Article

The BJP’s Welfare Schemes: Did Studies in Indian Politics

7(2) 219–233, 2019
They Make a Difference in the © 2019 Lokniti, Centre for the
Study of Developing Societies
2019 Elections? Reprints and permissions:
in.sagepub.com/journals-permissions-india
DOI: 10.1177/2321023019874911
journals.sagepub.com/home/inp

Rajeshwari Deshpande1
Louise Tillin2
K.K. Kailash3

Abstract
In this article, we use data from the 2019 NES post-poll survey to assess the impact of BJP’s welfare
schemes on voting behaviour. We demonstrate that compared to earlier elections, voters are more
likely to give credit to the central government as opposed to state governments or local politicians
for welfare schemes. This centralization is especially the case for some of the BJP’s new welfare pro-
grammes such as Ujjwala and the Jan Dhan Yojana. However, even earlier Congress-era schemes such
as the Mahatma Gandhi National Rural Employment Guarantee Scheme and the Awas Yojana are now
more associated with the central government. Schemes such as the Public Distribution System (PDS)
and Old Age Pensions are still more likely to be associated with state governments. At the all-India
level, we find some evidence that voters who received benefits under Ujjwala, Jan Dhan Yojana or Awas
Yojana schemes were more likely to vote for the BJP, whereas recipients of pensions or MGNREGA
were less likely to support the BJP.

Keywords
Welfare schemes, Ujjwala, BJP vote, NES 2019, centralization of credit, programmatic politics

The scale of the BJP’s victory in the 2019 Lok Sabha elections at a time of faltering economic growth,
rising agrarian distress, and increasing unemployment has led most observers to assume that voters
supported the BJP despite their material interests. The dominance of national security and strong
leadership as campaign themes, the more overt deployment of appeals to the majority Hindu community
compared to the 2014 campaign, the BJP’s ability to saturate the public sphere with sympathetic media
reporting and a massive imbalance in campaign finance have all loomed larger in ex-post analyses than

1
Savitribai Phule Pune University, Pune, India.
2
King’s India Institute, King’s College London, England.
3
University of Hyderabad, Hyderabad, India.

Corresponding author:
Rajeshwari Deshpande, Savitribai Phule Pune University, Pune, Maharashtra 411007, India.
E-mail: [email protected]
220 Studies in Indian Politics 7(2)

material factors in explaining the vote. However, this was also an election in which welfare schemes—
both claims of credit for and promises of—loomed large.
The first term of the Modi administration (2014–2019) saw the launch of several flagship schemes
(such as Ujjwala Yojana; Beti Bachao, Beti Padhao; Swachh Bharat; Jan DhanYojana; and Ayushman
Bharat) and rebranding of others (Pardikar, ‘Art of Political Branding’). These and existing schemes
were also integrated with Aadhaar. PM Narendra Modi and his government relentlessly claimed credit
for these initiatives throughout their first term of office. In the run-up to the 2019 elections, the central
government also launched a new cash transfer directed at farmers—PM-Kisan—in response to the rising
agrarian distress that was perceived to have contributed to the BJP’s defeat in the state elections in late
2018.4 In return, the pledge to introduce a new basic income scheme, Nyuntam Aay Yojana (NYAY) was
the centrepiece of the Congress manifesto. In many states, regional parties also focussed on welfare
schemes in their election campaigns.
We know very little about how such existing—and promised—welfare schemes affect voting decisions.
It is incontestable that the Indian state continues to function in a discretionary rather than rule-bound
fashion when it interacts with citizens. Local intermediaries or brokers play a crucial role for many
citizens in enabling citizens to access services or programmes which, in theory, they should be able to
access as a right (Auerbach & Thachil, 2018; Bussell, 2019). The substantial role of discretion helps
to sustain the notion that India is a ‘patronage democracy’ in which voters make decisions in the
expectation of securing future access to state resources (Chandra, 2004). However, a growing literature
contests the thesis of a quid pro quo exchange of benefits for votes. If quid pro quo exchanges are
happening, brokers need to be able to monitor how voters behave. Yet research shows that brokers
struggle to do so (Chhibber & Verma, 2018, p. 120; Schneider, 2016). Urbanization and population
growth make it harder for party workers and brokers to discern voters’ actions and make it more efficient
for political parties to rely less on party machines and more on programmatic forms of political
communication or unmediated appeals to voters by political leaders (Stokes, Dunning, Nazareno, &
Brusco, 2013, p. 21).
Furthermore, even if the discretionary disbursement of benefits to party supporters continues,
patronage itself is not sufficient to win elections. For this reason, political leaders have become more
inclined to adopt what Manor (2010) describes as ‘post-clientelist’ schemes, which are implemented
more effectively and are not directed along partisan lines. Studies in several states have pointed to the
growing importance of such ‘post-clientelist’ or more ‘programmatic’ policies in explaining election
outcomes (Elliott, 2011; Sinha, 2016; Wyatt, 2013).
Contrary to the earlier readings of Indian democracy as a patronage democracy, the NES tradition of
empirical studies have not found much evidence of clientelism as a determining factor in how Indians
vote. Instead, NES surveys have reported the increasing relevance of governance and performance of the
governments as explanatory factors in Indian elections (Verma, 2012; Yadav & Palshikar, 2014). Other
studies have also pointed to the emerging significance of issue-based voting, or economic voting in
recent elections (Suri, 2009; Vaishnav & Swanson, 2015).Yadav and Palshikar (2014) discussed a
gradual shift from identity politics to that of politics of governance since 2004, which coincided with the
UPA’s agenda of programmatic politics. At one level, the shift occurred in the context of the waning of
social cleavages based on religion and caste in the decade of the 2000s. In the period since then, political
parties have been unable to activate these cleavages further and to keep them at the centre stage of
politics. Even at the state level, political parties had very little success in building cleavage politics on an
enduring basis. Yadav and Palshikar argued that the receding politics of social cleavages forced political

