coursework
1
What explains the changes in the level of electoral volatility in India?
Introduction
Since India became a democracy, the level of electoral volatility has
fluctuated with there being no consistent trend over time. This is
important as it makes it difficult to determine what is behind India’s level
of electoral volatility as if there was more of a clear trend it would be
easier to pinpoint factors. It is also necessary to come up with an
explanation for the changes in India’s electoral volatility because the
changes go from one extreme to the other and so it is important to
determine such a large disparity in electoral volatility. Furthermore, the
stability of a party system is important in terms of its consequences for
democracy. There are those such as Mainwaring (1999) who believe that
party system institutionalization is necessary albeit not sufficient for the
healthy functioning of democracy. Also Mainwaring and Zoco (2007: 157-
158) argue that systems with high volatility create greater uncertainty
about who will govern and political outsiders come in to power more
easily, increasing the risk of authoritarian leaders. While democracy in
India has remained stable despite its changes in electoral volatility, it is
still worth coming up with explanations for the changes, as they could be
useful in terms of consolidating democracy in other countries across the
world.
I will try to answer the question by looking at three different factors,
which are party system fragmentation, cleavage strength and the time
since becoming a democracy. I will look to see whether there is a
correlation between these factors and the level of electoral volatility in
India between 1952 and 2004 in terms of whether the data supports my
hypotheses. The essay will be structured by starting off with a literature
review, looking at what scholars have said about the impact of the factors
on electoral volatility generally and from this I will formulate my
hypotheses. The second section will look at data and measurement,
looking at why I have chosen the case and method of analysis as well as
2
explaining how the independent variables and dependent variables are
going to be measured. The third section will present and analyze the data,
determining whether or not the data supports the hypotheses and explain
why this is the case. In the final section I will come to a conclusion based
on the findings and the literature review.
Literature review
There is a lot of literature on the effect which party system fragmentation
has on electoral volatility, with the widespread assumption being that
there is a positive relationship between the two. Several scholars find that
an increase in party system fragmentation leads to an increase in
electoral volatility (Pedersen 1979; Bartolini and Mair 1990; Mainwaring
and Zoco 2007; Powell and Tucker 2014). This empirical evidence is partly
explained by the argument that voter choice is characterized to some
degree by randomness and so an increase in the number of competitors
means an increase in random choices and therefore electoral volatility
increases (Gherghina 2014: 24). But the dominant explanation is that
when there are more effective parties in a party system, the ideological
space between parties decrease and this means that voters are more
likely to shift from one party to another due to having more options close
to their ideological preference, thus increasing electoral volatility
(Pedersen 1979; Bartolini and Mair 1990; Mainwaring and Zoco 2007:
1962). This is a strong argument as it makes sense for voters to not
attach themselves strongly to a particular party if there are lots of like-
minded parties who they can feel equally attached to. However, it is not
always necessary for there to be a large number of effective parties for
there to be a low level of polarization as there can be a small number of
effective parties that are ideologically close together, albeit this tends not
to be the case. Mainwaring and Scully (1995 cited in Roberts and Wibbels
1999: 578) also argue that an increase in party system fragmentation
leads to an increase in electoral volatility due to there being a lack of
coherent legislative majorities which means that governments are more
likely to be unstable and thus support for the parties in government is
more likely to fluctuate. On the other hand, there are some scholars who
3
dispute the claim that an increase in party system fragmentation leads to
an increase in electoral volatility. Shair-Rosenfield (2008: 14) argues that
an increase in electoral volatility doesn’t necessarily require a large party
system as smaller party systems can still have a high level of electoral
volatility in specific electoral contexts, such as poor economic performance
that sees voters punish the parties in government. While it is important to
note that the positive relationship between party system fragmentation
and electoral volatility is far from inevitable, most of the empirical and
theoretical evidence suggests that this relationship exists.
H1: As party system fragmentation increases, electoral volatility
increases.
