There is no crime rate for South Delhi. Not a high one, not a low one. No government body publishes crime figures for South Delhi, South East Delhi or South West Delhi, and even if one did, there is no usable population figure to divide them by. Every article that ranks South Delhi colony by colony, as people actually use the name from safest to least safe is ranking them on nothing.
What does exist is citywide, and it is better news than the headlines suggest. Delhi Police recorded falls in every major crime category in 2025, including a 31% drop in snatching since 2023. Snatching is still the crime a renter is most likely to meet, and it is the one the police solve least often: 64% of cases, against more than 95% for murder, robbery and rape.
This page sets out what the data says, what it cannot say and what you can check about a specific address instead. Where it does arithmetic, it shows the arithmetic. It does not tell you which locality is safest, because no honest page can.
Who publishes this, and why that matters here. This page is published by South Delhi Rentwala, which lists rental property in South Delhi and earns when a tenancy is agreed. That is a real interest in what you conclude, and a safety assessment from such a publisher should be read with it in mind. Two things follow. This page does not rank localities by safety, partly because the data cannot support a ranking and partly because a rental business ranking neighbourhoods on crime would be doing something it has no standing to do. And where it links to rental listings, it is linking to its own inventory.
Nobody who wrote this is a criminologist, a police officer or a safety consultant. What follows is a reading of published statistics, with every source named at the foot of the page so you can check the reading rather than trust it.
What is on this page. Which South Delhi the data counts · why no crime rate exists for it · the district figures that do exist and why to distrust them · what the citywide numbers show · what is wrong with the “crime capital” figure · the crime a renter is most likely to meet · safety for women · police strength · what to check before you sign · how to read the next safety ranking you see.
| The state of the evidence | Where it stands |
|---|---|
| Crime rate published for South Delhi | None, by anyone |
| District figures that exist | Five years, obtained under the Right to Information Act, unreliable |
| Delhi Police’s own statistics page | Citywide only, series ends July 2022 |
| Best current data | Citywide, Delhi Police annual review, released Jan. 21, 2026 |
Which South Delhi are we counting?
Before any number means anything, the boundary has to be settled. It is not.
“South Delhi” names three different areas depending on who is speaking: the Delhi government, which administers revenue districts; Delhi Police, which uses its own; and everyone else, who means a set of colonies that matches neither. Start with the government’s version. Three revenue districts carry the name: South, South East and South West. Here is what each is made of, as published by each district’s own administration.
| Revenue district | Its subdivisions, as its own government publishes them |
|---|---|
| South | Saket, Hauz Khas, Mehrauli |
| South East | Defence Colony, Kalkaji, Sarita Vihar |
| South West | Dwarka, Najafgarh, Kapashera |
Read that third row again. The South West revenue district is Dwarka, Najafgarh and Kapashera. None of the three is what anyone means when they say South Delhi. Meanwhile Vasant Vihar and Vasant Kunj, which everyone files under South Delhi, appear to sit in the New Delhi revenue district, while R.K. Puram next door to both sits in South West.
Those last placements come from encyclopaedic sources rather than an official gazetteer, and this page flags them as indicative rather than settled. District administrations publish subdivisions and villages, not colony names, and no official colony-to-district index could be found. That absence is the point: nobody publishes the mapping most people assume exists. The South and South East names travel better, and this site lists what sits inside the South East boundary and what sits inside the South West one as those names are commonly used.
Then there is the police, who use a different map again. Delhi Police has 15 districts of its own, in place since January 2019, and the published list states plainly that they differ from the revenue districts the government administers. Names overlap; boundaries do not. Delhi Police’s Southern Range covers South and South East, while South West sits under New Delhi Range alongside New Delhi district. Every district crime figure in this article comes from that police map. Every population figure comes from the revenue one.
Then it moved again. On Jan. 1, 2026 Delhi redrew its revenue districts, going from 11 to 13. The cabinet approved the change on Dec. 11, 2025, Shahdara was abolished, Old Delhi, Central North and Outer North were created, and every boundary was realigned to Municipal Corporation of Delhi zones, the New Delhi Municipal Council and the Delhi Cantonment Board. Subdivisions went from 33 to 39. Chief Minister Rekha Gupta gave the reason: for years the revenue boundaries had not matched the municipal ones.
