Someone sold your land. You found out from the neighbours.

Seller impersonation fraud does not target the house you live in. It targets the three acres you inherited, the lot you have been holding since 2009, the cabin you visit twice a year. The industry's answer has been to verify sellers harder at closing, which is a check performed on a stranger by people who have never met the real owner.

The phone call nobody is ready for

Here is the shape this takes, again and again, in county after county.

A woman inherits three acres from her father. It is not an investment and it was never meant to be. It is the field behind the house he grew up in, and she keeps it because selling it would feel like a decision she is not ready to make. She pays the property taxes every year by direct debit. She drives out maybe twice a year, usually in spring.

In April she drives out. There is a truck on the access track. There is a surveyor's stake at the corner of the parcel. There is a concrete foundation, poured and cured, where the treeline used to start.

The man on the site is polite and completely bewildered. He bought the land in November. He has a deed. He has title insurance. He has a closing file with an executed grant deed, a notarial certificate, a settlement statement, and a wire confirmation. He has been paying a construction loan for five months. As far as every record in the county is concerned, he owns the ground she is standing on.

She has never spoken to him. She never listed the property. She never signed anything. She did not know it was for sale.

Somewhere in the four months between the listing and the recording, an entirely fictional version of her contacted an agent, produced identification, answered questions, signed a deed in front of a notary, and collected roughly two hundred thousand dollars, which left the country before the wire was a day old.

What follows is years. Her title insurer, if she has one, may not cover her, because she is not the one who bought a policy in this transaction. His insurer will defend his title. The county recorder cannot simply unrecord a document. The dispute becomes a quiet title action, and the person who did nothing wrong pays lawyers to prove she still owns land she never sold.

What is seller impersonation fraud?

Seller impersonation is a fraud in which somebody who does not own a property presents themselves as its owner, executes a sale to an innocent buyer, and takes the proceeds. It is sometimes called deed fraud, title theft, or home title theft, and the marketing name has drifted enough that the terms now blur together.

It is worth separating it from two adjacent crimes it gets confused with. It is not the same as closing wire fraud, where the sale is genuine and the attacker redirects the disbursement. And it is not mortgage fraud, where a real owner lies about income or occupancy. In seller impersonation, the transaction itself is a fiction from the first phone call, and the only real participants are the buyer and the professionals who processed it.

The distinguishing feature, and the reason it is so hard to unwind, is that everybody in the chain behaved reasonably. The listing agent met a seller. The title company ran the checks it runs. The notary observed a signature. The recorder accepted a document that was facially valid, which is generally what a recorder is legally obliged to do. There is no negligent party to point at, which is exactly why the loss lands on whoever can least afford to litigate.

Why does this happen to vacant land?

The concentration here is not subtle, and it is the single most useful fact in this whole subject.

The National Association of Realtors reported in its 2025 Deed and Title Fraud Survey that a large majority of title fraud cases in the prior year involved vacant land, with owner occupied homes making up only a small minority, and that around six in ten association advocacy professionals had encountered title fraud in their market. Earlier work from the American Land Title Association pointed the same direction, with vacant land accounting for the overwhelming share of seller impersonation incidents. Different surveys, different methodologies, same finding.

Once you see why, the whole crime becomes legible.

Nobody is standing on it

The single control that protects an occupied home is that somebody lives there. A stranger cannot sell your house out from under you easily, because at some point a buyer, an agent, an inspector, or an appraiser wants to walk through the property, and there you are. Occupancy is a continuous, free, extremely effective identity check performed by the property itself.

Vacant land has no such thing. Nobody is on it. Nobody visits. A buyer can walk the lot at their leisure and meet nobody, because meeting nobody is what buying vacant land is like.

The target list is published, for good reasons

Property ownership records are public, and they are public for sound reasons that nobody should want to change. Public records are how you find out who owns the parcel next to yours, how liens get discovered, how the entire system of marketable title works.

