chief minister was chosen from this party. Thus, even though the CPI lost these elections, the makeover of the state was still socialist in nature. By 1965 (the third assembly elections), the CPI had given birth to a new faction, the CPI(M), just as the Congress had split to form the Kerala Congress. The hung seat share and political rivalry prevented a government from being formed, and President’s Rule was imposed yet again.

The fourth assembly elections in 1967 proved to be decisive. The CPI(M) swept off the results, securing almost 40% of
the seat share, and the CPI came in second with 14.7%. The Congress was left far behind at less than 7% of the seats this time.3 Setting aside their differences over China, the CPI and CPI(M) came together to form the United Front, along with a few other parties (including the Muslim League), and formed the government. However, the coalition could not withstand the pressure of internal distrust, and by 1969, the majority was lost. With the support of the Congress, the government finished its term and fresh elections were called for in 1970. The CPI, with Congress support, appointed C Achutha Menon the chief minister until 1977. The 1977 assembly elections brought the Congress to the top although the government formed was an unstable coalition with the CPI—in a span of two years, Kerala was ruled by four different chief ministers from the Congress, the CPI, and the Muslim League. The assembly was dissolved due to lack of a proven majority in the house yet again, and in the 1980 assembly elections, the baton was passed jointly to the CPI(M) and the Congress once more. The CPI(M) had the highest number of seats this time too, but the coalition became unstable again, and fresh elections were called for in 1982.

Since 1982, the United Democratic Front (UDF; led by the Congress) and Left Democratic Front (LDF; led by the CPI[M]) have been grabbing power alternately without supporting each other. The Congress, through the UDF, has led the state during the assembly years of 1982–87, 1991–96, 2001–06, and 2011–16. The LDF (with the CPI[M] as the leading party) has
secured the alternate spots in 1987–91, 1996–2001, and 2006–11 and won the fourteenth assembly elections in May 2016.

Data: Variables and Model

We retrieved election data for West Bengal and Kerala from the website of the Election Commission of India (nd) to investigate the extent to which the people’s preferences are reflected in the election results, and borrowed and constructed a select set of variables. Two of them use the seat share and vote share of the parties,4 helping us construct disproportional representativeness and electoral competition (Kaushik and Goyal 2015; Kaushik and Pal 2012). Seat share is the ratio of the number of seats acquired to the number of total seats in
the assembly, while vote share is the ratio of the number of votes polled in favour of a party to the total number of votes polled. Note that in a first-past-the-post system, a party
with the highest seat share can have a low vote share and still come to power.

Disproportionality index (Gallagher 1991) is a measure of how representative a government is or how much it represents the preference of the voters. This is measured using the difference in seat share and vote share because a perfectly representative government will be one where seat share and vote share are the same and is given as follows:



(Here, sj denotes the number of seats of the political party j in the elected body; vj denotes the number of votes received by the j-th political party; and P denotes the set of political
parties contested.)

Clearly, DISPR takes the value zero if the political system
is perfectly representative. On the other extreme, if there is dictatorship, DISPR takes the value 100. Therefore, DISPR takes values between 0 and 100  [0, 100]. It is easy to observe that a higher value of DISPR indicates a lower representation of the people’s preferences.

Electoral competition measures the effective number of parties contesting in the elections. It is not merely a measure of the number of parties contesting because not all parties are equally strong. Hence, the seat share (or vote share) would better capture the proxy for the size of the strength of that political party. Thus, the effective number of political parties (ENP), which is defined as follows, can be considered as an appropriate measure of electoral competition (Chhibber and Nooruddin 2004; Laakso and Taagepera 1979):5

ENP can be interpreted as the number of hypothetical equally sized parties that would create the same electoral competitiveness as with actual parties of unequal size. It is evident that the lowest possible value of ENP is 1, when
only one party wins all the seats. On the other extreme, if all the parties win an equal number of seats, ENP is equal to the number of parties. Clearly, a higher value of ENP indicates more intense competition among political parties in elections. ENP is therefore a reliable measure of electoral competition. We use these two variables to see (i) how competitive the electoral politics is, and (ii) how representative the governments have been in West Bengal and Kerala by analysing
election data for both West Bengal and Kerala from 1982
until 2016.

