For nearly five years now, Election Metrics has maintained that in a straight fight (“two-cornered contest”), even small differences in vote share can lead to big differences in seat share (bit.ly/2zVOEWC). With the “right distribution” of vote share and in the presence of many smaller parties and independents, this column has frequently argued, it is possible for a party to get a significant lead in terms of seats even with a lower vote share.
A party with fewer votes can get more seats and win the election
Possibly the most famous instance in recent times of a “party” with fewer votes getting more “seats” and winning the election is that of Donald Trump winning the US Presidency in 2016, when he polled fewer votes than Hillary Clinton but won the election because his votes were “better distributed”. Tuesday’s results have shown a similar effect in Madhya Pradesh, where the Congress got fewer votes on aggregate than the Bharatiya Janata Party (BJP), but trumped it (no pun intended) when it came to seats, winning 114 against the BJP’s 109 (in a 230-member House).
It was a close election, and not just in the final aggregate tally. Seventeen seats were decided by a margin of less than 1 percentage point, and another 13 by a margin between 1 and 2 percentage points. The BJP and the Congress were roughly equally matched in their share of close victories. Where the Congress did much better than the BJP was where it seemingly had no chance of winning.
The BJP got less than 20% of the votes in only two seats. The Congress got less than 20% of the votes in 11 seats. Putting it another way, the Congress managed to concentrate more of its vote share in winnable seats, while the BJP ended up spreading itself too thin, thus being unable to translate its superior vote share into an electoral victory.
While getting a higher number of seats with a lower vote share is a reasonably common occurrence in India, what makes this occasion special is that the beneficiary of the “superior distribution” is the Congress. Being the oldest major party, the Congress has a presence in virtually all corners of the country, implying that it frequently runs into the problem of its votes being “spread out too thin”. Newer parties such as the BJP, on the other hand, have built up a presence in certain parts of the country and certain parts of states, giving their votes greater “efficiency” in terms of conversion to seats.
Elsewhere, Rajasthan gave further credence to the concept that in a contest with two major parties, a small difference in vote share can lead to a big difference in seat share. There, the Congress vote share was barely 50 basis points higher than that of the BJP, but this resulted in a difference of 26 seats, with the Congress nearly getting an absolute majority with 99 out of 199 declared seats (polling in one seat was postponed).
The elections in Madhya Pradesh and Rajasthan also showed up the problems with forecasting two-cornered contests. With small changes in vote share leading to large changes in seat shares, and the distribution of votes being especially important, it is not easy for pollsters to predict such elections.
Yet another reason why exit polls find it hard in a market like India is the presence of a large number of small parties and independents. These small parties (as far as Madhya Pradesh or Rajasthan are concerned) are virtually impossible to model in any statistical fashion, and they can get left out of predictions. Twenty-seven of the 199 seats in Rajasthan went to these small parties, something no exit poll had been able to predict