4  
See, for instance, Hindustan Times (2018).
Deshpande et al. 221

parties to adopt ‘catch-all’ policies. There is also a recognition by scholars that Indian elections take
place around multiple issue-axes and Indian voters should not be seen as single-issue voters (Sinha,
2016).
Rather than indicating the resilience of clientelism or contingent exchanges of benefits for votes (and
vice versa), the prominence of welfare schemes in Indian elections since the 2000s raises the question as
to whether voters retrospectively reward governments or political parties for policies that have provided
them with a material benefit or whether they weight the potential benefits they could receive from a
different party when making voting decisions. Policy-related retrospective voting is unlikely to be the
determining factor driving election outcomes. However, it could make a difference in informing voter
assessments about the overall performance of governments, especially among certain groups of voters.
There is evidence from Latin America, for instance, that the new generation of social policies introduced
in countries such as Brazil and Mexico since the early 2000s have had an electoral impact, even if they
have not caused long-lasting political realignment (De La O, 2013; Hunter & Power, 2007; Zucco,
2013).
Making this kind of argument is complex in a multi-level polity such as India’s where chains of
electoral accountability are opaque. Voters may struggle to identify, which level of government is
responsible for which policy. If a voter cannot correctly identify chains of accountability, we might ask
whether they can reward/punish the correct party or tier of government. The constitutional set-up further
intensifies this problem, where responsibility for some areas of social policy is shared between the
centre and the states. Furthermore, even where the central government initiates programmes, or
predominantly funds them, state governments—or panchayats in the case of a policy like MGNREGA—
may be the primary administrative tier responsible for service delivery. In the past, this has allowed state
governments to steal the credit for policies initiated and primarily funded by the central government
(Tillin & Pereira, 2017).
In this article, we look at evidence from the NES post-poll survey on welfare and voting in the 2019
Lok Sabha elections. The post-poll survey was conducted in 26 states after polling (12 April to 21 May
2019). A total of 24,236 voters were surveyed in 211 parliamentary seats.5 The survey included a series
of questions on welfare schemes, some of which have been asked in previous national and state election
surveys providing comparative historical data. Using this data, we ask three principal questions: (a) How
many people benefit from government welfare programmes? (b) Who do beneficiaries credit for welfare
schemes: central government, state government or local officials? (c) Is there any evidence of retrospective
voting?
We find first that the welfare schemes introduced by the BJP since 2014 have achieved relatively high
levels of coverage—34 per cent of respondents had received gas cylinders under the Ujjwala6 scheme,
for instance. Furthermore, some schemes introduced by previous governments have also achieved
extensive coverage. The PDS is most notable—nationally an average of 44 per cent of respondents
accessed the PDS, but, at the state level, the proportion reached between 80 and 90 per cent of respondents
in Andhra Pradesh, Telangana and Odisha—all states ruled by regional parties.
Second, we uncover a pronounced centralization in the attribution of credit for welfare schemes by
beneficiaries in 2019 compared to 2014. Compared to earlier elections, there has been a substantial
centralization in the likelihood that voters give credit to the central government (as opposed to state

5  
For full details of the survey methodology see https://www.lokniti.org/media/PDF-upload/1565073104_34386100_method_pdf_
file.pdf. (accessed July 20, 2019).
6  
The Pradhan Mantri Ujjwala Yojana was launched in 2016 to provide free liquefied petroleum gas (LPG) cooking connections
to poor women.
222 Studies in Indian Politics 7(2)

governments or local politicians) for welfare schemes. However, there is variation across schemes, with
voters more likely to credit state governments with responsibility for agricultural loan waivers, pensions
and PDS—all schemes at which state governments have been at the forefront of funding and
implementation. The central government receives the lion’s share of credit for other programmes,
reflecting the aggressive centralization of credit claiming for welfare programmes by the Narendra Modi
government since coming to office in 2014. Interestingly, the central government also received a larger
proportion of the credit for the MGNREGA—a UPA-era scheme—than in previous elections. This may
suggest that voters are becoming more aware of the chains of electoral accountability for different
schemes.
Third, we look at the most difficult question: whether social policy has an impact on the way people
vote. Specifically, we look for evidence of whether voters rewarded the BJP for their receipt of welfare
schemes. We find evidence in NES 2019 to support the claim that the BJP received some benefit for its
new welfare schemes. Beneficiaries of schemes that were more associated with state governments were
somewhat less likely to vote for the BJP. However, these findings should be interpreted carefully. Given
the number of schemes in operation regionally and nationally, more research is necessary to understand
how voters interpret the very complex web of accountability for different government activities.
Furthermore, receipt of welfare benefits was unlikely to have been a determining factor in people’s
voting decisions. The evidence we discuss here, however, suggests that it may not have been an
immaterial part of the connection between voters and the BJP.

