A 30-year title chain is a bank credit-policy convention, not a legal requirement: no Indian statute anywhere mandates a 30-year chain. So when the seller's oldest paper is a 1998 deed, that is almost always a problem with the file, not with the title. The pre-1998 links the seller cannot hand you are reconstructed from the Sub-Registrar's own register, a 30-year Encumbrance Certificate plus certified copies of the older deeds under Section 57 of the Registration Act, 1908: and from the revenue records, not from the seller's cupboard. A short chain is only genuinely fatal when the root of title itself cannot be established.
In short: The panel advocate is asking for 30 years of comfort, not 30 years of paper the seller must physically own. A 30-year EC pulled from the SRO shows every registered transaction on the property regardless of who holds the deeds, including the 1996 or 1997 sale that fed the 1998 deed. Read that EC, note the older registration numbers, and pull certified copies of those deeds from the same SRO. A certified copy is admissible to prove the contents of the original (s.57(5), Registration Act, 1908). The chain gets rebuilt from the register; the thin physical file does not decide the title.
Is a 30-year title chain actually a legal requirement?
No. There is no statute in India that says a title must be traced for 30 years, or for 13, or for any fixed number. The period a bank asks for is a prudence convention, and it is built on two unrelated ideas:
- The 12-year figure comes from the Limitation Act, 1963. Article 65 gives a person 12 years to sue to recover possession of immovable property based on title; after that, an adverse possessor's claim can harden. A search that safely clears the last 12–13 years covers that risk window.
- The 30-year figure comes from an evidentiary presumption, now in Section 92 of the Bharatiya Sakshya Adhiniyam, 2023 (the successor to Section 90 of the old Evidence Act): a document 30 years or older, produced from proper custody, may be presumed genuine. Trace back 30 years and the oldest deed in the chain tends to carry that presumption on its own.
Banks pick 30 as the conservative, belt-and-braces number. It is policy layered on prudence, not law. That distinction is the whole answer to your predicament: the advocate wants title established, and title is not measured in calendar years.
Myth: "30 years is the law." ✗ No Indian statute mandates a 30-year chain. It is a lender convention resting on the Limitation Act's 12-year window and a 30-year document presumption.
The seller's oldest deed is from 1998: is the title defective?
Usually not. A short chain is not the same thing as a broken chain. What the advocate actually needs is the root of title: the deed at which ownership first vests cleanly and everything after it flows in an unbroken line. If the 1998 document is that root, then 1998 is where the chain legitimately begins and there is nothing "missing" at all.
A 1998 deed is frequently the root when it is:
- an original government grant, patta, or allotment (from a development authority, a housing board, or a revenue grant), title is created here, so nothing precedes it;
- the first sale after a land conversion (agricultural to non-agricultural / NA), where the converted parcel is a fresh unit of title;
- a partition deed or family settlement that carved out the specific plot; or
- the first registered sale after a re-survey or layout formation, where the plot number itself is new.
In each of these, asking for pre-1998 deeds is asking for the history of a thing that did not exist yet. The advocate's job is to identify and verify the root, not to fill an arbitrary 30-year box. Where the root is older than 1998 and the seller simply doesn't hold the earlier paper, that is the reconstruction problem the rest of this post solves, almost always solvable.
Myth: "No original deed means no title." ✗ A sale passes title only through a registered deed (Suraj Lamp & Industries v. State of Haryana, (2012) 1 SCC 656), which means the register already holds the chain, whether or not the seller kept the paper.
Where do the missing pre-1998 years actually live?
In the register, not in the seller's file. This is the single most useful fact for your situation.
Under the Registration Act, 1908, every Sub-Registrar Office keeps Book No. 1, the "Register of non-testamentary documents relating to immovable property" (Section 51), which holds the running record of registered dealings in immovable property. An Encumbrance Certificate is simply a certified extract of the index to Book No. 1 for one property over a chosen period, the registered sales, mortgages, gifts, partitions and releases, in date order.
Crucially, the EC is generated from the SRO's records, not from the seller's documents. So a 30-year EC (say 1996–2026) will show the 1996 or 1997 transaction that conveyed the property to whoever then sold it in 1998, even though the seller has never seen that older deed. That entry gives you the registration number, date, SRO and book/volume reference; with it, the "missing" deed is a request form away.
How do you get deeds the seller doesn't have?
You buy them from the register. The Registration Act makes Book No. 1 and its index "open to inspection by any person" (Section 57(1)), and any person may obtain certified copies of the entries. A certified copy is "signed and sealed by the registering officer, and… admissible for the purpose of proving the contents of the original documents" (Section 57(5)). In evidence law, that certified copy is secondary evidence (Section 58, Bharatiya Sakshya Adhiniyam, 2023). In plain terms: the copy you pull from the SRO stands in for the original the seller lost.
Here is the concrete reconstruction sequence a competent advocate (or a title-search platform) runs:
- Order a full-period EC: cover the whole window the bank wants (e.g. 30 years), not just 1998-onwards. This is the map.
