Method to the Madness

Chetan Bhagat clearly had it all wrong. He thought that two states were problematic. What then, would he say of twenty four?

Twenty four states, two thousand five hundred and fifty three unique values, all to be sorted into a 290 x 24 matrix. This was the December that Shiva and I faced. I’m happy to say we are still standing at the end of it all, armed with a notebook full of pertinent observations to boot. Shiva gave you his take on the exercise last week, so to follow up, here I am with my thoughts.To quickly recap, the Rule of Law project (henceforth the ROLP) is collating a database of all the cases pending before the higher judiciary of this country. We seek to make the ROLP database accessible to the legal fraternity, government bodies, independent researchers and most importantly the general public. The database will contain the case numbers of pending cases. As Kavya has mentioned earlier, these unique case numbers are made up of a case type abbreviation, a number and a year.

To give an example, in the case number WP12345/2006, WP stands for Writ Petition, 12345 is the number and 2006 is the year. What’s that you say? That sounds easy enough? Well my friends, that’s merely the tip of the iceberg. Remember that number we mentioned earlier, a certain two thousand five hundred and fifty three? That is the collective number of case types found in the nomenclature systems of twenty four High Courts across India. Not only do we have the abbreviation WP, we have CP and OP, WA and WTC, and every permutation and combination in between.

For us at the ROLP, it was crucial that we created some sort of key or index for these case types. If lawyers themselves did not know case types, corresponding or otherwise, in states where they did not practice, what hope was there for an ordinary Ram who wanted to do some research on our database? It was at this juncture that Kishore uttered those four words, the ones which filled our mails, our thoughts, our talks and at times our dreams for the next three weeks. ‘Case Type Standardization Table’, or as I like to say, the method to the madness.

We got to work at once. The first stumbling block was waiting for us, right by the start sign. There were no readymade list of case types, each set had to be painstakingly listed out from a case status page, available on the website of the respective parent High Court. At the end of two days, with our aggregated table in front of us, Shiva and I were stupefied by the numbers we were dealing with.
Calmness came in the form of Harish, who told us to take it slow and steady. So we started off, choosing Bombay as our reference court (it had the most case types – 289) and splitting the remaining twenty three courts between us. The process we followed was to match each High Court’s set with the Bombay list.  Day one was rather traumatic. Our conversation was filled with exasperated questions ‘How can Bombay not have a Death Reference?’  or ‘Where do I put Contempt Appeal, under contempt cases or appeals?’. However, it did get easier, and we got better.

In a week we knew all the facts off the top of our head, how many case types each High Court had, which court had a specific type of case, which didn’t, we had our own little KBC[1] going on. In a couple of weeks we had finished the process of matching to Bombay, and heaved a sigh of relief. There was still more to come, with what I call ‘un-matchables’, case types that we could not correlate with the Bombay list. So we rounded up all the un-matchables, categorized and matched them amongst each other and added another hundred odd case types to the Bombay list.

There you had it, three hundred and eighty seven case types. They filled in Column ‘A’ of the (in)famous 'Case Type Standardization Table’. We categorized them broadly as part of the exercise, coming up with a hundred categories they could be sorted into. All of a sudden, we were finished.


There is clearly lots to take away from the operation we carried out. It was exhilarating to work on a pioneering effort like this, and to create a key that has not been publicly available before. The flipside was the bewilderment as to why there is no uniformity in the system. To give you an idea, at the Supreme Court level, which is the only place cases from the High Courts can go for appeal and is common to all, there are just twenty nine[2] distinct case types. How does two thousand five hundred magically morph to twenty nine?

The level of variance that we have in case types at the judiciary is confounding. To draw parallels with the other two arms of the government, it would be as if we had vastly inconsistent state rules under a central legislation or varied departments and officers under each state executive.


Perhaps even more perplexing than the variance itself, was the fact that the variance was variant. What I mean by this is that there was:

•  Variance in Type – There were case types that featured on a particular High Court list, that were nowhere to be found on another list. 

• Variance in Form – Writ Petition was not given the same name across lists, nor were its sub categories uniformly named. Like this there were many cases with numerous nicknames and alibis. 

• Variance in Local Flavour – There was a clear local flavour to each list, case types unique to each would show up. For example Bombay had numerous case types dealing with Parsi Cases, and Kerala had a number of Dewaswom Board related case types. 

