The Aam Aadmi Party has scored a massive victory in Delhi with results that do not offer much nuance or variation, as most of the data points in the same direction. It is useful, however, to try to look behind the big numbers to see what the data can tell about the election and about the transformation of AAP from an anti-system movement into a conventional political party.

In this piece, based on data collected from a variety of sources, including fieldwork conducted during the campaign, we look at trends in participation, candidates and party performance, at the geography of the results and at the profile of the previous and new assembly.

Participation down

After a brief period of cacophony from the Election Commission, the participation rate in the Delhi election has been measured at 62.6%, five points below the 2015 turnout but an identical turnout to the last general election.

Delhi has always registered a relatively high turnout among India’s large cities, barring 1998 when it went below the 50% mark. The 2015 elections was unusual, as AAP erupted and swept the election on the back of a wave of popular support. Thus, the turnout in the 2020 election is back to a regular level for a more conventional Delhi election.

The Election Commission is yet to provide constituency-wise turnout data so at this point, we cannot tell anything about turnout variations across seats or seats category.

In this election, 742 individual belonging to 96 different parties or running as independent candidates contested for seats. In most elections, around 75% of the candidates lose their deposit since they secure less than one-sixth of the total votes polled in their seats. This year, the figure climbed to 81% due to the absence of an effective third party. Most voters opted for either AAP or the Bharatiya Janata Party. Few were willing to waste their vote on another candidate.

As in other states, the total number of parties contesting is on the rise. The phenomenon in Delhi is accentuated by the fact that many regional parties field candidates in the Capital. The Bahujan Samaj Party contested 68 seats and the Rashtriya Janata Dal four, in alliance with Congress. The Janata Dal (United) contested two seats, the communists six. The Shiv Sena and the Nationalist Congress Party contested each five seats and the Lok Janshakti Party one.

Most of these candidatures prove to be perfectly futile, as the cumulative vote share of the two and now three main parties in Delhi keeps rising. In 2008, one Delhi voter out of four opted for a different party (the Bahujan Samaj Party scored 14% of vote share that year). In 2015 and 2020, there are less than 4% of total voters casting their vote for a small player. This contrasts with a state like Maharashtra, where the share of minor player is as high as 30%.

Vote share: AAP way ahead

In terms of vote share, the data reveals how decisive AAP’s victory was. For the first time in decades, an incumbent government wins a consecutive term with a vote share superior to 50%. One has to go back to the Congress in Gujarat in the 1980s to see that kind of performance. Even the BJP in Gujarat did not have such staggering numbers. Compared to 2015, AAP thus maintains its vote share at 53.4%, losing less than 1% compared to five years ago.

This performance is even more impressive when one compares the fate of AAP only eight months ago, in the last general election. In 2019, AAP finished third with 18% of vote share, behind Congress. Even more strikingly, AAP failed to win a single assembly segment in the general election, as the BJP won 65 segments and the Congress five.

Granted, this was a different election. The 2019 election was like a national plebiscite in favor or against the prime minister and the campaign, of course, was led on national themes. We are now accustomed with a number of voters who opt for one party in one type of election and for another in another type.

In this case, the gaps are enormous. Between spring 2019 and winter 2020, AAP won 35% of the vote share, winning 62 seats. This comparison of AAP’s performance helps understand the tactical choice that party made to avoid antagonising voters in Delhi who may support Narendra Modi at the Centre but could still vote for AAP in the assembly election.

As a result, AAP maintained its vote base largely intact and succeeded in getting additional votes, notably from former Congress supporters. The maps of AAP’s performance in 2015 and 2020 are nearly identical. AAP registers its best scores in Central and South Delhi, as well as in reserved seats.

Constituency-wise vote share of AAP in the 2015 Delhi State Assembly Election. Source: TCPD-IED (Indian Elections Dataset)
Constituency-wise vote share of AAP in the 2020 Delhi State Assembly Election. Source: TCPD-IED (Indian Elections Dataset)

In 2020, AAP’s average vote share in reserved seats is 58%, just 1% below where it was in 2015. This is a five-point gap compared to its average vote share in 2015.