South, South East and South West all survived the redraw by name. Their edges did not stay put.
Three maps, then, and the newest of them is eight months old. Every crime figure below comes from the police map, every population figure from the revenue one, and the colonies in your search are on neither. That is not a preamble to the answer. It is most of the answer.
Why no one can give you a South Delhi crime rate
A crime rate is a fraction. Crimes on top, people underneath. South Delhi has a problem with both halves.
The numerator
Delhi Police publishes citywide counts and nothing smaller. Its statistics page carries three files: crime by head for 2001 to 2010, crime by head for 2011 to 2022 and crimes against women for 2013 to 2022. Every one of them is a table of crime categories against years for Delhi as a whole. None contains a district column. None contains a rate, despite one being titled as though it does. The most recent series stops at July 15, 2022, which is roughly four years ago.
- The annual review is citywide too. Delhi Police holds a press conference each January and releases the previous year’s figures. The 2025 review, released Jan. 21, 2026, gave eight crime categories with solving rates. It gave them for Delhi. No district appeared in it.
- NCRB treats Delhi as one unit. The National Crime Records Bureau publishes Crime in India annually, and its metropolitan chapter covers 19 cities. Delhi is one row. There is no sub-city breakdown anywhere in it.
- The Delhi government publishes district population but no district crime. Its Directorate of Economics and Statistics puts out district-wise population, literacy, density and sex ratio, all from the 2001 and 2011 censuses. It publishes no crime data at district level at all. Nor do the three southern districts’ own government websites.
The denominator
This half is worse.
At the 2011 census Delhi had nine districts. South Delhi was counted at 2,731,929 and South West Delhi at 2,292,958. South East Delhi was not counted, because it did not exist. It was carved out in 2012, taking Delhi from nine districts to 11.
So the 2011 figure for South Delhi describes the territory that is now South and South East together. It is not a denominator for either.
South East has never been measured on its own. Its district government website publishes no population figure of any kind. No official post-2012 estimate exists. The projections that circulate online are arithmetic extrapolations run by the websites that publish them, and those sites say so.
There was no 2021 census. And then the boundaries moved again in January 2026.
Even if both halves existed, they would not fit together. The only district crime counts anyone has are for Delhi Police’s 15 districts. The only district populations anyone has are for the government’s revenue districts. Same names, different maps. Dividing one by the other produces a number with no meaning. As of July 2026, aligning the two was still under discussion, and it needs Union Home Ministry approval because policing in Delhi is a central subject.
| Source | Publishes district crime? | Publishes district population? |
|---|---|---|
| Delhi Police statistics page | No | No |
| Delhi Police annual review | No | No |
| NCRB, Crime in India | No, Delhi is one unit | No |
| Delhi Directorate of Economics and Statistics | No | Yes, 2001 and 2011 only |
| District government websites | No | South West yes, South East none |
| Census of India | No | 2011, nine districts, no South East |
That is the whole answer to the question people actually type. Not “South Delhi is safe” and not “South Delhi is dangerous”, but: nobody has the numbers, and the ones who claim to have not checked. Nor does wealth settle it, whatever the assumption behind searches for the assumption that money and safety travel together. None of this is helped by the fact that how Delhi’s land and zoning boundaries were drawn has its own long history of lines that do not agree with each other.
The district numbers that do exist
There is one exception, and it needs handling with tongs.
Read the caveats before the table. The figures below came from Right to Information requests filed with Delhi Police by the Praja Foundation and the Commonwealth Human Rights Initiative, and published in their white papers on policing in Delhi. They cover 2014, 2015, 2016, 2021 and 2022. They are the only district-level Delhi crime figures in public. They are also, on the evidence below, not reliable enough to compare one district against another, and this page publishes them only so a reader can see what the ceiling of available evidence actually looks like.
