Analysing the Differential Voting Behaviour of TN Voters in 2019 LS & By-election Polls

As soon as the election results for 2019 Lok Sabha (LS) polls and 22 by-elections in Tamilnadu (which were held together) were announced, it was apparent that people voted differently between the Lok Sabha elections and the Assembly elections. This is clear from the fact that while ADMK alliance managed to win 9 / 22 assembly elections that went to polls, they managed to win only 1 / 38 Lok sabha that went to polls.  

Research Objective and Questions 

The objective of this post is to share a more critical detailed data analysis of the 20 constituencies where people voted for both Assembly by-elections as well as Lok sabha polls. Please note Gudiyattam and Ambur Assy by elections have to be excluded, because they come under Vellore LS where the Lok Sabha election has been postponed and voters only voted for Assy elections. 

 Our interest is to determine answers to the following questions :

1. What is the quantified swing in vote % towards DMK (or ADMK) when it came to Lok Sabha polls ?
2. Was EPS right in tying up with PMK , DMDK and PT for the Lok sabha polls? 
3. What is the effect of the swing when BJP candidate contested in LS polls ? How did it impact the the performance of the alliance? 
4. What is the effect of the swing when DMK candidate directly contested in LS polls?  How did it impact the performance of the alliance? 
5. What is the effect of the swing when Congress candidate directly contested in LS polls?  How did the performance of the alliance? 
6. What is the effect of the other parties (MNM, NTK and AMMK) on this differential voting? How did the performance of the alliance? 

Study Design and Model 

 We took the published voting details of the 20 assembly constituencies that went into polls as our sample. We took the by-election figures for DMK+ alliance and ADMK+ alliance in these constituencies and the Assembly segment figures for the Lok Sabha polls. We ignored the voting figures of all other parties because they can always be normalised in the study. 

We normalised all the vote numbers to percentage figures so that comparison is meaningful across constituencies. 

We took the dependant variable(s) as follows for the models we were building :- 

•  % Swing towards DMK+ in the LS polls 
•  % Swing towards ADMK+ in the LS polls 

We tried a variety of explanatory variables, but in the final model, the following variables were the only biggest explanatory variables that need to be included. The list is as follows :- 

• Whether it was Theni (=1) where OPS Son contested and ADMK+ won in the LS polls 
• Whether BJP candidate contested (= 1) on behalf of ADMK alliance in LS polls 
• Whether DMK candidate contested (= 1) on behalf alliance in LS polls 
• Whether Congress candidate contested (= 1) on behalf of DMK alliance in LS polls 
• The % of votes that went to the Other parties in one of these polls. 

Please note we tried whether ADMK, PMK or DMDK or Communist or IUML candidate candidate contested was significant and they did not significantly affect the base case. So those indicator variables were dropped. 

The final Regression model results were as follows :- 

%Swing towards DMK+ in LS Polls = -8.9 - 6.9 (Theni) + 10.5 (BJP) + 18.9 (DMK) + 10.4 (Congress) + epsilon 

Adjusted R-square = 69%, Model F gives p-value ~ 0%, All coefficients are with p-value <  5% 

%Swing towards ADMK+ in LS Polls = -75 + 7.9 (Theni) - 10 (BJP) - 17.3 (DMK) -10.2 (Congress) -74.4 (Others %) + epsilon

Adjusted R-square = 66%, Model F gives p-value ~ 0%, All coefficients are with p-value <  5% 

Analysis 

The model confirms the following key observations which were apparent earlier. 

EPS Effect 

We first examine whether EPS did the right thing by tying up with PMK, DMDK, PT for the Lok Sabha polls and was it useful to boost their performance in Lok sabha polls compared to their performance in assembly by-elections. The result is somewhat counter intuitive.

Ceteris Paribus on all other parameters, there was a 7.5% swing towards ADMK alliance in Lok sabha elections compared to their performance in Assembly polls, when any non-BJP candidate contested on behalf of them. 

Ceteris Paribus on all other parameters, there was a minus 8.9% swing towards DMK alliance in Lok Sabha elections compared to their performance in Assembly polls, when any non-DMK or non-Congress candidate contested on behalf of them. 

This shows the existence of "EPS effect" where people favoured the alliance of ADMK + PMK + DMDK when compared to the DMK allies of VCK, IUML, and Communists. This confirms that ADMK got the right allies for contesting the Lok Sabha polls. 

BJP Effect 

We next examine how did Tamilnadu people accept the presence of BJP in the ADMK alliance when it came to Lok sabha polls. 

Ceteris Paribus on all other parameters, there was a minus 2.5% swing towards ADMK alliance in Lok sabha elections compared to their performance in Assembly polls, when any BJP candidate contested on behalf of them. 

