Bengaluru: The vote in favour of the Aam Aadmi Party (AAP) in the 2013 assembly elections in Delhi was tentative. People wanted to give the then fledgling party a chance, but were unsure of whether enough other people would support it. This meant that not everyone who wanted to vote for the party possibly voted for it. Thus, in those elections, which were the party’s debut, the AAP got 29.5% of the votes and 27 seats in the Delhi assembly.

Prior to some state elections, the Association for Democratic Reforms (ADR) in association with Daksh (a Bengaluru-based not-for-profit) conducts a comprehensive survey of voters in the state going to polls trying to understand the issues that voters consider most important. Among the questions that Daksh-ADR ask in such surveys is a rather innocuous looking question as to whether the respondent voted for the winning candidate in the earlier elections.

In the survey conducted just prior to the Karnataka state assembly elections in 2013, a staggering 75% of respondents mentioned that they voted for their incumbent MLA (member of Legislative Assembly) in the previous elections.

Going by data from the 2008 assembly elections from Karnataka, though, the median vote share of a winning candidate was 43%, much lower than the 75% that claimed to vote for the winning candidate. So the responses of the survey suggest one of two things—either the sample that ADR-Daksh interviewed was biased (unlikely since it’s a scientifically designed survey) or that voters wanted to be seen as voting for the winning candidate.

It is the latter point that is of interest to us here, in the context of what is known in the theory of elections as the bandwagon effect. In an election, people don’t want to vote for a candidate who they perceive to be unwinnable. Rather than voting for one such candidate, people are more inclined towards voting for a second-best candidate who has a reasonable chance of winning. From this perspective, an important feature of any political campaign is a show of strength, which gives the voter confidence that there are enough other people voting for this party (this earlier piece on the Sweep Index touched upon this: bit.ly/1jG7pRd)

There are various ways in which parties can exhibit a show of strength—the campaign blitz and “Mission 272" campaign by the Bharatiya Janata Party (BJP) in the run-up to the 2014 general election helped convert the marginal voter to the BJP, for example. The best way to show strength, however, is to have a rehearsal, a kind of mock election that tells voters which way other voters are thinking.

For example, two months prior to the state assembly elections in Karnataka in 2013, most of the state’s urban local bodies went to the polls. The Congress party put up a strong performance in these polls, which had the additional impact of convincing the voters that the Congress party was on a strong wicket, and that the party had considerable strength in the state at that point in time. A consequence of this was that the hitherto marginal voter tilted in favour of the Congress and the party was able to form the government comfortably in Karnataka.

There was no immediate rehearsal prior to the just-concluded Delhi elections, but it can be argued in hindsight that the 2013 assembly elections in Delhi effectively played the role of the rehearsal in this case. Back then AAP was a fledgling start-up, and people were unsure of the party’s strength (despite the party’s attempts to convince the electorate otherwise by means of party-sponsored opinion polls). Thus, the marginal voter moved away from the party and its performance in 2013 was understated.

The 2013 assembly elections, however, established one fact— that there was widespread support for the AAP (getting 30% of the vote on first try is no joke). This establishment of widespread support for the AAP and the decimation of the Congress in the 2014 general election (both in Delhi and in the rest of India) meant that the marginal voter tilted towards the AAP (and away from the Congress) in the just-concluded elections.

Figure 1 here shows the distribution of the constituency-wise vote swing in favour of each major party between the 2013 and 2015 elections (swing defined as vote share in 2015 minus vote share in 2013). The remarkable thing to notice is that the BJP simply held on to its votes—there is very little change between 2013 and 2015 elections in the constituency-wise vote shares for the party. The median swing towards the AAP is in excess of 20 percentage points, while the median swing away from the Congress is nearly 20 percentage points.

Essentially the story of the 2015 elections is that the marginal voter who tilted towards the Congress in the 2013 assembly elections moved away from the Congress to the AAP in 2015, on the back of the rehearsal in 2013 which showed the party’s significant strength in the state.

There is one possible fly in the ointment—if the AAP had shown its strength in 2013, why did it not win a single seat in the 2014 general election in Delhi? The answer there has to do with the fact that while the AAP had been established as a strong contender in Delhi, it was not seen as a strong national party (for the record, the party contested over 400 seats in those elections, with all but 19 of its candidates losing their security deposits), which pushed the median voter in the 2014 election to the BJP.

Look at Figure 1 carefully. You might notice that there was not a single constituency where the AAP lost vote share from the 2013 to the 2015 elections. On a similar note, there was not a single constituency where the Congress gained vote share from the 2013 to 2015 elections.

Close