The BJP and Welfare

The BJP’s position on welfare is unlike any other political party in the country and on the face of it
appears to be founded on a set of contradictory ideas. On the one hand, the party is highly critical of the
type of welfare model adopted and practised in India. On the other hand, the party has also not abandoned
the idea that governments should provide some form of welfare to the people. Its campaign and
mobilization rhetoric beginning with the 2014 elections continuing through the term of the NDA-II and
the 2019 campaign itself has revolved around these two precepts.
For the BJP, it is not politically prudent to do away with state-provisioning of welfare, and it has had
to find some space in an already crowded market. From two rounds of the State of Democracy in South
Asia (SDSA) (De Souza, Palshikar, & Yadav, 2008; Shastri et al., 2017), we know that most people not
only in India but across South Asia associate welfare and justice as integral dimensions of democracy.
Nearly everyone regardless of education and class linked democracy with some ‘substantial outcome in
terms of equality and welfare’. Furthermore, when people were asked about the most essential element
of democracy, people underlined welfare, unlike in say the industrialized West, where the procedural
dimensions of democracy were prioritized.
Given the large numbers of poor people, and this particular understanding of democracy, welfare and
the poor are a strong component on the agenda of almost all political parties. Moving away from the poor
and welfare would not only cede space to political opponents, but it would be the ‘equivalent of political
suicide’ (Chhibber & Verma, 2018). By 2014, the BJP had learnt from its experience. The Congress, the
BJP’s primary challenger, has almost always campaigned on a pro-poor and development agenda. In
2004, when the BJP pitched its tent on the ‘India Shining’ platform, the Congress batted for the underdog.
Without debunking the economic reforms, the Congress brought the ‘aam aadmi’ (common man) to the
foreground and pushed for a more humane agenda revolving around the idea of inclusiveness (Arora &
Deshpande et al. 223

Kailash, 2014). Consequently, the BJP has put great effort to both signal that it is in the welfare game but
more importantly, that it has a slightly different approach compared to the others.7
With its broader goal of a Congress-mukt Bharat, the BJP has challenged the Congress on multiple
fronts including the welfare dimension. The BJP framed its pro-welfare campaign around the idea that
the developmental strategy of the Congress was weak, not pro-poor and that it dulled the aspirations of
the people, especially the young. The campaign compared the past (Congress rule) with the current
government to create an image that development in real terms is only taking place now. By speaking of
welfare like never before, the BJP was going into the Congress terrain. It was a vote-maximizing strategy
aimed to squeeze the votes of its opponents.
Redistribution and equalizing opportunities that would allow for a greater social mobility have not
been the forte of the BJP. Traditionally, for the party, welfare meant economic investment geared for
economic growth and fiscal prudence. The party has tweaked this position slightly to adapt to the
competitive environment. To position itself as a welfare agent, the BJP has argued that welfare
programmes were being systematically abused and there is a need to direct them to the most needy. Here,
the party was implying that the previous distribution mechanisms were faulty and the current government
has set it right.
The two points that invariably appeared in the 2019 campaign of the BJP included, first that welfare
distribution under the Congress regime was expensive and undeserving sections of society had
appropriated the benefits. Second, the Congress favoured particular sections of society and did not treat
all citizens equitably. The party consequently argued that the welfare conception and distribution has to
be reworked to something that approaches near-universality so that ‘everybody’ benefits from the
development. It is around this idea that the party campaign slogan of sab ka sath, sab ka vikas revolves.
It has gone to great lengths to underline the point that the use of technology and direct benefit transfers
has helped weed out fake beneficiaries who were receiving various subsidies. This line is very similar to
the European right, who too aim to ‘reclaim’ those services from the ‘bureaucrats’ and ‘welfare
scroungers’ that ‘abuse them’ at the cost of their ‘rightful owners’: the common man who is falling on
hard times (de Koster, Achterberg, & Van der Waal, 2013). The government claimed that when non-
existent beneficiaries were removed, public money was being saved and it also showed the corruption
free nature of this government.
Under Modi, the use of Aadhaar (biometric identification) to identify beneficiaries, which began
under the UPA, has intensified. Aadhaar was intended to promote access to government welfare schemes
by providing a unique identification number to every Indian, as well as weeding out ‘fake’ beneficiaries
via ‘de-duplication’ of beneficiaries. There is evidence (including from the 2019 NES post-poll survey)
that Aadhaar has had the unintended consequences of generating new exclusions as a result of technical
errors and over-zealous implementation (see, for instance, Khera, 2015). However, since 2014, Aadhaar
has been presented as part of a package of technology-enabled reforms to governance that is helping to
eradicate corruption and improve the delivery of government programmes to the poor. The Modi
government has also focussed fastidiously on top-down management of programme implementation of
its flagship new schemes in priority districts (Aiyar & Tillin, 2019).
During his campaign rallies in 2019, the prime minister made the point that his government was
committed to strong and equitable development and is ‘dedicated to poor and all schemes are being

7  
The role of the BJP’s affiliates in the Sangh Parivar in providing social welfare was also an important part of the party’s ability to
break through in new regions in the 2000s (Thachil, 2014).
224 Studies in Indian Politics 7(2)

formulated to give maximum benefits to this lot’.8 In the election campaign of 2019, the BJP thus
attempted a re-positioning of the PM as a leader of the poor rather than mere VikasPurush of the 2014
version. The election outcomes suggest that these strategies paid off for the BJP when it consolidated its
electoral support in these elections. In this article, we try to understand with the help of our empirical
explorations of the NES data, whether and to what extent its new welfarist agenda contributed to the
victory of the BJP.