- Read the EC backwards: find the transaction before the seller's 1998 deed. Note its registration number, date and SRO.
- Pull certified copies under s.57 of every deed the seller cannot produce, the 1996/1997 sale, the earlier mortgage release, the partition, straight from the SRO where each was registered.
- Cross-check the revenue records: RTC / pahani / khatian / jamabandi and the mutation register, to confirm the same succession of owners appears on the revenue side.
- Run the two searches the EC will never show: CERSAI (for equitable mortgages) and a name-based litigation search (for pending suits). More on these below.
Steps 1–3 alone typically rebuild the chain from a 1998-only file to a full-period, evidence-grade chain, using nothing the seller had to keep.
What if the register itself has no record before a certain year?
This is the real edge case, and it comes in two flavours.
Online floors are not record floors. Many state portals only serve digitised records from a cut-off date. Karnataka's Kaveri Online, for example, issues digitally-signed ECs for 1 April 2004 onward; anything before that needs a separate manual/offline application at the SRO (the "EC before 01-04-2004" service). Skipping that older window is the most common way a 1990s mortgage or partition stays invisible in a Karnataka search. The record exists; the website just doesn't reach it. The fix is a back-office manual search at the SRO, not a conclusion that "there's nothing there."
Genuine record loss is rarer. If the SRO's own volume is destroyed or illegible, you reconstruct from the surviving corroboration: the EC/index entries that do exist, the revenue mutation trail, and any certified copies held elsewhere in the chain. And where it is the owner's copy of a deed that is lost (not the register's), the recognised lost-deed protocol is: file a police complaint / FIR (or NCR) → publish a public notice in newspapers → swear a notarised indemnity affidavit → obtain the certified copy from the SRO. (We cover this end-to-end in our guide on recovering lost property documents and certified copies.)
Do mutation and revenue records prove the chain?
They corroborate it, they do not, by themselves, prove title. This is a line the advocate will hold firmly, and rightly.
The Supreme Court has been consistent for decades: in Jitendra Singh v. State of Madhya Pradesh (2021), it held that a "mutation entry does not confer any right, title or interest… the mutation entry in the revenue record is only for the fiscal purpose," reiterating Balwant Singh v. Daulat Singh (1997) that mutation "neither creates nor extinguishes title… nor has it any presumptive value on title." The Court said the same again in Jamnabai v. Vasudev (2026 INSC 900): a revenue mutation entry, by itself, cannot extinguish a person's proprietary rights.
So use RTC/pahani/jamabandi and the mutation register as a second, independent trace of who owned the land when, powerful for older and rural parcels where registration records thin out, but always anchored back to the registered instruments. Mutation says who paid the land tax; the registered deed says who owns the land. The reconstructed chain needs both, in that hierarchy.
Is a reconstructed chain as good as the original deeds?
For tracing title, largely yes, on two counts.
First, evidentiary weight. A certified copy from the SRO is admissible to prove the contents of the original (s.57(5), Registration Act) and qualifies as secondary evidence (s.58, BSA 2023). A chain rebuilt from certified copies is not a weaker chain; it proves what the lost paper said, although the 30-year presumption of due execution attaches to an original produced from proper custody and not to a certified copy (Lakhi Baruah v. Padma Kanta Kalita, (1996) 8 SCC 357).
Second, what "clear title" even means. A registered deed is not, in itself, a certificate of ownership. The Sub-Registrar's role is ministerial: in K. Gopi v. Sub-Registrar (2025 INSC 462), the Supreme Court held it is "not the function of the Sub-Registrar… to ascertain whether the vendor has title," and struck down a rule making registration conditional on proving title. Title is established by the whole picture, the unbroken chain of registered instruments, the EC, the revenue trail, not by any single deed being physically present. That is exactly why reconstructing the missing links from the register can support a sound title opinion.
When is a short chain genuinely fatal?
Most 1998-file cases are fixable; a few are not. Being honest about the difference is what keeps a title opinion trustworthy.
| Situation | Verdict | Why |
|---|---|---|
| Short chain rooted in a 1998 government grant / patta / partition / post-conversion sale | Fixable, often already complete | 1998 is the root; nothing precedes it |
| Seller holds no physical pre-1998 deeds, but the EC shows the transactions | Fixable | Pull certified copies under s.57 from the SRO |
| Pre-2004 (or pre-portal-cutoff) entries missing online | Fixable | Manual/offline EC search at the SRO reaches them |
| Mutation lags behind a registered sale | Fixable | Registered deed governs; get mutation updated |
| Owner's original deed lost, but registered | Fixable | Lost-deed protocol → certified copy |
| A real break: EC shows a conveyance to a party who then never re-conveys | Needs more / may be fatal | The chain literally stops; no secondary evidence can invent the missing link |
| A contested or litigated link (disputed partition, pending suit on ownership) | Needs more | Title is sub judice; resolve before lending |
| A deed the register cannot produce and no revenue/EC corroboration exists | Fatal without more | Nothing to reconstruct from |
The test is not "how many years of paper?" It is "can the root of title be established, and is every link after it present in, or reconstructable from, the register?"