Incidentally, the areas in which AAP registers its highest vote share today happen to overlap significantly with the areas in which the Congress performed the most in 2008. This confirms that AAP occupies largely the space left by the Congress, not just in terms of aggregate vote share but also in terms of geography.

Constituency-wise vote share of Congress in the 2008 Delhi State Assembly Election. Source: TCPD-IED (Indian Elections Dataset)
Constituency wise winning parties in the 2008 Delhi Assembly Election. Source: TCPD-IED (Indian Elections Dataset)

Glass half-full, half-empty for the BJP

The BJP finds itself in a curious situation. It scores its second-worst performance in terms of seats (from three seats in 2015 to eight seats in 2020) with its second-best ever performance in terms of vote share. One has to go back to 1993 to see the BJP doing better, when it came in power under the leadership of Madan Lal Khurana.

The BJP can find solace in registering a small bump of vote share (+6%) but the seats’ gains did not follow. The increase in vote share of the BJP probably comes from the extreme bipolarisation of this election, in which the Congress completely collapsed.

Historically, the BJP always had a base of voters of at least 32%, no matter what. A small increase can come from the absence of third party and an influx of new voters. It performed well in North-East Delhi, a sub-region with a large migrant population from the Hindi Belt.

Capital distress for Congress

One hesitates to throw salt on the Congress’ open wounds. The not so grand old party suffered yet again a humiliating defeat. Not only did the party crash down to 4% of vote share, but its candidates fail to reach the number two position in any of the seats. Worst even, Congress candidates lost their deposits in 63 of the 66 seats it contested (it gave four seats to the RJD).

The list of problems that plagued the Congress is long and well-known: weak local leadership, unpopular national leadership, a half-hearted campaign. The party remained in dormant mode between the election and woke up to the campaign at the last minute. They campaigned against the BJP but forgot to offer something for Delhi at the same time. AAP has replaced the Congress in Delhi for the foreseeable future, if not for good.

These variations in vote-share performance have led to AAP sweeping Delhi. It now occupies 88.5% of the seats, against 95% of the seats five years ago. The BJP increased its seat share from 4.3% to 11.4%, hardly a consolation.

The electoral map is fairly easy to read. AAP wins 62 seats across the city except in the North-East, where the BJP wins seven seats. The BJP also retains its bastion of Rohini. North-East Delhi covers the two Lok Sabha seats occupied by Gautam Gambhir and Manoj Tiwari, both of whom played a visible role in the campaign. These are seats with high proportion of recent migrant population from Uttar Pradesh and Bihar.

Constituency-wise winning parties in the 2020 Delhi Assembly Election. Source: TCPD-IED (Indian Elections Dataset)

Of the three seats it held in 2015, the BJP retained two and lost one, in Mustafabad.

Constituency wise winning parties in the 2015 Delhi Assembly Election. Source: TCPD-IED (Indian Elections Dataset)

The geography of the BJP hasn’t changed much since 2015. Its vote share gains are fairly uniform across the map. The BJP gained votes in most sub-regions, except in Central and South Delhi, in important large constituencies like Rajinder Nagar (an old Jan Sangh bastion), Delhi Cantonment (which has a high military population), Mehrauli and Malviya Nagar. It stagnated in the industrial constituency of Okhla, adjacent to Jamia Millia Islamia.

Constituency-wise vote share of BJP in the 2020 Delhi State Assembly Election. Source: TCPD-IED (Indian Elections Dataset)
Constituency-wise vote share of BJP in the 2015 Delhi State Assembly Election. Source: TCPD-IED (Indian Elections Dataset)

Not a close election

A rapid examination of victory margins reveals that this was not a closely fought election. Only 13 candidates won with a victory margin inferior to 5% and only four with a margin below 2%.