We dig deeper into the electoral indicators and discover
the important figures of vote share. Vote share—however
irrelevant in representative democracy—conceals a powerful message, that of the extent of likeability of the party. For both West Bengal and Kerala, we take winners and losers (those second to the winners; the runners-up) by calculating the number of seats; the party with the highest seat share is
the winner and that with second-highest seat share is the
loser. We then calculate the ratio of the vote share of the
loser to that of the winner for both states separately. Our intuition is that in a system where preference aggregation is leakproof, the total votes accrued by the losing party should be fewer than that accrued by the winner. Thus, in some sense, the lower the value of this ratio, the truer the representation is. For instance, if the value of this ratio is greater
than 1, it indicates that the loser got more votes than the winner despite getting fewer seats. Vote share ratio (VSR) is given as follows:


(Here, vlc represents the number of votes of the losing party [second in the seat share list] in constituency c [belonging to the set of constituencies C], and vwc is the number of votes accruing to winning party in constituency c.)

Thus the vote share ratio is the total votes received by the loser divided by the total votes received by the winner. Similarly, seat share ratio (SSR) will be:


(Here, sl represents the total number of seats achieved by the loser and sw represents the total number of seats achieved by the winner in the assembly.)

Note that SSR is, by construction, lower than 1 because the loser will definitely have fewer seats than the winner. VSR, on the other hand, need not necessarily be less than 1, because in a representative democracy, the loser might get more aggregate votes than the winner.

In reality, a greater-than-one value of the ratio is not unusual. In fact, that is one of the criticisms of the first-past-the-post system.6 In other words, if the runner-up has a few votes more than the winner, however far from ideal this may be, it does not really raise a red flag. But if the difference between seat share and vote share is huge while VSR is greater than 1, then the system merits special attention. It indicates that a high degree of dissatisfaction must be brewing in the public, who voted for the runner-up but could not secure the government. To give this ratio context, we find the value of the ratio of loser-to-winner vote ratio to loser-to-winner seat ratio. We do this for every assembly election from 1967 until 2016 in both the states. We call it the V–S Ratio.

V-S Ratio will be high if the votes for the loser relative to the winner are high compared to the seats secured by the loser relative to the winner. It relates positively to the votes of the loser and the seats of the winner and negatively to the votes of the winner and the seats of the loser.

Next, we construct another set of variables to map the
winning margins of the top three parties in the constituencies in which they won. We select the political parties that emerged as the top three in each assembly election in the
past five decades (1967–2016), in West Bengal as well as in Kerala. These parties are the top three parties based on seat share. We take the party with the highest number of seats and measure the average margin of votes between this party
and the party with the second-highest seat count in the
constituency where the former is the winner. For the party that has secured second position in the state chart, we look at its margin of victory for the seats where it has won. We do the same exercise for the party that has the third-highest
number of seats in the state. We call this the average margin of superiority (AMS).



(Here, n is the number of constituencies in which party f won, denoted by t; vft is the number of votes secured by the first party f in constituency t; vst is the number of votes received
by party s in constituency t, which is second to party f; and vft represents total votes in constituency t.)

AMS is a very important indicator of how “heavy” the victory of the party is. If the first party’s margins are smaller than the second party’s, then we infer that the second party is in a more “comfortable” position with a higher level of loyal voters in its constituencies. In other words, values of AMS are unique windows for parties to understand the degree of their superiority, wherever they are superior. A high value of AMS (in a way) permits the party to be politically complacent in the constituencies of their victory. Note that AMS does not help us compare a given set of parties, because it is not an inter-party variable.7 It selects a party and looks at its relative position with respect to the party just beneath it in the seat tally of the state. Therefore, AMS is a valuable indicator of the extent of power a party wields in a constituency it has won. Such an indicator lets us analyse how monopolistically a party may behave. If AMS is high, the party may not be as deliberative with its constituency citizens as another party with low AMS. Lower AMS values would signify volatility in political preferences amongst people and encourage parties to work harder if they are to retain their power in forthcoming elections.