There is also a lot of literature on the effect that cleavage strength have
on electoral volatility, with the main argument being that stronger
cleavages help stabilize party systems. Many scholars argue that the
stronger the cleavages in a country, the lower the level of electoral
volatility will be (Lipset and Rokkan 1967; Bartolini and Mair 1990). Lipset
and Rokkan (1967) argue that party systems stabilize because voters
develop strong attachments to parties on the basis of their social location
and so this makes voters less likely to switch between parties. Bartolini
and Mair (1990) find empirical evidence to support this argument, finding
that high levels of working class organization, ethnic and religious
diversity tended to lower electoral volatility in Europe between 1885 and
1985. It is also argued that it is important to consider time when looking
at the effect that cleavages have on electoral volatility. Birnir (2007) finds
when looking at new democracies since 1945, that cleavages such as
language stabilize the vote immediately, whereas others such as race and
religion take more time and so a stabilizing effect is not evident in early
elections. On the other hand, a lack of cleavages in society leads to a
higher level of electoral volatility as parties lack stable support bases
(Evans and Whitefield 1993 cited in Gherghina 2014: 25). For example,
Shair-Rosenfield (2008: 10) suggests that there are limitations to
applying social cleavage arguments to countries in the post-industrial
global economy, pointing towards the differences between the industry-
4
defined social classes of the late 19th century and early 20th centuries in
Europe and the less cohesive structures of late 20th century democracies.
Furthermore, the expectation that cleavages will reduce electoral volatility
is based on the assumption that different parties will represent distinct
social groups, so when this doesn’t happen one would not expect
cleavages to lead to electoral stability (Madrid 2005: 3). So in party
systems where there are catch-all parties that seek to represent many
different groups in society, one shouldn’t expect there to be low levels of
electoral volatility. But at the same time, the number of parties
representing specific groups in society are in itself a reflection of cleavage
strength and so if there are strong cleavages in society, one would also
expect there to be lots of parties representing distinct groups.
H2: As cleavages become stronger, electoral volatility decreases.
Another factor that can affect electoral volatility is the time since a
country became democratic, with the argument being that electoral
volatility decreases over time. Converse (1969) argues that party systems
become more stable over time because the attachment between voters
and parties strengthens. The deepening attachment to parties over time is
explained by the fact that voters support parties for longer and have
greater exposure to elections as more are held. Furthermore, Mainwaring
and Zoco (2007: 161) argue that over time one might expect that parties
build a stable support base from different groups in society as they
routinize their appeals. This demonstrates the link between cleavage
strength and time, in terms of cleavages becoming stronger over time.
Empirically there is evidence to support the theory, with (Lupu and Stokes
2010: 94) pointing out that various cross-national studies consistently
reveal much higher electoral volatility in new democracies than in old
ones. But on the other hand, there are arguments against the theory of
electoral volatility reducing over time, in terms of the fact that this doesn’t
appear to happen in post-1978 democracies. The assumption that most
citizens are satisfied enough with their preferred parties to continue to
support them over time is wrong in many post-1978 democracies,
because a large number of citizens believe that parties persistently fail
5
them when they get in to government rather than delivering on their
promises (Mainwaring et al. 2006 cited in Mainwaring and Zoco 2007:
169). Because citizens are so dissatisfied with their parties, politicians
have little reason to remain loyal to party labels and thus are more likely
to defect to another party or indeed form new parties and so that explains
why electoral volatility is so high in post-1978 democracies (Zielinski et al.
2005 cited in Mainwaring and Zoco 2007: 169). On the other hand,
Mainwaring and Zoco (2007: 171) argue that in earlier democracies, the
elites relied on parties to win elections and as a result had incentives to
invest in party-building, leading to stronger attachments between voters
and parties, which explains the lower levels of electoral volatility in these
countries. Therefore, it is clearly important to consider when a country
became democratic when looking at the effect that time has on electoral
volatility. While the empirical evidence may be mixed when looking at the
effect time has on electoral volatility, Lupu and Stokes (2010: 95) argue
that there are problems with studies that claim to show that volatility
doesn’t decline over time, most notably they typically only look at a small
number of elections and party identification is something that increases
gradually. This demonstrates the importance of looking at a large number
of elections when looking at the effect of time on electoral volatility.
H3: As the time since a country became democratic increases, electoral
volatility decreases.