• Variance in Level of Classification – While the High Courts of Bombay and Jharkhand both had excellent classification, the former list had 289 entries whereas the latter had just 59 entries. So there was huge contrast in the level of detail each High Court showed.

• Duality – There were case types that could clearly fit well into two separate categories/ case types, but were placed under one type as there was no cross referencing between types.

There you have it, our Case Type Standardization Saga. Right now we are four weeks from where we started and have completed our first Case Type Standardization Table, the first of its kind in this sphere. We’ve found a method to the madness, and next up, hope to render the variance invariant. 

[1] Kaun Banega Crorepati – The Indian television game show
[2] http://sc-efiling.nic.in/sc-efiling/helpdemo.pdf . Refer page 24 of the document for a list of case types that are accepted while making e-filings before the Supreme Court of India. 

The Matchmaking Exercise

As a part of the Rule of Law Project, Ramya and I were assigned the task of doing something which had never been done before. 24 High Courts, about 2000 different case types (each case is classified into a particular type by the Court. For example, a case which deals which contempt is usually classified as a contempt petition. There is a high degree of variance in this nomenclature as different High Courts may assign different names/case types). Standardising these case types across the 24 High Courts in India and coming up with one comprehensive list which would cover all the High Courts. Each of the 24 High Courts have a few common case types and several uncommon ones. This exercise looked to make a single table to see, for example, what a Writ Petition would be named across the 24 High Courts*. Seems manageable? Read on and you will see how this task can make understanding Christopher Nolan’s movies seem simpler.

The task began with extracting the different case types from the websites of the respective High Courts. While the popular High Courts, such as Bombay, Delhi, Madras, Karnataka, Gujrat had the expansions of case types, a few others like Andhra Pradesh, Gauhati, Tripura, Meghalaya did not have the expansions readily. The expanded case types are indispensable in such an exercise as a ‘Co. Pet.’ can be read as a Company Petition as well as a Contempt Petition and an ‘Appl.’ and ‘Apl.’ can be read as an application or as an appeal and a ‘Rev.’ can be read as review as well as a revision.  About 80% of the expansions of case types readily available and about 10% were on a scanned page which had been uploaded as a part of the High Court Rules of the respective states. For the rest of the states, such as Jammu and Kashmir which did not have the case types nor the expansions, we had to visit the cause lists (every court lists out cases which to be heard on that particular day) of those High Courts for about a week daily and extract and expand the case types. While most of the High Courts had the expansions of case types readily and some as a part of the rules, High Courts such as Jammu and Kashmir did not feel the need to put them up anywhere.

After the mammoth exercise of finding the different case types of the 24 High Courts, it was time for the real deal. As Bombay had the highest number of case types, it was chosen as the base list with which the case types of other High Courts would be matched. At this point we noticed some expected similarities as well as many glaring differences in the case types. This trailer showed us that the ride was going to be a super bumpy one. Here are a few instances of the vast variance in the case types of various High Courts.

Variance 1: Let’s take something as simple as a Public Interest Litigation (PIL) to start off with. A few courts decided to have them under the case type of Writ Petitions, the others decided to have a special case type for them, named mostly as ‘PIL’. Some High Courts such as the Bombay High Court went a step further and had a ‘Criminal Public Interest Litigation’ case type as well, other than the ‘Public Interest Litigation’ type.

Variance 2: Some courts had a case type for bail application, many others did not even have a separate case type for bails! While Uttarakhand had first, second, third, fourth, fifth and a sixth case type for bail applications.

Variance 3: While only Calcutta and Bombay had a special case type for BIFR (Board for Industrial and Financial Reconstruction) cases, none others had a special case type for it.

Variance 4: Orissa had 3 specific case types for Matrimonial related matters [MATA-Matrimonial Cases (Appeals), MATCAS-Matrimonial Case (Suits/Petitions), MATREF-Matrimonial Case (References) and MREF-Matrimonial Reference] while many High Courts did not even have a different case type for matrimonial matters. Gujrat had a special case type for petitions under the Christian Marriage Act and Bombay for Parsi Matrimonial Matters, but none of the other courts reflected anything remotely close.

It was heartening to see that there was uniformity in case types relating to Arbitration, Company and Contempt matters. The entire exercise showed us how incoherent the case types of various High Courts are. This exercise lasted a little over 3 weeks, with myriad calls/mails to members of the legal fraternity all over the country and many disturbing dreams while asleep on the intricacies of the high degree of variance.

