Sunday, April 19, 2009

Analyzing the SMS effect on 2009 Election Result


The  BJP has launched an agressive SMS campaign aiming to send SMS to 250 million cell phone subscribers urging them to vote for BJP candidate in the upcoming Lok Sabha elections. I am not aware of the Congress or Third Front having efforts on a  similar scale. Can this change the outcome of the election ?  A first cut analysis is presented here.


I believe that an SMS message is unlikely to change the party choice of an individual, but it CAN increase turnout amoung people likely to vote for the BJP anyway. Assuming that 80% of the SMS reach valid voters.  This  implies an SMS reaches 200 million voters. Assuming further that  40% of  voters  receiving SMS could be inclined to vote for the BJP.  This is a reasonable  assumption given that the BJP captured around 28% vote nationally and BJP is likely to do better than that  in subsection of the population that owns cell phones. So 80 million voters who might vote for BJP could get SMS reminding them to vote. Assuming that 50% of these voters would vote anyway, then we are left will 40 million SMS to voters who could be inclined to vote for BJP but are not voting.


Scenario 1: 


Assuming 25% effectiveness this could result in 10 million extra votes for BJP.  The SMS  is likely to reach more voters  in the urban constituencies compared to  rural ones.  Assuming  50% of  the SMS went to the 100 urban constituencies where BJP is a  factor, then the 50,000 extra votes for BJP in each of these consituencies. This can easily have the effect of switching 10 seats from the Congress to  BJP  i.e. a 20 seat swing which can dramatically alter the post poll equations.


Where can we look for this effect ? The results in Delhi  (a predominantly Urban state) would be a good indicator for effectiveness of SMS campaign. Most predictions today show either a 5-2 or 6-1 split in favour of  the Congress and vote percentages of  the Congress and  BJP  were quite close in last year's  assembly elections. Thus, if the BJP achieves aa 5-2 or 4-3 result in Delhi it might indicate that the SMS campaign played a key role. Other Urban constituencies with traditionally close results might also be good indicators.


Can we detect such a things by looking at turnout ?  Actually, it might be hard to dectect the effectiveness of SMS using turnout as the increase would only be 1-2%.


Scenario 2:


On the other hand if SMS method was only 2.5% effective,  it would translate into only 1 million votes nationwide.  In that case, it may not be able to switch the result in more than 1-2 seats and be an overall  non factor.




Wednesday, April 1, 2009

2009 Election Prediction

1st  April  2009:


UPA +  SP:   221

NDA:   201

Third Front + BSP:  115


There will be a truly hung parliament,  with  it  being  difficult  to  form any stable  government


19th  April  2009:


Updating  by  separating  the  4th  front  from  UPA:


UPA:  198

NDA:  191

Third Front: 117

Fourth Front:  30


Situation is quite similar to current parliament, but it seems there is too much bad blood between left and UPA  to form a stable government.



7th May  2009


STOP THE  PRESS:  NDA  to  form  next  government


Tamil Nadu, Andhra Pradesh, Maharashtra  to  be  drag  on  UPA.  NDA  does  better than  expected  in  UP.


UPA: 156

NDA: 217

Third Front:  134

Fourth Front:  34


NDA will form goverment  with  support  of  AIADMK  and  allies  (30 seats)   +  TDP and alliens (22 seats)











Sunday, March 15, 2009

Handicapping 2009 LS Polls


The details are still in flux,  but as the dust settles we are getting a clearer picture of the contest in the 2009 Lok Sabha.  We have 3 fronts,   the UPA,  the NDA and the 3rd Front fighting it out.  Who will come out victorious ?


For the sake of  my  analysis I the  following,  which  are  strictly  not  true.  The SP is in effectively  alliance  with  the UPA,  the BSP and  BJD are  part  of  the  3rd  Front.


All three fronts  will  have a hard  time  cobbling  together  the  numbers to form a  governments. The only front that can get an  absolute  majority  even  best case scenario is the UPA which could hit 280 seats.  The  NDA  can  only  win  225  seats  in  its  best  case  scenario,  while  the 3rd  Front  can  win  a  maximum of 170 seats  in the  best  case.  I have  included spreadsheets reflecting the best case scenario for each alliance.



However,  this  does  not  necessarily  mean  that  the  UPA  is most  likely  to  form  the  goverment  after the  elections.  This  is  because  the  UPA  also  has the  highest  "magic  number".  The  "magic  number"  is  defined  as  the  minimum number  of  seats  the  alliance  needs  to  win to  pull  together  a  post poll  coalition  that  has  a  majority. 