Trends from the NES So Far

The National Election Studies has routinely conducted detailed enquiries on the extent and perceptions
of social welfare schemes since 2009. These enquiries have given us three-pointers about the nuanced
connection between welfare and vote. First, levels of voter awareness of social welfare schemes have
been increasing. In the 2009 Lok Sabha elections, around 40 per cent voters were aware of the Mid Day
Meal Scheme of the UPA and 30 per cent voters from the poor, and very poor sections had heard of the
National Rural Employment Guarantee Scheme. Three among every four of parents of school-going
children were aware of the Mid Day Meal Scheme. At that time, the proportion of beneficiaries of these
schemes was much smaller, though not insubstantial (Yadav & Palshikar, 2014, pp. 90–92). As we report
below, the extent and awareness of welfare schemes continued to grow over the years and underlined the
importance of the distributive dimension of Indian democracy.
Second, there was some evidence of a weak connection between welfare benefits and vote for the
ruling party. In 2009 elections, the Congress party had a net advantage of over 15 per cent votes over
the BJP among those voters who had heard of the UPA flagship welfare schemes and had benefitted from
them. Even among those who had heard of the schemes but did not benefit the Congress party enjoyed
around 5 per cent net advantage over the BJP (Yadav & Palshikar, 2014, p. 93). Beneficiaries of NREGA
voted more for the Congress party than those who did not benefit. The same was the case with schemes
like Mid Day Meal and Old Age Pension. On the basis of his analysis of the 2009 NES data and that of
state assembly election studies during the years 2011–2012, Kumar (2013) made two observations. First,
the ruling party or alliance generally tended to gain electoral support from welfare scheme beneficiaries.
Second, it was state-level ruling parties rather than the central, which were better able to appropriate this
support. Thus, the earlier advantage to the Congress among welfare beneficiaries gradually disappeared
in the 2014 Lok Sabha elections as regional parties claimed credit for social service delivery. The Congress
could still retain a slightly more than its average support among the National Rural Health Mission
Scheme (NRHM) and Pension scheme beneficiaries. However, the NREGA beneficiaries preferred the
regional parties over the Congress and the BJP in 2014 (data provided by the CSDS [Centre for the Study
of Developing Societies] data unit).
The pattern of voters crediting state governments for welfare benefits was also evident in state
assembly elections after 20149 (Kumar, Deshpande, & Kailash, forthcoming). Surveys across the cycle
of state elections underscored that states continue to matter in understanding the relationship between
welfare and vote. However, in terms of electoral advantage to the ruling party, welfare politics had only
a moderate impact in these elections. The impact of welfare programmes on election outcomes was

See https://english.newstrack.com/lead/pm-narendra-modi-says-he-has-shut-shops-of-middlemen-4261.html
8  

The studies were supported by ESRC grant (ES/K005936/1), the ICSSR Research Program Grant (G-27/ICSSR/2016-17/PRS)
9  

and by the S P Pune University under its Universities with Potential for Excellence (UPE) scheme. We gratefully acknowledge
their support.
Deshpande et al. 225

mediated by the nature of party competition in which BJP’s entry into the region as a new political player
also mattered. Third, the impact of welfare on election outcomes was also affected by the record of
implementation. There was a definite advantage to ruling parties among welfare beneficiaries in states
where strong welfare regimes existed and where the presence of the BJP was marginal. On the other
hand, in states like Uttar Pradesh, rolling out social welfare schemes did not help the incumbent ruling
part(ies) win the state elections. Our analysis of the survey data from these studies suggested that
the citizens’ level of satisfaction with the state government and more importantly their assessment of the
overall performance of the state governments seem to have mattered most in deciding the nature of vote
of welfare beneficiaries.
The National Election Studies of 2019 confirmed and strengthened some of these earlier observations.
However, they also marked a shift in terms of attribution of credit for social welfare policies. So far, in all
our earlier studies, respondents mainly credited the state governments for the overall implementation of
the welfare policies. In 2019, on the other hand, as we show below, voters were more likely than
in the past to hold the central government responsible for many welfare schemes. Moreover, voters were
also able to distinguish more accurately between schemes and have attributed credit to respective state
and central agencies accordingly.

Analysis of the 2019 NES Data

We have already noted how the scope of welfare schemes has expanded over the years to cover a large
number of beneficiaries. Among the new schemes introduced by the NDA government, Ujjwala and Jan
Dhan Yojana had a reasonably decent coverage whereas the schemes for farmers had a moderate reach.
As reported in Table 1, the schemes like Awas Yojana or Pension scheme, which were introduced by
earlier Congress governments underwent considerable expansion during the past 5 years. Besides, large
sections of people availed the benefits of the PDS, one of the oldest social welfare schemes implemented
in the country.
An exception to this trend is that of the employment guarantee scheme under MGNREGA. There is a
5 percentage point decrease in the number of beneficiaries under this scheme during the last decade
despite the rural employment crisis looming large. In total, 33 per cent of the rural respondents of
the NES 2019 feel that, compared to earlier days, fewer jobs are available under the employment

Table 1.  Extent of Benefits Under Different Welfare Schemes

Percentage of Beneficiaries
Scheme 2019 2014 2009
Housing scheme (Awas Yojana) 21 15 15
MGNREGA 21 20 26
Pension scheme (old age, widows, disabled, etc.) 25 17 21
Public Distribution System 44 – –
Ujjawala Yojana 34 – –
Ayushman Bharat/National Rural Health Scheme (2009, 2014) 17 21 14
Jan Dhan Yojana 22 – –
Agricultural loan Waiver 13 – 18
Income support scheme for farmers 13 – –
Source: Compiled by the Lokniti, CSDS data unit based on the NES 2009, 2014, and 2019.
226 Studies in Indian Politics 7(2)