Data and measurement
I have chosen to conduct a small-N analysis looking at one country, which
is India. India has a population of 1.3 billion making it the world’s second
most populous country and the largest democracy in the world. India is
also an incredibly diverse country, with 29 states, 18 official languages
and many different religions and India’s large population, together with its
extreme diversity makes the survival of its democracy very impressive.
Kohli (2001: 1) also argues that the success of democracy in India is
impressive because it defies prevailing theories that specify preconditions
for democracy in terms of India not being a developed economy and
6
Indian businessmen and middle classes not fully controlling the country’s
politics. Furthermore, Adeney and Wyatt (2004: 1) point out that India is
a successful democracy, despite the fact that most other countries in
South Asia and indeed South-East Asia are not. This partly explains why
India is a country worth looking at, as it is clearly a very unique case.
Despite democracy in India remaining stable, the stability of the Indian
party system has greatly fluctuated over the years, which also explains
why India is a country worth studying. The lack of an obvious trend in
electoral volatility suggests that there are a number of factors that affect
the stability of the Indian party system. Also there is lots of available data
on all the elections in India, which means that it is a country that can be
analysed in good depth. The method of analysis will involve looking to see
whether there is a correlation between the independent variables and the
level of electoral volatility. Through looking at the correlation or lack of
correlation, I can determine whether the data supports the hypotheses
and also what independent variable is most important in explaining the
changes in electoral volatility.
The outcome variable electoral volatility will be measured using the
Pedersen index, looking at general elections between 1951 and 2004.
Electoral volatility is the degree of change in voting behaviour between
two consecutive elections. Pedersen developed the index in 1979 and it is
a widely used measure of electoral volatility, with Powell and Tucker
(2014: 124) pointing out there is near enough a consensus amongst
scholars that it is the best way of measuring electoral volatility. The
Pedersen index conceptualizes electoral volatility at an aggregate level as
opposed to an individual level. Bartolini and Mair (1990: 19) argue that
aggregate electoral volatility is better than individual level electoral
volatility in capturing electoral change between two consecutive elections,
due to the fact that is has the advantage of being generalizable which
makes it useful for looking at data over a long period of time. This is
particularly important for my study, as I look at the changes in electoral
volatility in India over a long period of time and so this makes the
Pederson index suitable. Another advantage of the Pedersen index is the
fact that it extends beyond voter’s values in providing explanations for
7
shifts in votes as it allows institutional and contextual factors to be
included such as cleavages (Gherghina 2014:55). Again this is important
for my study due to the fact that factors I look at in the case of India are
largely institutional and contextual. In terms of how the Pederson index is
calculated, it is done by looking at the vote shares of parties between two
consecutive elections. The index adds all absolute differences in the
electoral results of parties and given that wins for some parties
automatically equal losses for other parties, this total amount of change is
divided by two with the scale ranging from 0 to 100.
The independent variables I will be using to explain the changes in
electoral volatility in India are cleavage strength, party system
fragmentation and the time since becoming a democracy. Party system
fragmentation will be measured by looking at the effective number of
electoral parties in India at each election between 1951 and 2004. The
effective number of parties, introduced by Laakso and Taagepera in 1979,
is the most widely used measure of party system fragmentation and
refers to the relative strength of the parties in each system in terms of the
number of parties that are likely to affect political outcomes. High values
of ENP generally represent a diverse party system with control of
government shared by many political parties and low values represent a
party system where control of government is concentrated in the hands of
only one or a few political parties. I have chosen to look at the effective
number of electoral parties because in India a plurality voting system is
used for national elections, which produces a lack of proportionality
between votes and seats and so the vote share of each party is a more
accurate representation of the effective number of parties than seat
share. Cleavage strength will be measured by comparing the vote share of
national parties with state parties in India at general elections between
1951 and 2004. I have chosen to measure cleavages in this way because
Lipset and Rokkan (1967: 5) argue that parties reflect the main cleavages
of a given society. There are also those that go as far as arguing that
parties don’t just reflect cleavages, they also produce them through their
organizational coverage and pressure (Sartori 1969: 89 cited in Gherghina
2014: 12; De Leon, Desai and Tugal 2009). National parties tend to be
8
broader in their appeal, whereas regional parties represent more specific
groups and in the case of India state parties represent specific ethnic,
linguistic or religious groups and so their vote share is a good indicator of
how strong cleavages are. The higher the vote share for state parties and
the lower the vote share for national parties, the stronger the cleavages
are and vice-versa. Finally, the time since becoming a democracy will be
measured by looking at the number of years between each election and
the year in which India became democratic, which is 1950 when its
Constitution became effective.
