Deceased identity fraud: how it works and where to stop it

Why a dead person is the perfect applicant

Fraud controls are built to find things that look wrong. A name that does not match an address. A date of birth invented to fit a credit thin file. A Social Security number that was never issued. A deceased identity has none of those problems. The name is real, the date of birth is real, the number was issued properly decades ago, and there is a long, well behaved credit history sitting behind it.

Most of what fraud teams call identity verification is really consistency checking. Does this bundle of attributes hang together, and has it been seen before in a way that makes sense? A stolen deceased identity answers yes to both. It was a real person with a real footprint, right up until the moment they died.

The industry calls this ghosting. It sits somewhere between straight identity theft and synthetic identity fraud. Unlike synthetic fraud, there is no fabricated element to detect. Unlike ordinary identity theft, there is nobody to notice a strange letter, check a statement, or call your fraud line. The victim cannot complain, and the family usually finds out months later, if at all.

Where the identities come from

Obituaries are public. Funeral home notices, memorial pages and local news all publish full names, ages, home towns and surviving family. Combine that with a data broker record or an old breach dump and an operator has enough to reconstruct a usable identity within minutes.

The second source is closer to home. Household members, carers and people with access to mail or filing cabinets already hold the missing pieces. A meaningful share of deceased identity abuse is committed by someone who knew the person, which is one reason it is rarely reported.

The third source is professional. Identity packages are traded, and a recently deceased identity carries a premium because the buyer knows there is a window before it burns.

The exposure window, and what actually closes it

The window opens on the date of death. It closes when your decisioning systems can see that the person has died. Everything in between is unpriced risk sitting inside a portfolio that looks healthy.

That window is not one length. It is a chain, and each link adds days. A death is certified locally. It is registered with a state vital records office, often through an electronic registration system with varying levels of adoption. The state then reports onward. Downstream aggregators pick it up, normalise it, and publish it into a commercial file. Only then does it reach a screening decision at your end.

This is why the useful question to ask a data provider is not how many records they hold. It is how many days pass between a death and that record being available to you, measured from the date of death rather than the date they loaded the file.

• Ask for time to availability measured from date of death, not from ingestion date
• Ask what proportion of records arrive within the first fourteen days
• Ask how state level coverage varies, because a nationwide claim can hide a weak state
• Ask what the provider does when a death is later corrected or rescinded

Four places to put the check

There is no single control point. The right answer is usually two or three of the following, sequenced so that the cheapest check happens first.

Origination is where you prevent loss rather than chase it. A deceased check at application, before approval, converts a future charge off into a declined application. This is the highest value placement and the one most institutions add first.

Portfolio monitoring is where you find what is already inside. A periodic sweep of the existing book against fresh death data surfaces accounts that were opened legitimately and later taken over, plus accounts where the customer died and activity continued.

Servicing and payment events are the third point. A change of address, a new device, a payout request or an unusual withdrawal on an account belonging to someone who has died is a much stronger signal than the same event on a live account.

Outreach suppression is the fourth. Before any collections, marketing or retention contact leaves your building, deceased records should be removed. This one is about complaint volume and reputational damage more than fraud loss, and it is the easiest to justify internally.

What a good match looks like

A deceased match is a high consequence decision. Acting on a false positive means freezing a living customer's account, which is a service failure and, in some products, a regulatory one. Acting on a false negative means paying a fraudster.

That means match quality has to be transparent. You need to see which elements agreed, how strongly, and what the confidence looks like, rather than being handed a yes or no. A match on name and date of birth alone is not the same as a match on name, date of birth and last known address.

You also need date of death, not just a deceased flag. Date of death tells you whether an account was opened before or after the person died, and that single fact separates a legitimate estate matter from an outright fraud case.

• Confidence score on every match, not a binary flag
• Date of death returned so you can time bound the activity
• Last known address to support a second point of agreement
• Clear behavior when only a partial identity is submitted

Making the business case

The case is easier than most fraud investments because the cost model is simple. If you pay only for confirmed matches, the cost of running the check across a population is bounded by the number of deceased people actually in it, which is a small fraction of any healthy book.

Frame the benefit in three buckets. Avoided credit loss on applications you would have approved. Recovered exposure on accounts already open. Removed complaint and reputational cost from contacting families of people who have died.

The third bucket is often the one that gets the project funded, because it is the one your executive team hears about directly.

Common questions