The Congress defeat in Madhya Pradesh


back to issue

AFTER ten years in power, the Congress Party was soundly thrashed in Madhya Pradesh. On the highest turnout in the state’s history (67.41%), the BJP won a huge majority of 171 seats out of 228 declared (74.6%). The Congress, which had won a modest but stable majority of the seats at the last election in 1998, ended up with just 39.1 It suffered serious losses in every region except the Gwalior-Chambal division, but even there it only matched its poor 1998 showing by winning 9 of 38 seats. The BJP received 42.6% of the vote, a small increase of 3.7% over 1998. But the Congress share plummeted by 8.8% to 31.8%, mainly because it lost votes to smaller parties which fragmented the remaining 25.6%.2

The BJP outpolled Congress in all social groups except adivasis (among whom Congress’ traditional advantage shrank drastically to 2%), Muslims and dalits. Among numerically powerful OBCs, the BJP won 50% of the votes, against 26% for Congress. 22% of those who supported Congress in 1998 swung to the BJP, while the latter lost only 8% of its 1998 votes to Congress.3

Media assessments focused almost exclusively on two issues: roads and power (electricity). The BJP (guided by private polls) constantly emphasised these things, but this explanation oversimplifies grossly. Roads and especially power were extremely important, but their main role was in crystallising the significant but initially rather unfocused discontent with the Congress government. To understand this election adequately, we must explain the origins of that discontent. They lie in the political structures crafted by Chief Minister Digvijay Singh – which unintentionally led to poor government performance (that inspired discontent), blinded him to much of it, and prevented him from tackling it effectively.



Political structures and popular discontent:4 Given the BJP landslide, it is important to stress that discontent with the government was significant but far from total. A private BJP poll in August/September found that 55% of respondents thought that the government was ‘poor’, but 45% regarded it as ‘good’ or ‘okay’.5 A November poll in The Week reported that 39% saw the government as good, 37% as poor and 19% as average. Views of the chief minister were almost identical.6 These are not damning numbers. Discontent intensified by polling day, but only somewhat – 49.2% were not satisfied with the work of the government, and 35.6% were satisfied.7

Voters saw the BJP as better able than Congress to tackle only one problem – ‘infrastructure’ – whereas Congress was seen as better at tackling their other four main concerns (‘unemployment’, ‘communal harmony’, ‘religious interests’ and ‘price rise’). Unfortunately for Congress, 45% or respondents rated infrastructure as most crucial – while the other topics combined drew comment from only 35%.8



But if the discontent had clear limits, it was still sufficient to send the government down to a severe defeat. To understand it, we must examine the evolution of political structures in the state over the previous decade. Other things mattered too, but this was the core problem.

When Digvijay Singh considered how to make his influence penetrate down into society and how to obtain information from below, he had three main options. First, he could have developed accommodations with civil society organisations. After exploratory overtures during his first term, he largely9 rejected this option. Second, he could have built up his party’s organisation. But to do so was to risk both factional strife which he had largely contained, and unwelcome interventions from national-level Congress leaders. He therefore rejected this too.

This left him only one choice – to work through the formal state machinery, and through official programmes. He did this vigorously and developed a remarkable array of imaginative, sometimes visionary initiatives derived from the sort of arguments advanced by Amartya Sen. Some were routed through conventional bureaucratic channels, but several were pursued through special administrative instruments – Rajiv Gandhi Missions – that bypassed ministries which he regarded as inefficient or in some cases unsympathetic. The chief minister dominated policy design, assisted by picked civil servants and advisors. Other ministers and legislators (MLAs) played little or no role.



The great danger in relying on official channels was that he might be left with too little information – on what was happening, and on how ordinary people saw things – from sources other than MLAs and bureaucrats. For a time, he had a major alternative source of political intelligence from within the government – the members and chairpersons of zilla panchayats. Those people also had some power to tackle bureaucrats who were performing poorly.10 But in April 2001, a system of ‘District Government’ was introduced which – as it evolved11 – disempowered zilla panchayats and made district ministers and collectors dominant at that level. Senior government spokesmen presented this change as a step forward for democratic decentralisation, but that is the opposite of the truth.