guarantee scheme now. It connects with the initial reluctance of the NDA government to take ahead the
MGNREGA. Moreover, the scheme has a sizeable coverage of more than 20 per cent of the population.
The percentage of beneficiaries of MGNREGA increases only marginally among the poor sections
(22%). Table 2 sets out the social profile of beneficiaries. This data suggest relatively wide access to
welfare schemes, including a significant level of access among richer groups to welfare schemes. We see
considerable variation in the state-wise patterns of implementation of this scheme. In states such as
Chhattisgarh, Uttarakhand, Meghalaya, Telangana and West Bengal, more than 40 per cent of the
population benefitted from the employment guarantee scheme (Table 3). On the other hand, its coverage
remained very limited in BJP-ruled states such as Maharashtra, Gujarat and Uttar Pradesh.
The state-wise variations in the overall implementation of welfare schemes indicate the critical role
that state governments continue to play as mediating agencies. As we have argued in our earlier work,
the nature of regional welfare regimes is decided by a plethora of political, economic and social factors
that shape the local constellations of power (Tillin, Deshpande, & Kailash, 2015). These constellations
seem to influence the ways in which different welfare schemes are prioritized by state governments.
In the post-2014 phase, patterns of party competition and the consequent relationship between centre
and state(s) appear to be important factors in the selection of schemes at the state level. Thus, the coverage
under the popular central scheme Ujjwala is very low in states such as Andhra Pradesh, Punjab, Telangana
or West Bengal that are ruled by non-BJP parties. However, these states seem to have taken a more active
role in implementing schemes such as MGNREGS, PDS and pension schemes where they retain agency
to themselves. The BJP, it seems, was keen to push the central schemes in states under its rule but more
than that in states where it wanted to consolidate its dominance. Thus, we see substantial coverage under
Ujjwala and Jan Dhan in states like Rajasthan and Madhya Pradesh. The historically strong welfare
regimes such as Tamil Nadu, Kerala and Odisha, which had a record of active engagement with welfare
politics, have also implemented select central and state-level schemes in a substantive manner. By

Table 2.  Social Profile of Beneficiaries

Ujjwala MGNREGA Pension Jan Dhan

Overall 34 21 25 22
Class
Poor 32 33 28 27
Lower 35 34 35 34
Middle 22 22 24 24
Rich 12 11 14 14
Caste
Upper castes 19 17 20 21
Upper OBC 21 20 23 22
Lower OBC 14 12 13 12
Dalits 16 18 14 16
Adivasis 11 12 09 09
Muslims 12 11 12 11
Gender
Male 53 52 52 54
Female 47 48 48 46
Locality
Rural 80 90 74 78
Urban 20 10 26 22
Source: NES, 2019 data made available by CSDS Data Unit.
Deshpande et al. 227

Table 3.  State-wise Variation in the Implementation of Welfare Schemes (state-wise percentage of
beneficiaries)

Ayushman
State Name Ujjwala Awas MGNREGA Pension PDS Jan Dhan Bharat
Andhra 20 13 38 59 88 22 08
Assam 24 12 24 11 09 06 23
Bihar 33 20 09 19 39 20 13
Gujarat 33 16 04 10 31 25 19
Haryana 23 07 08 31 34 15 12
Himachal Pradesh 14 10 35 22 45 12 20
Jammu and Kashmir 27 10 19 14 30 19 04
Karnataka 34 28 07 45 34 25 13
Kerala 45 15 26 43 73 04 16
Madhya Pradesh 62 38 28 34 50 36 22
Maharashtra 20 11 06 03 08 08 04
Meghalaya 04 18 42 08 10 00 15
Odisha 59 20 13 28 91 25 07
Punjab 10 05 12 18 18 06 11
Rajasthan 54 34 36 52 37 50 26
Tamil Nadu 29 44 33 34 46 24 37
Uttar Pradesh 36 16 13 16 46 24 15
West Bengal 26 20 40 11 52 17 19
Jharkhand 61 47 34 32 45 45 43
Chhattisgarh 45 18 49 14 66 19 36
Uttarakhand 32 06 45 38 57 31 22
Telangana 26 14 43 48 81 29 10
All India average 34 21 21 25 44 22 17
Source: NES, 2019 data made available by CSDS Data Unit.

contrast, BJP governed states such as Maharashtra, Gujarat or Haryana retain their weak to moderate
character as regional welfare regimes. In other words, despite the centralization of welfare politics, we
see a significant variation in the implementation of welfare programmes at the state level in the past 5
years. This variation is likely to be an important factor in understanding the relationship between welfare
and vote.
Despite the state-wise variation in the implementation of different welfare schemes, the NES 2019
depicts a clear trend in terms of attribution of credit to different government agencies. As shown in Table 3,
schemes such as Ujjwala, Jan Dhan, Awas Yojana and Ayushman Bharat are identified as central schemes
by the voters, whereas PDS and pensions are perceived as state-level schemes. Strikingly, there has also
been a centralization of credit attribution for MGNREGA compared to 2014. As shown in Table 4, the
number of MGNREGA beneficiaries who credit the central government for the implementation of
the scheme has almost doubled between 2014 and 2019. The pension and Awas schemes are equally
striking cases. In 2014, less than 20 per cent of the beneficiaries of pension scheme attributed its credit
to the central government. There is a more than 10 percentage point rise in it in 2019. There is a marked
shift from states to the centre over the last 5 years in case of Awas. If nearly half of its beneficiaries
attributed credit to state governments in 2014, the same percentage gives credit to the central government
in 2019. These reversals symbolize centralization of not only the credit claiming mechanisms but also
that of the overall welfare discourse. At one level, these trends run parallel with the overall changes in
the nature of the party system and subsequent changes in the patterns of political competition. Therefore,
228 Studies in Indian Politics 7(2)

Table 4.  Who Gets the Credit for the Implementation of Welfare Schemes? (opinions of beneficiaries of
respective schemes)

2019 2014
Scheme Centre State Centre State
Ujjwala 72 14 – –
MGNREGA 50 32 27 42
Pension 31 52 19 54
PDS 28 57 – –
Jan Dhan 72 13 – –
Awas 51 32 22 50
Ayushman Bharat/NRHM (in 2014) 54 30 20 56
Source: NES, 2019 data made available by CSDS Data Unit.