Analysis
Fig.1. Electoral volatility and party system fragmentation in India 1951-
2004
Source: Volatility data- Niokolenyi, C. (2014) ‘Party System
Institutionalization in India’
(https://www.cambridge.org/core/services/aop-cambridge-
core/content/view/F1F4AB40A71BDF5BBD599F591CC982C4/9781107300
385c8_p189-211_CBO.pdf/party_system_institutionalization_in_india.pdf)
Fragmentation data- Kumar, S. (2010) ‘Regional Parties, Coalition
Government, and Functioning of Indian Parliament’
0
1
2
3
4
5
6
7
8
0 5 10 15 20 25 30 35 40 45 50
19 51
19 57
19 62
19 67
19 71
19 77
19 80
19 84
19 89
19 91
19 96
19 98
19 99
20 04
Ef fe ct iv e n um
be r of e le ct or al p ar ti es
P ed er se n v al ue
Election year
Electoral volatility
ENEP
9
(http://www.lokniti.org/pdfs_dataunit/publications2010/Journal%20of%2
0Parliamentry_studies_Sh[1]sanjay_kumar_article.pdf).
Fig.2. Electoral volatility and vote share of national parties v others 1951-
2004
Source: Volatility data Ibid. Vote share data- Adeney, K. (2017) ‘PSI in India’ Fig.3. Electoral volatility and the time since India became a democracy 1951-2004
0 10 20 30 40 50 60 70 80 90 100
0 5 10 15 20 25 30 35 40 45 50
19 51
19 57
19 62
19 67
19 71
19 77
19 80
19 84
19 89
19 91
19 96
19 98
19 99
20 04
P er ce n ta ge v ot e sh ar e
P ed er se n v al ue
Election year
Electoral volatility
National Parties
State Parties and Others
0
5
10
15
20
25
30
35
40
45
50
1 7 12 17 21 27 30 34 39 41 46 48 49 54
P ed er se n v al ue
Years since becoming a democracy
10
Source: Volatility data Ibid.
The party system fragmentation data overall doesn’t support my
hypothesis that an increase in fragmentation should lead to an increase in
electoral volatility. Initially there is some support for my hypothesis as
between 1957 and 1967 an increase in the effective number of electoral
parties produces an increase in electoral volatility in India. Also between
1967 and 1971 as the effective number of electoral parties decreases,
electoral volatility decreases, which further supports the predicted positive
relationship between the two. However, there is a large general increase
in effective number of electoral parties between 1977 and 2004, which
leads to a large general decrease in electoral volatility. Therefore, for
most of the period as party system fragmentation increases, electoral
volatility decreases and this clearly isn’t what I expected to see based on
my hypothesis.
The data on cleavage strength overall supports my hypothesis that an
increase in cleavage strength should lead to a decrease in electoral
volatility. Between 1957 and 1962 cleavages in India become weaker as
the vote share of national parties increases and the vote share of state
parties decreases. This produces an increase in electoral volatility, which
supports my hypothesis, as it is a negative relationship. However,
between 1962 and 1967 cleavages become stronger as the vote share of
national parties decreases and the vote share of state parties increases.
This produces an increase in electoral volatility, which doesn’t support my
hypothesis. But between 1971 and 1977 the data supports my hypothesis
because the vote share of national parties increases and the vote share of
state parties decreases, and at the same time there is a large increase in
electoral volatility. Furthermore, between 1980 and 2004 the vote share
of national parties generally decreases and the vote share of state parties
generally increases and this produces a general decrease in electoral
volatility. Therefore, the data does largely support my hypothesis on
cleavage strength, as there appears to be a negative relationship between
the strength of cleavages and electoral volatility in India.