Table 1 spells out the disproportionality and political competition in West Bengal and Kerala (the latter is expressed for both seat share and vote share, but our focus is on seat share in this case, denoted by ENP). Coincidentally, the assembly elections since 1982 have been taking place in both the states in the same years. We observe that electoral competition
is considerably higher in Kerala than in West Bengal and that Kerala also has fairly low levels of disproportionality in
representation. These are diagrammatically represented in Figures 1 and 2.

The figures indicate the strongly divergent behaviour of the two states as regards their political landscape and mandates. Political competition in Kerala has always been significantly higher than in West Bengal. High ENP in Kerala—even though a decreasing trend—suggests that people in Kerala have a more diverse choice architecture. In West Bengal, on the contrary, the value of ENP has remained consistently between 2 and 3, indicating a duopoly in the electoral market. Even in West Bengal, trends suggest the competition levels will come down in time. At the same time, disproportionality in Kerala is much lower.

The ratio of the loser’s vote share to that of the winner’s (VSR), coupled with the loser’s seat share to that of the winner’s (SSR), reveals very interesting insights for both the states (Table 2). The table also contains V-S Ratio values. For West Bengal, the value of VSR is greater than 1 in 1986 and 1996 and very close to 1 in 1982 and 1991. For Kerala, this happened in 1972, 1987, and 1996 and came very close in 2011. This means that there is a significant proportion of the population that favoured the loser over the winner in several of the
assembly elections in both the states (for West Bengal, this was more so during the height of communist rule in the state). As suggested, this may not be a very unusual observation. But what is remarkable here is the enormous difference between the VSR and SSR in West Bengal as compared to that in Kerala (Figures 3 and 4). Meanwhile in Kerala, even though the loser had more votes than the winner, the loss in seat share was
not unusually high. In other words, the seat share was trying to mimic and follow the mandate of the vote share. In West Bengal, however, this difference is bafflingly enormous; the loser was devastated in the seats even when it had more votes than the winner.

Finally, we come to AMS. In the case of West Bengal
(Figure 5, p 43), the average winning margin of superiority for the top three parties in the respective years since 1982
has remained consistent at their respective levels. The
year 1977 was peculiar with high AMS for all the top three
parties. During the 34 years of communist rule (1977–2006), when the CPI(M) was in unchallenged power, its margin
hovered at an approximate range of 14%–17%, which was
also the margin of the parties securing third positions in those elections. In 2011 and 2016, the CPI(M) stood at third position (the Congress was at second), and the figure
shows extremely low values of AMS for the party (approximately 6%). The point here is not only the decline of the CPI(M) but the sustenance of high AMS for the CPI(M) during its victorious years. Another valuable insight from the figure is the substantial difference between AMS values for the first and the second party during communist rule.

In Kerala (Figure 6), the picture is very different. First, we observe that the values of AMS for the top two parties in Kerala are significantly low. Second, the difference in AMS values for the first and second parties has not been as substantial as in West Bengal. Third, the winner’s AMS has an alternating rate. And fourth, the AMS of the party securing the third spot is often much higher than that of the first and second parties. The fourth point—however catchy—may not be as useful for us (the third party has often been the Muslim League or the CPI) much like in West Bengal since the vast majority of seats would be distributed between the first three parties, but the first
two offer very meaningful insights. We discuss these results
in the next section.