they have wider implications for the dynamics of welfare politics and the working of Indian federalism.
We discuss some of these implications in the concluding section of this article.
In the remaining section of the paper, we probe further to understand the possible impact of welfare
schemes on vote choices. Given the limits of the survey research, this is a somewhat ambitious enquiry.
The survey data can provide us with associational patterns but not with causal explanations. Besides, as
we have noted earlier, elections in India are woven around multiple issue-axes. Sections of voter citizens
develop their perceptions around a grid of issues. In this context, it would be difficult to assign a decisive
role to any one factor that influences people’s political choices. Therefore, welfare benefits cannot
be seen as the only decisive factor in deciphering voter choices even among the beneficiaries of one or
more welfare schemes.
However, as Table 5 suggests, there appear to be some interesting patterns among beneficiaries of
different welfare schemes in the 2019 elections. The BJP and allies enjoy a more pronounced support
among Ujjwala, Awas Yojana and Jan Dhan beneficiaries. These are all schemes for which voters
attribute credit to the central government. However, among the MGNREGA beneficiaries, the BJP gets
less than its average support. These beneficiaries do not vote for the Congress though, suggesting that
welfare schemes may not create long-lasting associations with the party that introduced them (as studies
in other countries have also found—see, for example, Zucco (2013). MGNREGA does not appear to
have played a significant role in stemming the attrition of the Congress vote. Instead, the vote of
MGNREGA beneficiaries is dispersed among regional parties across states. Despite centralization of the
credit attribution in these elections, more than one-third of welfare beneficiaries do not vote for the
national parties thus underlining the continued relevance of regional dynamics in welfare politics.

Table 5.  How Beneficiaries Voted?

Scheme Congress Congress Allies BJP BJP Allies Others

Overall vote share 20 07 37 07 29
Ujjwala 20 06 43 06 25
MGNREGA 20 06 32 05 37
Pension 23 06 34 06 31
PDS 19 06 36 05 34
Jan Dhan 18 06 44 06 26
Awas 21 09 38 08 24
Ayushman Bharat 19 10 40 07 24
Source: NES, 2019 data made available by CSDS Data Unit.
Deshpande et al. 229

Table 6.  How Farmers and Farmer Beneficiaries Voted?

Farmer Beneficiaries of the Income Farmer Beneficiaries of Agricultural

All Farmers Support Scheme for Farmers Loan Waiver Scheme for Farmers
Congress 20 16 22
Congress Allies 06 04 03
BJP 39 42 39
BJP Allies 08 08 07
Others 27 30 29
Source: NES, 2019 data made available by CSDS Data Unit.

Among PDS and pension beneficiaries, regional parties see a slightly higher share of votes. Of course, a
proportion of voters receive multiple benefits, and this is not reflected in the table.
In the wake of the deepening agrarian crisis, farmers’ issues remained at the centre stage of public
debate during the last 5 years—especially since 2018, when the farmers rallied to the parliament with
their demands. In response, the central government introduced an income support scheme (PM-Kisan)
and several state governments introduced agricultural loan waiver schemes. Looking at the farmers’ vote
in the 2019 elections, it seems that these schemes may have helped to neutralize farmer discontent. As
depicted in Table 6, the BJP enjoys marginally more than its average support among farmers who
received PM-Kisan, while the Congress and regional parties saw more support among loan waiver
beneficiaries. However, it is impossible to tell from survey data what the direct impact of welfare schemes
was on farmers’ perceptions. As Lokniti’s report in the Hindu notes,10 for many farmer voters, the overall
agenda of development mattered more in these elections rather than the specific issues of their welfare.
Despite their massive protests before the elections, nearly 70 per cent farmer voters expressed their
satisfaction with the performance of the NDA government at the centre.
Within the limitations of the survey data, we finally try to understand the relative (statistical)
significance of the welfare-related concerns in deciding the nature of vote in 2019 elections with the help
of the regression analysis presented below. Although policy-related retrospective voting is still unlikely
to play a determining role in election outcomes, the increased coverage of welfare programs and their
centrality in the election campaign make it a relevant exercise.
The dependent variable we use here is the NDA vote. The pattern remains more or less the same if
vote for the BJP is used instead. The regression analysis suggests that both the BJP and its allies clearly
gain support among the beneficiaries of welfare schemes even when other social and political variables
are statistically controlled. There is thus a significant yet nuanced pattern of support for the BJP among
welfare beneficiaries. Beneficiaries of centrally sponsored schemes like Ujjwala, Jan Dhan and
Ayushman Bharat are more likely to vote for the BJP. However, those of MGNREGA and pension
schemes are less likely to vote of the BJP and even its allies. News reports analyzing the election
outcomes argued that women in general and women beneficiaries of Ujjwala scheme, in particular, have
voted for the BJP.11 Our analysis of the NES data complicates this claim. As shown in Table 7, more men
than women are likely to support the BJP even when other variables influencing their voting decisions

10  
See  https://www.thehindu.com/elections/loksabha-2019/farmers-issues-were-not-centre-stage/article27266699.ece  (accessed
July 20, 2019)
11  
See  https://www.thehindu.com/elections/lok-sabha-2019/when-schemes-translate-into-votes/article27256139.ece  (accessed
July 20, 2019)
230 Studies in Indian Politics 7(2)

Table 7.  Whether Welfare Mattered?

