Delhi Results: Storm in a tea cup

                                     For many of us, who had ferociously supported Modi-led BJP in 2014 General elections for the Lok Sabha and later got back to our mundane desk jobs, Delhi results came in a complete shocker. How possible could, a party which swept all the 7 seats in LS elections just 8 months back, be defeated so badly.  This was worse than Indian cricket team getting whitewashed 4-0 in England and Australia just after winning the World Cup in 2011. And the opponent, a puny 2 year old party, which was badly decimated in the Lok Sabha elections. How could they stage such a stunning comeback? 

I decided to study the results deeper to figure out what exactly happened. Thanks to www.indiavotes.com, all the election data is readily available in a easily usable format. 

First, we should put things in perspective. For all the mind space and news space that it occupies, Delhi has limited significant in a national picture. It has 7 seats out of 543 (1.3%). In comparison, Mumbai & Thane together have 9 LS seats and Chennai / Kolkata both have around 5 seats. But the other 3 metros don't quite get the same level of attention that Delhi gets. In terms of Voter base, Delhi has around 1.3 Crore voters, a puny 1.5% of the total India base of ~84 Crores. It's ranked 19th in terms of size of electorate amongst the States/UTs of India. So though Delhi occupies the pride of place as the Capital of India, electorally, its far less important than UP, Bihar or even Chhattisgarh.

Fig 1: Rank of states/UTs by their Voter base

Vote Shares

While a lot of BJP supporters have put forth the point that BJP has managed to salvage its vote share and maintain it almost the same level of 2013 elections, taking a slightly longer-term view, its clear that from 2008 to 2015 - BJP, Cong and others have all lost vote shares - to Aam Aadmi Party. 

Fig 2: Vote share from 1993-2015

In the last 6 elections, BJP has its best vote share in 1993 - 42%. The 32.2% share that it secured in 2015 elections has been its worst performance ever. Also to be noted, any party which has secured a majority in Delhi election has always needed more than 40% vote share.  In 2013, when none of the parties could manage 40% vote, it resulted in a hung assembly. Delhi polity has always been bi-polar with BJP and Cong leaving little space for the "other" parties. The Non-BJP, Non-Cong parties like BSP, NCP etc have managed to hold to a 20-25% share of the votes in every election. In 2013, bi-polar contest was turned into a tri-polar with entry of AAP. But in 2015, Cong was decimated, leaving the contest bi-polar again. However, the non-BJP, non-AAP parties managed to secure only 13% of the votes, rest of which was entirely subsumed into AAP vote share. That this vote share transfer could happen within an year, and to a party which has no credentials, and was recently dumped in the Lok Sabha elections held 8 months back, is absolutely unprecedented. And, extremely suspect. What are the factors and forces which are behind this shift? There's definitely a lot more than what meets the eye. 

Votes Polled

Fig 3: Votes polled by each party from 1993-2015

If we look at the votes polled chart, the rise of AAP in Delhi is nothing less than spectacular. With less than an year since inception, it managed to convince 23 lakh voters to vote for them in 2013. Post multiple U-turns and 2014 LS thrashing, it managed to double the number of voters to 49 lakhs!! In fact, in the Delhi LS elections, in which approx 82 lakh voters cast their voters, AAP managed to get 27 lakh votes, higher than 2013 Assembly. BJP sailed home with 38 lalh voters and Cong sunk to 12 lakh votes. So the strengthening of AAP voter base and abandoning of congress was very much happening during lok sabha elections as well. That the Delhi voter preferred AAP over congress in the Lok Sabha elections, just after the dharna/bhagoda drama must be an eye opener. The assumption that if the elections were held right after LS elections, BJP would have won easily, should be questioned. Delhi's fling with AAP started with Anna movement and has continued to get stronger with each election. It's best to let the people get a real hang of the kind of governance which AAP will bring. Sadly, the bill for Delhi electorate's greed for freebies will have to be borne by the entire country. 

Constituency Deep dive

In all, BJP secured 3 lakh more votes vs. 2013 elections. Consider the following points -

1. In 28/70 seats, BJP vote share increased vs. 2013. This included 21 seats where BJP has lost in 2013. 
2. Off the 28 seats where BJP vote share increased, they managed to win only two - 
• Vishwas Nagar - which they had won in 2013 as well, but is a traditional Congress seat. 
• Rohini - A new seat created after delimitation in 2008. BJP won handsomely in 2008 and AAP narrowly in 2013. 
3. Off the rest, where BJP lost vote share in 2015, it managed to win only one - Mustafabad. Interestingly, this is a traditional Cong seat. And the only seat where Cong managed to put up a good fight with 32% vote share. 
4. BJP in 2015 got more votes than in 2013 in all but 18 seats. Without doubt, BJP lost all these 18 seats. 
• 8 off these 18 seats (Ambedkar Nagar, Badli, Balli Maran, Deoli, Mangolpuri, Sadar Bazar, Sultanpur Majra, Vikaspuri) are congress bastions which have not become AAP assets. 
• 10 off the 18 seats where BJP got lower votes were won by BJP in 2013 elections. Off these 10, 5 constituencies - Krishna Nagar, Adarsh Nagar, Kirari, Rithala and Tughlakabad are BJP bastions. While this may point to a internal sabotage, it's limited to just 5 constituencies and there doesn't seem to be any evidence of large scale sabotage.
• In the rest of the 51 seats, BJP has secured 3.5 lakhs higher votes than in 2013. But managed to win only 3. 