Our findings reveal three broad strands of ideas. First, the idea that people in the two states have overwhelmingly preferred the CPI(M) during their victories is diluted significantly once we observe the differences in seat share and vote share and the patterns of elected coalitions. Second, we see that the effective political competition in Kerala has been consistently higher than West Bengal and that it has also had very low disproportionality between the seats and votes distribution. Third, the vote share victory margins of parties that rank first, second, and third have remained smaller in Kerala than in West Bengal. Taken together, these ideas suggest mono­polistic tendencies of political parties in West Bengal versus
a mature democracy in Kerala and throw light on the diverging social indicators of the two states. The rising margins
in Kerala in the last two decades and stagnant ones in West Bengal offer a unique window to predict the diverging directions of the type of politics expected in the near future in
the two states.

Our results also show three striking features of the political mandates in West Bengal and Kerala that allay our excitement in painting both states with the same brush. First, West Bengal’s high DISPR and low ENP compared to that of Kerala shows disconnected and disengaged governance in West Bengal and politically decentralised government in Kerala. Second, the high V-S Ratio in West Bengal compared to Kerala indicates strong repulsion of the CPI(M) by the people in West Bengal, compelling us to revise our notions of West Bengal as the CPI(M)’s stronghold. Third, AMS values in West Bengal and Kerala indicate higher levels of political complacency and capture in the CPI(M) in West Bengal than in Kerala. Overall, the two states do not look similar at any level in terms of political processes, and we do not find an overwhelming desirability for communist parties in these states from the voters’ perspective either.

It appears that Kerala’s democracy is more mature and healthy whereas West Bengal suffers from a vegetative and unresponsive leadership, pushing the states far apart in
how ideals of political representation and the people’s choice are reflected in the formation of government. In Kerala, the high values of ENP and low values of DISPR indicate an
advanced state of democracy. The people’s preferences get translated into political power without much “leakage” of the mandate. This also explains the alternating ascendancy to power in Kerala. Since 1982, Kerala has witnessed a U-turn
in the people’s preferences every five years. Repetition of
the coalition has happened every ten years, and not five, in this period.

Note that the disproportionality index in Kerala is far lower than that in West Bengal, suggesting better transfer of the people’s preference into the assembly. This means that the Malayali population has never given a free hand to the CPI(M), who have emerged winners only in alternate elections. Whenever the CPI(M) has assumed power, it has had to be relegated to defeat in the next elections. Can this evidence disable all claims of the laudatory Kerala model built on the foundations of the communist thrust of egalitarianism and the welfare state (Heller 1996; Jeffrey 2001; Ramchandran 1997)?

We are conscious of the fact that the structural make-up of Kerala’s politics may be independent of specific parties, given the progressive and socialist ideals exhibited by both the Congress and the CPI(M). We do not intend to claim something radically different. Our aim is to flag the idea that examining Kerala’s politics through the lens of policies alone glosses over the nuances of the representative outlook of the people, which is valuable in its own right. Where scholars credit the success of the progressive state politics of “the Kerala model” to communist forces (Jeffrey 2001; Parayil 2000), our findings support the general impression of political awareness leading to such social democratisation of politics in Kerala. Our aim is to
encourage scholars to attempt to locate the credit not necessarily in CPI(M) coming to power but in the shifting foci of voters in every assembly election.

In fact, in the show of such volatile preference, one can only expect parties in power to offer significant benefits to the people for them to be voted in, in the next term. This also means that people in Kerala must be well-informed to be able to think (and voice) their change of opinions if the party in power has not delivered, and this is corroborated by scholars (Nag 1989). Indeed, Kerala’s participatory democracy has been established extensively in literature (Sandbrook et al 2007). But at the same time, the sizeable literature criticising Kerala’s caste dynamics and politics of exclusion orchestrated by the left politics of the state compels us to think deeper (Deshpande 2000; Devika 2010; Lindberg 2001). Our findings indicate, then, that uncritical labelling of the state containing socially progressive politics may be sourced in the participatory and decentralised governance of the people more than in any communist strategy.

