Four in ten left off the draft in Hyderabad's Old City

Part I: The five constituencies

In short

The Special Intensive Revision, a fresh check of every name on the electoral roll, reached Hyderabad in 2026. The pre-revision rolls of five Old City constituencies, the rolls as they stood before the revision began, held 16,57,207 electors. The September lists left 6,62,063 of them off the draft, 40.0%. They kept 3,69,972 more on the draft roll, the first published version of the revised roll, only on condition that they attend a hearing. These lists hold every elector, so the rate for any group, the share of that group left off, is a count over everyone in it and involves no sampling.

A seat's winning margin is the number of votes by which its winner beat the runner-up. Every one of the five lost more names than its 2023 winning margin. In Yakutpura and Malakpet the names left off the draft exceed every valid vote cast in 2023. Yakutpura was won by 878 votes and lost 1,54,982 names. The heaviest burden fell on the oldest electors and on those who could not be traced to the 2002 roll, the roll of the last intensive revision. 74.0% of electors aged 80 and over were left off the draft, and electors with no 2002 link were left off the draft at twice the rate of those with one.

Exclusions and coverage

A polling part is the section of a constituency's roll that votes at one polling station, and it is the smallest unit the lists can be added up by. Malakpet's list is read by polling-station name, which gives 60 stations instead of parts, and it is incomplete. It holds 1,42,688 names left off the draft, while a separate count of the published Malakpet list finds 1,47,799. For the other four constituencies the two counts agree exactly.

Exclusions against the margin and the votes polled

Each constituency's names left off the draft are compared with two figures from the 2023 Assembly election, the margin by which the seat was won and the valid votes cast. The comparison makes no claim that these electors would have voted, or how. It measures the size of the revision next to the vote that decided the seat.

Across the five constituencies the names left off the draft equal 99.0% of the valid votes cast in 2023. In every constituency more names were left off the draft than the winning candidate received.

No election has been held on the revised roll, so no result has been changed. Against the 2023 margins, only Yakutpura is close. There the runner-up would need 50.8% of the electors left off who would vote, at 2023 turnout and leaving out those entered as dead, to overturn a margin of 878. In Charminar and Malakpet the runner-up would need 73.6% to 83.0%, and in Chandrayangutta and Bahadurpura no split of those left off would be enough. Muslim names are a smaller share of those left off than of the roll in all five seats, by 2.6 to 10.4 points, so where Muslim electors lean to AIMIM, the exclusions take fewer of its voters than of its opponents'.

The reasons recorded

The list of electors entered as absent, shifted, dead or duplicate, known as the ASDD list, holds every elector left off the draft and the reason an officer wrote against each name. 64.2% of exclusions were entered as "permanently shifted", 20.8% as "untraceable or absent", 8.7% as dead and 6.0% as enrolled elsewhere. Each is what the officer wrote down after a visit, or after failing to find the elector, and nobody checked it.

The untraceable share runs from 6.5% in Malakpet to 38.3% in Charminar. Those entries record only that an officer did not find someone.

Age

In every constituency, the older the elector, the likelier the exclusion. Between 70.1% and 76.6% of electors aged 80 and over were left off the draft, and 19.3% to 25.7% of those aged 18 to 24. Death explains part of the exclusions at the oldest ages, and 57,526 deaths were recorded. But the rate roughly doubles between ages 18 to 24 and 30 to 39, when almost nobody dies, so death does not explain it all.

Women and men

Women and men were left off the draft at similar rates, 39.9% of women and 40.5% of men. Inside the same polling part the gap runs from -2.2 to 2.4 points, one way in some constituencies and the other way in others.

Religion

The roll does not record religion, so it is guessed from each elector's name, and a separate classification agrees on 99% of names. Across the five constituencies 52.0% of Hindu-name electors were left off the draft and 35.3% of Muslim-name electors. Inside the same polling part, where the booth officer, the street and the housing are the same for everyone, Hindu-name electors were left off the draft 9.6 to 24.5 points more often, in 728 of 823 parts large enough to compare. The lists give religion and age separately and never together, so part of this gap may be a gap by age.

The 2002 roll

The revision asks each elector to show a link to the roll of the 2002 intensive revision, the last one held, through their own name or a parent's. Where the lists record it, 51.6% of electors with no link were left off the draft against 25.3% of those with one. Inside the same polling part unlinked electors were left off the draft 21 to 28 points more often, in 1,024 of 1,098 parts. Anyone who moved to the city after 2002, or whose family name is spelt differently on the old roll, starts at a disadvantage. Malakpet's list carries no linkage.

Hearings

3,69,972 electors kept on the draft roll were summoned to a hearing, 22.3% of electors on the pre-revision rolls. Each case was listed for a hearing slot, a set date and time at which a batch of cases is called, and there were 7,164 slots. Each summons rests on one or more rules, set out in the table below. Not being traceable to the last intensive revision, the no-mapping rule, was the commonest, 31.6% of flags. The others flag a name that does not match the old roll, six or more children recorded against one parent, or family ages that look impossible, such as a parent less than 15 years older than a child. The heaviest single slot listed 459 cases, in Charminar, where the typical slot listed 38. The lists do not record how long a slot lasted.