They are also, from the criminal's perspective, a searchable index of targets that includes the owner's name, the parcel description, the assessed value, whether there is a mortgage, and the mailing address for the tax bill. If that mailing address is in a different state from the parcel, the record has just told the criminal that the owner is remote. If there is no mortgage, it has told them there is no lender who will need to be paid off and no servicer likely to notice.

The perfect target is an unmortgaged vacant parcel whose owner lives far away. The public record identifies those in a single query.

The clock runs long

In most frauds, discovery time is short. A diverted paycheque is noticed on payday. A fraudulent card charge is noticed on the statement.

An absentee owner of vacant land might not visit for a year. Their tax bill still arrives, and if it is paid by direct debit it does not even prompt a glance. The first signal is often, quite literally, a neighbour or a construction crew. By the time the real owner knows, the money has moved through several accounts and probably several jurisdictions, the buyer has taken possession in good faith, and the legal position has calcified.

How does the fraud actually work?

People imagine something technically sophisticated. It is not. It is procedural, and understanding the sequence is what makes the missing control obvious.

Step one: selection

Search the public records for unmortgaged parcels with out of state owner addresses. This is a database query against records that are online in most jurisdictions. The criminal now has a name, a parcel, an assessed value, and confirmation that the owner is not nearby.

Step two: contact

Approach a listing agent by email or phone as the owner. The agent has no prior relationship with the real owner, so there is no baseline to deviate from. The story that explains everything else is straightforward: the seller is out of state and cannot travel.

Step three: shape the transaction

Push for the conditions that keep every human interaction remote and every timeline short. Cash buyers, because there is no lender to order an appraisal or ask questions. A price slightly under market, because that generates a fast offer and a motivated buyer who does not want to lose the deal. Remote closing, remote notarisation, and a wire to an out of state account.

Each of these individually is a normal feature of a legitimate transaction. Plenty of honest sellers are out of state, want a fast cash sale, and price to move. That is precisely why the red flag training does not work well: the indicators are all common, and agents who treat them as disqualifying will lose real business.

Step four: identity

Produce identification in the owner's name. This is the step everyone assumes is hard, and it is the step that has become dramatically easier. Forged documents are readily obtainable, and remote verification flows that rely on a photograph of a document plus a selfie face an attack class that has advanced considerably, as we covered in the piece on injection attacks against verification. The check is being performed on a stranger by people who have no independent knowledge of what the real owner looks like.

Step five: notarisation

Get the deed notarised. In the in person version, the notary observes the forged identification and the signature and certifies exactly what they observed, which is a person presenting a document.

Remote online notarisation deserves careful treatment here, because it is genuinely useful, it is not going away, and it should not be. It has made property transactions accessible to people who cannot easily travel, and the credential analysis and session recording built into good platforms are real controls that the in person process often lacks. But the identity assurance in a remote session rests substantially on a live video interaction, and a live video interaction is a weaker check in 2026 than it was in 2020. Vendors have added detection features in response, which is the same arms race documented across this series.

Step six: closing and recording

The transaction closes. Funds wire out. The deed is recorded. From this moment forward, the public record says the buyer owns the parcel, and every subsequent professional who touches this property will rely on that record, because relying on the record is what the system is for.

Step seven: the gap

Months pass. Then somebody drives out to the field.

Why don't current defences stop this?

Every existing control sits in one of two categories, and neither is positioned to prevent the loss.

Seller verification at closing checks a stranger

Title companies do verify sellers, with document checks, knowledge based questions, and increasingly a selfie match. The structural problem is not the rigour of the check but its position. The title company has never met the real owner and has no reference point for them. All it can do is assess whether the person in front of it is presenting a coherent identity, and a well prepared impersonator presents a very coherent identity. Knowledge based questions draw on data that has been breached repeatedly and is purchasable.

You cannot verify somebody against a record you do not hold. The one party who could reliably say "that is not me" is never in the room.

