Two Banks, Two Title Verdicts on One Flat: Whose Word Counts?

Picture the situation. You approach two banks for a home loan on one flat and hand both the same file. Bank one's advocate writes that the title is clear and marketable. Bank two's advocate writes that it is not. The flat is exactly what it was last week. Which bank is right, and how would you ever know?

Short version for 2026

Do not assume either verdict is correct, and do not read the split as proof that someone was sloppy. The opinion you receive is one panel advocate's professional judgment about a presumptive title. No government stands behind it. India records deeds. It does not record ownership. So two skilled advocates can look at one property and honestly differ.

Three questions come before any choice:

• Was it really the title that failed, or only the loan? Every lender checks the law side and, separately, the technical and valuation side. Refusals come as often from valuation, the age of the building, a negative-area list or loan-to-value as from a genuine title flaw. Saying no to a loan does not always mean saying no to the title.
• Whose work was broader and more exact? Lean towards the opinion that went back further, that pins the problem to a named register, document or section, and that also looked at the two things a clean Encumbrance Certificate (EC) can conceal, namely a pending court case and an equitable mortgage.
• Commission one neutral, standard title search that repeats the full exercise and states the flaw in plain terms. Then you are no longer picking between two banks' private formats by luck.

What follows explains each step.

Why the same title draws opposite answers

First, a fact that few property blogs mention. Registering a sale deed in India does not make the buyer the owner, and it does not promise that he is. The Registration Act, 1908 compels registration of an immovable property sale (Section 17) and restricts what an unregistered paper can prove (Section 49). The Transfer of Property Act, 1882 asks for a registered instrument in a sale (Section 54). But the register notes only that a deal occurred. It cannot confirm that the seller held what he sold.

The Supreme Court made this point in Suraj Lamp & Industries v. State of Haryana, and again in Mahnoor Fatima Imran v. Visweswara Infrastructure (2025). In its words, registration "gives notice to the world that such a document has been executed" yet is "not to confer an unimpeachable validity on all such registered documents." The Court added that a buyer "must verify the entire chain of ownership, not merely rely on registered documents."

Jurists describe India's approach as a presumptive title system. Under a Torrens system the state keeps a single register and stands behind what it shows. India places the work of proving title on the buyer and the bank that lends to him. Nobody in government guarantees the entry.

This single fact accounts for the disagreement. The opinion is not a machine's reading of a perfect record. It is the advocate's estimate of how likely the seller's title is to be faulty, based only on what that advocate saw and searched. Nothing national tells advocates how many years to search, which registers to pull, or at what level a doubt becomes fatal. Each bank has its own panel. Each panelist has a favourite format and a personal tolerance for risk. The same file therefore returns "clear" in one office and "not clear" in another. That is how an unstandardised system behaves. It does not mean one advocate got it wrong.

Check whether the bank is really faulting the title

Many people go wrong here, so clear this up before anything else. For every home loan a bank runs two checks at once:

• Legal check, by its panel advocate: the ownership chain for the last 13 to 30 years, the sale deed and earlier link deeds, the sanctioned building plan, the NOC, the occupancy certificate and the EC.
• Technical check, by the bank's approved valuer or civil engineer: a visit to the site to compare measurements with the sanctioned plan and to judge construction quality, layout violations and market value.

Technical or commercial reasons alone can end a loan application, even when the title has no defect. Lenders keep "negative areas" off their books because resale is weak or disputes are common. A building aged about 30 years or more has less life left, so it draws a shorter tenure or a lower valuation. An unauthorised extension may be subtracted from value, or may sink the application. And since banks generally finance only 75 to 90 percent of assessed value, not of your price, a low valuation may look like a refusal.

Hence, before you fret about two "legal opinions", confirm that both concern the title. Write to each bank and ask for the precise ground: legal opinion, technical valuation or credit policy. In many cases the flat is sound, one lender just does not finance that type of building, that locality or that loan-to-value, and there is no title issue to solve.

Search depth: 13 years at one bank and 30 at the other

Suppose both objections do concern title. The likeliest cause then is how far each advocate went back. Sound practice calls for an unbroken chain across 30 years. For a long time many banks took 13 years as good enough. The longer search is dearer and slower, and that is the reason lenders' policies are not alike.