This action, which was taken to placate MLAs and ministers who were jealous of zilla panchayat members’ powers, was politically unwise because it infuriated those members – who might otherwise have lent crucial electoral support. Digvijay Singh certainly suffered at the 2003 election from their continuing anger. But even more damaging was the destruction of his only major alternative source of political intelligence. Because officials often sent him rosy reports, he was flying blind for his last 22 months in power. He was acutely aware of the state’s problems with power and roads, but he was badly under-informed about wider discontent caused by under-performance of the bureaucracy (especially at lower levels).

The disempowerment of zilla panchayats also deprived him of allies at lower levels who – if they had remained influential – could have tackled bureaucratic shortcomings.12 In other words, his excessive dependence on collectors, police superintendents, district ministers and MLAs – who forged tight alliances at the district level – bred discontent, and undermined his capacity to see it and to tackle it. Other things also influenced the outcome. Let us consider them.



Power, roads – and water: The BJP began by stressing its mantra, bijli, sadak, pani (electricity, roads, water). It soon de-emphasised water because this was not a negative issue for Congress, except in one specific sense. Had the election been conducted a year earlier, after several years of drought, water would have loomed large. But the good monsoon of 2003 had eased the extreme shortages and – more to the point – the government’s energetic pani roko programme to capture water had been reasonably successful. Roughly half of the minor works constructed had survived, so when this year’s rains arrived, adequate water for drinking and for crops was available.13



The key problem was that to convey water to fields, most farmers needed power for electric pumps. Power cuts impeded that process, so farmers faced an excruciating irony. Water was (at last) available, but they could not deliver enough of it to their crops.14 This called further attention to the much more important issue of power, which already caused acute resentment.15

Madhya Pradesh would not have faced serious power shortages had the Chhattisgarh government not denied it access to electricity which it had consumed before the bifurcation of the two states in November 2000. Chhattisgarh Chief Minister, Ajit Jogi, like Digvijay Singh, was a Congressman. But owing to a long-standing feud between the two men, Jogi refused power to Madhya Pradesh in the same extravagantly aggressive way that he over-centralised within his state and wrecked his prospects of re-election.16 (He also failed to honour other aspects of the bifurcation agreement, denying Madhya Pradesh substantial resources to which it was entitled.)



Digvijay Singh could not blame Jogi too loudly, because that would raise questions about why Sonia Gandhi did not intervene on behalf of his state. Her failure to do so, despite repeated appeals, was a breathtaking error that caused huge damage to her party in Madhya Pradesh.17 This was also patently obvious to many. On 5 December, an autorickshaw driver in Bhopal stated that victory for the BJP in the two states meant that the party’s national leaders would soon have power flowing from Chhattisgarh.

Sonia Gandhi’s inaction persuaded several (but not all) perceptive observers in Bhopal that she wanted Digvijay Singh to lose (alongside victories in Chhattisgarh and Rajasthan). They reckoned that he was the most serious alternative to her as prime minister in a hung Parliament after the 2004 election, if other parties refuse to support Congress with her as premier.18 If they are correct, this is the first state election in India’s history in which national leaders of both major parties sought to demolish their own state-level leadership (for the BJP, see below).

The state government made strenuous efforts to ease the power shortage as the election neared. It diverted scarce funds from other sectors to purchase some from other states. It also developed a crash programme for a hydroelectric project, set to come on stream a month before polling day, but it was delayed by two months. Digvijay Singh could only promise full power ‘by 2005’ – an inadequate response to constant BJP denunciations on this issue.



Resentment also developed against charges for power introduced by the state government. (Non-payment often meant loss of supply.) Charges had less impact than the problem of supply, but they sharpened discontent over shortages because voters who now got less power also had to pay.19

On roads, the state government was more culpable. After being re-elected in 1998, Digvijay Singh promised action on roads ‘within six months’ – but his government was slow to act. By early 2002, it was clear that that the government was acutely vulnerable on roads.20 Significant funds were committed to road building during the 18 months before the election, but they left it late, and progress was slowed by disputes with contractors and heavy rains during the 2003 monsoon.21

But if the government was responsible for the roads problem, it had made substantial progress on water – and on power, it was a victim of Congress leaders outside the state.