Matchmaking is hard, isn’t it?

*For convenience, each High Court has been referred to by the name of the state/city mentioned in its actual name.

Ctrl+C and other stories.

In the process of unfurling our research method at The Rule of Law project it has been important to examine what data can help us frame information that can answer the question of what “delay” and "pendency" really means, outside of the anecdotal narratives that get thrown up both in the media and in reports of and by the courts themselves. If we were to generate statistics on the nature of judicial delay, what would we need to do? 

My fellow travellers and guides Harish and Kishore, a few weeks into letting me roam the lands of literature review, came up with an ingenious solution.

Look to the cause list.

I'm fairly certain, after several months of speaking of cause lists with something like obsession, that the lay person in our country will know little or next to nothing about what a cause list is.

A cause list is produced daily by every court in the country, detailing the case number, the litigants, the lawyers, for the reference of those who are in the process of litigation. This is a public document.

The cause lists, however, are a kind of endlessly disappearing archive of information, since they are put up each day for the use of lawyers and litigants, and removed once their use passes. Beyond a certain amount of time (a week, perhaps a fortnight), these lists get taken off the main website.

A cause list is, on some enquiry, a straightforward document: it tells you which case is heard before which judge on what day. In this, it is very valuable, since each case number that is in process in the courts will necessarily show up at some point in a year, during the various stages of life in the courts. As we said in our last post, there are 44,56,232 pending cases in the High Court. The cause lists will ensure that we are able to capture these cases in our data in the coming year.

And most crucially, what we can capture with this information is the case number itself.

The case number is, for example, WP13457/2004. A Writ Petition, a number, and a year.

A typical cause list contains the following:

·  Case Number

·  Date of Hearing

·  Judge

·  Hall Number

·  Court/Bench

·  Advocates for Petitioner

·  Advocates for Respondent

·  Stage

·  (for eg, Orders, Interlocutory Appeals, Preliminary Hearing.)

With each Case Number, we are able to access the following details:

· Date of Institution

· Date of Disposal in Lower Courts

· Case Status

· Causes of Adjournments.

· Number of Times Listed

· Details of what the case is listed for (for eg. Non-compliance of office objections, etc.)

· District of Filing of Case

How were we going to put this information together, everyday, across 24 High Courts of the country?

A woman's best friend in such a situation is really manual data entry. We started our work by spending up to a week copying and pasting each element of the data available in the cause list into an excel sheet: a case number and all its corresponding details.

There is, for example, no standard format for the digitised PDF of the daily cause list that is put up on a High Court's website. It is typically a type-written document, put together by a clerk, scanned and uploaded on to a High Court's website as a PDF. A document that consists of scanned images of text is inherently inaccessible because the content of the document is images, not searchable text. Sometimes, these PDFs are not scanned images but rendered with html; copy and paste a single line of this into an excel sheet and you will be left with a glorious garble that breaks down the tentative coherence in format effected by that judicial clerk who has compiled the cause list that morning.

Each High Court, too, has different formats for the way a cause list is organised. Each daily cause list typically runs into up to 50 pages, ranging between 800 to 2000 cases before a few judges in a day.

I started with three High Courts: Karnataka, Delhi, and Gujarat. In the case of most cause lists, the process of copy and paste was fairly straightforward, since the PDF allowed for it. However, there were those cause lists whose PDF did not allow for an easy cut and paste of information. What I discovered was that while it took me about an hour and a half to put together details of 250 cases into an excel sheet from a PDF from the Karnataka High Court, it took me about two hours to put together merely 20 cases from a cause list of the Gujarat High Court.

It became quite clear from this exercise that this manual data entry would not only be inordinately time consuming, but that it was necessary for us to begin to understand what a large amount of this data collated could show us about possible inferences.

At this stage, we turned to the data entry elves at a small service provider called Data Con Services. Quite close to Toll Gate in North Bangalore, Data Con services is a small house converted into an office space full of cubicles, populated by young men and women typing away and keying in data, row by laborious row. Having spent some time keying in this information myself, I knew that this work involves a mixture of tedium and close attention.

Our data entry elves were able to collate and compile the daily cause list for four High Courts and managed to put together well organised excel sheets for 20 days of cause lists in the span of two months. By the end of November 2014, we had data for 80,000 cases in hand, done all by the force of pure manual labour.