Polling parts

Part exclusion rates run from 13% to 80%, and 30 parts lost more than 60% of their roll (5 in Charminar and 25 in Yakutpura). Across parts, the Muslim-name share of the roll and the exclusion rate move in opposite directions in every constituency.

Against Hyderabad and Telangana

On the state's draft-roll figures, the five Old City seats lost 39.9% of their roll and Hyderabad's other ten seats 41.6%. Every seat in the district lost between 34.6% and 45.2%, and 2 seats outside the Old City lost more than any Old City seat. Across Hyderabad district the draft left off 41.0%, and in the rest of Telangana 18.6%. The exclusions are heavy across the whole capital region.

Main findings

• 6,62,063 electors were left off the draft, 40.0% of the five pre-revision rolls, and 3,69,972 more were summoned to a hearing.
• More names were left off the draft than the 2023 margin in all 5 constituencies, and more than all the valid votes cast in Yakutpura and Malakpet.
• Yakutpura was won by 878 votes and lost 1,54,982 names, 176.5 times the margin.
• The rest of the city lost as large a share. On the state's figures the five Old City seats lost 39.9% of their roll, Hyderabad's other ten 41.6%, and the rest of Telangana 18.6%.
• 74.0% of electors aged 80 and over were left off the draft, against 20.9% of those aged 18 to 24.
• Electors with no link to the 2002 roll were about twice as likely to be left off the draft, 51.6% against 25.3%.
• 20.8% of exclusions were entered as "untraceable or absent", ranging from 6.5% to 38.3% between constituencies.
• 459 cases were listed for a single hearing slot in Charminar.
• 30 polling parts lost more than 60% of their roll.

What this shows and what it does not

The lists show how many electors were left off the draft, from which parts, and who they were by age, sex, religion guessed from names, and the link to the 2002 roll. They do not show whether any individual exclusion was right. A reason code cannot settle that. A household that moved out of the Old City is rightly marked "shifted", and the lists cannot tell it apart from an elector who was at home when the officer missed them. Religion and age are counted separately, so a gap by religion may partly be a gap by age. The Malakpet list used here is incomplete. These are the September lists. The Commission extended the calendar on 7 September, with claims to 7 October, and again on 29 October, with hearings to 26 November and the final roll on 30 November, as Telangana Today reported, so names put back after hearings are not in these counts.

Part II: Constituency chapters

Charminar

Charminar had 2,37,483 electors on its pre-revision roll, in 198 polling parts. The lists left 94,174 of them off the draft, 39.7%, and summoned 50,168 more to a hearing, 21.1%. In 2023 Mir Zulfeqar Ali won the seat with 49,103 votes, by a margin of 22,853. The names left off the draft are 4.1 times that margin and 96.0% of the valid votes cast.

Officers entered 44.9% of the exclusions as "permanently shifted" and 38.3% as "untraceable or absent". 73.5% of electors aged 80 and over were left off the draft, against 20.9% of those aged 18 to 24. Women were left off the draft at 38.1% and men at 41.1%. Hindu-name electors were left off the draft at 46.6% and Muslim-name electors at 36.1%, and inside the same part the gap is +9.6 points.

Electors linked to the 2002 roll were left off the draft at 26.2% and those with no link at 53.1%. Inside the same part the gap is +26.0 points, and it runs that way in 195 of 198 parts.

Its hearings ran to 639 slots, with 38 cases in the typical slot and 459 in the heaviest. The part with the highest exclusion rate is 22 (Dood Bowli), at 80.3%.

Yakutpura

Yakutpura had 3,74,948 electors on its pre-revision roll, in 332 polling parts. The lists left 1,54,982 of them off the draft, 41.3%, and summoned 87,367 more to a hearing, 23.3%. In 2023 Jaffer Hussain won the seat with 46,153 votes, by a margin of 878. The names left off the draft are 176.5 times that margin and 110.4% of the valid votes cast.

Officers entered 66.8% of the exclusions as "permanently shifted" and 16.1% as "untraceable or absent". 73.5% of electors aged 80 and over were left off the draft, against 20.3% of those aged 18 to 24. Women were left off the draft at 40.3% and men at 42.4%. Hindu-name electors were left off the draft at 55.8% and Muslim-name electors at 34.5%, and inside the same part the gap is +24.5 points.

Electors linked to the 2002 roll were left off the draft at 29.2% and those with no link at 52.2%. Inside the same part the gap is +21.0 points, and it runs that way in 280 of 332 parts.

Its hearings ran to 1,276 slots, with 60 cases in the typical slot and 323 in the heaviest. The part with the highest exclusion rate is 297 (Banu Nagar Colony), at 77.4%.

Malakpet

Malakpet had 3,16,454 electors on its pre-revision roll, in 60 polling stations. The lists left 1,42,688 of them off the draft, 45.1%, and summoned 68,188 more to a hearing, 21.5%. In 2023 Ahmed Bin Abdullah Balala won the seat with 55,805 votes, by a margin of 26,106. The names left off the draft are 5.5 times that margin and 108.1% of the valid votes cast.