The  UPA  "magic number"  is  245.   If  the UPA  gets  less  than  245  it  will  struggle  to  find  allies  to  achieve  a  majority in  a  post  poll  scenario.   In  fact,  the  only possible  post poll  ally  is  AIADMK  (if  it  does  much  better  than  DMK).  The  Communists  and  BSP  do  not  seem  likely  and the rest  are  impossible.  So  the UPA  can  lose  about 35 seats from its best case scenario and still form the goverment.


However,  the  NDA  "magic  number"  may  be  as  low  as   215  as  the  AIADMK,  TDP,  BJP  and  even  possibly  BSP  can  help  the  NDA  cobble  together  a  majority.  So NDA can get 10 seats below its best case scenario and still form the goverment.


The 3rd Front has the lowest  "magic number"  of  200,  but  their  problem  is  that  the  3rd  Front  cannot  get  200  seats  in  even  the  most  optimistic  scenario.



Let the fun begin ...






Saturday, February 21, 2009

Lok Sabha and Demographic Shift


From 1951 to 1971 the number of seats in the Lok Sabha was updated each time to reflect the growth in population.  However, since 1971  the number of seats in the Lok Sabha has been fixed at 542.   In 1971 India's population was 547 million, today it  is 1.1  billion. In addition, the number of seats in each state has also remained fixed since 1971.


Due to variations in population growth in the different states in India over the past 40 years we have "significant" imbalances in  the  number of seats each state has in the Lok Sabha.   For example,  according to their current share of the population, Bihar, Madhya Pradesh,  Rajasthan and Uttar  Pradesh  (the so  called  BIMARU  states) should be allocated  193  seats  in  the  Lok  Sabha.  However, they only have a combined 175 seats, an under representation of 18 seats.  On the other hand, the southern states,  Kerala, Tamil Nadu, Karnataka and Andhra Pradesh should have  117 seats  based on their population share but  they actually have 129 seats, an over representation of  12  seats. 


In  addition, small  states have always  been  over represented  in  the  Lok Sabha. The  states and union territories with a population under  3 million should have a combined representation of 8 seats in the Lok Sabha but they actually have 19 seats. Laskhadweep, our smallest  Lok Sabha constituency has a population of only  60,000 people compared  to  the  nationwide average of about  2,000,000  people.  That  is  over representation by  a  factor of  40.


This imbalance can actually determine who comes  to  power  especially  in  a  coalition  scenario.   In  2004  the UPA and Left  Parties  gained an estimated  25  seat  swing  over  the  other parties  due  to  this  imbalance  because  they  did  well  in  the  over represented  regions  and  poorly  in the under  represented regions.  This played a crucial role in the UPA's ability to form the government.


This imbalance is not fair if we are committed to universal adult franchise.  It is the duty of the Election Commission to update the number  the seats in each state at least after each new census. 



P.S.  The case may be made that we are rewarding those states which have made the least progress checking the growth of their population.









Monday, February 16, 2009

A Political Junkie's "Excellent" Journey


Editorial Note:


This posting is more personal than usual,  I hope my daughter will get know something about her father when she is old enough to read this. I apologize if this is not of general interest.



The story begins in February 1977,  I had just turned nine (My daughter turns nine this year). India was in the grips of the Emergency, and Indira Gandhi had just declared elections. Our family was a staunch Congress household  and the Emergency had been explained to me as Indira Gandhi has put all the bad people in jail. Of course,  my father had no doubt that the Congress would be swept back to power. This sense of absolute certainty was passed on to me.  So when a friend of mine insisted that the Janata Party would win this time and offered a bet of an "orange bar" I felt that it was like taking candy from a baby though I had no idea how I would pay up if I lost.


On the day the results were to be announced,  I was  still  hopeful  as I listened to All India Radio continuously reporting about how the Congress had won 41 out of the 42 seats in the Andhra Pradesh  and silent about the results in the rest of the country.  By the next morning the papers were full  of the Janata Party riding a wave back to power and I had  sinking  feeling as I wondered how I would get 55 paise.  I dared not tell me father about  the bet.  It took me two weeks of scrounging around for spare (aka unguarded) change,  but I re-payed my debt using 11 five paise coins.  This traumatic experience led me to make a solemn promise to myself:  I will never blindly believe what someone tells me about a election,  I will always come to my own conclusions based on the data.  That month, a political "junkie" was born.


To be continued ...