In Burari, Sanjeev Jha of AAP scored the highest victory margin, distancing his opponent by 88,158 votes. In Okhla, Amanatullah Khan, also from AAP, beat his opponent with a 71,827-vote margin. In Matia Mahal and in Seema Puri, Shoaib Iqbal and Rajendra Pal Gautam, both of AAP, beat the BJP runner-up by more than 50,000 votes. At the other end, Bhupinder Singh Joon (Bijwasan, AAP) and Abhay Verma (Laxmi Nagar, BJP) won their seat with less than 1,000 votes margin.

The map shows clear demarcation in the distribution of victory margins.

Constituency-wise Victory Margins in the 2020 Delhi State Assembly Election. Source: TCPD-IED (Indian Elections Dataset)

A party-wise comparison of victory margins show that this was by no means a closely contested election. AAP’s average winning margin is 18% against 8.2% for the BJP. The gap has somewhat narrowed compared to 2015, but it remains a large one.

Renewal of Delhi’s political elites

The rise of AAP in 2015 has led to a profound renewal of Delhi’s political class. In 2015, 49 of the Assembly’s 70 MLAs had been elected for the first time. That number is now down to 21 as most of AAP’s re-running candidates won (40 out of 43).

There are currently only four politicians in the Delhi assembly who have been elected before the 2013 election. Shoaib Iqbal is a six-time MLA in Okhla, now contesting on an AAP ticket. Parlad Singh Sawhney, from AAP, and Mohan Singh Bisht, from the BJP are both five-time MLAs. The last survivor is the BJP Gujjar leader Ramvir Singh Bidhuri, from Badarpur, who has been elected for the fourth time. Otherwise, there are no other members that have served under Congress Chief Minister Sheila Dixit.

Turncoat candidates

In this election, 37 of the 742 candidates contested with a different party affiliation. Nine of them contested on AAP tickets, six on Congress tickets and five on a BJP ticket.

Two turncoat Congress candidates came from the BJP, two from AAP and one from the Nationalist Congress Party. Three of the BJP turncoats came from AAP and two from Congress. Five turncoats fielded by AAP came from Congress, only one from BJP and three from the BSP.

Nine of these 20 turncoats won. All were on AAP tickets except for Anil Kumar Bajpai from Gandhi Nagar. The new BJP member is the only AAP deserter to have won.

The following chart depict the status of AAP and BJP’s candidates. A square represent winners and the circle losers. The number in the circles and the squares indicate the number of terms these candidates have sured. Squares and circles with colours other than the party colorr indicate turncoats (green from Congress, mauve from the Bahujan Samaj Party and so forth). For more details on this, see TCPD’s incumbency visualisation tool here.

Incumbency profile of the 11th Delhi Vidhan Sabha. Source: TCPD-IID (TCPD Individual Incumbency Dataset)

In the context of such a vast shuffle of the political class, one can ask how the overall sociology of Delhi’s representatives has been affected.

Delhi is a particular state where caste and electoral behavior do not align neatly. Historically, parties have provided representation to all major demographic groups and have drawn support across caste boundaries. There are some historical preferential alignments, like the Punjabi Khatris and the BJP. But by and large, parties mobilise across castes.

In recent years, AAP has emerged as a party that is officially caste agnostic. It claims to mobilise and to work in favour of Dilliwalas irrespective of their caste affiliation. One source of popularity of AAP’s schemes and policies is not only the fact that they do address daily issues, but that AAP provides access to benefits without asking about the caste of the beneficiaries, or without targeting beneficiaries on the basis of their social identity.

The question therefore is how AAP fares in terms of providing representation to the various segments of Delhi’s population. But before that, we examine the long-term representational trends in the Assembly.

Skewed representation of upper castes

The first observation is that upper castes – who make between 30% to 35% of the population – dominate the state assembly, occupying between 45% to 50% of the seats. It is only in 2008, when the Congress aggressively pursued Jat voters, that the share of upper caste MLAs decreased. Upper caste MLAs currently make for half of the assembly.

