When someone buys land in India, what they are actually buying is not a guarantee. It is a claim, resting on the last recorded transaction in a chain of transactions that stretches back through sellers, and their sellers, sometimes for generations. Nobody in the process certifies that the chain is unbroken. The buyer’s job is to check it themselves, and if they miss a defect three owners back, that defect is now theirs.
This is not a footnote to India’s land problem. It is the mechanism underneath it. Everything that follows in this article, the disputes, the decades in court, the reforms that keep falling short, traces back to this one design choice: India registers transactions, not ownership. A discussion paper from PRS Legislative Research on land records and titles lays out how that choice was made, what it costs, and what fixing it would actually require.
A widely cited number that is not a current measurement
The figure most often used to describe the scale of India’s land dispute problem comes from a World Bank study published in 2007. It suggested that land-related disputes account for roughly two-thirds of all pending court cases in India. Nearly two decades on, that number is still the one policy documents reach for.
That is worth pausing on. A 2007 estimate is now approaching twenty years old, yet it remains the reference point cited by government papers, research institutions and, this article included, when the subject of land litigation comes up. The persistence of a single old number is itself a symptom of the data problem this article is about, not just background colour.
The PRS paper does not claim a current, measured share of pending cases devoted to land disputes, and neither does this article. For a sense of how large and strained the overall docket is today, our earlier reporting on the arithmetic of India’s case backlog is a more current starting point than reaching for an old percentage and presenting it as fresh.
Alongside the two-thirds estimate sits a second figure, this one from a NITI Aayog paper: land disputes take an average of about twenty years to resolve. Twenty years is not a delay in the ordinary sense. It is long enough for the person who filed the case to have died before it concludes, leaving the dispute to their children.
Part of why land cases run that long is structural, not specific to land law. We have separately reported on the vacancy rates across police, courts and prisons, where roughly one in four sanctioned justice-system posts sits empty. A judiciary short-staffed at that scale processes every category of case more slowly, land disputes included, and the twenty-year average has to be read against that backdrop of institutional capacity rather than as a land-law problem alone.
Why “I registered it” does not mean “I own it”
Most people who have bought a flat or a plot in India assume that once the sale deed is registered, the matter is closed. It is not. Indian land ownership operates on what the PRS paper calls a presumptive system: registration records that a transaction took place, but it does not certify that the seller had a valid, unencumbered right to sell in the first place.
Ownership under this system is established by tracing a sequence of past transactions, ideally back far enough to rule out any defect. If an earlier link in that chain was flawed, a forged signature, an unresolved inheritance claim, a sale made without the consent of a co-owner, the flaw does not disappear when the property changes hands again. It travels forward, waiting to surface.
The burden of finding that flaw before it surfaces falls on the buyer. In practice this means hiring a lawyer to conduct a title search, examining decades of prior deeds, mutation entries and encumbrance certificates, and still accepting a residual risk that something was missed or forged. No government body performs this check on the buyer’s behalf and then stands behind the result.
The table below sets out some of the assumptions people commonly bring to a property purchase, and what the presumptive system actually delivers instead.
| What people assume | What actually happens | Why |
|---|---|---|
| Registration proves I own it | Registration proves a transaction was recorded, not that the seller’s title was valid | India’s system is presumptive: ownership rests on a chain of past deeds, not a state guarantee |
| The record shows my name so it is settled | The record can still be challenged if an earlier transaction in the chain is later found defective | Title is established through a sequence of past transactions, any link of which can be disputed |
| The government checked before registering | The registering authority typically confirms the document’s form, not the underlying title | Verifying ownership history is treated as the buyer’s responsibility, not the state’s |
| A computerised record must be current | A record can be digitised and still not reflect the most recent change of ownership | Computerisation of existing records is a different task from updating mutation entries when land changes hands |
That last row matters more than it looks. A screen showing a clean, official-looking digital record feels more trustworthy than a dog-eared paper file, even when the underlying information is exactly as outdated. Digitisation without updating can make an old error look like a current fact.
A system built to tax possession, not to prove ownership
To understand why Indian land records work this way, it helps to look at what they were originally built to do. The record-keeping framework in use across much of the country today has its roots in the zamindari system, under which administrators needed a reliable way to identify who was cultivating or occupying a piece of land, so that revenue could be assessed and collected from that person.