The table below is unreliable data, published only to show the ceiling of available evidence. Do not compare districts using it. The two flagged cells are not credible, and five specific defects are listed underneath.
| Crime, as Praja labels it. Covers only 156 of Delhi’s 179 police stations | South 2021 | South 2022 | South East 2021 | South East 2022 | South West 2021 | South West 2022 |
|---|---|---|---|---|---|---|
| Murder | 29 | 43 | 36 | 135 ⚠ | 12 | 11 |
| Attempt to murder | 30 | 30 | 75 | 90 | 18 | 8 |
| Rape | 71 | 66 | 247 | 111 | 107 | 86 |
| Kidnapping and abduction | 133 | 109 | 492 | 105 | 195 | 174 |
| House breaking | 150 | 98 | 240 | 71 | 82 | 35 |
| Chain snatching | 196 | 214 | 52 | 485 | 62 | 89 |
| Robbery and dacoity | 96 | 107 | 177 | 82 | 48 | 36 |
| Theft | 328 | 638 | 2,588 | 258 ⚠ | 1,956 | 1,944 |
Five reasons not to build anything on that table.
A quarter of the police stations did not answer. Praja states in the report that it approached all 179 police stations in Delhi and received data from 156. The remaining 23 either refused or referred the request onward, even though it had been forwarded from the deputy commissioners’ offices. A count missing 23 stations is not a count.
One number is not credible on its face. South East district shows 135 murders in 2022. NCRB recorded 509 murders across the entire city that year. One district producing more than a quarter of the city’s murders, when Delhi Police’s own citywide series shows nothing of the kind, does not survive a sanity check.
One number moves by a factor of ten. South East theft falls from 2,588 in 2021 to 258 in 2022. That reads as a missing digit or a set of missing stations. It does not read as a 90% collapse in theft in one district in one year.
Two collections disagree about the same district and year. South district’s murder count for 2014 appears as 44 in Praja’s 2023 white paper, and as 39 or 40 in the 2014 white paper, depending on which of that report’s two tables you read, because one counts by principal offence and the other counts every section invoked. Two RTI exercises, one district, one year, three numbers.
The columns do not add up to the report’s own totals. Summing the 2021 figures in the table above gives 1,033 for South, 3,907 for South East and 2,480 for South West. The same report’s own summary totals for those districts and that year are 1,036, 3,915 and 2,481. The 2022 columns reconcile exactly; the 2021 ones do not. Three small discrepancies inside one document are not a scandal, but they are one more reason to treat the whole table as indicative rather than exact.
None of that is a criticism of the organisations involved. Both published their response rate and their counting method openly, which is more than the official series does. The point is what it says about the underlying record.
The districts hold what you would expect them to hold. Saket’s rental stock sits in the South district and the flats listed around Lajpat Nagar in South East, both squarely in the belt this page is about.
Vasant Kunj, whose district label depends on which map you open is the awkward one, for the reason the earlier section set out. What none of those addresses has is a defensible crime figure attached to it.
What the citywide data shows
Here is the evidence that does hold up. It describes Delhi, all of it, and it is genuinely informative.
Delhi Police released its 2025 review on Jan. 21, 2026, with three years of counts and the share of cases solved in 2025.
| Crime | 2023 | 2024 | 2025 | Since 2023 | Solved in 2025 |
|---|---|---|---|---|---|
| Snatching | 7,886 | 6,493 | 5,406 | down 31% | 64.22% |
| Molestation | 2,345 | 2,037 | 1,708 | down 27% | 95.20% |
| Rape | 2,141 | 2,076 | 1,901 | down 11% | 97.11% |
| Robbery | 1,654 | 1,510 | 1,326 | down 20% | 97.51% |
| Attempt to murder | 757 | 898 | 854 | up 13% | 98.13% |
| Murder | 506 | 504 | 491 | down 3% | 95.32% |
| Eve-teasing | 381 | 362 | 337 | down 12% | 89.02% |
| Extortion | 204 | 228 | 212 | up 4% | 63.68% |
“Eve-teasing” is the term Delhi Police uses in its own returns for sexual harassment in public. It is kept here because changing it would break the link to the published figure, not because it is a good description of the offence.