Sunday, February 8, 2009

The Case for Separate Electorates



While Mahatma Gandhi is the father of the nation, and the nation was shaped and nurtured in its infancy by Nehruvian principles it can be argued that B.R. Ambedkar was more prescient about the real nature of India. There are number of issues where he disagreed with Gandhi and history shown that his position might have been closer to the truth.  Ambedkar had written in 1946 about the practical dangers of  a two-state  solution with the issues  related of massive transfers of population (partition horrors) and unending border disputes (Kashmir).  Ambedkar also sharply disagreed with Gandhi on the "romantic" nature of village life in India and correctly suggested that the best way to escape your caste identity was to migrate to the cities. Today is it clear that the best way to escape both the caste system and grinding poverty is to migrate to the cities.


In 1931 Ambedkar proposed separate electorates for Dalits  i.e. there would be separate seats in the provincial assemblies whose electorate would consist exclusively of the  "oppressed classes". Gandhi was vehemently opposed to this proposal on the grounds that it would divide the Hindu community for the future generations and went on indefinite fast to protest it. Ambedkar relented under the pressure and the result was a compromise known as Poona Pact, whereby certain seats were "reserved" i.e. the candidates would be Dalit, but the electorate includes the entire population. This agreement got codified in our constitution in the form reservation of SC/ST seats in Lok/Vidhan Sabha elections. However, other than this concession we adopted a "Westminster" style winner take all parliamentary system.


The impact of our current parliamentary system was not very clear as long as the Congress Party dominated the electoral scene in the first 30 years after independence. However, as we move to increasingly multi-polar contests we are faced with a situation where all parties practise vote bank politics.  These vote bank politics exist because various communities,  especially those who fear disenfranchisement and marginalization,  will vote as blocs to retain some sort of influence in the political process.  Political parties try to exploit this fact to win power by either supporting or opposing specific caste or religion based vote-banks.  In fact,  a cleverly cobbled together coalition of vote banks totalling 30-35% can result in a party or front  achieving an  absolute majority. 


Vote bank politics can result in both under and over representation of certain sections of society in elections. For example, in the latest Vidhan Sabha elections in Gujarat, where the Congress attempted  unsuccessfully combat Modi with soft Hindutva there are zero Muslim MLAs even though 7% of the electorate in Muslim. On the other extreme in the 2007 UP Vidhan Sabha elections, Brahmins,  who consist of about 8% of  the electorate in UP, comprised over 13% of the elected MLAs.  They managed this astounding result because of clever aggregation of vote banks by the BSP and the multi polar nature of the contest.  Interestingly,  thanks to the Poona Pact, the SC/ST are guaranteed a 22.5% representation in the legislature.


Why do we have vote bank politics ? Was Ambedkar right  that  India is  really a nation that contains many "communities" within the same borders,  having different needs and aspirations ? Should every "community" be guaranteed fair representation in the legislature based on separate electorates ?


A "community" would be allocated a certain percentage of seats based on their population.  Electorates for these seats would consist exclusively of people  belonging to that community.  This would reduce vote bank politics as each community would not have vote as a "bloc" for fear of disenfranchisement.  Each community would have its fair representation in the legislature guaranteed and there would be fewer accusations of real and imagined bias  (aka appeasement) towards any particular community.


In addition, it will raise the standard of debate within each community where there would be competition between different visions for that particular community.   For e.g. if there were a separate electorate for forward castes surely there would be parties which opposed reservation that would have representation in the legislature.  It is also possible that such an approach could also give rise to debate and moderate voices in the Muslim community.  The  list could go on and on.


The argument against such an approach is that it would codify the divisions in society for generations  to come.   However, if these divisions already exist then it may be better to move forward after acknowledging them.










Sunday, January 25, 2009

S.C. Mishra's lost letter to Mayawati


Editorial Note:


Currently we are faced with a situation where all parties practise vote bank politics.  Parties try to get votes by either supporting or opposing  specific caste or  religion based vote-banks.  In addition,  our elections are getting increasing multi-polar so a cleverly cobbled together coalition of vote banks totalling 30-35% can result in a party or front  achieving an  absolute majority. The  "letter" below  represents  a  hypothetical but plausible realpolitik scenario.  The radical centrist admires Mayawati's chutzpah, but would not neccesarily support the strategy suggested below for he is sympathetic to the Mandal report.


P.S.   S.C. Mishra is a Brahmin BSP leader from UP considered the architect of Mayawati's Dalit, Brahmin, Muslim coalition which won an absolute majority with less than 30% of the popular vote in 2007.



Dear Behenji,


             We have achieved a great victory under you wise  leadership  when we won an absolute majority in the UP Vidhan  Sabha by creating  "new" coalition of Dalits, Brahmins & Muslims. We have taken the traditional Congress coalition and turned it upside down  achieving a great victory for the Dalit cause.


