The “Modi tujhse bair nahi” fallacy

Stunned by the 0-3 verdict in the Hindi heartland, many BJP supporters are hoping that Prime Minister Narendra Modi’s charisma will counter the anti-incumbency that led to the party’s defeat in Chhattisgarh, Madhya Pradesh and Rajasthan. Some point to the slogan “Modi tujhse bair nahi, Vasundhara teri khair nahi” as a sign that voters were angry with state leaders rather than with the PM, and that this setback was a one-off.

This is somewhat optimistic. There was undeniably a national element to the state campaigns. Many of the issues that afflicted the incumbent BJP governments were national – agricultural distress, demonetisation impact, GST. Congress President Rahul Gandhi campaigned at length on the alleged failings of the Modi government, and Modi also spent a lot of time criticising the Congress Party.

That said, issues and candidates do differ between state and national elections, so we can’t rule out the possibility of some voters in those states voting differently in the coming general election. However recent history is not encouraging for the BJP, at least when it comes to Chhattisgarh, Madhya Pradesh and Rajasthan.

Consider the average swing for the winning and losing parties in these three states over the past 15 years.

Screen Shot 2018-12-20 at 11.56.51 AM.png

The party that wins a state election on average gains 7-9 percentage points in the following general election. The party that loses also gains 1-2 points. The overall share of the vote going to smaller parties and independents usually shrinks between state and national elections, benefitting national parties.

If this pattern holds, the INC can expect to increase its lead over the BJP in these states in the coming general election. And even if there is a Modi effect on some voters, the state winner effect could neutralise it.

Some might argue that Modi is sui generis, and that his personal charisma will suffice to see the BJP through, especially in the Hindi belt. The historical record suggests otherwise. Consider 2013-14, when the INC was at a low point and dealing with serious anti-incumbency. In Chhattisgarh, Madhya Pradesh and Rajasthan, the party lost only 1-3 percentage points between the 2013 state election and the 2014 general election.

Screen Shot 2018-12-20 at 12.05.06 PM.png

This means that very few INC voters in the state election switched to the BJP – or at least those losses were offset by incoming voters who had previously voted for other parties or independents. The INC lost such few votes at arguably its lowest point that it seems unlikely that it will slide when it is resurgent, and when the BJP faces anti-incumbency after five years in office.

The BJP could still jump ahead of the INC, as occurred early in the Atal Bihari Vajpayee era. The former won Madhya Pradesh (which included Chhattisgarh then) decisively in the February 1998 and September 1999 general elections, even though in between it had lost the November 1998 state election to the INC. Again, the INC held on to its vote share but the floating vote moved towards the BJP.

The problem is that the momentum is now going the other way. The BJP could in theory reverse its losses in the Hindi belt, but it won’t be easy.

Advertisements

Watch out BJP, you’re trailing behind the UPA in your bypoll performance

How can Modi be beaten in 2019? Following the Bharatiya Janata Party’s byelection reverses last week, the consensus in the commentariat clearly leans towards a grand alliance (or mahagathbandhan) of opposition parties of the sort that humbled BJP in Bihar in 2015 and in three recent Uttar Pradesh parliamentary byelections. The subtext is that an “arrogant” Indian National Congress should be more generous towards regional parties, possibly even putting forward a non-Congress candidate as prime minister to cement a grand alliance.

https://twitter.com/sardesairajdeep/status/1002079250424434688

https://twitter.com/sardesairajdeep/status/1002079250424434688

But is it valid to extrapolate from parliamentary bypolls to national politics? Gilles Verniers and Rajkamal Singh argue not, saying “bypolls do not have a predictive value for the following state or general election” and that they reflect mostly local factors, such as the interplay among influential political families in Kairana. There does seem to be some truth to this — the United Progressive Alliance won 6 of 12 bypolls in 2012 and 2013 and the simple average of its vote rose 2 percentage points vis-à-vis the 2009 general election. Yet it was routed in 2014.

If you look closely at parliamentary – as opposed to state – byelections though (see charts below), there does seem to be a correlation in terms of vote swing between how the UPA did in the by-elections and in the subsequent general election. The simple average of the vote swing away from the UPA in the seats that had byelections is similar to the national swing in 2014, implying that they may have been representative of the national picture.

Screen Shot 2018-06-03 at 9.21.48 PM.png

Screen Shot 2018-06-03 at 9.54.25 PM.png

It is striking how poorly the BJP has done in comparison with the UPA, winning only one of the 13 seats that had parliamentary byelections in 2017 and 2018. It seems difficult to imagine the BJP reversing all of the ground it has ceded in those seats between now and 2019. Rajasthan, Uttar Pradesh and Maharashtra are all states that were central to Modi’s 2014 victory. It might be challenging for the BJP to make up these losses with evident gains in eastern states such as West Bengal. There is reason for the BJP to worry.

Do the byelection results support the proposition that the opposition cannot win in 2019 without a mahagathbandhan? Not really. Of the 15 seats that just had elections, only four were won by an opposition grand alliance. Non-BJP parties won three seats fighting alone and another five were won by longstanding state alliances in Bihar (RJD-Congress) and Maharashtra (Congress-NCP).

Screen Shot 2018-06-04 at 3.05.10 PM.png

Screen Shot 2018-06-04 at 3.07.20 PM.png

The case for a mahagathbandhan is obviously strongest in Uttar Pradesh. The BJP remains ahead of the Samajwadi Party and Bahujan Samaj Party individually by quite a margin. It suffered a major vote erosion only in Phulpur; in Gorakhpur and in Kairana it slipped but still polled strongly in the high 40s.

That doesn’t mean that a grand alliance is the answer everywhere. The BJP and its allies rule 12 of India’s 20 biggest states, and the INC is in a position to defeat it – either singly or with its existing allies – in all but Uttar Pradesh. The one ally that could make a difference is the BSP whose vote base overlaps with that of the Congress, as in Madhya Pradesh where an alliance seems to be in the works. In a close election, the addition of the BSP vote could make the difference between victory and defeat in Chhattisgarh, Madhya Pradesh and Rajasthan, states that contributed 62 of 65 seats to the BJP in 2014 and will hold elections in a few months.

Whether Mayawati is amenable to a broader alliance, and on what terms, remains to be seen. It is often said that the BSP suffers because allied parties do not transfer their votes even as they they benefit from the BSP’s vote bank. However this does not appear to be the case in the BSP’s biggest alliances. In 1993 and 1996, the BSP allied with the SP and the INC respectively in the Uttar Pradesh state elections. The BSP’s vote share in the seats it contested was only about percentage point lower than its partners’, suggesting that a smooth transfer of votes occurred.

To sum up, the BJP is facing a serious erosion of votes that is making the 2019 election much more competitive than in 2014. But with the exception of Uttar Pradesh and Karnataka, the results suggest that the INC should be judicious about ceding ground to potential allies rather than rushing into a mahagathbandhan. In 2004 and 2009, the INC became the single-largest party by winning 35% and 47% of the 417 and 440 seats it had contested respectively. If the party cedes more space to a mahagathbandhan and ends up contesting, say, 300 seats, it will be difficult for it to come close to the 150 mark needed to block the BJP’s claims and to form a stable coalition.

(Originally published on NewsCentral24x7.)