11
The data on the time since becoming a democracy overall supports my
hypothesis that as the time since a country became democratic increases,
electoral volatility increases. Between 21 and 27 years of becoming a
democracy, electoral volatility significantly increases in India, which
clearly doesn’t support my hypothesis. However, the fact that between 30
and 54 years of becoming a democracy, electoral volatility generally
decreases. Therefore, while initially support for my hypothesis is
inconsistent, eventually time does lead to a reduction in electoral
volatility.
The data shows that cleavage strength and time explain the changes in
the level of electoral volatility in India. The increase in party system
fragmentation in India post-1980 doesn’t lead to an increase in electoral
volatility as my hypothesis predicted, due to cleavages becoming stronger
over time. Gowda and Sridharan (2007: 15) point out that as the Indian
National Congress centralized under the leadership of Indira Gandhi, many
marginalized leaders and groups felt underrepresented in the party and so
left to form rival parties, which they argue explains the increase in party
system fragmentation in India. For example, strong regional parties
emerged Andhra Pradesh, Assam and Tamil Nadu and parties deriving
support form Sikh religionists emerged in the Punjab. Kumar (2010: 89)
also notes the increase in support for regional parties in India since the
1980s, pointing out that various surveys reflect the general feeling that
the Indian National Congress could not cater to the specific needs and
problems of people living in specific regions or from a specific community.
Therefore, despite the number of parties increasing electoral volatility
decreases as the kind of parties that emerged were regional parties with
stable support bases. Another reason for this happening that is less
important is the 1985 anti-defection law. The law restricted party
switching and so led to an increase in the number of new parties and this
has helped reduce electoral volatility as it has discouraged splits in the
large parties and also the splinter parties that have been formed are
small, meaning they do not have a noticeable impact on the level of
electoral volatility (Suri 2013: 215; Nikolenyi 2014: 205). The large
12
increase in electoral volatility between 1971 and 1977, despite the
effective number of electoral parties decreasing can be explained by lots
of voters giving their Janata Party in order to remove the INC from power,
after the state of emergency between 1975 and 1977 which had
suspended democracy.
Conclusion
In conclusion, cleavage strength and time explain the changes in the level
of electoral volatility in India. The data doesn’t support my hypothesis on
party system fragmentation as a post-1980 increase in the effective
number of electoral parties leads to a fall in electoral volatility. On the
other hand, the vote share of state parties increases and the vote share of
national parties decreases post-1980 and so in this sense the fall in
electoral volatility supports my hypothesis on cleavage strength. While
initially time doesn’t have a clear effect on electoral volatility as it
increases and decreases, the eventual fall suggests support for my
hypothesis. The explanation for this is that although stronger cleavages
may lead to a more fragmented party system, they stabilize party support
as voters develop strong attachments to party based on a specific
cleavage and so are less likely to vote for different parties. Time also
plays a role as it is thought that cleavages strengthen over time, which is
the case in India. Also naturally party systems become more stable over
time because the attachment between voters and parties strengthens due
to voters familiarizing themselves with different parties. This contradicts
previous findings on the effect that party system fragmentation has on
electoral volatility, which perhaps have looked at cases where the increase
in fragmentation wasn’t caused by cleavages becoming stronger. A
possible limitation to my study is that I didn’t look at a wide enough range
of factors that could explain the changes in electoral volatility in India.
Other possible factors I could have included are the effect of personality
politics and party organization, but due to the constraints of the word
count this wasn’t possible. Another possible limitation is the fact that I
didn’t carry out a logistic regression that would have provided more
detailed information on the relationship between the independent
13
variables and the dependent variable. Furthermore, in terms of looking at
whether the relationships between the variables are true more generally
as opposed to just India, this would have required me to look at a wider
range of countries because of course India could just be an exception.