The selection of Congress candidates: The chief minister prevailed in most decisions about Congress candidates. This angered other senior Congressmen – Kamal Nath openly dissociated himself from the process. Some curtailed support for the campaign. There was also a second problem. Digvijay Singh knew that at the 1998 election, many sitting MLAs – in both major parties – had been defeated. Anti-incumbency feeling at constituency level was an extremely potent force.22 He escaped severe damage then by dropping 41% of his MLAs, but in 2003, he only dropped 16%.23 In the event, a huge proportion of incumbent Congress MLAs were defeated, as were approximately two-thirds of ministers. Even BJP incumbents suffered. Despite the landslide, roughly one quarter were ousted. The chief minister persisted with so many legislators because he feared damage from factional strife and rebel candidates if he changed many faces. He was probably mistaken.

Digvijay Singh and Uma Bharti: Long before the election, Digvijay Singh had hoped to get Uma Bharti as the BJP’s campaign leader, mainly because he believed that her emotional volatility would stand in stark contrast to his own unflappability. The BJP obliged by imposing her, in order to marginalise its old-line leadership in Madhya Pradesh (as in Rajasthan).24

Singh tried to exploit this by suggesting a televised debate with her, but the BJP proposed India’s Law Minister Arun Jaitley instead. A poorly briefed Jaitley had made embarrassing misstatements on an earlier visit, so the chief minister – anticipating more such slips – agreed to the debate. When it occurred, however, Jaitley was extremely well prepared and in top debating form. At one point, he had Digvijay Singh fumbling for a reply to a point on poverty rates. Viewers who had never seen their chief minister at a loss found their confidence shaken by this.25 Meanwhile Uma Bharti avoided the fire-breathing hindutva themes for which she is famous – focusing on power, roads and appeals to fellow OBCs and women.



A meticulously managed BJP – or rather, RSS – campaign: It is impossible to overstate the meticulous character of the BJP’s campaign. This writer gained access to a pre-election strategic survey of the state, developed for the BJP by professional analysts at a think tank. In 31 years of studying state elections, he has seen nothing to rival it. It runs to 452 pages, excluding an extensive introduction and annexes. It contains extremely detailed information – constituency by constituency – on caste composition, local conditions and problems, other parties, voters’ perceptions, and the party’s chances of winning.26 The BJP used this formidable resource systematically in candidate selection and throughout the campaign.



But to call this a ‘BJP campaign’ is misleading. The Rashtriya Swayamsevak Sangh (RSS) played the pre-eminent role – which was essential to victory, especially but not only in tribal areas.27 The view of one Bhopal editor that the BJP’s organisation in the state is ‘very weak’ is widely shared.28 A senior RSS leader stated that the party’s organisation ‘was in a messy state’, so from mid-2002, ‘the RSS took over the reins of the election’. It lobbied for Uma Bharti as campaign leader, strongly influenced candidate selection, and used its penetrative, cadre-based structure (which the BJP lacks) for canvassing.29 The BJP’s organisation in Madhya Pradesh is no stronger than that of the Congress (which is quite frail), but with the RSS involved, Congress was at a severe disadvantage.

Government employees’ resentment: Uma Bharti won passionate support from among 500,000 Class II to IV state employees by promising to reverse unpopular decisions by the Congress government – the retrenchment of 28,000 daily employees, freeze on many promotions, non-payment of dearness allowances, etc. These measures were part of a fiscal stabilisation package which had kept the government from the near-bankruptcy that plagued many other states.30 Government employees were correct in stating that much of the money saved went to pro-poor social programmes.31 The BJP gained substantial votes from employees as a result, but it will now struggle to avoid the fiscal havoc wrought by their last government in Maharashtra.