While our team had spent a few months beginning to understand all the possible ways in which we could use the information in a cause list for our research, Kishore had spent his energy understanding the way the courts organised their data on the High Court websites. This proved to be an inordinately useful two-pronged approach, since our team was able to understand the ins and outs of the information even as softwares were being written to make the data available to us.

For example, our team undertook the task of picking a randomised data set from the Karnataka High Court of October 2014, to trace the date of institution for each case number. This involved another week spent with Ctrl+C, that gloriously simple yet tedious function: feeding in a case number into a High Court website, to trace it backwards and see when that case was instituted, and from which lower court it emerged. This process of tracing date of institution helped us understand that a simple case number points towards several layers of important data. 

It was by this time that softwares that defy the imagination of an ordinary sociologist came into force. Kishore had cracked for us a method that parsed all the data on the High Court websites into clean and crystal clear excelsheets: what had taken us a few months to understand, a software broke down in a matter of mere seconds. 

So it turns out that what we had considered virtually impossible, turns out to be possible. Armed as we are with a fast growing database, we will leave you again, to return later with more stories of our methods, of case numbers and case types and our ever growing, never ending relationship with excel sheets. 

“…the case went on for years and years and years”*

Just last week, the mother of all delayed cases – the Disproportionate Assets Case against former Tamil Nadu Chief Minister Jayalalithaa, under the Prevention of Corruption Act – has come to a head. Jayalalithaa was in the Supreme Court yesterday, which has deemed the necessity of a swift judicial process in the hearing of her appeals. An exclusive Special Bench will be constituted to address her defence, in the Karnataka High Court, with daily hearings scheduled to finish within three months.

The story of Jayalalitha- the 800 kg of silver, the 28 kg of gold and the 750 notorious shoes, has become the stuff of common parlance for what signifies corruption. In nearly two decades, the corrupting figure of Rs. 66.65 crores – crumbs in the broad expanse of corruption in the legislature – has been featured endlessly in the media.

It is, however, a true case of missing the wood for the trees.

An eighteen year old case, this judicial saga has seen a remarkable trajectory that serves to highlight one of the primary ailments of the Indian judiciary: delay.

For us at DAKSH, the question that is far more pressing is, has justice been done by the recent verdict in the Jayalalithaa case? It is true that this is a rare case of a Chief Minister being brought to justice by the judiciary, demonstrating the independence and possibility of the precept of “nobody above the law” encoded into our Constitution. However, is it justice if it takes eighteen years to deliver it?

Earlier this month, Chief Justice of India H.L. Dattu expressed concern over the high numbers of pendency in the Indian courts, as one of the biggest hurdles for the dispensation of justice in the country. On the same day, litigants who have seen 1500-1900 dates in court, formed a human chain to stage a protest outside the Delhi High Court, with the one statement that "We only want justice not delay." Justice A.P. Shah famously remarked in 2011 that it would take the court up to 466 years to clear the backlog in just the Delhi courts.

The problem of pendency -- all cases instituted but not disposed of, regardless of when the case was instituted in the courts -- we have found, is a much studied one, accompanied as it is with a widely held common sense that sees the judiciary as several steps behind its potential in being unable to deliver timely justice. “Justice hurried is justice buried” and “Justice delayed is justice denied”: commonly understood maxims that tell us that the quality of justice is very much dependent on the time taken to deliver it.

DAKSH’s new work, The Rule of Law project , looks at this problem of pendency and backlog in the Indian legal system. As a group of lawyers, sociologist and data analysts, we are excited to bring together a different register of critical energies to understand this problem.

Our first question at the outset of our study has been, what data do we begin to work with? What is pendency, and how do we begin to address it in our research?

We begin our pursuit of this study with a very preliminary and layperson’s perspective of the legal system. The legal system exists to ensure that in case of the violation of this personal freedom, or if a citizen has violated the legal code of our country, justice is meted out in a qualitative manner in accordance with the legal spirit of the country.

While we live in a country with a vast and sprawling network of well-ordered judicial institutions, these systems are densely congested with cases that they are unable to see through in a timely fashion. To put it more simply, while there exists a dynamic and extensive judiciary, the numbers of cases lodged in the system that have not yet seen the light of day, remain an immense challenge to the quality and delivery of justice.