Dependent Variable NDA Vote

B S.E. Wald df Sig. Exp (B)
Gender −0.069 0.035 4.001 1 0.045 0.933
Locality (rural–urban) −0.090 0.041 4.762 1 0.029 0.914
Caste 933.932 5 0.000
Upper castes 0.854 0.097 77.496 1 0.000 2.350
OBCs 0.816 0.095 74.146 1 0.000 2.262
Dalits 0.271 0.101 7.203 1 0.007 1.311
Adivasis 0.451 0.108 17.496 1 0.000 1.570
Muslims −1.264 0.115 121.282 1 0.000 0.282
Media exposure ( high to low) 0.041 0.016 6.271 1 0.012 1.042
Education 0.027 0.018 2.222 1 0.136 1.027
Ujjwala benefit 0.128 0.039 10.796 1 0.001 1.137
Awas benefit 0.252 0.045 31.012 1 0.000 1.287
MGNREGA benefit −0.522 0.045 133.613 1 0.000 0.593
Pension benefit −0.348 0.041 71.465 1 0.000 0.706
Jan Dhan benefit −0.203 0.044 21.166 1 0.000 1.225
Class −0.046 0.018 6.539 1 0.011 1.048
Ayushman Bharat 0.236 0.046 26.196 1 0.000 1.267
Satisfaction with central govt. 1.187 0.021 3.127E3 1 0.000 0.305
Constant 1.579 0.167 89.449 1 0.000 4.850
Source: Authors’ own based on data made available by the CSDS data unit.
Note: B: Regression coefficient; SE: standard error; Wald: Wald statistic; df: degrees of freedom; Sig: statistical significance; Exp
(B): odds ratio.

are statistically controlled. At the same time, it is a statistically significant possibility that beneficiaries
of Ujjwala are likely to vote for the BJP/NDA despite their gender identities.
The more important takeaways from this table are twofold. The first is about a distinctive connection
between welfare and vote in the 2019 elections. The second is about the ability of voters to distinguish
between central- and the state-level schemes, which may also have influenced their voting decisions.
Thus, if beneficiaries of Ujjwala rewarded the BJP, the beneficiaries of MGNREGA and pension chose
not only regional parties but non-BJP regional parties. As a result, both the BJP and its NDA partners
enjoy less support among these two groups of beneficiaries.

Conclusion
Using a first-cut analysis of the NES 2019 post-poll survey data, we argue in this article that welfare
politics did matter in the 2019 elections. However, it cannot be seen as the only factor or as the only
decisive factor that shaped the voting decisions of welfare beneficiaries. At the same time, the data
indicates voters were aware of questions of service delivery in the elections this time, held certain
expectations from parties and governments, were able to attribute credit to different government agencies
for different schemes and, therefore, may have opted for retrospective voting to certain parties.
We argue that, rather than a case of quid pro quo exchange, the welfare connect in the 2019 elections
marks a further step towards the expansion of programmatic politics in Indian elections. Given the
complex nature of Indian elections, it is a still indiscernible shift that would require careful observation
Deshpande et al. 231

in further research. Similarly, there is every likelihood that the expansion of programmatic politics may
get arrested as a result of centralization of the polity, consequent weakening of federalism and possible
erosion of the democratic space in Indian politics. While the distribution of material welfare benefits is
likely to continue to play an important role in electoral politics, wider conceptions of welfare as a
component of social citizenship may come under pressure. Notwithstanding this, the symbolic signi­
ficance of the connection between welfare and vote cannot be undermined.
Despite the presence of large majorities of poor, the welfare agenda has remained under-articulated in
Indian elections for a long time. This has led to characterization of Indian democracy as a patronage
democracy in which welfare claims mainly took the form of clientelist exchanges. In their empirical
theorizations on Indian elections and democracy, three generations of scholars associated with the NES
have often been argued against these readings. The NES findings, along with many other recent studies
of Indian democracy, have suggested a gradual shift in Indian democracy towards programmatic politics
leading to expansion of democracy albeit in an untidy manner. Our preliminary exploration of the 2019
NES data contributes to these arguments in two ways. One, we demonstrate that compared to earlier
elections, there has been a significant centralization of credit attribution for welfare programmes. Two,
we suggest that the BJP and its allies may have benefited (marginally) from new programmes that were
clearly attributed to the Centre.
Further research is necessary to better understand how schemes such as Ujjwala benefitted the BJP.
Since the scheme remains at an early stage of implementation and many beneficiaries have not yet
received more than their initial cooking gas connection, the impact of the scheme may lie more in the
realm of offering the promise or potential of future benefits rather than approval of benefits already
received. Such programmes may help make BJP promises of future delivery more credible than rival
pledges (such as Congress’ basic income manifesto commitment).12 In this sense, schemes such as
Ujjwala become less about retrospective ‘vote buying’ and more about signalling a particular reputation
for future service delivery.
The implications of some of the dynamics sketched in this article for the performance of social welfare
programmes will be important to watch. As research on other regions has shown, the centralization of
credit claiming for welfare initiatives can alter the incentives of subnational governments when it comes
to programme implementation (Niedzwiecki, 2018). Opposition-ruled state governments, such as West
Bengal or Delhi, have already sought to make political capital out of impeding central government
initiatives or implementing competing schemes, for instance, in the field of health insurance. The future
of welfare politics and Centre-state tensions in this field will have important implications for the quality
and universalism of public goods provision across India.13

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
The authors received no financial support for the research, authorship and/or publication of this article.