Recording alerts arrive after the fact

Many counties now offer free property alert services that notify an owner when a document is recorded against their parcel. These are genuinely good, they cost nothing, and every owner reading this should sign up today.

They are also, unavoidably, notification rather than prevention. The alert fires after recording. At that point the sale has closed, the money has moved, and a good faith buyer has an interest in the property. Your alert has converted an unknown catastrophe into a known catastrophe, which is a real improvement and is not the same as stopping it.

Title lock subscriptions notify, they do not block

A commercial category has grown up selling monitoring subscriptions, and the marketing frequently implies something stronger than what is delivered. These services watch public records and tell you when something is filed. They have no authority over the recorder and cannot prevent a filing. Several consumer protection bodies and county officials have pointed out that the underlying alerting is often available free from the county.

There is also an uncomfortable incentive worth naming plainly: a business model built on alerting has no particular reason to want prevention.

And the recorder is not permitted to be the check

This is the part outsiders find hardest to believe. County recorders generally cannot refuse a document that is facially valid and properly presented. They are not adjudicators of authenticity, and the recording acts do not give them that role. This is deliberate, and it exists for a good reason: a recorder empowered to decide which deeds are genuine would be a single point of arbitrary failure in the property system.

So the one institution with a complete view of the parcel is legally structured to be a filing cabinet.

What would actually stop it?

Everything above fails for the same reason. The system checks documents presented by whoever shows up, and it has no way to hear from the person whose property is actually being conveyed.

So give the owner a way to speak.

The proposal, and it is a proposal rather than a shipped product, is a parcel key. An owner enrolls once, binding a key they control to a specific parcel, registered with the title plant, the underwriter, or eventually the recorder. From then on, a conveyance of that parcel requires a fresh signature from that key over the specific instrument.

The analogy is the account you already have this on. A bank does not verify a large transfer by assessing whether the person requesting it seems like you. It requires a credential only you hold. Property is the largest asset most people own and it is conveyed on a weaker check than a two hundred pound bank transfer.

What is signed

conveyance_request = {
  "parcel_id":          "APN 4271-018-005",
  "jurisdiction":       "county_of_record",
  "instrument_type":    "grant_deed",
  "instrument_sha256":  "b71f4a02...",   // hash of the exact deed document
  "grantor_key":        "pk_owner_9c3e...", // registered against this parcel
  "grantee_name":       "...",
  "consideration":      "USD 240000",
  "requested_at":       "2026-09-13T14:22:08Z"
}

signature = Ed25519_sign(owner_private_key, canonical_json(conveyance_request))

Two properties matter. The signature covers a hash of the exact document, so it authorises this deed rather than deeds in general, and a substituted instrument fails verification. And it verifies against a published key without contacting anyone, which means a title examiner can check it in five years when the file surfaces in a dispute.

def check_conveyance(instrument_bytes, request, signature, registered_key):
    if sha256(instrument_bytes) != request["instrument_sha256"]:
        return "INSTRUMENT_SUBSTITUTED"
    if request["grantor_key"] != registered_key(request["parcel_id"]):
        return "NOT_THE_REGISTERED_OWNER"     # the impersonator stops here
    if not ed25519_verify(registered_key, signature, canonical_json(request)):
        return "SIGNATURE_INVALID"
    return "OK"

Run the fraud against this. Selection still works, because records stay public. Contact still works. The transaction can still be shaped. The forged identification still passes the title company's check, because that check is unchanged. The notary still observes a signature.

And then the conveyance requires a signature from a key registered years earlier by a woman who has never heard of any of these people, and the transaction stops. Not because anybody detected anything, but because a required input does not exist and cannot be manufactured.

How do the defences compare?

The objections a title professional will raise

Anyone who has worked in this industry is already composing a list. The objections are good ones and they deserve a real answer rather than a slogan.

