Elections to State Assemblies – a simple analysis

(1) A good friend had shared the Outlook piece by Praveen Chakravarty on the latest election results to the State Assemblies. Praveen is a friend too.

(2) After reading it, I prepared the following table:

Vote share in assembly elections.png


(i) https://en.wikipedia.org/wiki/Results_of_the_Indian_general_election,_2014

(ii) http://eci.nic.in/eci_main/StatisticalReports/AE2011/stat_TN_May2011.pdf

(iii) http://eci.nic.in/eci_main/StatisticalReports/AE2011/stat_KA_May2011.pdf

(iv) http://eci.nic.in/eci_main/StatisticalReports/AE2011/stat_WB_May2011.pdf

(v) http://eci.nic.in/eci_main/StatisticalReports/AE2011/Stats_Report_AS2011.pdf

(vi) http://eciresults.nic.in/

(vii) https://politicalbaaba.wordpress.com/tag/kerala-polls/. This site has different figures for P vote share in 2014 Lok Sabha elections than presented here for Kerala, Assam and WB. The numbers are 36.9%, 10.5% and 17.0% respectively, according to this site.

(3) In the State-to-State comparison vs. 2011, BJP have done well.

In terms of vote share (of total votes polled)

  • Congress has done poorly compared to 2011 Assembly polls
  • BJP has done well in all the four States, compared to 2011 Assembly polls
  • Compared to 2014 Lok Sabha elections, BJP has a statistically insignificant improvement in Kerala. But, lost ground elsewhere.
  • Compared to 2014 Lok Sabha elections, Congress has increased vote share in West Bengal and Tamil Nadu. Lost ground significantly in Kerala and somewhat statistically insignificant gain in Assam.

(4) Between Lok Sabha and State Assemblies, voters tend to choose the same party if the elections took place within six months. See Praveen Chakravarty in MINT after the Bihar Assembly elections. That drops as the gap between the two widens, presumably because lot of factors change. Simply, it is no longer a case of ‘all factors held constant’. Here, two years have elapsed. Hence, comparing BJP vote and seat share between 2014 National Elections and these Assembly polls is not persuasive.

(5) As a matter of historical interest, in any case, two major exceptions to the ‘six-month’ test: the gap between 1980 Lok Sabha Polls (January) and Tamil Nadu State Assembly elections (May) was only four months. DMK in alliance with the Congress won in the Parliament elections. ADMK won in the TN Assembly polls.

The gap between 1984 Lok Sabha Polls and the State polls in March 1985 (Andhra and Karnataka) was even shorter – less than three months. AP and Karnataka went to TDP and Janata Party respectively. INC had won in Karnataka (but not AP) in the Lok Sabha polls

(6) Also, statistically, it would be richer analysis if one compared elections that took place after 2 years (or, around midpoint) in the previous cycles and see if the ruling Party at the Centre has done well or poorly in those Assembly elections, how much vote share that they lost (or, gained) on average and how does that compare with what BJP has achieved in 2016 in these four States.

(7) Praveen Chakravarty draws the following conclusions in his ‘Outlook’ piece:

This is not about national versus state elections. There is enough evidence to show voters do not differentiate between the two, contrary to the popular narrative that they do. Increasing voter disenchantment with the national parties carries through state and national elections. The question then is not one of Congress-mukt Bharat or BJP-mukt Bharat but an Alliance Bharat.

(8) His last sentence might well ring true in 2019 National elections. Or, may not. It is too premature. Similar prognosis was made by the punditry in 2014. They were proven wrong. So, these are still early days. Second, the conclusion that voters do not differentiate between national and state elections is not yet settled. It is a work in progress.

(9) For now, the BJP can be happy that they have not received bad news as they did in Delhi and Bihar. But, there is not much justification for extended celebration.

(10) While doing this analysis, I came across a few interesting sources of reference for future analysis:


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s