Six of the eight categories fell. Two rose, and both rose from 2023 to 2024 before falling back in 2025. Snatching fell hardest in absolute terms, by 2,480 cases.
The solving rates are the more interesting column. Six categories sit above 89%. Two sit in the sixties, and they are snatching and extortion. That gap is not random, and the next section explains what it means for you.
NCRB’s own account of Delhi, from Crime in India 2024, published in May 2026, points the same way:
- 275,402 total cognizable crimes in 2024, down from 324,257 in 2023, a fall of about 15%.
- 180,973 thefts, which works out at about 496 a day across the city, and which is 73.3% of every theft case recorded in India that year.
- 504 murders.
- 5,580 kidnappings and abductions, at 25.5 per lakh against a national average of 6.8. One report of the same data gives 5,715 for 2023 where another gives 5,681; the 2024 figure is not in dispute.
- A chargesheet rate of 31.9%, and for crimes against children 31.7% against a national average of 61.4%.
That theft line is widely misreported, and the error is easy to catch. Several outlets printed the national theft total of 246,882 and the national property-crime total of 280,867 as though both belonged to Delhi. They are India’s, and an earlier draft of this page repeated the mistake before checking it. The giveaway sits in those same reports: 246,882 cases would be 676 a day, not the roughly 497 they all state alongside it. Delhi’s own count of 180,973 does divide to about 496 a day. The corrected figure is the more striking one, because it means roughly three of every four thefts recorded anywhere in India in 2024 were recorded in Delhi.
Two of those deserve flagging rather than burying. Delhi’s kidnapping rate is roughly four times the national one, and its chargesheet rate is roughly half. A city can register more cases and still take fewer of them to court, and Delhi does both. That combination is worth more attention than the headline total. It also applies to the whole territory without distinction, which is worth remembering whenever a number like this gets attached to a particular belt, whether that is Vasant Vihar or Kalkaji, which this data covers without describing.
The “crime capital” figure, and what is wrong with it
You will have seen the claim that Delhi has the highest crime rate in India. Here is the number behind it and here is what it is dividing by.
A crime rate is crimes divided by people, multiplied by 100,000. In India the same figure is usually written “per lakh”, and a lakh is 100,000, so the two phrases mean the same thing. Because a rate is a division, you can run it backwards: divide the crimes by the rate, multiply by 100,000, and you recover the population the rate used. NCRB does not print that population beside the rate. It can be recovered anyway, and it is the whole story.
The 2024 figure. NCRB put Delhi’s cognizable crime rate at 1,258.5 per lakh. Divide 275,402 crimes by 1,258.5 and multiply by 100,000, and you get 21,883,353 people. Check it against a second, unrelated line in the same report: 5,580 kidnappings at a rate of 25.5 per lakh gives 21,882,353. Two independent routes, a thousand people apart in 21.9 million.
So NCRB’s 2024 Delhi rate divides by the whole National Capital Territory, roughly 21.9 million people. It does not describe a city centre, a district or your street.
The 2022 figure, and why it cannot be set beside the first. NCRB’s metropolitan chapter for 2022 gave Delhi a crime rate of 1,832.6 per lakh, more than three times the 544 average across the 19 cities it covers. No other city came close.
| City, NCRB metropolitan chapter 2022 | Crime rate per lakh |
|---|---|
| Delhi | 1,832.6 |
| Jaipur | 916.7 |
| Indore | 767.7 |
| Kochi | 626.7 |
| Patna | 611.7 |
| Average of the 19 cities | 544 |
| Kolkata | 78.2 |
That table is a league position, not a safety measurement, and the bottom row shows why. Kolkata’s 78.2 does not make it 23 times safer than Delhi. A gap that size between two large Indian cities reflects recording practice long before it reflects danger. Read the column as a measure of how readily each force registers a complaint, and the rest of this section explains what else it is measuring.