"There are thousands of recording jurisdictions and they do not agree on anything"

Correct, and this is the largest practical obstacle. County recording in the United States is administered locally across thousands of offices with different systems, different statutes, different budgets, and different levels of digitisation. Any plan that begins with all of them adopting anything is not a plan.

This is why the near term path does not run through recorders. It runs through underwriters, of which there are few, who bear the loss, and who already impose requirements on their agents. An underwriter can require a parcel key signature for a defined risk category as a condition of issuing a policy, tomorrow, without a single statute changing. Recorder integration is a policy conversation for later, and the Property Records Industry Association is the venue where that conversation would happen.

"Owners will never enroll"

Most will not, and a design that requires universal enrollment fails. But universal enrollment is not the objective, because the fraud is not uniformly distributed. It concentrates heavily in one category. Enrolling a meaningful share of unmortgaged vacant parcels in the counties where this is active is a much smaller problem than enrolling every property owner in the country.

The natural enrollment moment is purchase, when the owner is already present, already verified, already signing a stack of documents, and already paying for title services. Add one step to that stack and the enrolled population grows with every transaction without anyone running a campaign.

"What about trusts, estates, LLCs and co-owners?"

This is the objection that most deserves respect, because a great deal of vacant land is held exactly this way, and a design that only handles a single natural person owner would miss much of the target population.

Entity ownership needs authorised signer keys with a defined update path when authority changes, which is a governance problem the corporate world already solves imperfectly. Estates are harder: the owner has died, authority passes to an executor, and the transition must be possible without becoming the fraud path itself. Co-ownership needs multiple required signatures, and threshold signing of that kind is not something Manav ships today. These are real gaps and should be stated as such rather than hand waved.

"What happens when somebody loses their key?"

This is the objection that could sink the whole idea, and it is the one to take most seriously, because a recovery path is just a second front door.

If recovery means "present identification to a title company", then the fraud simply moves there and the parcel key has achieved nothing. The recovery path has to be harder than the fraud it prevents, and it has to work for an eighty year old who has changed phones twice since enrolling.

The answer is continuity rather than re-proofing from documents: a prior enrolled device, a liveness check that binds to the original enrollment without storing a face, a waiting period with notification to the address of record, and for high value parcels a human attestation step. We wrote about why recovery is where strong authentication usually gets defeated, and property is the case where that lesson matters most. Any implementation that treats recovery as an afterthought will reproduce the problem it set out to solve.

"Underwriters might prefer to price the risk"

Possibly, and this is a legitimate commercial question rather than a technical one. If seller impersonation losses stay small relative to premium, the rational response is to price it and move on. Prevention becomes compelling when losses grow, when reinsurers ask, or when a state regulator starts asking why a known fraud pattern with a known control is still producing claims.

What to do this week

If you own vacant land, a second home, or inherited property:

1. Sign up for your county recorder's property alert service. It is usually free, it takes about ten minutes, and while it is notification rather than prevention, knowing in a week instead of a year materially changes your legal position.
2. Do not buy a title lock subscription before checking what your county offers free. Compare what you would actually receive against the county service. Frequently it is the same alert.
3. Put a physical eye on the parcel more often, or ask somebody to. A neighbour, the adjacent farmer, or a local agent. Discovery time is the variable that most affects your outcome, and it is the one you can change cheaply.
4. Check that your address of record with the assessor is current. Notices go there. If it is a house you moved out of in 2014, you have removed your own warning system.

If you are an underwriter, title agent, or recorder:

5. Pull your seller impersonation claims and segment them by property type and owner distance. If your book matches the survey findings, your exposure is concentrated in a category small enough to address specifically.
6. Write down what your seller verification actually proves. Not what the vendor's brochure says, but the specific claim you could defend in a deposition. For most workflows the honest answer is that a person presented a coherent identity, which is not the same as the owner authorising the sale.
7. Pilot owner enrollment at purchase for one risk category. Vacant land with an out of area owner is the obvious first band. Enrollment at closing costs one extra step in a process that already has forty.
8. Design the recovery path before the enrollment path. If you cannot describe a recovery flow that is harder to abuse than forging an identity document, you are not ready to deploy the enrollment flow.