Depth alters the outcome. Take a search that begins in 2013, for instance. If the flaw lies in a transaction of 2006 (a mortgage that was never formally released, a partition that omitted a coparcener, or a gap in the chain) the shorter search will look spotless. A search covering 30 years reaches it. Thus "clear" over 13 years and "not clear" over 30 years may both be truthful and yet conflict. When they do, trust the longer one more.

Two further scope gaps produce the same split:

• What papers each advocate was given. One may have read the whole mother-deed chain, and the other just the seller's latest sale deed. Fewer papers means less risk in view.
• The survey number behind the EC. An EC means little unless the number searched is the right one. A NIL EC for a wrong survey number looks clean because the wrong parcel was searched. If the two banks used different numbers, they effectively searched two properties.

So set the opinions side by side and look at scope first: the years searched, the papers said to have been examined, and the precise survey or property number used for the EC.

Why a clean EC can still sit beside a warning

Two of India's gravest title risks never turn up on an Encumbrance Certificate. An advocate finds them only by going elsewhere, and not all do.

• A court case in progress. A suit over the property, a partition claim, an injunction or a dispute over ownership is never entered in the sub-registrar's records, which is why the EC cannot list it. Section 52 of the Transfer of Property Act (lis pendens) binds a buyer who purchases during a pending suit to whatever result the suit reaches. A search of the relevant courts by party name and by village or survey number is the only way to find it. One advocate searched eCourts and raised the suit. The other relied on the clean EC.
• An equitable mortgage. A borrower who leaves the original title deeds with a bank creates a mortgage (Section 58(f), Transfer of Property Act) that frequently does not appear on the EC. It is supposed to be recorded with CERSAI, the central charge registry. Anyone can search it for about ten rupees, either by the asset or by the borrower's PAN. A live charge that the EC hides completely would be visible to an advocate who checked CERSAI.

Therefore the fact that both ECs were clean proves little. Usually it means that one advocate ran the extra searches (eCourts and CERSAI) and the other did not. Put your trust in the opinion that checked more.

How to choose between the two opinions

Do not weigh the verdicts. Weigh the work. Request each bank's opinion, or at least its scope, and line the two up on five points:

• Years searched. Thirty beats 13. A clean result over a longer period proves more.
• Precision over vagueness. "Title not clear" is useless. A statement such as "a partition suit, O.S. 123 of 2019, is pending, so Section 52 lis pendens attaches" names the problem and shows if it can be cured. Prefer an opinion that cites the register, document or section to one that offers a single-line conclusion.
• Registers consulted. Did both advocates go beyond the EC to eCourts and CERSAI, and did they compare the deed's survey number with the revenue record? An opinion rests on the databases behind it.
• Verdict type: clear, conditional or not clear. Advocates often accept a title "subject to conditions", such as proof of release of an old mortgage or one specific no-objection. Conditional acceptance is a different thing from refusal. Check if bank two found a real defect, or merely a condition that bank one had already written into its sanction.
• Date of issue. When the opinions are weeks or months apart, a new charge, suit or attachment could have arisen in the gap, so the later opinion may just be up to date.

Do this exercise and most of the time you will find that the opinions do not truly clash. One went deeper, or consulted a register the other ignored, or was talking about valuation and not title, or attached a condition that the other bank silently accepted.

A lasting fix: one standard search in place of rival opinions

You face a choice between verdicts because each was prepared for its own bank, in its advocate's format, at its bank's level of risk tolerance. Nobody drafted them to be compared. That is where the gap sits.

To close it, get one independent title search that is not tied to either lender. It should run a complete 30-year ownership chain, pull the EC on the correct survey number, include revenue records, search CERSAI for charges, search eCourts for litigation, and run the prohibited-property and 22A checks, all in one place. It should end with a standard rating that names every defect and grades the risk. In place of "clear" from one advocate and "not clear" from the next, you would have a single report you can compare, which repeats everything a bank's advocate does and tells you exactly what is wrong and if it can be put right.

LegiScore reports are designed for this purpose. A report runs the searches that a bank's panel advocate would run, across all the government databases, and returns one standard title rating where every risk is named and sourced. You stop depending on whichever of two bank formats happened to be more thorough.

When two banks give you two answers, take one independent report and decide on facts.

Get an independent standardised title search report