Ceteris Paribus on all other parameters, there was a + 1.6% swing towards DMK alliance in Lok Sabha elections compared to their performance in Assembly polls, when any BJP candidate contested against them. 

So we can see, not only did all the good work of EPS effect (base case) got vanished, but the alliance with BJP led to a 10% vote share loss to ADMK+ and a 10.5% vote share gain to DMK+ (the balance 0.5% coming from other parties) when it came to Lok sabha polls. 

This is as clear as it gets that Tamilnadu people did not like ADMK and BJP alliance and hence punished ADMK in the Lok sabha polls in the order of loss in 10% vote share. We can all argue how their perception is wrong about BJP etc., but perception is reality and that is what hit ADMK the most in the Lok Sabha elections! 

Stalin Effect 

We next examine the DMK effect where whether there was any positive impact when DMK contested the election directly, as opposed to giving the seat to the allies. There was a lot of criticism that Stalin was extremely generous and gave far too many seats to the allies. 

Ceteris Paribus on all other parameters, there was a minus 9.8% swing towards ADMK alliance in Lok sabha elections compared to their performance in Assembly polls, when any DMK candidate contested against them. 

Ceteris Paribus on all other parameters, there was a + 10% swing towards DMK alliance in Lok Sabha elections compared to their performance in Assembly polls, when any DMK candidate contested on behalf of them. 

So we can see, not only did all the good work of EPS effect (base case) got vanished when DMK contested directly by putting their candidate, but DMK would have gains 10% vote share when they directly contest the seat. 

This is clear that Stalin made an error in giving far too many seats to its allies whereas it was the perception of DMK among the voters that actually tremendously helped the UPA alliance. 

Rahul Effect

We next examine whether there was any impact when the congress candidate contested in the seat. 

Ceteris Paribus on all other parameters, there was a minus 2.7% swing towards ADMK alliance in Lok sabha elections compared to their performance in Assembly polls, when any Congress candidate contested against them. 

Ceteris Paribus on all other parameters, there was a + 1.5% swing towards DMK alliance in Lok Sabha elections compared to their performance in Assembly polls, when any Congress candidate contested on behalf of them. 

So we can see, not only did all the good work of EPS effect (base case) got vanished when Congress contested directly in that seat, but UPA would have gains of 1.5% vote share when they directly contest the seat. But the corresponding ADMK loss was 2.7% indicating some of the ADMK vote loss went to "Other parties" when congress contested there. 

This is clear that there is a positive Rahul effect in Tamilnadu, but this is not as strong as the DMK effect. In fact when congress candidate was there, some of the people switched to other parties instead of voting for DMK alliance. 

Theni Effect 

 Theni went different to other seats because there ADMK won the Lok sabha seat. This may be due to OPS son contesting there as ADMK candidate, or because the opposition candidate EVKS Elangovan was a complete outsider imported from Erode district. But whatever was the reason there was a Theni effect which took Theni in the opposite direction. 

Ceteris Paribus on all other parameters, there was a + 15.4% swing towards ADMK alliance in Lok sabha elections compared to their performance in Assembly polls, in Theni. 

Ceteris Paribus on all other parameters, there was a minus 15.8% swing towards DMK alliance in Lok Sabha elections compared to their performance in Assembly polls, in Theni. 

So we can see that the EPS effect was accentuated in Theni with a OPS effect presumably and the swing was 15.4% in Lok sabha elections (compared to the base case of  7.5%. In fact DMK suffered 15.8% vote share loss here, indicating that 0.4% of the DMK votes went to other parties (presumably AMMK) instead of going to EVKS Elangovan. 

Other parties Effect 

We next examine whether the presence of the other parties (MNM, AMMK, NTK etc.) had any impact on the differential voting between assembly by-elections and Lok sabha polls. 

There was no impact of the presence of other parties on DMK alliance performance. 

But this impacted ADMK alliance performance by a degree of  minus 0.744 % for every 1% swing that went towards the other parties. 

Since these numbers are materially insignificant, we can say the presence of other parties did not matter when it came to differential voting. 

Conclusions 

The summary of this differential voting behaviour reveals some interesting insights. 

• EPS did the right thing by roping in PMK, DMDK, PT etc for the Lok Sabha polls as it gave them a positive advantage over their performance in Assembly by-polls. 
• Presence of BJP in the alliance hit ADMK alliance very hard when the candidate was a BJP candidate. People punished ADMK alliance in LS polls more when BJP contested. 
• The performance of DMK alliance is largely due to the perception people have on DMK. In fact Stalin was wrong in giving far too many seats to his allies. 
• There was a positive perception against Congress as well. It may not be as strong as DMK, but it exists. 
• Theni seat witnessed an accentuation of EPS effect. We can explain it as EPS + OPS effect or a EPS + EVKS effect. Whatever it is, Theni was different from others.  
• The presence of other parties like MNM, AMMK, NTK in the lok sabha seats was immaterial in explaining the differential voting. Their voters are their voters. 