Officers entered 81.4% of the exclusions as "permanently shifted" and 6.5% as "untraceable or absent". 76.6% of electors aged 80 and over were left off the draft, against 25.7% of those aged 18 to 24. Women were left off the draft at 45.1% and men at 47.4%. Hindu-name electors were left off the draft at 53.6% and Muslim-name electors at 38.2%, and inside the same part the gap is +16.4 points.

Malakpet's list carries no 2002 linkage and no part numbers, and 7,891 electors on its roll are in none of the three outcomes. The list is incomplete, so read its figures beside the state's own count for Malakpet, 1,47,799 names left off the draft, 44.8% of its roll.

Its hearings ran to 1,489 slots, with 45 cases in the typical slot and 284 in the heaviest. The station with the highest exclusion rate is GHMC Community Hall, at 58.9%.

Chandrayangutta

Chandrayangutta had 3,77,299 electors on its pre-revision roll, in 305 polling parts. The lists left 1,48,869 of them off the draft, 39.5%, and summoned 80,246 more to a hearing, 21.3%. In 2023 Akbaruddin Owaisi won the seat with 99,776 votes, by a margin of 81,660. The names left off the draft are 1.8 times that margin and 96.8% of the valid votes cast.

Officers entered 61.4% of the exclusions as "permanently shifted" and 25.4% as "untraceable or absent". 73.9% of electors aged 80 and over were left off the draft, against 20.8% of those aged 18 to 24. Women were left off the draft at 40.0% and men at 39.0%. Hindu-name electors were left off the draft at 51.7% and Muslim-name electors at 36.1%, and inside the same part the gap is +20.2 points.

Electors linked to the 2002 roll were left off the draft at 24.8% and those with no link at 51.7%. Inside the same part the gap is +27.5 points, and it runs that way in 290 of 305 parts.

Its hearings ran to 1,707 slots, with 50 cases in the typical slot and 299 in the heaviest. The part with the highest exclusion rate is 158 (C Block Main Road, Harmain Masjid), at 57.2%.

Bahadurpura

Bahadurpura had 3,51,023 electors on its pre-revision roll, in 263 polling parts. The lists left 1,21,350 of them off the draft, 34.6%, and summoned 84,003 more to a hearing, 23.9%. In 2023 Mohammed Mubeen won the seat with 89,451 votes, by a margin of 67,025. The names left off the draft are 1.8 times that margin and 84.0% of the valid votes cast.

Officers entered 59.2% of the exclusions as "permanently shifted" and 24.4% as "untraceable or absent". 70.1% of electors aged 80 and over were left off the draft, against 19.3% of those aged 18 to 24. Women were left off the draft at 35.8% and men at 33.4%. Hindu-name electors were left off the draft at 44.2% and Muslim-name electors at 33.4%, and inside the same part the gap is +12.4 points.

Electors linked to the 2002 roll were left off the draft at 21.7% and those with no link at 49.6%. Inside the same part the gap is +28.2 points, and it runs that way in 259 of 263 parts.

Its hearings ran to 2,053 slots, with 50 cases in the typical slot and 150 in the heaviest. The part with the highest exclusion rate is 208 (Tadban), at 51.6%.

Conclusion and what the Election Commission should do

In five constituencies the draft left off 40.0% of the electors on the rolls, more names than the margin that decided every one of the five seats in 2023, and in two of them more than every valid vote cast. The burden fell on the old, and on those the process could not trace to a roll prepared 24 years earlier. The final roll has not yet been published, and the hearings are still being held.

What the Election Commission should do before the final roll:

1. Record and publish the basis for every exclusion entered as "untraceable or absent", 1,37,825 names in all, including the visits made and the notice given.
2. Publish how exclusions of electors with no link to the 2002 roll were decided, since the link is the revision's own test and the gap it produces is the largest in the lists.
3. Give each elector served a notice enough time to be heard, online or in person. Some slots in the September lists carried several hundred cases.
4. Explain the exclusions in the 30 parts that lost more than 60% of their roll.
5. Publish restorations after hearings, part by part, so that the final count can be set against these lists.

Method and sources

• The rolls before the revision, the list of the absent, shifted and dead, and the discrepancy and no-mapping lists come from the Chief Electoral Officer, Telangana, the Election Commission's officer in charge of the state's rolls, for the five constituencies, September 2026, added up by polling part. The elector-by-elector lists were compiled for these five seats only.
• Religion is guessed from each elector's name because the roll does not record it, and a separate classification agrees on 99% of names.
• Hearing schedules come from the same lists, 28 August to 12 October 2026.
• 2023 Assembly result: Election Commission of India.

The exclusion rate is the electors on the list of the absent, shifted and dead divided by the electors on the roll before the revision, for the group in question. The within-part difference is the rate of one group minus the other inside each part, pooled across parts with at least 20 of each and weighted by the smaller group. Part counts add up exactly to the constituency totals. No name, EPIC number or address is published.