Five issues of which little was made: Hindutva The BJP downplayed this issue, partly because it lacked salience with voters. Only 5% in one poll said that it would be an issue.32 But the chief minister also did much to neutralise it. No significant communal clash had occurred during his decade in power. His appeal to the prime minister in early 2003 for more attention to cow protection had armed him with a telling rejoinder when accused of ignoring Hindu values. And when the arrest of a VHP incendiary for a communalist speech produced no serious protest, Hindu extremists had been embarrassed. Voters in one poll rated the Congress as better able than the BJP to look after ‘religious interests’ (48% to 42%).33 The BJP was shrewd to dodge this issue.

Panchayati raj The chief minister referred in many speeches to democratic decentralisation, but the message was so familiar after nine years of panchayati raj that other things received more stress. Ironically, given that the state has achieved much in this sector by international standards, the BJP turned the issue partly against the Congress – by promising to abolish District Government if elected.

It is remarkable that Congress neutralised hindutva while the BJP neutralised panchayati raj in this election – but that is what happened.



Corruption – Uma Bharti emphasised corruption very occasionally, but mainly stressed the core themes of power and roads. BJP polls indicated that corruption failed to register among leading concerns of voters – even though corruption in the state government was serious.34 So even before the Judeo tape was shown, the BJP had decided not to stress corruption.35

The Dalit Agenda: The state government had a more formidable programme for dalits than any in Indian history, after Digvijay Singh committed himself to all 21 recommendations of a dalit conference which he had convened. At the moment of decision, a visibly emotional chief minister told an aide that this was an issue worth risking electoral defeat upon.36



There were, however, political dangers in relying entirely on the bureaucracy to pursue an agenda crafted by dalit intellectuals from outside the state. Most Congressmen were hostile, and dalit organisations within Madhya Pradesh scarcely existed. The BJP – which convened its own dalit conference – made little of this issue. But it is widely agreed that non-dalits – especially the numerically powerful OBCs – reacted negatively. A significant number of dalits obtained land, but the backlash from other groups, poor implementation in some areas, and unintended outcomes in others actually depressed dalit support for Congress.37

The Education Guarantee Scheme (EGS) – The state’s extraordinary EGS, which provided schools to 26,571 habitations (and 12.33 lakh students) that had not had them before,38 went uncriticised by the opposition. That is hardly surprising since the central government is developing a national programme on this model. (Nor were there complaints about state government initiatives to train large numbers of para-professionals, akin to EGS teachers, to provide more services in other sectors.) Further research is needed to determine whether the EGS won many votes – the strong showing of the BJP in the tribal belt suggests that it did not – and whether para-professionals supported Congress.



Digvijay Singh’s basic aims were to make government more responsive and to broaden the base of Congress. He decentralised powers to give people at the grassroots and their representatives influence over development projects. He increased the reach of government services, despite fiscal constraints, by training large number of para-professionals to deliver them. Some programmes (like panchayati raj) were intended for the population in general. Others (like the EGS and the Dalit Agenda) were aimed at stimulating demands from poorer groups who had previously made few claims on government, and at responding to those demands. Taken together – despite some failures and many ambiguities – these initiatives constitute an impressive developmental record.

They succeeded (as intended) in raising people’s aspirations, but this entailed political dangers if aspirations were not met. By stressing social sectors, the government fell short on tangibles like roads, and in providing electricity (although others were mainly to blame for the latter). This enabled the BJP – using a very narrow definition of ‘development’ to imply power and roads – to argue that the government had failed. Some programmes were poorly implemented, but after the unwise imposition of district government, the chief minister could neither learn enough about this nor do much to tackle it. That was his fundamental problem and it mainly explains his defeat.



1. The Election Commission of India’s website –, and CSDS analyses. (Two seats remain officially undeclared at this writing, which news agencies awarded to the BJP.) The comparison between this and the 1998 elections – if we only consider seats remaining in the state after the creation of Chhattisgarh are:

1998 2003

Congress 124 39

BJP 83 171

Others 23 18

2. CSDS election study in The Hindu, 10 December 2003, and Hindustan Times (Bhopal) 6 December 2003. Minor parties drew only small shares of the state-wide vote, but they damaged Congress in every sub-region except Malwa.

