Bihar Assembly elections 12 October 2015 - 5 November 2015

[Clarko95 📚💰📈]

Good. The BJP must be defeated everywhere they can be.

[jaichind]

NDTV getting it wrong in the early phases of the vote count.





As did India Today





At the same time CNN-IBN was much closer to the mark





and then

[jaichind]

Premature celebrations at BJP headquarters when various TV channels projected NDA victory.

[jaichind]

It seems other than issues with various news stringers at the vote count, another reason why the the initial trends were wrong has to do with postal votes.  Postal votes are counted first and it seems in retrospect the postal votes when very heavy in favor of NDA.  That they were counted first gave a false impression that NDA was ahead when in fact it was not. 

[Filuwaúrdjan]

Well this is pleasing

[🦀🎂🦀🎂]

This is the first gain for the InC in Bihar since 1985, from what I can find.

I'm guessing the BJP have been hurt from all these cow stories?

[jaichind]

This is the first gain for the InC in Bihar since 1985, from what I can find.

I'm guessing the BJP have been hurt from all these cow stories?

Correct on first count.  First net gain in terms of seats for INC in Bihar since 1985. 
I think what did BJP in was a lack of Hindu consolidation in fact of clear OBC Muslim consolidation.   BJP's allies did very poorly.  They were suppose to bring in the OBC and Dalit votes and they clearly failed.   

[jaichind]

All seats called
                            
                           Contested          Won          Vote Share
Grand Alliance           243             178              41.9%
   RJD                        101               80              18.4%  
   JD(U)                     101               71              16.8%
   INC                          41               27                6.7%

NDA                          243             58               34.1%
   BJP                        157              53               24.4%
   LJP                          42               2                  4.8%
   RLSP                       23               2                  2.6%
   HAM                        21               1                  2.3%

Left Front                                      3                 3.5%
   CPI(ML)(L)              98                3                 1.5%
   CPI                         98                0                 1.4%
   CPM                        43                0                 0.6%

Third Front                                    0                  2.9%
   SP                         135               0                  1.0%
   JAP                        109               0                  1.4%
   NCP                        41                0                  0.5%

BSP                          228               0                  2.1%
SHS                            73               0                  0.6%
Independent                                  4                  9.4%

Note some of these numbers are too low since they are not normalized by the 2.5% NOTA vote.

[jaichind]

Lalu and Nitish celebrates at Lalu's home.

[jaichind]

NDTV came out with a election result chart based on candidates with criminal background.  It seems Grand Alliance and NDA nominated about the same number of candidates with criminal background.  Those with criminal background won at the same rate as the alliance they belong to won at.

Alliance        Candidates   Winners   Strike Rate
Grand Alliance    142              99            69.72
NDA                   139              37            26.62
Others               754                6              0.80

[ag]

NDTV will take a while to get over this reporting disaster.

[jaichind]

As to why all the surveys and exit polls were way off  Axis APM, there are several theories. 

Famous political analyst Yorendra Yadav (who used to be part of famous 2015 AAP Delhi Assembly landslide against BJP until he was expelled) insisted the day before counting of votes that his reading of exit polls told him that Grand Alliance was headed to a comfortable victory with a seat count somewhat over 130 seats (which was the most pro-Grand Alliance exit poll other than  Axis APM.)  Yadav's argument was that exit polls in Bihar tend to under estimate support for the "OBC party", which is Grand Alliance in this case, by 2%-4%.  So if exit polls has Grand Alliance and NDA neck-to-neck then Grand Alliance should win by a conformable margin.     

I made a similar projection of a Grand Alliance victory of 140 seats but more based on the momentum these exit polls and survey had for Grand Alliance relative to similar surveys before the voting started.  My rule of thumb is that if you see such momentum in multiple surveys then most likely is is underestimated.   So if exit polls had the two sides neck-to-neck then it must be a solid victory for Grand Alliance.

It seems what took place was both these factors at the same time which pushed the Grand Alliance to an even large victory than the 140-150 seat variety. 