As preliminary leg work for our project, our team has carried out an extensive review of literature to understand why this is the case. What does “delay” in the courts really mean? How do we quantify pendency, and what amount of time is too much, when a case is in process? Can there be a predetermined mandate for how long the resolution of a single case ought to take in a court?

It is with these questions that we have come to understand at DAKSH that a simple number, statistic, or even prescription cannot hold much value in looking at this problem.

We understand from the literature made available by the Supreme Court itself in Court News, that there are 64, 330 cases of pendency in the Supreme Court, 44,56,232 in the High Courts, and 2,68,39,256 in the District and Subordinate Courts, as of 31-12-2012. What do these numbers indicate? If each case in this system has been in the system for only six months since its admission, the number appears less daunting; however, if these cases have been in the system for years, they indicate a wholly different tendency.

So, what is the tendency of pendency? In a country where the judiciary can vary across different states, where the profile of litigants and cultures of litigation differ, what common index can help us understand the difference in what backlogs indicate for different contexts?

In response to various studies that have studied the problem of pendency, DAKSH has undertaken the quantitative task of collating and collecting primary data that is made publicly available by the courts themselves, to allow for further public debate and research that will hopefully be able to throw light on the specificities thrown up by this issue.

It has been seen that it is crucial to produce reliable statistics of the kinds of delay in the various courts in India. Official statistics do exist, but it is not entirely clear what these numbers indicate. More cases are registered in the courts than are resolved, and the movement of disposals in a responsive and qualitative manner will depend, in the future, on a management of pendency within the courts.

Our work at The Rule of Law project is to collect and collate data on the High Courts in the country at an all-Indian level that can be used by researchers, lawyers, even the courts, to study the systems of litigation in our country. As a starting point, The Rule of Law project is collecting primary data from the daily cause list that provides a detailed profile of cases, from different High Courts in the country. Over the course of the coming year, we will be collecting this data for all 24 High Courts and the Supreme Court.

Stay tuned for upcoming blogposts where we will tell you more about our method, which has involved a lot of minute work understanding the courts, and we will present a series of reflections and experiences in collating and organising our data, and the challenges they pose to the study of delays at the court in our country.

*  The blog title is taken from Chief Justice of India H. L. Dattu's comment in the Supreme Court, can be found here: <http://www.thehindu.com/news/national/as-it-happened-a-minutebyminute-account-of-jayalalithaa-bail-hearing/article6510700.ece>

DAKSH-India Together Fellowship

DAKSH and India Together (www.indiatogether.org) have instituted the DAKSH-India Together Fellowship to encourage data journalism centered around elections. Two of our scholars have written their first pieces- I found both of them to be extremely insightful. In this age of "opinion journalism", this is certainly welcome trend. Congratulations to the editorial team of India Together in starting this initiative and partnering with DAKSH.
http://indiatogether.org/low-presence-of-women-in-lok-sabha-government
http://indiatogether.org/fragmentation-in-indian-party-system-analysed-government

Caste- myths and realities

In election season, everybody is a pundit and has a view about the factors that will determine the outcome of the election. The diversity of the electorate ensures that everyone can get their voice in and can claim credit for predicting something right!  One of the all-time favourites during election is the argument around caste and its importance.  For the national English media, the relevance of caste is almost a holy grail, partly because they are unable to understand the intricacies of caste at a local level; it is one of the “others”, but an “other” they believe to be important because everybody including politicians tell them it is the most important other in politics. Regional media thrive on caste; in Karnataka, my home state, veteran journalists and political analysts cry themselves hoarse about various sub-caste intricacies and how one important sub-caste is important or has been hurt and is bound to impact the result dramatically.  And yet, after each election, politicians and the media go to town talking about how the vote was for change (positive or negative) or for good governance or development!

In the DAKSH and Association for Democratic Reforms (“ADR”) survey, we asked the respondents whether caste was important to them when voting in an election. I have tried to capture some of the results in this short entry.

The big numbers- 58% of our respondents (nearly 2,50,000 of them) said that caste of the candidate is not an important factor for them while voting in elections; 24% of our respondents said it was an important factor and 14% said it was a very important factor for them. So, 38% of the respondents said caste of the candidate is important for them when voting in an election. So, caste, whether we like or it is still relevant.

Let us look at some of the other highlights- the biggest surprise was Gujarat. 60% of the respondents said that caste was very important for them and 24% said it was important, making 84% of the respondents saying that it was an important factor.  This was the highest in the country by a long long way! What does this say about the development and good governance NaMo? I will let you make your own conclusions.