The recorder as a verifier

Property records are one of the oldest pieces of trust infrastructure in continuous operation. The system works remarkably well, and it works because the record is authoritative and because everybody can rely on it. That reliance is the asset.

The weakness is that the record has never had a way to hear from the owner. It records what is presented. For centuries the practical protection was friction: you had to physically appear, in the county, with documents, in front of people who might know you. Distance, digitisation and remote closings removed that friction, for reasons that were mostly good, and nothing replaced the assurance the friction was quietly providing.

A parcel key is not a new institution or a replacement for the recorder. It is a way for the owner to be heard by a system that currently has no channel to them, using the same mechanism your bank already uses for a wire transfer. The technical parts are unremarkable: enrollment, a key on a device, a signature over a document hash, an offline verifiable receipt. Manav ships those pieces, and the documentation covers the payload and verification path, with a working flow in the signing demo.

The hard parts are not cryptographic. They are the estate, the trust, the lost phone, the eighty year old, the three thousand recorders, and the underwriter's decision about whether preventing a loss beats pricing it. Those are the parts worth arguing about, and they are the parts that will determine whether the woman standing in the field ever gets a better answer than a lawyer's retainer.

Frequently asked questions

How do you prevent someone from selling your property without you? Today, the best available step is registering for your county recorder's free property alert service, which tells you when a document is filed. That is notification, not prevention. The control that would actually stop a sale is an owner registered key bound to your parcel, so that a conveyance requires a fresh signature only you can produce.

What is seller impersonation fraud? A criminal identifies a property, usually vacant land with an absentee owner, poses as the owner to a listing agent, produces forged identification, executes and notarises a deed, sells to an innocent buyer, and wires the proceeds away. The real owner typically discovers it months later, often when construction starts.

Do title lock services actually stop deed fraud? No. They monitor public records and notify you when something is filed, which happens after recording. They have no authority over the recorder and cannot block a filing. Most counties offer equivalent alerting free, so check yours before subscribing to a paid service that resells it.

Why is vacant land targeted so much more than occupied homes? Because occupancy is a free and continuous identity check. Somebody lives in a house and will be encountered during a sale. Vacant land has nobody on it, the ownership records that identify absentee owners are public, and discovery can take a year, by which time the money and the legal position have both moved.

Can the county recorder refuse to record a fraudulent deed? Generally no. Recorders accept documents that are facially valid and properly presented, and they are not empowered to adjudicate authenticity. That limitation is deliberate, since a recorder with discretion over which deeds are genuine would be an arbitrary single point of failure in the property system.

Does title insurance protect the real owner? It protects the party who bought the policy, which in a fraudulent sale is the innocent buyer. The defrauded owner is frequently not covered in that transaction and ends up bringing a quiet title action at their own expense, which is why prevention matters far more here than in frauds where an insurer simply reimburses.

What happens if the owner loses the key? This is the hardest design question, because a weak recovery path becomes the new attack. Recovery has to rest on continuity rather than re-presenting documents: a prior enrolled device, a liveness check tied to the original enrollment, a waiting period with notice to the address of record, and human attestation for high value parcels.

Sources

1. National Association of Realtors, 2025 Deed and Title Fraud Survey and related research: nar.realtor
2. American Land Title Association, seller impersonation fraud research and member guidance: alta.org
3. FBI Internet Crime Complaint Center, annual reports including real estate fraud loss reporting: ic3.gov
4. Federal Bureau of Investigation, field office public service announcements on deed and quitclaim fraud: fbi.gov
5. Property Records Industry Association, eRecording standards and recorder practice: pria.us
6. CertifID, wire fraud and real estate fraud research: certifid.com

Property is the largest asset most people own, and it is conveyed on a weaker identity check than a small bank transfer.