[jaichind]

NDTV will take a while to get over this reporting disaster.

Yep.  Their exit polls were off and the report on counting was a disaster.  I also think CNN-IBN did well by being the main English language outlet that first called the election for Grand Alliance and had in the end more accurate trends.  Somehow they must have used a different set of news stringers then that was used by NDTV.  What could have made CNN-IBN coverage perfect is if they actually went through with showing the shocking and eventually dead on exit poll of Axis APM instead of burying it.

[FredLindq]

BJP at 24.4% is +7.9% since 2010.
It is also the best results ever for the party in regional elections in Bihar.
1990 11.6%
1995 13.0%
2000 14.6%
2005 Feb. 11.0%
2005 Oct. 15.7%
2010 16.5%
2015 24.4%

[Famous Mortimer]

What percentage did the Grand Alliance lose compared to the combined JD(U) and RJD last time?

[FredLindq]

RJD 18.4% in 2015 (-0,4). 18,8% in 2010, 23,5% in oct. 2005
JD(U)  16.8% in 2015 (-5,8). 22,6% in 2010, 20,5% in oct. 2005
INC 6.7% in 2015 (-1,7). 8,4% in 2010, 6,1% in oct. 2005

[Famous Mortimer]

So who were all these new BJP supporters voting for last time?

[jaichind]

So who were all these new BJP supporters voting for last time?

Since BJP contested 102 seats in 2010 and contested 157 seats this time, BJP's greater vote share is partially explained by the BJP Upper Caste vote base now can get to vote for BJP in more seats.  BJP also gained in the Youth vote, the Dalit vote, and the EBC vote.  The latter two are more about the BJP is now THE anti-RJD party (Yadav and Muslim) so it will pick up the vote of the social base that is opposed to the RJD social base.

[jaichind]

It is also instructive to see how the each party fared against each other as that might give us an idea on how the Grand Alliance won.

INC vs BJP                  INC wins 19-9
INC vs BJP allies          INC wins  8-3

JD(U) vs BJP               JD(U) wins 29-25
JD(U)U vs BJP allies    JD(U) wins 42-2

RJD vs BJP                 RJD wins 54-19
RJD vs BJP allies         RJD wins 26-0

INC which tended to run Upper Caste candidates in an attempt to retake some of the Upper Caste vote against BJP seems successful as the JD(U) and RJD seems to be able to transfer their OBC vote to INC against the "Upper Caste" BJP.  Given how weak BJP allies (LJP RLSP HAM) performed, INC did not do as well in a relative sense against BJP allies since BJP allies tend to project an image of OBC/Dalit social base.

JD(U) ran as an ally of BJP in 2010 and because of the incumbent affect did not end up running against BJP as much.  Where it did, BJP did fairly well relative to the overall results.  This seems to be because in 2010 JD(U) was THE anti-RJD (Yadav) party and now that anti-RJD vote will go to BJP.  RJD and INC seems less effective at transferring their base to JD(U) these districts.  JD(U) destroyed the BJP allies since they (LJP HAM RLSP) have done business with RJD in the past so it is hard to project themselves as THE anti-RJD party and at the same time the BJP base did not transfer over completely over to them.

RJD did very well against BJP.  In 2010 the BJP got the anti-RJD vote anyway so its not going gain more this time as being THE anti-RJD party.  But the RJD was able to get some transfer of support from JD(U) and as a result soundly defeated BJP.  RJD ran a campaign of anti-BJP polarization to consolidate the anti-Upper Caste vote.  It seems to have worked.   In RJD versus BJP allies the social bases of the two sides seems to overlap and RJD won since the Grand Alliance coalition is just bigger than NDA.   

Overall this election is an election of negatives.  It became how effective the NDA candidate is at being THE anti-RJD party and how effective the Grand Alliance candidate is at either co-opting the BJP Upper Caste base OR being the anti-Upper Caste candidate by pulling in the Dalit and EBC vote.