Digvijay Singh wanted to make pre-poll alliances with minor parties. But although the Congress had committed itself at Shimla to such alliances, he was told that this applied only at the national level. The same thing happened and may have deprived Congress of a majority in Rajasthan. Economic Times, 13 December 2003. Salman Khurshid’s post-election claim that ‘the CMs’ opposed alliances (Outlook, 15 December 2003, p. 40) is thus inaccurate. The high command is to blame.

The ‘understanding’ of the Congress with the Bahujan Samaj Party (BSP) did not prevent the BSP from taking vital votes from Congress. The Samajwadi Party (SP) unexpectedly hurt Congress in many seats – and once Congress had an ‘understanding’ with the BSP, agreement with the SP became impossible. In the six constituencies of the Gond tribal belt, the Gondwana Ganatantra Party (GGP) polled between 22% and 39% of the votes – even though its state-wide share was under 2%. The GGP won only one seat there, but damaged Congress. Hindustan Times (Bhopal) 7 and 8 December 2003. The RSS and BJP had been assisting/funding the GGP for a long period to undercut Congress. Interview, Askari Zaidi, Bhopal, 5 December 2003.

3. CSDS election study in The Hindu, 10 December 2003. I am especially grateful to the CSDS election team for sharing details of their findings.

4. Given limitations on space, this section is presented with extreme brevity. A detailed analysis of these issues will appear in a book on Brazil, Uganda and Madhya Pradesh, being written by Marcus Melo, Njuguna Ng’ethe and this writer.

5. Society for Development of Humanity, Jabalpur, ‘MP Pre-Election V(idhan). S(abha). Study’, p. xiii.

6. Other respondents offered no opinion. The Week, 30 November 2003, p. 39.

7. The CSDS exit poll.

8. Ibid., p.38. The CSDS exit poll found that 65% saw power/roads as the main problem, which dwarfed all others. The Hindu, 10 December 2003.

9. The main exception was the Ekta Parishad, a Gandhian organisation pressing for land rights for poor people – with which Singh developed an understanding as the 2003 election approached.

10. It has been clearly established that when district-level panchayats are generously empowered, they can play these roles. See the discussion of Karnataka in R. Crook and J. Manor, Democracy and Development in South Asia and West Africa: Participation, Accountability and Performance (Cambridge University Press, Cambridge, 1998). A discussion with the head of the Madhya Pradesh zilla panchayat chairpersons association in Sehore, 22 March 2002, plainly indicated that they had played this role there as well.

11. Digvijay Singh did not initially intend zilla panchayat members to lose so much power. But a nexus between collectors, district ministers, police superintendents and MLAs soon developed and excluded others. Interviews with three knowledgeable sources, Bhopal, 5, 6 and 10 December 2003.

12. Another major change in January 2002, the introduction of gram swaraj (direct democracy at the village level), curtailed the power of village sarpanches. This also alienated key potential election supporters of the Congress at the grassroots, but it was less important than the imposition of District Government in damaging information flows to the chief minister.

13. Voters repeatedly told interviewers that drinking water was not a problem, thanks to the pani roko programme. See for example, Central Chronicle (Bhopal), 2 December 2003.

14. The CSDS exit poll found 66.2% of respondents rated drinking water supply the ‘same as’ or ‘better than’ before, while 32.0% rated it ‘worse than before’. But irrigation facilities were seen as ‘worse’ by 30.7%, ‘same’ by 30.7% and ‘better’ by 15.0%.

15. The CSDS exit poll found that electricity supply was seen as ‘worse than before’ by 78.9%, ‘same’ by 13.3% and ‘better’ by 6.%.

