The verdict is unambiguous: a majority of the country’s electors have not even heard about the possibility of mid-term polls and hardly anyone wants these polls. But if it’s forced on the voters, the Left and the NDA are the ones who need to worry. For, despite the first signs of anti-incumbency, the UPA is set to come back to power, just short of a majority but leaving the NDA far behind.

And, no, it’s not the Indo-US nuclear deal that has exhausted the patience of the aam admi. The voters are either unaware or benignly indifferent or mildly supportive of this deal. If the aam admi is beginning to be unhappy about one thing, it’s the economy. He calls it by various names: price rise or unemployment or deteriorating conditions of the farmers. Just like the last time, this time too, the rich are seen as the prime beneficiaries of economic policy. Add to it strong popular anxieties on corruption and terrorism and you know why even the UPA can’t take the voter for granted.

The good news for the UPA is that this resentment does not translate into big loss of votes and seats. Part of the reason is leadership: Sonia Gandhi and Manmohan Singh continue to be rated and trusted more than Advani or any other Opposition leader in the electoral fray. The logic of state-level electoral cycle happens to favour the UPA. The BJP’s incumbency in some of its key states is old and beginning to hurt, while the tide has not turned against the UPA in the states that it rules.

Above all, what’s working for the UPA is its alliance arithmetic. If the current alliances persist in Maharashtra and Tamil Nadu and if the UPA can repeat the 2004 alliances in Bihar and Jharkhand, the combined votes of the alliance partners simply outnumber everyone else. While the UPA has held on to its large coalition, the NDA has been losing key partners like the TDP and the AIADMK. If the mid-term polls were held in the first week of September, when this nation-wide survey took place, the Congress and its existing allies would have secured 267 seats, just short of a clear majority, leaving the NDA way behind at 133 seats.

This would have meant a slight fall from the tally of 300 that the State of the Nation Poll series carried out by the CSDS had projected for the UPA in January this year. In the last six months, the UPA has shed about 4 percentage points of the national vote share, while the NDA has gone up by one point. Still the popular vote for the UPA is three points above what it had in 2004, while the NDA remains some seven points below its vote share in the last elections. Yet such a projection would mean a substantial gain for the UPA and a big decline for the NDA since the last elections held in 2004. If this happens, it would be the first time since 1991 that the Congress would cross the 200 seat mark and the first time since 1989 that the BJP would fail to touch three digits.

This would change the political equation radically. The UPA may not need a major ally at all or may be able to substitute the Left by the BSP as its partner.

Major losses in Kerala would bring the Left’s tally down to 45 seats. The BSP would improve its tally to 35 seats and register significant gains in some states outside UP. With only 49 seats, the newly formed UNPA would have fallen far short of its ambition of playing the kingmaker.

The poll shows that the biggest losses for the NDA are likely to come from the BJP-ruled states like Rajasthan, Madhya Pradesh and Chhattisgarh where the wheel of incumbency has turned against the ruling party. Gujarat continues to defy anti-incumbency as the BJP appears clearly ahead of the Congress in the parliamentary and the Assembly election race.

The BJP’s allies are likely to help the NDA hang on to its current position in Orissa, Punjab and Bihar, but the alliance in Karnataka does not seem to be yielding much dividends for the Lok Sabha elections. The Congress is likely to suffer some losses in Andhra Pradesh, Haryana and Delhi but the tide has not turned fully against the ruling party.

The UPA allies are poised comfortably in Tamil Nadu and Maharashtra. The situation in Uttar Pradesh has not changed much since the assembly elections, except that the two national parties are likely to gain some ground at the expense of both the big regional parties if the Lok Sabha elections are held.

Poll methodology

THE findings of the Indian Express-CNN IBN-CSDS Poll reported today are based on the Mid-Term Poll Verdict conducted in 18 states of the country between 1-4 September 2007. During the survey, 18,750 respondents were interviewed by 489 field investigators and Supervisors spread across 940 polling stations from a sample of 217 Lok Sabha constituencies.

A three-stage stratified random sampling technique (Probability Proportionate to Size (PPS) method) was used to select the respondents for this purpose. The respondents who were interviewed during the survey were drawn randomly using the electoral rolls of the sampled polling booths.

In a face-to-face situation the respondents in all the 18 states were interviewed at their place of residence using a structured questionnaire in the local dialect of the state in which he or she resided. The sample comprised 45 per cent women, 79 per cent Hindus, 11 per cent Muslims, 18 per cent Dalits, 7 per cent Adivasis and 73 per cent rural respondents.

The fieldwork for the survey was coordinated and conducted in the states by P. Narasimha Rao (Andhra Pradesh), Sandhya Goswami (Assam), Rakesh Ranjan (Bihar), Baba Mayaram (Chhattisgarh), Kinjal Sampat (Delhi), Priyavdan M Patel (Gujarat), Dr Harish Kumar (Haryana), Harishwar Dayal (Jharkhand), Padmavathi B. S. (Karnataka), Sajad Ibrahim (Kerala), Ram Shankar (Madhya Pradesh), Nitin Birmal (Maharashtra), Surya Narayan Mishra (Orissa), Jagroop Singh Sekhon (Punjab), Sanjay Lodha (Rajasthan), Gundapuneni Koteswara Prasad (Tamil Nadu), A. K. Verma (Uttar Pradesh Central), Sudhir Kumar (Uttar Pradesh East) and Suprio Basu (West Bengal). Sanjay Kumar of Centre for the Study of Developing Societies (CSDS), Delhi, directed this all-India survey. The Central team, which designed, coordinated and analysed the survey, comprised Yogendra Yadav, Praveen Rai, Sanjeer Alam, Dhananjai Joshi, Vikas Gautam, K. A. Q. A Hilal, Himanshu Bhattacharya, Kanchan Malhotra of CSDS and Rajeeva Karandikar of Cranes Software International Limited.