The second observation is that the share of representation of upper caste has increased alongside the rise of AAP. Otherwise, the trajectory of most other groups has remained somewhat stable. Sikhs have seen a decline of their representation (currently at 3%). Muslims’ representation has been stable at 7% (four to five seats). No Scheduled Caste candidate has ever been elected in a non-reserved seat.

Among the upper castes, Brahmins have always had a prominent place in Delhi politics, occupying on average 15% of the seats, a few notch above their demographic weight (which stands at 10% to 12% of the population). Their representation peaked in 2015 and 2020, with 14 seats won in both elections.

The second upper caste group well represented in the assembly are the Banias (7-8% of the population), who make for a third of all upper caste MLAs in Delhi. Upper caste Punjabis, and notably Punjabi Khatris (7%-9% of the population) have seen their representation sharply decline since the BJP peaked in the 1990s. Their association with the BJP in fact cost them in terms of representation, the party having been out of power since 1998.

Among the intermediate and Other Backward Classes, there are only three groups represented. The Jats (5% of the population) make 13% of the MLAs over time. Their representation peaked in 2008 at 18%, during the third mandate of Sheila Dixit. The two other groups represented are the Gujjars (5%) and the Yadavs (a smaller number).

It is striking that no other OBC group has ever been represented in the state assembly. Barring Surender Saini in Minto Road in 1972, all backward classes MLAs have belonged to these two groups.

A third observation is that the eruption of AAP has not fundamentally changed the social composition of the assembly. The trends clearly indicate a lot of continuity in this regard.

Brahmin premium in AAP

A comparison of AAP, BJP and Congress candidates however reveals some variations between parties. Notably, the share of representation of Brahmins among AAP candidates is significantly higher than among the BJP and the Congress. Banias, often associated to the BJP, also finds good representation in AAP but not in Congress.

Among intermediate and backward classes, Jats find equal representation among the three parties. AAP nominated slightly more Yadav candidates that its adversaries, while the Congress nominated more Sikh candidates than the other two parties.

Finally, both AAP and Congress distributed seven tickets to Muslim candidates, in constituencies where they have an important demographic presence.

Caste distribution of candidates in the Delhi 2020 election. Source: SPINPER Project – Indian Legislators and Candidates Dataset

What is the signification of this information, particularly in a state not known for divisive caste-based politics?

On the one hand, the over-representation of demographically significant groups make sense. All the castes represented in the assembly are demographically significant, if not dominant, in a number of seats. Some castes are more concentrated (like the Gujjars and the Jats) and other more dispersed geographically (like the Rajputs or the Banias). The relationship between geographic concentration and representation is not straightforward though. Punjabi Khatris are concentrated in specific parts of the city and yet, their presence in the Parliament has been declining, as AAP does not particularly woe them through tickets.

One important implication is, be that as it may, electoral politics in Delhi excludes most castes. Arguably, a large part of the population does not belong to any of those major groups and they do not get any representation, in any of the state’s three main parties.

The case of AAP is both interesting and revealing, as this is a party that claims to be blind to caste differences. Yet, it provides preferential representation to traditional and elites to a large extent: Brahmins and Banias. Non-traditional business elites like Jats and Gujjars also find good representation in AAP. In this assembly, nearly half of AAP’s 62 MLAs are upper caste and half of them are Brahmins. This does not happen by coincidence.

This could mean two things. One, that AAP had adapted to Delhi terrain and selects its candidates on the basis of political and social geography, like any other party. On the other hand, it is also highly possible that the party’s official caste blindness leads it to recruit its candidates in a sociological pool that contains large numbers of upper caste elites: fewer businessmen and more educated professionals with foreign degrees and private-sector experience. Thus, AAP’s caste agnosticism becomes the mechanism through which traditional elites find representation in the ranks of the party.

This does not make AAP a casteist party. It merely signifies that its sociology does not differ much from the Congress or the BJP and that the choices that AAP made in terms of definition of its identity have an impact on its sociology.

Low representation of women

Another aspect in which AAP does not distinguish itself is the low representation of women. AAP fielded only nine women candidates, ahead of BJP who fielded only eight women. Eight of those nine AAP women candidates won and none of the BJP women candidates won.

