What do the most recent state elections tell us about the likely fortunes of the Bharatiya Janata Party (BJP) in the 2019 national election?

The BJP lost the elections in all three of the Hindi belt states that went to the polls in November and December, including two states (Chhattisgarh and Madhya Pradesh) it had controlled for 15 years. This continued a string of weaker-than-expected electoral performances (e.g., Bihar, Gujarat) in states where the BJP had a dominant performance in the 2014 national election.

BJP partisans remind us that national elections are very different from state elections and that one should not extrapolate from these results — Modi’s charisma and popularity will loom large in 2019, they argue. Meanwhile, the Opposition sees these results as unmistakeable evidence that the BJP will be greatly weakened in 2019.

Political punditry aside, what do state elections really tell us about national elections?

It should be clear that state-level voter preferences play a major role even in national elections. Several state-level political parties such as the Trinamool Congress (TMC), the Biju Janata Dal (BJD) or the Telangana Rashtra Samithi (TRS) are likely to win many seats in the Lok Sabha, even though they do not have clearly articulated national agendas. This is a natural consequence of the fact that a party’s capacity to deliver benefits and infrastructure such as roads and schools locally and the strength of the local party organisation are among the most significant determinants of a popularity (even in national elections) — and strong regional parties often perform well on these criteria. At the same time, charismatic leaders such as Indira Gandhi or Narendra Modi, and specific issues or policies like the Ram Janmabhoomi movement or the popularity of the National Rural Employment Guarantee Scheme (NREGS) have helped the BJP or the Congress disproportionately in national elections.

In some ways, a party’s performance in Chhattisgarh, Gujarat, Madhya Pradesh and Rajasthan is likely to be more predictive of the outcome even in national elections; they also do not have the coalition dynamics of Karnataka or Uttar Pradesh and are largely straight BJP/Congress contests. The BJP won state contests in Chhattisgarh, Gujarat and Madhya Pradesh in 2007 and 2008, and also won these states in the 2009 national election, even in the face of a commanding performance by Congress nationally. The BJP won all four states in 2012 and 2013, and swept them in the 2014 national election. Of the 91 parliamentary constituencies (PCs) in these four states, the BJP won an eye-popping 88 seats, a strike rate of 97%. About 31% of all seats the BJP won in 2014 came from just these four states.

Data from the Election Commission of India (ECI) allows the analysis of national election results down to state assembly constituency (AC) level.

Chart 1 displays the AC-level vote share for the BJP in 2012 and 2013 state elections with the AC-level vote share for the BJP in the 2014 national election. There is clearly strong statistical relationship — state electoral results predict national results. For every 10% increase in AC-level vote share for the BJP in a state election, there is a roughly 5.7% increase in vote share in the 2014 national election. To be sure, the 2014 result was more than a repeat of past state results. It was a wave election, and Narendra Modi had a major effect. In these four states, the BJP had an average vote share of 45% in state elections but an average of 55% in the 2014 national election.

This implies a Modi effect — the extra bump to BJP’s vote share in going from state to national elections — of about 10%. This 10% was not distributed uniformly, some areas saw much larger increases than 10% and some areas saw much less.

Cut to the present situation. In 2017 and 2018, the average vote share of the BJP in state elections in these four states dropped to 41%. Given the poorer performance in these four states (it retained Gujarat and lost the other three), and the uneven distribution of the Modi effect, how will the BJP perform in 2019?

Chart 2 displays the predicted number of seats in these four states in the 2019 national election using a statistical model accounting for the relationship between state and national results and the scale of the Modi effect (and its uneven distribution) — all based on how the 2012 and 2013 state election data predict the 2014 results.

If there is a Modi effect at 2014 Modi wave levels, then the BJP can be expected to win about 83 seats in 2019, only five less than in 2014. But if the Modi effect drops to about 5% (that’s still a fairly large effect), then the BJP will only win about 65 out of the 91 seats in these four states.

If there is no Modi effect, then the party will win just 37 seats.

When combined with a stronger opposition coalition in Uttar Pradesh, the BJP will have a tough time returning to power in such a scenario.

The BJP will need to repeat its commanding performance in these four states if it is to win again. No one can really say what will be the electoral impact of the BJP’s announcement of reservations for economically backward classes or the Pulwama attack.

But this analysis underscores one simple point. The BJP will have to manufacture another Modi wave if it is to return to power in 2019.

First Published: Feb 25, 2019 08:03 IST