Delhi’s total in that chapter was 298,988 cases. Run the same division: 298,988 divided by 1,832.6, times 100,000, gives 16,314,962 people.
That number is worth pausing on. Delhi’s urban agglomeration at the 2011 census was 16,314,838. The population NCRB’s 2022 metropolitan rate divides by, recovered from its own published figures, lands 124 people away from a census figure that was already 11 years old when the report came out.
| The two Delhi rates in circulation | Rate per lakh | Population it divides by | What that population is |
|---|---|---|---|
| NCRB metropolitan chapter, 2022 | 1,832.6 | 16,314,962 | The 2011 urban agglomeration |
| NCRB, 2024 | 1,258.5 | 21,883,353 | The whole territory, projected |
Both denominators are this page’s own back-calculation from NCRB’s published totals and rates. Every step is above, so you can repeat them with a calculator.
Set 1,832.6 against 1,258.5 and you appear to find a 31% fall in Delhi’s crime rate in two years. There was no such fall. There were two different denominators, one of them a decade out of date, and neither headline says so.
Then there is the deeper problem, which no arithmetic fixes. A crime rate built from registered cases measures registration. A police force that records complaints readily will post a higher number than one that does not, and the higher number may mean it is doing its job better. Researchers at the Observer Research Foundation name three distortions in Indian crime statistics: crimes citizens never report, cases police decline to register and the principal offence rule, under which a case involving several offences is counted only once, under the most serious.
One further note on the 2024 rate: at least one aggregator site, lyfsmile.com, gives it as 1,688 rather than 1,258.5. This page uses 1,258.5, which is the figure carried by the agency copy and the one that reconciles with NCRB’s own kidnapping rate. Where two numbers circulate, the one that survives a cross-check is the one to keep.
The upshot is narrow and worth stating exactly. Delhi’s figures are a reasonable guide to how Delhi’s own numbers are moving over time, which is what the previous section used them for. They are a poor guide to whether Delhi is more dangerous than Mumbai or Bengaluru, and no guide at all to whether your street is safer than the next one. That is a different question from what your rent buys, which a monthly income of ₹1 lakh, set against what this part of the city costs sets out separately.
The crime that should concern a renter
Put the eight-row table next to the question a person moving to South Delhi is really asking, and one line answers it.
Snatching. At 5,406 cases in 2025 it is the highest-volume violent street crime Delhi Police reports, four times the robbery count and 11 times the murder count. It is also the category police close least often, at 64.22%. A snatching is over in three seconds, usually involves a phone or a chain, usually happens on a road rather than inside a building, and usually leaves nothing behind to investigate. That combination is why the volume is high and the solving rate is low, and it is why this is the crime you are statistically most likely to encounter.
Theft. NCRB’s 180,973 thefts in Delhi in 2024 work out at about 496 a day, and they amount to 73.3% of every theft recorded in India that year. Most of it is not confrontational. Vehicles, parcels, phones, things left in cars.
Between them, snatching and theft account for the overwhelming majority of what actually happens to ordinary residents. Neither is what dominates the coverage.
Now the counterweight, because proportion matters in both directions. Delhi recorded 491 murders in 2025 across a territory of roughly 22 million people. That is a real number attached to real families, and it is also a rate low enough that it should not be the thing shaping where you rent.
What follows from that, and what does not. Two things follow from the definitions of the offences themselves, and are not inferences about particular places:
- Snatching happens outdoors and in motion. That is what the offence is. The exposure this category describes is on a street rather than inside a building, which is a statement about the crime, not about any neighbourhood.
- Theft at this volume is largely non-confrontational. Vehicles, parcels, phones and things left in cars. Again, that is what the recorded category consists of.
What does not follow is where in South Delhi any of it concentrates. No published dataset breaks Delhi’s crime down by street type, by market, by metro exit or by parking arrangement, and this page will not invent that breakdown from the shape of the offence. Anyone who tells you which streets or which stations are the risky ones is working from impression.