In West Bengal, the story is fairly clear. Except during 2001, the disproportionality index has remained at very high levels. Taken together with a low value of ENP, this indicates that power is entrenched within the hands of a select number of party candidates in the party’s stronghold. This explains the CPI(M)’s characteristic presence in the state. West Bengal’s low levels of social development indicators (Mukherjee et al 2014) compared to Kerala, despite being lauded as a socially progressive state, can be explained by the monopolistic tendencies of the CPI(M) during its rule when political competition was so low.

In fact, direct interpretations of the results support the view that West Bengal and Kerala, even though commonly identified as communist states, have significantly diverging “party formations” with stronger roots in civil society in Kerala and weaker in West Bengal (Desai 2001). The decentralised development in Kerala (Harriss 2000) can be explained by low DISPR and high ENP just as a low level of political incorporation of people in West Bengal (Mallick 1993) is evident from its high DISPR and low ENP. These values also show a deep rootedness of the CPI(M) government in West Bengal.

V-S Ratios assert this idea further. In a first-past-the-post system of electoral choice, vote shares do not influence government formation on the surface. In a representative democracy, a candidate winning by a 1% margin has equal weight in Parliament as another who has secured all the votes. How much a candidate is liked or disliked, then, does not matter. However, this degree of likability (how much people want a particular party) is extremely meaningful. In effect, vote share is a more accurate measure of desirability of a party, even if winners and losers are decided by seat share. The V-S Ratio hides within it a psychological mirror of the voters’ minds and holds crucial information on electoral mindset.

Forming the government with low vote share is not something unusual even though it signifies a weak mandate. But securing a rather low vote share simultaneously with considerably high seat share confirms the huge leakage in the mandate. The V-S Ratio for West Bengal is very high, three to four times that in Kerala. Figure 7 illustrates this for both states. The low V-S Ratios for Kerala appear in striking contrast with the overhanging curve of West Bengal. Kerala’s political structure also gives us a strong reason to doubt the seamless transfer of people’s aggregate preferences to Thiruvananthapuram, but then the reflection of the people’s choice was evident even in the seat shares. For most politicians, the focus on vote share has been rather blurry. Therefore, when seat shares fail to capture the people’s intent, it sends a wrong signal to the ruling party—as happened in Kolkata. We doubt that even the CPI(M) cadres have noticed this. If they had, they would have realised that their crushing defeats in 2011 and 2016 were simply time bombs ticking for the preceding four decades, leading to unbearably huge losses.

The AMS values bolster this narrative. AMS holds information on how strong (on average) the party is in its winning constituency. If the party wins in say just 10% of the total constituencies, AMS tells us how powerful the win was in those 10% constituencies. It does not tell us how strong the party is in the entire state or country. In other words, AMS captures information on the heaviness of victories and victories alone (without adjusting for losses). Thus, high AMS values of the winning party tell us that the party was highly superior to the next best party in the winning constituencies. Also, since AMS measures strength on average, it is biased in favour of fewer victories but those with high margins.

This implies that with high AMS, the party is in a psychologically complacent position. West Bengal winners have
considerably higher AMS values compared to those in Kerala. This is then reflective of highly entrenched left party dominance in West Bengal. Bhattacharyya (2010) calls West Bengal a case of party–society where the left parties have remained in permanent incumbency. By the early 1990s, it was clear that West Bengal was suffering from huge rural inequality and dismal economic realities in all sectors, including agriculture (Chattopadhyay 2005). Yet the patriarchal leadership lacked any will or imagination for social services to the poor (Beg 2011). Scholars have also shown how deeply entrenched clientelistic and patronage politics in West Bengal is, which also glues seats to power (Bardhan et al 2009; Ruud 1999). Panchayats in West Bengal came under partisan control
(Bhattacharyya 2010), and party cadres would adopt mafia-type extortion for the party, with “CPI(M) goons” becoming
a menace (Beg 2011: 85).