Word Count: 3,776
14
Bibliography
Adeney, K. and Wyatt A. (2004) ‘Democracy in South Asia: Getting beyond the Structure-Agency Dichotomy’, Political Studies, 52/1: 1-18. Adeney, K. (2017) ‘PSI in India’, lecture notes, University of Nottingham, delivered 14 March 2017. Bartolini, S. and Mair,P. (1990) Identity, Competition and Electoral Availability Cambridge: Cambridge University Press. Birnir, J.K. (2007) ‘Divergence in Diversity? The Dissimilar Effects of Cleavages on Electoral Politics in New Democracies’, American Journal of Politics Science, 51/3: 602-619 Converse, P. (1969) ‘Of time and Partisan Stability’, Comparative Political Studies, 2/2: 139-171. De Leon, C., Desai, M. and Tugal, C. (2009) ‘Political Articulation: Parties and the Constituion of in the United States, India and Turkey’, Sociological Theory, 27/3: 193-219. Gherghina, S. (2014) Party Organization and Electoral Volatility in Central and Eastern Europe: Enhancing Voter Loyalty London: Routledge. Gowda, M.V.R and Sridharan, E. (2007) ‘Parties and the Party System, 1947-2006’, in Ganguly, S., Diamond, L. and Plattner, M.F. (eds.) The State of India’s Democracy, Maryland: John Hopkins University Press, pp. 3-25. Kohli, A. (2001) ‘Introduction’, in Kohli, A. (ed.) The Success of India’s Democracy, Cambridge: Cambridge University Press, pp. 1-20. Kumar, S. (2010) ‘Regional Parties, Coalition Government, and Functioning of Indian Parliament: The Changing Patterns’, Journal of Parliamentary Studies, 1: 75-91. Lipset, S.M. and Rokkan, S. (1967) ‘Cleavage Structures, Party Systems, and Voter Alignments: An Introduction’, in Lipset, S.M. and Rokkan, S. (eds.) Party Systems and Voter Alignments: Cross-National Perspectives, New York: Free Press, pp. 1-64. Lupu, N. and Stokes, S. (2010) ‘Democracy, interrupted: Regime change and partisanship in twentieth century Argentina’, Electoral Studies, 29/1: 91-104.
15
Madrid, R. (2005) ‘Ethnic Cleavages and Electoral Volatility in Latin America’, Comparative Politics, 38/1: 1-20. Mainwaring, S. (1999) Rethinking Party Systems in the Third Wave of Democratization: The Case of Brazil Stanford: Stanford University Press. Mainwaring, S. and Zoco, E. (2007) ‘Political Sequences and the Stabilization of Interplay Competition in Old and New Democracies’, SAGE Publications, 13/2: 155-178. Nikolenyi, C. (2014) ‘Party System Institutionalization in India’, in Hicken A. and Kuhonta, E.M. (eds.) Party System Institutionalization in Asia [e- book], Cambridge University Press, pp. 189-211. Available at: https://www.cambridge.org/core/services/aop-cambridge- core/content/view/F1F4AB40A71BDF5BBD599F591CC982C4/9781107300 385c8_p189-211_CBO.pdf/party_system_institutionalization_in_india.pdf (Accessed 05 May 2017). Pedersen, M.N. (1979) ‘The Dynamics of European Party Systems: Changing Patterns of Electoral Volatility’, European Journal of Political Research, 7/1: 1-26. Powell, E.N and Tucker, J.A. (2014) ‘Revisiting Electoral Volatility in Post- Communist Countries: New Data, New Results and New Approaches’, British Journal of Political Science, 44/1: 123-147. Roberts, K.M. and Wibbels, E. (1999) ‘ Party Systems and Electoral Volatility in Latin America: A Test of Economic, Institutional, and Structural Explanations’, American Political Science Review, 93/3: 575- 590. Shair-Rosenfield, S. (2008) Assessing the Causes and Effects of Electoral Volatility: Party System Fragmentation, Time and Executive Turnover Arizona: The University of Arizona Press. Suri, K.C. (2013) ‘Party System and Party Politics in India’, in Suri, K.C. and Vanaik, A. (eds.) Indian Democracy [e-book], Oxford University Press, pp.210-252. Available at: http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780198084 952.001.0001/acprof-9780198084952-chapter-4?print=pdf (Accessed 08 May 2017).