That is a narrower question than “who owns this land.” A possession record answers “who do we collect from this season.” It does not need to resolve inheritance disputes, verify a chain of sale deeds, or guarantee that the occupant has a valid legal right against all other claimants. For the purpose it was designed for, it worked well enough.
The trouble is that this possession-tracking apparatus was never rebuilt from the ground up when land became something people buy, sell, mortgage, inherit and litigate over at modern scale. Instead, the old framework was carried forward and layered with new legal and administrative processes, while its core logic, record who holds it now, stayed largely intact.
A system engineered to answer a revenue question cannot, without deliberate redesign, answer an ownership question. Asking it to do so anyway is a large part of why so many disputes reach the point of litigation: the records themselves were never built to settle the question being fought over in court.
The gap inside the good news: 86 percent versus 47 percent
Government progress reports on land records modernisation tend to lead with a strong headline number. As of September 2017, 86 percent of land records in India had been computerised, according to figures cited in the PRS paper. On its own, that sounds like a programme well past its halfway point.
The same snapshot from September 2017 recorded something less encouraging sitting underneath that headline: only 47 percent of mutation records had been computerised. Mutation is the process that updates a land record to reflect a change of ownership, after a sale, an inheritance, a partition, or a court order. It is the part of the record that actually needs to move when land changes hands.
The gap between those two figures is the gap between digitising what already existed and keeping what exists accurate. A record can be carefully computerised and still describe an owner who sold the land, or died, years earlier, if the mutation entry that should have updated it was never made or never processed.
- A digitised record with a stale mutation entry is not a smaller version of the old problem; it presents old, incorrect information with the visual authority of an official database.
- Buyers, lenders and courts that treat a computerised record as inherently reliable can be misled precisely because it looks more reliable than a handwritten register.
- Computerisation targets that count “records digitised” without separately tracking “mutations kept current” can hit their headline number while the underlying accuracy problem persists untouched.
None of this means digitisation was pointless. It means digitisation and accuracy are two different jobs, and reporting progress on the first can create the impression that the second is also being solved, when the September 2017 figures cited in the PRS paper show the two moving at very different speeds.
What the modernisation programme has, and has not, delivered
The main vehicle for fixing this is the Digital India Land Records Modernization Programme, generally known by its acronym, DILRMP. Its goal is broad: computerised, updated, verified land records linked to accurate maps, available in something close to real time. The PRS paper reports several progress figures against that goal, and they are worth setting out plainly rather than folded into the 86 percent headline.
| Component | Reported progress | What the gap means in practice |
|---|---|---|
| Computerisation of land records | 86% (as of September 2017) | The headline figure, but it measures digitisation, not accuracy or currency of the record |
| Computerisation of mutation records | 47% (as of September 2017) | Over half of ownership-transfer entries were not yet reflected in computerised form at that time |
| Real-time updation of Records of Rights and maps | 15% | Most records were not being kept current as changes occurred, undermining trust in even the computerised share |
| Spatial data verification | 39% | Under two-fifths of land parcels had their physical boundaries checked against the record, leaving most boundary claims unverified |
| Cadastral maps linked to Records of Rights | 26% | Most parcels still lack a working link between the map showing where the land is and the record showing who holds it |
| Villages surveyed | 9% | A fresh, modern survey, the base layer a reliable titling system would need, had reached under one in ten villages |
Read together, these figures describe a programme that has made real progress on the easiest task, converting existing paper into digital form, while the harder tasks, keeping records current, verifying them against the ground, and connecting maps to ownership records, lag well behind. Nine percent village survey coverage in particular suggests the foundational work is still at an early stage.
It is also worth being honest about the age of these figures themselves. They are reported in the PRS paper as a status snapshot, and a modernisation programme is, by definition, a moving target. Where the programme stands today is not something this article can state without a more current, verified source, and it does not invent one.
The proposed fix: conclusive titling
The reform most often proposed to address all of this is a shift from India’s current presumptive system to what is called conclusive titling. The idea is straightforward to state, even though it is difficult to implement.
Under a conclusive titling system, the government would guarantee the accuracy of registered land titles. If it later turned out a title was wrong, in a way that caused loss to someone who relied on it, the state would pay compensation rather than leaving the loss to be fought out between private parties in court for twenty years.
