NEW DELHI : As the Bharatiya Janata Party’s (BJP’s) celebrations recede, the analysis begins. The emerging wisdom is that BJP’s historic victory was driven by a Modi wave: Narendra Modi’s popularity, more than concerns about the economy or jobs, drove voters to elect him. In this Modi wave, how important were his rallies?

One approximation could be to examine how voters reacted to Modi rallies in their constituencies. At these rallies, Narendra Modi mobilized party workers, attacked the opposition and wooed voters.

If he was successful in doing all this, then the constituencies where Modi held rallies should experience greater vote swings to BJP. Conversely, in constituencies where Rahul Gandhi held rallies, a similar Gandhi rally effect could be approximated by examining changes in vote swing to the Indian National Congress (INC).

Taken together, both leaders held 245 rallies across 210 of the country’s constituencies in the 50 days before the last day of polling (17 May). Of the 135 rallies attended by Modi across 134 constituencies, BJP won 99 constituencies. In the 110 rallies attended by Gandhi across 102 constituencies, the Congress won 14. In the 26 constituencies where both held rallies, BJP was dominant: Winning 21 seats compared to Congress’ 3. A constituency victory, though, could simply reflect other factors, including historical leanings towards a party. For instance, more than half the constituencies Modi visited had voted for BJP in 2014 (see chart 1).



Graphics: Ahmed Raza Khan & Sriharsha Devulapalli/Mint
Graphics: Ahmed Raza Khan & Sriharsha Devulapalli/Mint


A better measure of rally performance would be to analyse the difference in the party’s vote share between 2019 and 2014 (the vote swing), and compare this to the corresponding vote share difference at the state level. For instance, on 29 March, Modi spoke at Kurnool in Andhra Pradesh, one of the few states, which actually saw the vote share for the BJP fall by 6.2 percentage points between 2019 and 2014.

However, in Kurnool itself, the BJP increased its vote share by 2 percentage points—the difference between the party’s constituency swing and the state swing (8.2 percentage points for Kurnool) is what is considered the rally effect in this analysis. For BJP, Kurnool was one of the constituencies where the Modi rally effect was high.

Across all the constituencies where Modi rallies were held, the average difference between constituency swing and state vote swing was around 1 percentage point, suggesting there was indeed a Modi rally effect.

The equivalent figure for Rahul Gandhi though was higher (1.5 percentage points) than Modi’s.

The higher rally effect for Gandhi could be driven in part by the limited swing for the Congress in several states, lowering the denominator, and inflating the rally effect. (see chart 2). In contrast, BJP made significant state-wide gains in vote share in West Bengal and Odisha which may have made Modi’s rally effects less pronounced. The Modi rally effect was the most pronounced in parts of Bihar such as Araria (32 percentage points) and Darbangah (29 percentage points) (see charts 3a and 3b).




The timing of rallies also may have affected voter decisions. Research has shown that a significant portion of voters make their voting choice in the days leading up to the polls . In the three days leading up to voting day, over the seven polling phases, Narendra Modi was particularly active, holding 21 rallies. These late campaign efforts seemed to have helped the BJP’s cause with a rally effect of around 1.5 percentage points.

In contrast, Rahul Gandhi’s 9 rallies held three days before voting actually had a negative effect on vote swing (see chart 4). For both leaders, the optimal period to hold poll rallies was around four to seven days before voting—both Modi’s and Gandhi’s rally effects were the highest during this time period.

The timing of rallies also may have affected voter decisions. Research has shown that a significant portion of voters make their voting choice in the days leading up to the polls . In the three days leading up to voting day, over the seven polling phases, Narendra Modi was particularly active, holding 21 rallies. These late campaign efforts seemed to have helped the BJP’s cause with a rally effect of around 1.5 percentage points.

In contrast, Rahul Gandhi’s 9 rallies held three days before voting actually had a negative effect on vote swing (see chart 4). For both leaders, the optimal period to hold poll rallies was around four to seven days before voting—both Modi’s and Gandhi’s rally effects were the highest during this time period.


vishnu.p@livemint.com

Close