What the two facts above are good for is deciding what to inspect in a specific flat: where a vehicle would sit overnight, and what the approach on foot is like at the hour you would use it. Both are properties of an address, and both are in the checklist further down.
Safety for women, specifically
Many readers arrive at this page asking a narrower question than the title. It deserves a straight answer rather than a reassuring one, and the caveat has to come first, because without it the numbers below will mislead you.
Every crime statistic in India counts cases registered, not crimes committed. A city where more women are willing to report, and where police register more readily, produces a higher count than a city where neither is true. A high number is partly a measure of how many women were harmed and partly a measure of how many were heard. Nobody can currently separate the two. Any page that tells you Delhi is definitively more or less dangerous for women than another Indian city is going past its evidence, and so is any page that reads a high count as proof of either.
With that in front rather than behind, two things are true at once.
Delhi records more crimes against women than any other Indian metropolitan city. NCRB’s 2023 data put the figure at 13,366. That is the highest of the 19 metropolitan cities the report covers, and it is not a small margin.
And the citywide counts have fallen for two consecutive years. On Delhi Police’s figures, rape cases went from 2,141 in 2023 to 2,076 in 2024 to 1,901 in 2025. Molestation went from 2,345 to 2,037 to 1,708. Recorded eve-teasing went from 381 to 362 to 337. NCRB’s 2023 total of 13,366 was itself down 5.59% from 14,158 in 2022. Delhi Police reports solving 97.11% of rape cases and 95.20% of molestation cases registered in 2025.
What this page will not do is rank South Delhi localities on this dimension. The data does not exist, and a wrong ranking here would do more damage than a wrong ranking anywhere else on this site.
What exists. Delhi Police runs 112 as the single emergency number, which reaches police, fire and ambulance and has a mobile app of the same name. Its citizen services portal, listed in full further down this page, takes complaints online without a station visit. Every police station has a women’s help desk, and the force publishes station-level contact numbers through its district pages.
What is checkable sits at the level of a building and its approach rather than a colony: whether the last 200 metres are lit and in use at the hour you would walk them, whether the entrance is overlooked by occupied windows, who controls the gate after midnight and whether that person is awake, and which metro exit the address is nearest, since gates on the same station can open onto very different streets. One station’s exits, compared side by side shows how much that varies within one station. The same question is worth putting to any address near a nightlife cluster, including the streets around Hauz Khas Village’s restaurants and bars. Go and look at the hour you would actually be walking it, because nobody publishes that either.
How many police there are
Delhi Police is authorised 92,044 personnel and was 9,248 short as of Nov. 30, 2025, leaving roughly 82,796 working. Those figures come from a Rajya Sabha answer by Minister of State for Home Nityanand Rai in December 2025.
This page’s own arithmetic, not an official statistic: set 82,796 working officers against the same 21.88 million population base NCRB’s own rate implies and you get about 378 officers per lakh people. It counts every working post as equivalent, from constable to commissioner, and it includes officers in specialist units who never appear on a residential street.
The shortfall is uneven by rank, and it falls hardest in the two places a resident actually meets the force. Sub-inspectors, who investigate most cases, are 1,039 short of 8,087 sanctioned. Head constables are 3,057 short of 23,724 and constables 4,591 short of 50,946, which is where beat patrolling comes from. Assistant commissioners are 125 short of 346, a 36% vacancy at the rank that supervises several police stations at once, and by some distance the worst gap in the table.
Put those together and the picture is not a force with no officers. It is a force that is roughly a tenth short overall, with the supervisory tier thinner than the rest.
For the southern districts, the only staffing figures ever published by district are a decade old, from the 2014 white paper, and they are for police districts rather than the revenue districts named earlier. They are set out here because nothing more recent exists, and they should be read as history.
Do not read the table below as a comparison between the three districts. They differ in area, population and the kind of ground they cover, none of which appears here.
| Delhi Police district, 2014 | Police stations | Sanctioned | Working |
|---|---|---|---|
| South | 16 | 4,067 | 3,647 |
| South East | 17 | 4,016 | 3,507 |
| South West | 15 | 2,923 | 2,972 |
The figures come from a single RTI exercise a decade ago. A district with more officers is not thereby better policed, and one with fewer is not thereby worse.