The blind spot even a full 30-year chain won't cover
Extending the search to 30 years buys you the registered history. Two serious encumbrances are, by design, not registered instruments: so an EC, however long, can miss them:
- Equitable mortgages (mortgage by deposit of title deeds, s.58(f), Transfer of Property Act, 1882). These can be created without a registered deed (an EC shows one only where a memorandum of deposit was registered, as is common in Tamil Nadu and Karnataka), which is why banks record them at CERSAI, the central registry of security interests. A CERSAI search, alongside the EC, is what catches a subsisting bank charge on the property.
- Pending litigation / lis pendens (s.52, Transfer of Property Act, 1882). A suit over the property is not, by itself, entered in the SRO register, so it does not appear in the EC unless a court has communicated an attachment. A name-based search of the litigation courts is the only way to surface it.
A "30-year chain" that skips CERSAI and the litigation courts is long but not complete. (See also why two banks can reach different legal opinions on the same property, the searches they run, not just the years, differ.)
Can the loan still go through?
In the overwhelming majority of 1998-file cases, yes, once the chain is reconstructed and documented. The advocate does not need the seller to own 30 years of paper; the advocate needs a title opinion that establishes the root and shows every subsequent link, backed by certified copies and a clean EC/CERSAI/litigation search. That is a document-gathering exercise, not a dead end.
Where a lender has already declined, get the reason in writing first: a "short chain" objection is fixable; a "broken chain" objection needs a different response entirely. Our guide on what to do when a legal opinion sinks your home loan covers reading that objection precisely.
This is also where a standardised title-search platform earns its place. In Central Bank of India v. Prabha Jain (2025 INSC 95, 9 January 2025), the Supreme Court urged the RBI and stakeholders to "collaborate in developing a standardized and practical approach for preparing title search report before sanctioning loans" and to set "standard guidelines for fees and costs", because reports are too often "obtained cheaply." (The RBI has been consulting banks on such a framework since around April 2026, including a minimum search period; as of 1 October 2026 it has not issued a direction.) LegiScore applies one consistent method to every search: it reads the EC, the registered deeds and the revenue records together, flags the links and certified copies still missing from the file, and sets the CERSAI and litigation-court results beside them. A thin physical file does not mean a thin title. LegiScore is a technology provider that augments the panel advocate's opinion; the legal call remains the advocate's.
FAQ
Is a 30-year title chain legally mandatory in India? No. No statute mandates a 30-year (or any fixed-year) chain. It is a bank credit-policy convention resting on the Limitation Act's 12-year adverse-possession window (Article 65) and a 30-year document-genuineness presumption (Section 92, Bharatiya Sakshya Adhiniyam, 2023). The law requires that title be established, a question of the root of title, not calendar years.
The seller only has a 1998 deed. Can I still get a loan? Usually yes. If 1998 is the root of title (a grant, patta, partition or post-conversion sale), the chain is already complete. If the root is older, the pre-1998 links are reconstructed from a full-period Encumbrance Certificate plus certified copies of the older deeds from the Sub-Registrar, the seller does not need to physically own them.
How do I get a deed the seller lost? Obtain a certified copy from the Sub-Registrar Office where it was registered. Under Section 57 of the Registration Act, 1908, the register is open to any person, and a certified copy is admissible to prove the contents of the original. Use the Encumbrance Certificate to find the registration number, then apply for the certified copy.
Is a certified copy as good as the original deed? For tracing the chain, yes. A certified copy is admissible to prove the contents of the original (s.57(5), Registration Act) and is secondary evidence under Section 58 of the Bharatiya Sakshya Adhiniyam, 2023. A chain reconstructed from certified copies proves the contents of the lost deeds; the 30-year presumption of due execution applies to originals, not to certified copies.
Does the RTC / mutation record prove ownership? No. Revenue and mutation records corroborate the chain but do not confer title, the Supreme Court has repeatedly held (Jitendra Singh v. State of MP, 2021; Balwant Singh v. Daulat Singh, 1997; Jamnabai v. Vasudev, 2026) that a mutation entry is for fiscal purposes and neither creates nor extinguishes title. Use them as a second trace, anchored to the registered deeds.
Why can't I see pre-2004 entries online for a Karnataka property? Because Kaveri Online serves digitally-signed ECs only from 1 April 2004. Pre-cutoff records exist but require a manual/offline application at the Sub-Registrar Office. The absence of an online entry is not an absence of a record.
When is a short chain a genuine problem? When there is a real break the register cannot bridge, a conveyance to a party who never re-conveyed, a contested or litigated link, or a deed the register cannot produce with no EC or revenue corroboration. A short chain rooted in an old grant, or a missing physical copy the register still holds, is not a break.
Related reading: 13-year vs 30-year title chain, what the difference actually means · How to trace a property's chain of title in India · Only a photocopy of the mother deed? Verify title before paying the advance · Your own lawyer's title report vs the bank's empanelled advocate