            We know your destiny is to become  the Prime Minister of India by winning the national elections.  However, as results in a number  of other state elections have shown,  the BSP can be a spoiler (mainly for the Congress)  on many occasions may be difficult for the BSP to  make a major mark on the national stage. I humbly submit that we can  achieve our goal and  transform national  politics  forever,  if  you can take a single bold step.  


            The BSP should take a stand AGAINST the recommendations of  the  Mandal commission !


            It  appears shocking at  the outset that  a party inspired by Ambedkar would take a stance against caste based reservation.  In fact,  every major party  across the political spectrum whether it is the BJP, Congress or the left  supports  the  recommendation of  the  Mandal commission because no party can ignore the fact that 40%-50% of the population of India are OBCs (The main beneficiaries of  Mandal).   No party can expect to survive ignoring this vote bank. 


             However,  there exists an estimated 20% Anti-Mandal vote bank,  consisting of the upper castes.  Currently,  no political party represents this  "merit" vote bank because everyone supports Mandal.   If  a political party promised to roll back Mandal it would get the en-bloc support of  this vote bank.  In addition, you will achieve "instant" credibility among the intelligentsia who dominate the media. I am sure there will be many people hailing you as India's "Barack Obama".  Of course we will lose some OBC support, but we don't get that much OBC support anyway. We just need to change our slogan Jai Bhim, Jai Parasuram to Jai Bhim, Jai Shri Ram.

             By taking an anti-Mandal stance, and retaining the  BSP's  20% dalit vote bank,  we can cross  the  40%  threshold  in a  national  election which should be sufficient for an absolute majority in a three way race.   Of course,  BSP would need to carefully select candidates   using caste and  "merit" considerations but this is something which we are already doing.


             I am sure you will lead the BSP to a great  victory in the next Lok Sabha elections, and fulfill your destiny by becoming the Prime Minister of India.   

Forever you servant,

S.C. Mishra











Saturday, January 10, 2009

On Government Salaries & Services


Let us begin with three observations regarding  government employees.


It is clear that the salaries received by government employees  in spite of the pay increases specified in the sixth pay commission have  not kept track with the salary increases in the  private sector.  In fact, salary discrepancies have reached dangerous proportions with serious impact on corruption and stability of the social and moral fabric of the nation.


On the other hand  a significant portion of government expenditure is spent on salaries of government of government employees.  It  is estimated  that  the  recommendations  of the sixth pay commission alone will cause a 0.5% impact on the GDP of the nation.  [Need  to  find  a  good  number  as  to  the  percentage of  Central/State Budget  spent  on  employee  salaries].  This  number  is  sure  to  go up  with continued  pressure from every interest  group  which  feels  shortchanged.  The  trend of salaries taking up  ever larger percentages of total expenditure is a recipe for fiscal disaster.


A third observation,  is  that most of us are frustrated with the corruption,  slow pace  and lack of flexibility exhibited in our interactions with government employees for the simplest of tasks.   Most of us would not tolerate such service in the private sector,  but  when it comes to government we use a combination of  bribes,  influence  and  complaining acceptance  to deal with the situation.


Is  there a common sense way to solve these three issues with a single stroke,  i.e.  increase the compensation the government employees can earn,  reduce the impact on the government expenditure and improve the quality of government services received by the common man.  The answer is,  yes.


If every government employee was able to earn up to four times their gross salary by providing "value added services" to the public then we could increase the potential compensation of every government employee by four times.  These services could include anything  that  they could "legitimately" provide to the public.   This will have an impact  on both  "dishonest"  and  "honest" government  employees who would see a legitimate way to augment their income.  For  simple  straightforward  services  one  could create an efficient online market  where there could actually be bidding for the lowest quote  to  allow  market  forces  to  rationalize  the  amount  required  for  a  particular service. 


In addition,  if  the government employee  had  to  pay  a  12.5%  service  tax  on  this  compensation  then  the  government  expenditure  on  these employees  would actually go down by  50%.   The transparency of the system  could  also be  ensure  by requiring  any  service  for which the  charge   was  more  than  a  threshold  amount  (say  Rs. 10,000) needed  to  be  available  online  on  the  public  record.


This system will not eliminate corruption  (especially  large  scale  corruption)  but  it  will provide  a  serious  disincentive for small scale  corruption  and  improve the  user experience  of  the  private  citizen  with  government, while  improving  compensation of  government  employees and reducing  government  expenditure.