Staffing is one input into what an address receives, alongside street lighting, waste collection and everything else that varies block to block rather than colony to colony.
What you can check before you sign
This is the part that beats any ranking, because it is about your address rather than an average.
Delhi Police runs its citizen services online, and the exact names and addresses matter because search results are full of imitations. These are the services as listed on the force’s own site.
| Service | Where |
|---|---|
| Lodge a complaint | cctns.delhipolice.gov.in/citizenicms |
| Check complaint status | cctns.delhipolice.gov.in/citizen |
| Theft e-FIR | propertytheft.delhipolice.gov.in |
| Motor vehicle theft e-FIR | mvt.delhipolice.gov.in |
| Lost report | lostfound.delhipolice.gov.in |
| View an FIR | delhipolice.gov.in/viewfir.aspx |
| Missing person or child registration | cctns.delhipolice.gov.in/citizenservices |
| Domestic help and tenant registration | cctns.delhipolice.gov.in/citizenservices |
| Senior citizen registration | seniorcitizen.delhipolice.gov.in |
| Missing or stolen mobile phones | zipnet.delhipolice.gov.in |
| Police clearance and character verification | pcccvr.delhipolice.gov.in |
One practical warning. At the time of writing, the www.delhipolice.gov.in address was serving an expired security certificate, so some browsers show a warning before the page loads. The bare delhipolice.gov.in address and the individual service subdomains above were working normally. A certificate warning on a genuine government domain is a nuisance rather than a danger, but it is exactly the situation that trains people to click through warnings, so check the address bar rather than the reassurance.
Tenant registration is worth understanding rather than resisting. It is the landlord’s job to file it, and it puts your name, photograph and identity document on record with the local police station. Tenants sometimes treat it as surveillance and try to avoid it. That is the wrong instinct. A registered tenancy is evidence that you live where you say you live, which matters if anything ever goes wrong, and a landlord who does not want to file it is often a landlord whose paperwork will not stand up in other ways either. Claims circulate online that skipping it carries a specific criminal penalty under a named section; this page could not verify any of those citations from a primary source and does not repeat them. Ask your landlord to file it because it is in your interest, not because of a threatened fine.
Then run these five checks on the actual address:
- Find the police station the address falls under, and how far it is. Ask the landlord, then confirm it at the station rather than taking the answer on trust.
- Walk the last 300 metres after dark, on a weekday. Not the drive. The walk from wherever the auto or the metro actually leaves you.
- Look at where a vehicle would sit overnight. Street, stilt, basement, gate. This decides more of your practical risk than the colony’s reputation does.
- Find out who controls the entrance at night and whether anyone is awake. A gate with nobody behind it is a gate.
- Ask two neighbours what has happened on that street in the last year. Not “is it safe”, which gets you a reassurance. Ask what happened. Ask specific and you get specific.
How to use the answers. None of the five is a pass or fail on its own, and no address will be perfect on all five. Weight them by what you will actually do every day. If you come home after dark on foot, checks two and four decide it. If you keep a vehicle, check three does. If you travel and leave the flat empty, checks three and four matter more than one and two. An address that fails the check tied to your own daily routine is the one to walk away from, whatever it scores on the others, and an address that fails only a check you will never encounter is not a problem.
Older stock needs a sixth check, because grilles, gates and lighting were usually retrofitted rather than designed in, and the inspection an older builder floor deserves covers what to inspect. The same applies further out, where plot sizes and boundary walls differ from the planned colonies, as renting in Mehrauli, start to finish sets out.
Entrances vary more than floor plans do, which is worth weighing when you compare studio flats across South Delhi at similar rents. Where a listing comes through the brokers and owners this site works with, put the same five questions to them.
How to read the next “safest areas” list
You will meet another one within a week. Four questions settle almost all of them in about a minute, and they work on any safety claim about anywhere, not just this one.
What is on top of the fraction, and who published it?