[jaichind]

Once we accepted that the social bases of RJD JD(U) and INC did seems to meld well this election, from a pure vote share point of view NDA did not do that badly relative to they impressive performance of LS 2014. 

The key thing to realize is that in LS elections the vote for independents are a lot smaller than in Assembly elections.  The main reason is since the district sizes for an LS election district is much larger than in an Assembly election from a game theory point of view, it is a lot more likely for an independent candidate (often a rebel from one of the larger parties) to have a chance of winning ergo a lot more of them contest and drive up the vote share for independents.  The best way to deal with this is to normalize the vote share of election blocs to the non-independent vote. 

In LS 2014, NDA (BJP+LJP+RLSP) won 39.5% of the vote when independents won 4.3% of the vote.  This means NDA won 41.3% of the non-independent vote in LS 2014.  Likewise, UPA (RJD+INC+INC) and Third Front (JD(U)+CPI) won 30.2% and 17.2% of the vote respectively in LS 2014.  We now must take into account that in 2015 Assembly election NCP and CPI did not contest as part of Grand Alliance.  A good way to measure NCP and CPI natural support rate is to look at 2010 Assembly election results where NCP and CPI both ran by itself and won 1.8% and 1.7% of the vote respectively.  So the Grand Alliance parties in 2014 LS won (30.2%+17.2%-1.8%-1.7%) = 44% of the vote or 46% of the non-independent vote.

In 2015 Assembly election NDA (BJP+LJP+RLSP+HAM) won roughly 35.1% of the vote once we normalize for NOTA.  Independents won around 9.6% of the vote.  So NDA won around 38.8% of the non-independent vote.  Grand Alliance won around 43% of the vote which translates to around 47.6% of the non-independent vote.

So from 2014 LS to 2015 Assembly the NDA went down by 2.5% and Grand Alliance went up by 1.6% of the vote.  We do have HAM defecting from JD(U) to NDA which might mean around 1%-2% of the vote.  Once we take that into account the NDA lost around 3.5%-4.5% of the vote from the Modi peak.  While this is significant this is much higher than anything NDA has accomplished before SAP/JD(U) joined NDA back in the 1990s.  BJP has clearly established itself a a significant power in Bihar.  This is bad news for INC as the chances of the INC recovering its Upper Caste based is fairly low and unless there is another reshuffle of the political deck INC will continue to a second rate political party in Bihar as a junior ally of JD(U) and/or RJD.

[jaichind]

One of the reasons aside from momentum the NDA needed to win in Bihar is the issue of the Upper House.  Currently NDA has 64 out of 245 seats.  It could in theory try to rope in AIADMK with 11 seats and maybe BJD with 6 seats.  But even then it is far from a majority.  Had the NDA won a 3/4 majority in Bihar that over time will be able to help NDA slowly gain a plurality in the Rajya Sabha.  Just to look at the outlook for NDA improvement in Rajya Sabha for 2016 one can look at all the seats in Rajya Sabha up for re-election between now and end of 2016.

Nominated - 7 (here since the President is from the INC the 7 that will replace the 7 up for reelection are no more likely to back the NDA agenda.)