12  
The NES 2019 survey found that 56 per cent of respondents had heard of Congress’ NYAY scheme but that only 35 per cent of
those believed that Congress would be able to fulfil its promise.
13  
For an exploration of these questions in India and Latin America see Acosta and Tillin (2019).
232 Studies in Indian Politics 7(2)

References
Acosta, A. M., & Tillin, L. (2019). Negotiating universalism in India and Latin America: Fiscal decentralization,
subnational politics and social outcomes. Regional & Federal Studies, 29(2), 115–134.
Aiyar, Y., & Tillin, L. (2019, May). The problem: The union and the states. Seminar, 717, 10–13.
Arora, B., & Kailash, K. K. (2014). Strengthening legislative capabilities of the Indian parliament: The national
advisory council. In S. Pai & A. Kumar (Eds.), The Indian parliament: A critical appraisal (pp. 189–230).
Hyderabad: Orient BlackSwan.
Auerbach, A. M., & Thachil, T. (2018). How clients select brokers: Competition and choice in India’s slums.
American Political Science Review, 112(4), 775–791.
Bussell, J. (2019). Clients and constituents: Political responsiveness in patronage democracies. Modern South Asia,
Oxford, UK and New York, NY: Oxford University Press.
Chandra, K. (2004). Why ethnic parties succeed: Patronage and ethnic head counts in India. New York, NY:
Cambridge University Press.
Chhibber, P. K., & Verma, R. (2018). Ideology and identity: The changing party systems of India. New York, NY:
Oxford University Press.
De Koster, W., Achterberg, P., & Van der Waal, J. (2013). The new right and the welfare state: The electoral relevance
of welfare chauvinism and welfare populism in the Netherlands. International Political Science Review, 34(1),
3–20.
De La O, A. L. (2013). Do conditional cash transfers affect electoral behavior? Evidence from a randomized
experiment in Mexico. American Journal of Political Science, 57(1), 1–14.
de Souza, P. R., Palshikar, S., & Yadav, Y. (2008). State of democracy in South Asia: A report. New Delhi: Oxford
University Press.
Elliott, C. (2011). Moving from clientelist politics toward a welfare regime: Evidence from the 2009 assembly
election in Andhra Pradesh. Commonwealth & Comparative Politics, 49(1), 48–79.
Hindustan Times. (2018, December 12). BJP pays price for agrarian distress, employment crisis. Hindustan
Times. Retrieved from https://www.hindustantimes.com/assembly-elections/bjp-pays-price-for-agrarian-distress-
employment-crisis/story-rZ4JpVPr91M1UEOPR9LBQM.html
Hunter, W., & Power, T. J. (2007). Rewarding Lula: Executive power, social policy, and the Brazilian elections of
2006. Latin American Politics and Society, 49(1), 1–30.
Khera, R. (2015). UID: From inclusion to exclusion. Seminar, 672, 78–81.
Kumar, S. (2013, September). Welfare policies of the government and election: Does party in power gain electorally
from populist policies? Paper presented at the Comparative State Politics and Public Policy in India workshop,
King’s India Institute, King’s College London.
Kumar, S., Deshpande, R., & Kailash, K. K. (Unpublished). Inequalities, welfare and democracy: Studies of politics
policy interface in Indian states. Report submitted to the ICSSR.
Manor, J. (2010). What do they know of India who only India know? The uses of comparative politics. Commonwealth
& Comparative Politics, 48(4), 505–516.
Niedzwiecki, S. (2018). Uneven social policies: The politics of subnational variation in Latin America. New York,
NY: Cambridge University Press.
Pardikar, R. (2019). Art of political branding: Why 18 NDA schemes carry ‘Pradhan Mantri’ label. Retrieved from
https://www.news18.com/news/india/art-of-political-branding-why-18-nda-schemes-carry-pradhan-mantri-
label-2033221.html
Schneider, M. (2016). Do brokers know their voters? A test of guessability in India (SSRN Scholarly Paper
ID 3007407). Rochester, NY: Social Science Research Network. Retrieved from https://papers.ssrn.com/
abstract=3007407
Shastri, S., Palshikar, S., Kumar, S., Hachhethu, K., Tawfique M. S., Mehboob, A. B., & Peiris, P. (2017). State of
democracy in South Asia. Report II. Bengaluru: Lokniti-CSDS, Delhi and Jain University Press.
Sinha, A. (2016). Why has development become a political issues in Indian politics. Brown Journal of World Affairs,
23, 189–203.
Deshpande et al. 233

Stokes, S. C., Dunning, T., Nazareno, M., & Brusco, V. (2013). Brokers, voters, and clientelism: The puzzle of
distributive politics. New York, NY: Cambridge University Press.
Suri, K. C. (2009). The economy and voting in the 15th Lok Sabha elections. Economic and Political Weekly,
44(39), 64–70.
Thachil, T. (2014). Elite parties, poor voters: How social services win votes in India. Cambridge: Cambridge
University Press.
Tillin, L., & Pereira, A. W. (2017). Federalism, multi-level elections and social policy in Brazil and India.
Commonwealth & Comparative Politics, 55(3), 328–352.
Tillin, L., Deshpande, R., & Kailash, K. K. (2015). Politics of welfare: Comparisons across Indian states. New
Delhi: Oxford University Press.
Vaishnav, M., & Swanson, R. (2015). Does good economics make for good politics? Evidence from Indian States.
India Review, 14(3), 279–311.
Verma, R. (2012). What determines electoral outcomes in India? Caste, class, or voters’ satisfaction with government
performance? Asian Survey, 52(2), 270–297.
Wyatt, A. (2013). Combining clientelist and programmatic politics in Tamil Nadu, South India. Commonwealth &
Comparative Politics, 51(1), 27–55.
Yadav, Y., & Palshikar, S. (2014). Between fortuna and virtu: Explaining congress’ ambiguous victory in 2009. In S.
Palshikar, K. C. Suri, & Y. Yadav (Eds.), Party competition in Indian states: Electoral politics in post-congress
polity (pp. 42–99). New Delhi: Oxford University Press.
Zucco Jr., C. (2013). When payouts payoff: Conditional cash transfers and voting behavior in Brazil 2002–2010.
American Journal of Political Science, 57(4), 810–822.