16. For a useful, detailed analysis of this, see Hindustan Times, 5 December 2003.

17. Mrs. Gandhi was warned twice about Jogi’s damaging behaviour by the official Congress observer, former Maharashtra Chief Minister Vilasrao Deshmukh – and she was further warned even by a senior figure in the NDA, probably Brajesh Mishra, speaking for the Prime Minister. Hindustan Times (Bhopal), 8 December 2003 and Economic Times, 15 December 2003.

18. Those who take this view point to speeches during the campaign by Arjun Singh – former chief minister of the state, and now an advisor to Sonia Gandhi’s circle. He apologised for the Congress government’s failure to provide power, and thereby focused blame on the chief minister. The BJP made much of this.

19. The Election Commission prevented Digvijay Singh from waiving electricity bills since 2001 for farmers and the urban poor in the weeks before polling day.

20. This is based on interviews with leading bureaucrats and Congress politicians in Bhopal in March 2002.

21. The CSDS exit poll found that the condition of roads was seen as ‘worse than before’ by 62.0%, ‘same’ by 16.7% and ‘better’ by 19.3%.

22. It has been underestimated. The CSDS exit poll found that 40.3% of respondents gave the greatest importance to the candidates, against 12.7% who stressed stage government performance.

23. India Today, 17 November 2003, p. 39. He did so partly under pressure from senior state Congress leaders, some of whom largely sat out the campaign.

24. I am grateful to Amitabh Singh for stressing this point. The RSS pressed for Uma Bharti in order to side-line state BJP leaders, and retained huge influence after the election. She chose her small cabinet ‘after extensive consultations with the RSS leadership.’ This meant that ‘very few of the remaining [non-RSS] factions have got representation in the list.’ Hindustan Times (Bhopal), 8 December 2003.

25. This is based on numerous discussions with a diversity of people who witnessed the debate, Bhopal, 3-10 December 2003.

26. ‘MP Pre-Election V(idhan).S(abha). Study.’

27. There was a 14% swing in the tribal belt against Congress – greater than the state average. The RSS had been working in tribal areas for years, and it reportedly paid monthly stipends to 400-500 RSS activists in the tribal areas of Madhya Pradesh, Rajasthan and Chhattisgarh for six months prior to polling day, for full time election work. This was in addition to swayamsevaks who were already on the ground. Economic Times, 15 December 2003.

28. Interview with Pushpendra Solanki, Bhopal, 6 December 2003. Numerous interviews indicated that the BJP lacked analytical capacity, and that it was too distracted by hindutva in early 2003 to reach out to sarpanches who were unhappy about gram swaraj and to para-professional teachers seeking formal status as government employees.

29. Interview with kshetra pracharak Narmohan, Hindustan Times, 6 December 2003. The words quoted are those of the reporter, summarising Narmohan’s views. More than 172 BJP candidates had attended RSS shakas and retained ties to the organisation. The RSS proposed candidates for all but 60 seats, most of which were accepted. ‘The entire media and election management was handled by RSS cadres.’ One reason for the extensive involvement of the RSS was that VHP and Bajrang Dal members lent little support to the BJP in this election. Central Chronicle (Bhopal), 4 December 2003.

30. Interview with Planning Commission officials, New Delhi, 7 March 2003.

31. Hindustan Times (Bhopal), 6 December 2003.

32. Interview with two BJP organisers, Bhopal, 4 December 2003; and The Week, 30 November 2003, p. 38.

33. Limited communal clashes occurred after polling in only one part of Indore, triggered by violence at one polling centre between Congress and BJP activists. A curfew was imposed, and normalcy was soon restored. Central Chronicle, 3 December 2003.

34. The CSDS study found that 61% saw the government as fully or somewhat corrupt. The Hindu, 10 December 2003.

35. Hindustan Times (Bhopal), 3 December 2003.

36. Interview with Amar Singh, Bhopal, 6 December 2003.

37. Dalits – who have recently given Congress less support here than in other states – gave Congress 31.4% of their votes (as against 37.6% in 1998), and 28% to the BJP, 19.2% to the BSP and 21.3% to others (the CSDS exit poll).

38. Interview with Amita Sharma, Bhopal, 4 December 2003.