AP - 4 (2 INC, 1 BJP, 1 TDP) -> (2 TDP, 2 YSRCP).  Net gain of 0 for NDA.
Assam - 2 (2 INC) -> (1 INC 1 BJP) Assuming narrow BJP narrow win for 2016 Assam Assembly election.  Net gain of 1 for NDA.
Bihar 5 (5 JD(U)) -> (2 JD(U), 1 INC, 1 RJD, 1 BJP).  Net gain of 1 for NDA.
Chhattisgarh 2 (1 BJP, 1 INC) -> (1 BJP, 1 INC). Net gain of 0 for NDA.
Haryana 1 (1 BJP) -> (1 BJP).  Net gain of 0 for NDA.
Jharkhand  1 (1 INC) -> (1 BJP). Net gain of 1 for NDA.
Karnataka 4 (1 Ind, 1 INC, 2 BJP) -> (2 INC, 1 JD(S), 1 BJP).  Net loss of 1 for NDA.  Really net loss of 2 since the Ind is pro-BJP but I will still count it as net loss of 1.
Kerala 3 (2 CPM, 1 INC) -> (1 INC, 2 CPM). Assuming narrow LDF win in 2016 Kerela Assembly elections.  Net gain of 0 for NDA.
Madhya Pradesh 3 (2 BJP, 1 INC) -> (2 BJP, 1 INC).  Net gain of 0 for NDA.
Maharashtra 6 (2 INC, 2 NCP, 1 SHS, 1 BJP) -> (2 BJP, 2 SHS, 1 INC, 1 NCP).  Net gain of 2 for NDA.
Nagaland 1 (1 NPF) -> (1 NPF).  Net gain of 0 for NDA.
Odisha 3 (2 BJD, 1 IND) -> (3 BJD).  Net gain of 0 for NDA.
Punjab 7 (3 SAD, 1 BJP, 3 INC) ->  (3 SAD, 1 BJP, 3 INC). Net gain of 0 for NDA.
Rajasthan 4 (2 INC, 2 BJP) -> (4 BJP).  Net gain of 2 for NDA.
Tamil Nadu 6 (3 AIADMK, 2 DMK, 1 INC) -> (3 AIADMK, 1 DMDK, 2 DMK). Assuming close race in 2016 TN Assembly election with narrow win for AIADMK.  Net gain of 0 for NDA.
Telangana 2 (1 TDP, 1 INC) -> (1 TRS, 1 INC).  Net loss of 1 for NDA.
Tripura  1 (1 CPM) -> (1 CPM). Net gain of 0 for NDA.
UP 11 (6 BSP, 3 SP, 1 INC, 1 BJP) -> (6 SP, 3 BSP, 1 INC, 1 BJP). Net gain of 0 for NDA.
Uttarakhand 1 (1 BJP) -> (1 INC).  Net loss of 1 for NDA.

So NDA will have gain of 7 and loss of 3 for a net gain of 4 seats by the end of 2016.  The 2 YSRSP and 1 DMDK could perhaps be added to the parties that could be pro-NDA although TDP cannot be with YSRCP and AIADMK cannot be with DMDK.  There is also the issue of SHS perhaps defecting from NDA.  

So NDA could look forward to having 68 Rajya Sabha seats by end of 2016.  Still not enough to push through its agenda without working out deals with UPA and/or parties like SP, BSP, AITC, JD(U), NCP etc etc which currently does not support its agenda.

[jaichind]

Looking at the decline of the NDA vote relative to the LS performance gave me the idea to compare this drop-off relative to other Assembly elections at the same time or after the 2014 LS NDA landslide victory.  I am going to use the comparison of share of non-independent vote to get an apples-to-apples comparison to take into account that Assembly elections tend to have higher independents vote share.   As I look at the numbers I will also take into account the concept that if Party A and Party B contest separately their combined vote share will be higher than if Party A and Party B contest as an alliance.

The first 3 assembly elections took place on the same day as the LS election.  The numbers are in terms of non-independent vote share

Orissa.            2014 LS NDA 22.3%  2014 Assembly 19.2%  Drop off around 3%
AP                  2014 LS NDA 48.4%  2014 Assembly 48%.  Pretty much no drop-off.
Telengana       2014 LS NDA 23.6%  2014 Assembly 22.9%.  Pretty much no drop-off.

The fact that drop-off is low makes sense since the election took place the same day.  It seems that in Orissa the local BJP organization is weaker but the Modi brand is strong so even though the election is the same day.



Then we had a couple of elections in Oct 2014.