Did Dravidianisation of Tamil Hindu Practices contribute to Jallikattu ban?

The traditional bull taming or bull hugging sport of Tamil Nadu, called Jallikattu, is now back in the news. This is an ancient 5,000-year-old sport that is deeply enmeshed in the rural way of life and worship of local deities. Sangam literature (2nd century CE) has several references to the sport of bull hugging (ஏறுதழுவல்). It is usually conducted as a part of week-long festivities associated with the Hindu festival of Pongal (Makara Sankranthi).

The day after Pongal is dedicated to bovine farming animals when they are decorated, celebrated and worshipped. However, based on a petition by animal welfare/rights groups, the Supreme Court banned Jallikattu in May 2014. Since then the people of Tamil Nadu have felt victimised over this emotive issue and have perceived this ban to be a grave assault on their longstanding culture, tradition, and worship patterns.

Although the people and politicians of Tamil Nadu blame everyone from animal welfare groups to the Supreme Court, to foreign corporate firms, did the native Dravidian movement wittingly contribute to this degeneration? Can simple apolitical farmers of Tamil Nadu (and elsewhere) understand the machinations of globally organised NGO networks and the Idea-of-India lobby? Can they learn from the vulnerability fostered by more than five decades of Dravidian movement? What can they do to prevent being easy pickings for more of such cultural assaults?

To read more, click here for the original article in Swarajya magazine  

Is the Net Neutrality and Telecom Pricing Dilemma for real ?

Indian Telecom service providers (TSPs) such as Airtel, Idea and Vodafone have been complaining about over-the-top service providers (OTT SPs) such as Whatsapp, Facebook and Viber for more than a year now. Their complaint is two-fold. First of all, these OTT SPs are able to provide competing services of voice calls/text messages without having to pay spectrum license fees, termination charges to other service providers or abide by other security/statutory regulations. Secondly, these OTT SPs are rapidly growing at the expense of, and as if to add insult to injury, on top of the very telecom infrastructure created by these TSPs.

Last year, Airtel proposed a policy of price discrimination, to charge data usage differentially for these communication services. But Airtel was taken aback by the big public outcry especially on social media on the principle of “Net Neutrality” and deferred the price discrimination scheme. Meanwhile, Telecom Regulatory Authority of India (TRAI) seized on the issue and produced a white paper. TRAI has invited suggestions/responses from public on questions raised by the white paper. The objective of this article is to review some of the grievances expressed by Indian TSPs and analyze some of the solutions discussed in the white paper / proposed by policy experts.

For reading the complete article from Swarajya Magazine click here 

Part-III: The Vicious Trap of LARR Resettlement & Rehabilitation

In the first part of this series we explored the need for government intervention in the land acquisition process. In the second part we addressed the ethical question of who owns the premium accruing out of change of land use purpose. In this third and final part we discuss some of the important clauses in the Rehabilitation and Resettlement (R&R) part of the LARR Act 2013. The LARR act 2013 is different from the Land Acquisition Act 1894 because of the inclusion of rehabilitation and resettlement clauses within the same act. Hence the R&R part is perceived by activists as the crowning jewel of social justice. In this article we discuss some of the darker sides of the act and explore whether such a crowning jewel of social justice, truly stands up to the welfare of all sections of the society. 

For reading the complete article click here. 

Part-II : Who Owns the premium accruing out of change of land use purpose? An ethical perspective

The first part of this series focused on the need for a land acquisition act. In this part we focus on a substantive issue on one particular component within the compensation and rehabilitation package. The premium accruing out of change of land use purpose from agriculture to the amended purpose. The LARR act 2013 passes this entire premium to the titleholder. This is as much as 2 times for urban land and 4 times for rural agricultural land. Now is it fair that the title owner claims ownership of the entire premium? Who exactly owns this premium, considering the history of land tenure in India? Who exactly owns this premium, by taking an ethical perspective of value addition?  

 Click here to read this complete article. 

Part-I : Why do we need a Land Acquisition Act ? - A Game theoretic perspective

Agricultural land is much more than a mere capital asset owned by farmers. It is not only a source of livelihood but also a source of financial security, lifetime memories and emotional strength. Hence any parting with farmland, whether by volition or by force is always an emotional act for the farmer. That is why many argue, even if the farmer was compensated with a land of equal size and arability, things will no longer remain the same in his life. But for a rapidly growing and urbanizing country like India, industrial growth and urban expansion is inevitable. 