The chart below is misleading is it seems to indicate progress. Eleven percent of women MLAs is nothing to brag about. In this election, 71 women contested, although only 25 on major party tickets. The Congress did marginally better than the other two party, by fielding ten women candidates.

This could be dismissed as business as usual, but since AAP claims to work for all people without consideration of caste, gender or class, there is no reason that it should not apply the same standard to its cadres.

The map below reveals also that no women were elected in the periphery of the city, barring Atishi Marlena in Kalkaji.

Gender-wise Winner in the 2020 Delhi Assembly Election. Source: TCPD-IED (Indian Elections Dataset)

Age no bar

On other socio-demographic variables such as age, there is no substantial difference between AAP and the BJP. AAP candidates were on average five years younger than their BJP counterparts. There were significantly more MLAs below the age of 40 in the 2015 assembly (25) compared to the current one (14). In part, AAP recruited slightly older candidates this year. But it also reveals an ineluctable fact of life, which is that young people don’t stay that way for very long. The large number of successful incumbents necessarily pushes the average age of legislators upward.

Educational level

In terms of education, there are no major differences between the 2015 and 2020 assembly. Fifty eight percent MLAs are graduates or more. Only nine MLAs have not completed their 12th standards.

There are more graduates and graduate professionals among BJP candidates and more postgraduates among AAP candidates. This is not surprising as AAP has always been inclined to make space to highly educated individuals, including candidates from a private-sector professional background.

Businessmen vs professional politicians

An examination of the distribution of occupations among MLAs and candidates is not particularly enlightening, since it is self-declared information, prone to misrepresentation. For instance, most AAP re-running MLAs declared politics to be their main occupation. That may be so since they were elected, but that does not inform us about their actual occupational background.

The BJP recruits its candidates significantly more among self-declared businessmen (39 out of 67). There are only few lawyers and other liberal professions.

Another variation in this election is that AAP’s candidates are markedly more affluent than five years ago. Thirteen MLAs have net assets greater than Rs 1 crore. Eleven belong to AAP. But while the median wealth (net assets) of AAP candidates grew from 1.4 crores to 2.3 crores, this was still much less than the median assets of candidates fielded by INC and BJP in both elections.


All the information above gives us insights about AAP and about the 2020 elections, but are insufficient per se to propose a general explanation for AAP’s victory. At best, it helps to introduce some nuance to the general discourse about AAP and its purportedly new form of post-identity politics.

What this data and other evidence from the field reveal is that AAP has evolved profoundly since it emerged out of the anti-corruption movement of 2011 and 2012. In 2015, it was brought to power through a wave that was meant to transform the way politics was done in Delhi, if not in India. Five years later, the results in that regards are very modest. AAP has become a conventional party, committed to delivery and effective governance, for sure. But it also reneged or renounce on many of its ambition of transforming political practices.

As any other conventional party, it has become hierarchical, centralised, devoted to the cultivation of its leader. It no longer publicises its sources of funding and engages into local caste calculations with regard to political selection of candidates.

What is remarkable is that while it operate this transformation, it did not lose a lot of support in the process. Some early supporters are discontented with the transformation AAP has undergone. Some have left. Some have been expelled. But new supporters rise, among new voters and among the few voters that were still supporting Congress.

In doing so, AAP is drawing the contours of a distinct model of doing politics: focus on local daily life issues and avoid engaging adversaries or take positions on controversial issues. This model may work for AAP in Delhi but is unlikely to export well to other states, where larger political issues do matter and where the infrastructure to actually deliver effective governance and welfare delivery does not match does the quality of Delhi’s.

Gilles Verniers is Assistant Professor of Political Science at Ashoka University, Co-Director of the Trivedi Centre for Political Data and Senior Visiting Fellow, Centre for Policy Research. Basim U Nissa, Saloni Bhogale and Mohit Kumar are Research Fellow at TCPD. The caste Data collected by Basim U Nissa during the campaign. Views are personal.