Two things follow from that guarantee, and both are demanding:
- Registered titles would need to become the primary evidence of ownership, rather than one input among a chain of past deeds that a buyer must independently verify.
- The underlying records would need to be clear, updated and verified before the state could responsibly stand behind them, since a guarantee only works if what is being guaranteed is actually accurate.
That second requirement is the hard part, and it is why conclusive titling has been discussed for years without becoming the default nationwide. A government cannot credibly guarantee a title, and cannot budget for the compensation liability that guarantee implies, while mutation records are only 47 percent computerised and real-time updation sits at 15 percent, per the figures above.
Conclusive titling is not primarily a legal reform. It is a data-quality precondition wearing a legal reform’s name. The law can be rewritten relatively quickly; bringing the underlying records to a state where the government can safely guarantee them is the slower, more expensive piece of the work, and it is the piece the DILRMP figures show is still incomplete.
What conclusive titling would change, in brief
- The document of primary legal weight shifts from a chain of past sale deeds to a single, government-verified title.
- The burden of proving ownership shifts away from the buyer having to trace history themselves.
- The state accepts financial liability for errors in the record it now guarantees, which is what forces it to get the record right before switching over.
What twenty years of litigation actually costs
The numbers so far describe a systems problem. It is worth being specific about what that systems problem does to the people caught inside it, because a land dispute is rarely just about the land.
- Land tied up in a dispute generally cannot be used as collateral for a loan, which can cut a family off from institutional credit for farming, education or a medical emergency, for as long as the case runs.
- Inheritance disputes rooted in unclear title can outlast the original parties, passing from one generation to the next as an unresolved case rather than an asset.
- Uncertain title discourages investment in the land itself, since improvements made to disputed property can become part of what is being fought over.
- A twenty-year average case length falls hardest on people who cannot afford to fund two decades of legal fees, or who cannot afford to walk away from the dispute either, because the land is their main asset.
This is also where the wider pattern of the justice system’s capacity problems becomes relevant again. Land cases do not sit in a separate court system with its own dedicated pace. They move through the same overstretched courts, staffed at the same reduced levels, described in our reporting on justice-system vacancies.
A twenty-year land case and an overcrowded undertrial population, which we have also reported on in the context of pre-trial detention, are different symptoms of the same underlying capacity strain running through Indian public institutions.
There is a broader pattern here too, of laws and systems that promise a service on paper without the machinery to consistently deliver it. Our reporting on India’s right-to-service laws, and how unevenly they are actually enforced two decades on, describes the same shape of problem in a different domain: a well-intentioned framework, a real gap between the promise and the delivery, and citizens left to absorb that gap.
Making policy on old numbers
Step back from the individual figures and a pattern emerges. The two-thirds estimate that anchors most discussion of land litigation in India is from 2007. The computerisation snapshot most often cited, 86 percent overall against 47 percent for mutations, is from September 2017. Both are described in the PRS paper as the figures available, not as figures that have been recently re-measured.
That is not a minor caveat. It means that when policymakers, researchers and journalists discuss the scale of India’s land dispute problem today, they are frequently reasoning from a measurement taken before smartphones were common in rural India, and describing the state of digitisation using a snapshot that is itself now several years old.
It is possible, even likely, that both numbers have shifted since they were recorded, in either direction. Digitisation coverage has almost certainly grown past its September 2017 levels somewhere in the years since. Whether the underlying share of court cases rooted in land disputes has moved much from the 2007 estimate is genuinely not known, because nobody appears to have repeated that specific measurement at comparable scale.
A country cannot design a conclusive titling system, one that requires the state to guarantee millions of individual titles and accept liability for the ones it gets wrong, without current, granular, verified data about where its records actually stand today, parcel by parcel and district by district. The PRS paper’s own figures illustrate the size of that gap: DILRMP survey coverage at 9 percent of villages is not a rounding error away from the kind of ground-truth baseline conclusive titling would need.
The honest version of this story is not that India’s land records are a crisis frozen in 2007, or that they were solved by a modernisation programme measured in 2017. It is that the country continues to make consequential policy choices about land, credit, inheritance and litigation using numbers old enough to need a date attached every time they are used, and that the absence of a fresher, more complete national data set is itself one of the central obstacles standing between the presumptive system in place today and the conclusive one that has been proposed to replace it.