The deep-rooted irony in the neglect of social welfare of the population by a party that based itself on the platform of
social justice can be explained by looking at election data and at AMS figures in particular. If AMS is low, the party would need to be “on its toes,” so to speak, and must engage constructively in all its constituencies if it wants to retain its seats. This is the case in Kerala. The social apparatus in
Kerala, which prioritises social reform, was a result of the high level of political competition in the state more than the left’s influence. In fact, even the alternating failures of the left parties in the elections could not dilute the legacy of their
socially-oriented policies because the competition was high; the Congress could not have continued to be in the race if they had diverged from the democratic decentralisation that was constructed in Kerala’s social fabric since the pre-independence era (Harriss and Törnquist 2015). Heller (2005) has located the difference between Kerala’s social development model and predatory rent-seeking activities of North India on the demand side (the people) rather than the supply side (the
political parties). Note that in Kerala, the AMS rate for the winner as well as a small difference between AMS for the first- and second-ranking parties fluctuates, which signifies a strong electorate. The lack of hegemony in Kerala and its presence in West Bengal—which explains their differing models of development—is in turn explained by election data in the most persuasive form.

When we estimated the median values of marginal superiority (instead of average, we calculated the mean), we observed a negatively skewed curve (which is also somewhat evident in Figures 5 and 6) for both West Bengal and Kerala. This suggests that the average is heavy because of high-margin wins for a small number of seats. Clearly, the mandate was not as strong as it appeared, but party leadership was interested only in what it appeared to be. In Kerala, the skewness is not as clear. This ensured that the left parties in the state were not misled into believing that they have a stronghold. They had to work harder each time.

Read with results on ENP and DISPR, V-S Ratio is a powerful explanation of differential party structures and formations
regardless of their communist make-up on the surface. Desai (2001) in particular has addressed these differences. In our view, the differences in the two states are so massive, particularly from the point of view of the demand side of the political market (Heller 2005), that putting both the states in the same communist basket is unwise. Also, the dynamics on the same demand side, surfacing in the poll data, show that neither has a communist make-up in the manner scholars often fall into the temptation of assuming.


Elections are sensory experiences, inventory of preferences and expressive of political desires. Their variables cannot be stripped off from any serious political examination, for it is
in election figures that much of the truth about society’s aggregate value system emerges. In fact, the mighty literature on demo­cracy and political freedom and rights comes at a stark contrast with meaningful numerical analysis of election data that goes beyond the surface. It comes as no surprise then that when we construct such variables, we are able to discover visions of predictive and explanatory power with little effort.

The extent to which our findings support the literature on Kerala’s democratic development and decentralised polity on one hand, and patronage politics in West Bengal on the other, is noteworthy. Our departing points lie at the mainstream view of preference of the left rule in these states. While we do not intend to criticise the foundations of such a view, we assert that there may be some merit in viewing the policy in West Bengal and Kerala through a what-people-want lens.


1 The Congress did come to power in the sixth and seventh assembly elections (in 1971 and 1972), but it was clear that West Bengal was not its playground because even in 1971, the CPI(M) secured the highest number of seats.

2 Even though the AITC split from the Congress, it partnered with the Congress in 2001 in the United Progressive Alliance. Politics indeed makes strange bedfellows.

3 Note, however, that in 1967, the Congress had the largest vote share at 35.4%, while the CPI(M)’s vote share was a mere 23.5%.

4 Note that our study is based on party (rather than coalition) politics because it is the party that is the residual bearer of the people’s preferences directly during elections.