 In all, two theories 

1. BJP would have won had it held elections right after Lok Sabha elections
2. BJP lost because of internal sabotage or strategically lost

don't seem to be true. BJP did try to win, but ran a very weak campaign with poor strategy and leadership. Congress gave up mid-way through the campaign and voters seem to have shifted enmasse to AAP. Hopefully the voters would see through the hollowness of socialist promises by then and BJP would be able to come up with a better strategy to divide the opposition vote. 

 

Vote share in seats contested

Vote shares of a party is a proxy metric which is used to estimate the level of support the party enjoys across the country. Vote share is calculated by dividing the total votes secured by the party by the total votes polled across the country in all the 543 seats

 Vote share of BJP in 2014 elections is 31% - a huge 11% gap over the next highest Cong. Over the last 30 years, since the inception of BJP, this is how the vote shares have moved:

 

Fig 1: Vote share from 1984 - 2014 for BJP & Cong

If we notice, what started as a 40% gap in 1984 elections, started slowly tapering off as congress started weakening and BJP started gaining with each election. The inflection point was 1998 - when BJP and Cong vote shares were equal. Interestingly, BJP's vote share started going south again and dropped in 1999, 2004 and 2009, before finally taking a leap in 2014. The most weak performance however was in 2009 - when there was a huge gap of 10% between Congress and BJP. The flux years of 1998 to 2004 - when Cong & BJP vote shares was +/- 5% of each other offer an interesting insight, if we look at the numbers of seats secured by both the parties -

Fig 2: Seats won from 1984 - 2014 for BJP & Cong

Even though Cong has higher vote shares in 1998, 1999 - BJP managed to win 41 more seats in 1998 and 68 more in 1999. The reason is that BJP's votes are concentrated in fewer seats, and also from 1998 onwards BJP contested with alliance partners, hence contested from lower number of seats. So while BJP concentrated on seats where it was relatively stronger and managed to convert more seats, whereas Congress spread itself thin and converted fewer seats. 

Since vote share, as its traditionally calculated, is dependent on seats contested, a party which is contested from more seats stands to gain votes (and hence vote share) but not necessarily seats. I've looked at vote share from a different angle - Vote share in seats contested. I define this as follows -

  

BJP was born in 1980 and the first Lok Sabha elections it participated in was in 1984. Now, the 1984 election was a huge wave election for Congress, in terms of vote share and seats secured, it was its best performance, even better than 1952 performance. BJP contested in 229 (and won only 2) but its VSSC was at a respectable 18% even back then.

Fig 3: Seats contested (Bar) & VSSC (line) from 1984 - 2014 for BJP

Since then, BJP has steadily built through the 3 decades with a minor blip in 1991 (behind a huge 2X expansion in number of seats contested) and a decline in 2004 and 2009. The graph above confirms the fact that 2009 was one of BJP's worst performance ever. In fact, if we exclude the wave elections of 1984, 2009 is THE worst performance ever.   

To note, in years 1999 and 2014, BJP secured almost the same level of VSSC - ~40%, but ended up winning 100 more seats in 2014 vs 1999. This is largely behind the increase in number of seats contested - 339 in 1999, and 428 in 2014. Essentially, this means BJP has successfully expanded its footprint across a much larger part of the country in 2014 compared to 1999. However, there's one more factor, which contributed immensely to the record-breaking 2014 mandate. 

Fig 4: Seats contested (Bar) & VSSC (line) from 1984 - 2014 for Congress

 

In 1999, Congress which won 114 seats, had 33% VSSC in the 453 seats it contested. By 2014 however, VSSC had shrunk to 22%, even though the number of seats contested was largely the same. Of particular note, is the steady decline in Congress  VSSC from 1984 to 2014. Even though there was a slight recovery in 1999-2009, it was never able to attain even the 1991 level, though it contests in far fewer seats than it did till 1996. 

Fig 5: VSSC comparison from 1984 - 2014 for BJP & Congress

A comparison of VSSC from 1984-2014 shows that Congress is almost in the same condition that BJP found itself in till 1991. In other words, polity in India has remained unipolar - with BJP occupying the dominant pole position and congress relegated to a distant second. 1996-2009 was the period when both BJP and Congress had similar levels of VSSC, and what can be called a bi-polar polity. In the long range, BJP has taken up the pole position in Indian politics which Congress has enjoyed so far. If Modi's govt delivers well, BJP stands to consolidate its pole position and occupy it for a significant period of time. 

This should also lay to rest all aspersions of BJP enjoying support of only 30% of the electorate. BJP is enjoying as much support, as any Cong govt post 1984. In fact, if we add the VSSC of all NDA partners as well, then NDA VSSC will be higher than even that of Rajiv Gandhi's monstrous mandate of 1984.  

Let's hope Modi delivers 100% of the manifesto, so that BJP consolidates its position for the next 20 years.