Maharastra    2014 LS NDA 53.5%.  2014 Assembly 31% (BJP+) + 20.5% (SHS+) = 51.5%
Haryana        2014 LS NDA 41.6%   2014 Assembly 37.2% (BJP+) + 4% (HJC) = 41.2%

Here in both cases the NDA broke up from the LS elections with BJP and SHS separating in Maharashtra while BJP and HJC breaking up in Haryana.  So we should take into account the split affect that the sum of the vote shares of the two NDA blocs overestimate the real vote share.  In Maharashtra this affect is mitigated by the split of INC and NCP.  So taking that into account it seems the drop-off is around 2% for this round.



Then we have a couple of elections in Dec 2014

J&K              2014 LS NDA 34.8%                             2014 Assembly 24.9%
Jharkhand    2014 LS NDA BJP 42.1% AJSU 3.9%     2014 Assembly 38.1%

In Jharkhand in the 2014 LS elections BJP and AJSU ran separately but AJSU joined NDA by the time of assembly elections.  J&K BJP local organization is weaker so there was a drop-off of almost 10%.  This drop-off is exaggerated a the BJP usually runs much stronger in LS elections in J&K relative to assembly elections.  It historically pulls in numbers like the 25%-35% range of  the non-independent vote in LS elections while in assembly elections it usually runs in the 12%-15% range of the non-independent vote.   In Jharkhand the drop-off is in theory 8% but using the split factor it is more like 6%.   In addition the 2014 LS the anti-NDA blocs were INC-JMM and JVM while in the Assembly INC and JMM also split so NDA faced 3 opposition blocs.  While this helped NDA win seats it acted to push its vote share down.    One way or another by Dec 2014 one can see the Modi factor wearing off quickly.



Then we have the 2 2015 Assembly elections

Delhi          2014 LS NDA 48.1%               2015 Assembly 33%
Bihar          2014 LS NDA 41.3%               2015 Assembly 37.7%

Here the drop-off in Delhi was a massive 15%.  But this is an exaggeration of the NDA drop-off.  BJP always seems to do much better in LS elections in terms of non-independent vote share in Delhi just like J&K.  In 1996 1998 1996 LS elections BJP would get above 50% of the non-independent vote even as in Assembly elections of 1998 2003 the BJP would get around high 30% in non-independent vote.     In Bihar it was more like 3.5% which most likely is an underestimation since we have the HAM defection factor which should had added to the NDA vote share plus the fusion of the Grand Alliance which should have reduced the overall Grand Alliance vote share to the benefit of NDA.  it should be more like 5%.

What one can derive from this is the NDA drop-off from 2014 LS results is increasing over time.  The amount of drop-off is most likely around 5% if not greater if there is a strong anti-NDA front facing NDA.

  

[BundouYMB]

Thanks for the great analysis jaichind.

[Frodo]

Hopefully the election results in Bihar shouldn't have too much of an impact on PM Modi's efforts to structurally reform and modernize India's economy:

BJP’s defeat in Bihar elections unlikely to have economic implications: Fitch

By: PTI | New Delhi | November 9, 2015 12:42 PM

You must be logged in to read this quote.

[jaichind]

Looking back, one reason for the NDA defeat is they goofed up the communal polarization card.  The lesson from 2014 LS seems to be, yes polarization by community does work but it is the second mover advantage.  In 2014 the NDA sent subtle signals of Hindu consolidations which provoked UPA into active polarization of minorities to vote UPA.  It did not work and it merely provoked a silent Hindu counter-polarization against UPA.  

So to play the communal card one must use jujitsu which is to use the opponents energy against themselves.    The BJP seems to forgot this lesson or perhaps felt they had no choice since they were behind anyway.  What the BJP did it to run against "Jungle Raj" of Lalu Yadav which everyone can read as anti-Yadav consolation by EBC and Dalits.  Also BJP came out with ads like



talking about cow protection which is an explicit attempt at Hindu consolidation.  When you have to run ads with a girl hugging a cow you know you must be in trouble.  It did not work as the Dalit and EBC vote were split and it merely provoked a silent Muslim and Yadav consolidation in favor of Grand Alliance.