5 When the ENP is estimated using vote share (instead of seat share), we denote it by ENPv. Note that ENP and ENPv need not be the same unless vote share and seat share of all parties are the same. But they should move in the same direction. ENPv is given as follows:

6 Representative democracy converts an analogue preference of people into a binary “winning or losing” divide. This is not necessarily the ideal way to aggregate preferences. These shortcomings of the first-past-the-post system are commonly acknowledged and we do not intend to re-emphasise them (Kaushik and
Pal 2012). Our aim is to locate the blind spots in our imagination of politics in West Bengal and Kerala, and in doing so, steer the direction of popular narratives in a slightly different

7 The top three parties do not need to remain the same in our formulation.


Bardhan, Pranab, Sandip Mitra, Dilip Mookherjee and Abhirup Sarkar (2009): “Local Democracy and Clientelism: Implications for Political Stability in Rural West Bengal,” Economic & Political Weekly, Vol 44, No 9, pp 46–58.

Bardhan, Pranab, Sandip Mitra, Dilip Mookherjee and Anusha Nath (2014): “Changing Voting Patterns in Rural West Bengal: Role of Clientelism and Local Public Goods,” Economic & Political Weekly, Vol 49, No 11, pp 54–62.

Basu, Subho and Auritro Majumder (2013): “Dilemmas of Parliamentary Communism: The Rise and Fall of the Left in West Bengal,” Critical Asian Studies, Vol 45, No 2, pp 167–200.

BBC News (2011): “India: Mamata Banerjee
Routs Communists in West Bengal,” 13 May,

Bhattacharyya, Dwaipayan (2010): “Left in the Lurch: The Demise of the World’s Longest Elected Regime?” Economic & Political Weekly, Vol 45, No 3, pp 51–59.

Bhaumik, Subir (2011): “Defeat Rocks India’s
Elected Communists,” Al Jazeera, 13 May,
2011/ 05/2011513143311330487.html.

Bag, Kheya (2011): “Red Bengal’s Rise and Fall,” New Left Review, Vol 70, pp 69–98.

Chakrabarty, Bidyut (2011a): “The 2011 State
Assembly Election in West Bengal: The Left Front Washed-out!” Journal of South Asian
, Vol 6, No 2, pp 143–67.

— (2011b): “The Left Front’s 2009 Lok Sabha Poll Debacle in West Bengal, India: Prospective Causes and Future Implications,” Asian Survey, Vol 51, No 2, pp 290–310.

Chatterjee, Jyotiprasad and Suprio Basu (2009): “West Bengal: Mandate for Change,” Economic & Political Weekly, Vol 44, No 39, pp 152–56.

Chattopadhyay, Apurba Kumar (2005): “Distributive Impact of Agricultural Growth in Rural West Bengal,” Economic & Political Weekly, Vol 40, No 53, pp 5601–10.

Chhibber, Pradeep and Irfan Nooruddin (2004): “Do Party Systems Count? The Number of
Parties and Government Performance in the Indian States,” Comparative Political Studies, Vol 37, No 2, pp 152–87.

Desai, Manali (2001): “Party Formation, Political Power, and the Capacity for Reform: Comparing Left Parties in Kerala and West Bengal, India,” Social Forces, Vol 80, No 1, pp 37–60.

Deshpande, Ashwini (2000): “Does Caste Still
Define Disparity? A Look at Inequality in Kerala, India,” The American Economic Review, Vol 90, No 2, pp 322–25.

Devika, J (2010): “Egalitarian Developmentalism, Communist Mobilisation, and the Question of Caste in Kerala State, India,” The Journal of Asian Studies, Vol 69, No 3, pp 799–820.

Election Commission of India (nd): “Election
Results—Full Statistical Report,”

Gallagher, Michael (1991): “Proportionality,
Disproportionality and Electoral Systems,” Electoral Studies, Vol 10, No 1, pp 33–51.

Harriss, John and Olle Törnquist (2015): Comparative Notes on Indian Experiences of Social
Democracy: Kerala and West Bengal, Simons Papers in Security and Development
, No 39, School for International Studies, Simon Fraser University, Vancouver.

Harriss, John (2000): “Renewing Development and Deepening Democracy,” Frontline, Vol 17, No 17, https://www. fl1717/17170750.htm.

Heller, Patrick (1996): “Social Capital as a Product of Class Mobilization and State Intervention: Industrial Workers in Kerala, India,” World
, Vol 24, No 6, pp 1055–71.