The objective of this article is restricted to one controversial aspect of the land acquisition process viz. Why do we need government intervention in the land acquisition process in the first place? Why can’t we let the buyers deal with the sellers directly via market mechanisms? 

Click here to read this full article that was published in Swarajya Magazine. 

Review of the Research Paper - Does Affirmative Action Affect the Productivity of Indian Railways?

A research study appeared recently that analysed the impact of affirmative action (caste based reservation system in India) on the productivity of a PSU (Indian Railways in this case). The study was conducted by Ashwini Deshpande & Thomas E. Weisskopf in 2011. A link to the pre-published version of this paper is available (here). This blog post is a review of this paper (strictly) based on the version hyper linked here. 

The objective of the study was to prove that caste based reservation system does not impact the productivity of an organization. A systematic study would not only shut the voices of people raising this concern against the reservation system, but also provide valuable supporting evidence for the Supreme Court of India to make a decision on several different Mandal Commission related cases pending before it. Although the claim of the authors may be true, is the study really producing such a clinching evidence to make such a claim in support of caste based reservation system? 

Here are some of my critical observations of the study :- 

1. If you read the statement linking Footnote 4 and what the actual footnote 4 reads, you will get a good idea of how exactly the study flows in terms of linking things.
2. Firstly at a conceptual level to argue that Labour mix significantly impacts how many passenger KMs are produced or not is a stretch for an asset intensive, capital good intensive industry like Railways. Secondly it is a monopoly and certain amount of captive demand will get filled irrespective of what the Labour mix is. And this captive demand will be increasing over time because of population explosion and increase in economic mobility in the country. This captive demand hike will NOT be captured by fixed effect time variable during regression. We need a dynamic time effect that is increasing with time to capture this effect. A better proxy is to use GDP of that year or for that state(zone) as a control variable. They have not controlled for this aspect. 
3. Taking passenger KM as a measure of output assumes that demand is infinite and capacity utilization is not a function of demand. It is only the other production factors that influence how many passenger KMs are produced. I am not sure how Railways demand can be modeled this way de-linking from the demand. This is a critical concern that needs to be addressed. 
4. Railways output in terms of Passenger KM is a function of how many new trains are launched and in which routes (which are budget announcements). One can argue time controls for this during regression. But a pure fixed effect regression may not adequately capture this. We need a better control variable. 
5. Railways policy of how "crowded" their lines can be has evolved with time. With better technology (and some say reduction in safety standards) they have allowed for more "crowding" i.e. they run two consecutive trains on the same track with lesser time gap. One can argue technology controls for this effect. There is another issue with technology variable they have chosen and we will come to it later. 
6. They are using FUEL QUANTITY as an independent variable. This is normally a good proxy for technology in many capital intensive industries e.g. Iron & Steel. But during regression this factor should pop up with a negative sign i.e. the quantity of fuel consumed for covering the same Passenger KM should decrease with time. This has not happened and this is already triggering alarm bells on how technology has been controlled for!!! 
7. Any employee with time "learns" on the job. The rate of learning may be different for different employees. if critiques of caste reservation policy are to be taken seriously, they may argue this rate is different for different categories too. However the model makes no effort to consider the learning effect of employees both general and SC/ST over time. The existing model has a secularly increasing figure of %SC/ST over time. This variable combines older employees with newer employees and can produce a spurious result due to the presence of learning effect over time! This wont be captured by fixed effect time variable as it an interaction variable. 
8. The first approach to comparison is some what naive and I dont lend too much credence to it. The second approach is a more appropriate to study a problem of this type. But what I don't understand is why has time adjustment not been made in the 2nd step regression in the second approach ? That has potential to throw spurious results. We need to see the exact equation they have used.
9. Normally such problems should be strictly studied in a "difference in difference" approach by having a dataset that covers zones/time periods where the policy was not introduced with the ones after it was introduced. I understand they have data limitations in attempting to do anything like this. But at the same time they should give benefit of doubt to their limitations. 

To summarize the following red herrings stand out to me :- 

1. Choose an industry such as Public sector Banks which are service businesses and the quality/productivity/performance will be strongly linked to the human resources they deploy. Furthermore banking sector is competitive and hence if people are unhappy, they will move to another competitor. This will make a more convincing case to study than an asset intensive monopoly like Railways.  
2. Unlimited Demand assumption. This is invalid and needs to be accounted for. 
3. Improper use of Fixed time effect as it does not capture dynamic captive demand surge
4. Not controlling for Govt of India announcements and Railway "crowding" policy changes. 
5. Not having a good proxy for Technology. In fact having a bad proxy. 
6. Not accounting for the learning effect of the employees. 

It would be nice if they can overcome these limitations and produce better version of analysis so that we can all appreciate the results better.