Second, of the respondents with a background in the defence sector, nearly 22% said caste is very important and 25% said it is important making 47% of those with a defence background saying that caste is important for them when they vote in elections. Again, this is a shocker given the defence services emphasis on unity and integrity.  Contractors came next with around 42%.

Another surprise is the contrast (or the lack of it) between those with an IT background- nearly 39% say caste is important (of this 16% said it is very important)- and those in Agriculture- for which the number is 36% (of this only 12% said it is very important). What does this say- caste has more or less the same role, whether in rural areas or urban areas and whether you are in a modern profession or a traditional profession.

Let us next look at educational background- irrespective of the level of education (i.e., postgraduate, graduate, high school, primary etc) or lack of it, approximately similar percentage of people feel that caste is important (the % varies between 36 and 41 with high school educated being the highest and primary school educated being the lowest).   Similar even patterns are evident when we look at wealth distribution- the % of people who say caste is important varies between 30 to 35%.

The biggest surprise however is in how people from different castes feel. The % of people who believe that caste is important for them during elections in the OBC, SC and ST category varies between 31 and 35, but is a whopping 43% for people belonging to the General category. Again, I will let you come to your own conclusions on this as well.

We should only remember one aspect, there is no single constituency in the country where a candidate can win with support from only one caste or indeed, the lack of support from any caste (a point that Dipankar Gupta has lucidly explained in his op-ed in The Hindu on March 21http://www.thehindu.com/opinion/lead/the-caste-bogey-in-election-analysis/article5811003.ece). Further, if any caste is dominant (not only in numbers but because of their influence based on wealth, land holdings and patronage) in a constituency, all political parties put up candidates belonging to the same caste. Analysts argue that in such a case, the caste of the state level or the national level leader becomes relevant! 

We will come up with more analysis on this shortly.

Why do people vote the way they do?

Caveat emptor: I wish I knew. And so do hundreds of wannabe-MPs right now, I'm sure. Karthik Shashidhar's very interesting analysis of our data in Mint earlier this week documents some of the reasons we heard from nearly 2.5 lakh people across the country. In all our past surveys, as in this one, the Candidate is the most common selection: that people vote for the Candidate instead of his/her Party (the second most chosen factor), his/her Party's PM candidate, etc. In all our surveys, again, we've noticed a pattern that Karthik panned us about: 86% of the people who voted said their candidate won. That's obviously not possible, as he's pointed out. Then what's at work here? Let me see if I can address that here.
To understand this, let's look at the most common reasons for people voting they way they do:

• Ideology-based. This is typically a Leftist and Rightist position: that one votes for a candidate based on the candidate's ideology, if it is clearly present and aligned with one's own. In recent years, people who were just left or right of center have been pushed further along their direction, thanks to strident rhetoric on both sides.
• Issue-based. Not so common in India (yet!), this is typical in state-level elections in the US, where a specific issue (mostly clothed as a Proposition) becomes a significant factor to drive voters to specific candidates.
• Caste-based. Traditional wisdom in India is that the electorate is caste-driven, to a very large degree. Our surveys have consistently reported the complete opposite, but survey-theorists will tell you that that's because people are not comfortable telling you that they vote by caste, so they pick something else in the survey. It is also true that the political parties make a big deal about this, choosing candidates from the same caste against in a given constituency.
• Patronage-based. Particularly in semi-urban and rural areas, "powerful" politicians establish a network of patronage, through everything from granting contracts to paying a visiting farmer a little bit of cash. This pays off at the hustings, since the recipient's "gratitude" overwhelms even the most heinous behaviour of the candidate.

Yes, these are the major reasons, but I think the most important reason comes out of the lacuna that Karthik has pointed out: people vote for the winner. The way I see it, people say they voted for the winner because, in the absence of the reasons above, they want to vote for the candidate that they believe will win. And this is a fact that politicians know very well, hence their strutting in public, staking claims of popularity and "winnability" well beyond the realms of possibility. Seasoned politicians know that they need to be seen as the winner, to be the winner. And this is borne out in personal discussions for all of us, too: people regularly reject candidates with the comment "Oh, but s/he can't win", as if that decides the reason to vote for someone.
So there you have it: you want to win an election? Be seen as the winner!