— (2005): “Reinventing Public Power in the Age of Globalisation: Decentralization and the Transformation of Movement Politics in Kerala,” Social Movements in India: Poverty, Power,
and Politics
, Raka Ray and Mary Fainsod Katzenstein (eds), Oxford: Rowman and Littlefield, pp 79–106.

— (2007): “Kerala: Deepening a Radical Social Democracy,” Social Democracy in the Global
Periphery: Origins, Challenges, Prospects,
Richard Sandbrook, Marc Edelman, Patrick Heller, and Judith Teichman (eds), Cambridge: Cambridge University Press, pp 65–92.

Jeffrey, R (2001): Politics, Women and Well-being: How Kerala Became “a Model,” New Delhi:
Oxford University Press.

Kaushik, Arun Kumar and Yugank Goyal (2015): “Political Dynamics in Bihar: Electoral Competition and Representation,” Economic & Political Weekly, Vol 50, No 39, pp 18–22.

Kaushik, Arun and Rupayan Pal (2012): “How Representative Has the Lok Sabha Been?”
Economic & Political Weekly, Vol 47, No 19, pp 74–78.

Kohli, Atul (1990): “From Elite Activism to Democratic Consolidation: The Rise of Reform Communism in West Bengal,” Dominance and State Power in Modern India: Decline
of a Social Order
, Francine R Frankel and M S A Rao (eds), Vol II, Delhi: Oxford University Press.

Laakso, Markku and Rein Taagepera (1979): “The ‘Effective’ Number of Parties: A Measure with Application to West Europe,” Comparative Political Studies, Vol 12, No 1, pp 3–27.

Lindberg, Anna (2001): Experience and Identity: A Historical Account of Class, Caste, and Gender among the Cashew Workers of Kerala, 1930–2000, Scania, Sweden: Lund University.

Mallick, Ross (1993): Development Policy of a Communist Government: West Bengal since 1977, Cambridge: Cambridge University Press.

Menon, Dilip M (1994): Caste, Nationalism and Communism in South India: Malabar, 1900–48, Cambridge: Cambridge University Press.

Mukherjee, Sacchidananda, Debashis Chakraborty and Satadru Sikdar (2014): “Three Decades of Human Development across Indian States:
Inclusive Growth or Perpetual Disparity?” Working Paper Nos 2014–139, National Institute of Public Finance and Policy, New Delhi.

Nag, Moni (1989): “Political Awareness as a Factor in Accessibility of Health Services: A Case Study of Rural Kerala and West Bengal,” Economic & Political Weekly, Vol 24, No 8, pp 417–26.

Parayil, Govindan (2000): Kerala: The Development Experience: Reflections on Sustainability and Replicability, London: Zed Books.

Radhakrishnan, P (1989): Peasant Struggles, Land Reforms and Social Change: Malabar 1836–1982, New Delhi: Sage Publications.

Ramachandran, V K (1997): “On Kerala’s Development Achievements,” Indian Development: Selected Regional Perspectives, Jean Dreze and Amartya Sen (eds), New Delhi: Oxford University Press, pp 205–356.

Ruud, Arild Engelson (1999): “From Untouchable to Communist: Wealth, Power and Status among Supporters of the Communist Party (Marxist) in Rural West Bengal,” Sonar Bangla? Agricultural Growth and Agrarian Change in West Bengal and Bangladesh, Ben Rogaly, Barbara Harriss-White and Sugata Bose (eds), New Delhi: Sage Publications India Pvt Ltd, pp 253–78.

Sandbrook, Richard, Marc Edelman, Patrick Heller and Judith Teichman (2007): “Kerala: Deepening a Radical Social Democracy,” Social Democracy in the Global Periphery: Origins, Challenges, Prospects, Cambridge: Cambridge University Press, pp 65–92.

Updated On : 7th Sep, 2018


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top