If a page gives a number of crimes for a locality, ask which body produced it and in which document. In Delhi there is no locality-level publisher. There is no colony-level publisher either. If the sourcing is “police sources” or “as per data”, with no report named and no year given, what you are reading is somebody’s recollection formatted as a statistic.
What is underneath it, and does it cover the same ground?
A rate needs a population measured on the same boundary as the crimes. Delhi’s crime counts come from police districts and its census populations from revenue districts, and the two are different maps with overlapping names. A page that divides one by the other has not noticed. A page that gives a rate without naming a population base cannot be checked at all, which is why the section above spent four paragraphs recovering NCRB’s.
What date does it describe?
Delhi Police’s published statistical series ends in July 2022. The census is from 2011. The revenue district boundaries changed on Jan. 1, 2026. A page offering a “current” locality crime rate for South Delhi is offering something that does not exist in current form, whatever the copyright line at the bottom says.
Does it separate crimes reported from crimes committed?
Every crime statistic in India measures the first. A source that treats a high count as proof of a dangerous place, without mentioning registration rates even once, has skipped the single most important caveat in the subject. A source that mentions it and then ranks places anyway has mentioned it decoratively.
A claim that survives all four is worth your attention. Most will not survive the first, and the failure is usually not dishonesty. It is that ranking places is easy to write and checking numerators is not.
The same four questions are worth carrying into any confident claim about a place, not just a safety one. A page telling you an area is well connected, quiet or well served should be able to say from what, measured how and when. This site’s own pages on the public services a Greater Kailash address actually gets and the corridor that carries most of South Delhi’s through traffic are written to be checked the same way.
Applied to the genre you will actually meet: a “10 safest areas in South Delhi” listicle has no numerator, no denominator, no date and no distinction between reported and committed. It is four failures out of four, and the confident tone is doing all the work.
Frequently Asked Questions
Is South Delhi safer than other parts of Delhi?
Unknown, and not knowable from published data. The district figures that exist are unreliable for the five reasons set out above, and no rate can be built from them. What can be said is that the question is asked far more often about South Delhi than answered, and that every confident answer you find rests on impression rather than evidence.
Which is the safest area in South Delhi?
This page will not name one, and you should distrust pages that do. The honest replacement for that question is the address-level checklist above, which produces better information about the flat you are actually considering than any ranking could produce about the colony around it.
Is it safe to walk at night?
The data cannot answer that for a specific street, which is why the checklist tells you to walk it yourself. What the data does say is that street crime dominates the numbers and snatching dominates street crime, so the answer depends far more on how busy and lit a particular stretch is than on which colony it sits in. Two streets 400 metres apart can differ completely on that, and only one of them is the one you will use. Within a single colony the variation can be just as wide, which is why Safdarjung Development Area’s rental stock is worth walking block by block rather than judging as a whole.
Do gated societies actually help?
Nothing in published Delhi data separates gated from ungated addresses, so this page cannot claim a difference. What is observable is that snatching and theft, which dominate the statistics, mostly happen outside a building rather than inside it, so what matters is the approach and the parking rather than the gate itself. The same logic applies to open public space, whether that is a park, a market or the public garden off Said-ul-Ajaib: what changes across a day is how many people are present, and no published dataset records that either.
Is Delhi getting safer, or is reporting just changing?
Both interpretations survive the evidence, and honest reading holds them together. Six of eight categories fell in 2025, and total NCRB-recorded crime fell about 15% between 2023 and 2024. Falls that consistent across categories are hard to explain by reporting changes alone. But a fall in registered cases can also reflect fewer cases being registered, and Delhi’s chargesheet rate of 31.9% suggests plenty of room for the system to shed cases along the way.
What do I do if something is stolen?
For a phone, a vehicle or property, use the e-FIR services listed above rather than waiting for a station visit, and file the same day. The lost-report service covers things you have mislaid; the theft services cover things taken. For a vehicle, file the motor vehicle theft e-FIR immediately, because insurance claims usually depend on the FIR date. For anything involving a person, go to the station or call 112.