A glossary of private lending fraud terms.

Definitions for every term used in the Private Lending Fraud Index, in my insights, and in on-record quotes. Written for reporters, operators, and new analysts. Free to cite with attribution.

Draw fraud

Draw request

A borrower's request to release a portion of a construction or rehab loan, supported by photos, invoices, and a stated percent-complete against the approved scope of work.

Draw decisioning

The process of evaluating a draw request and issuing a pass, review, or fail verdict before funds are released. Combines scope verification, photo evidence, invoice review, and inspection signals.

Recycled photo

Collateral imagery a borrower has submitted before — on the same loan, a prior loan, or pulled from a public listing — repurposed to support a new draw or progress claim.

Inflated scope of work

A budget that overstates the cost or quantity of work to be performed, used to enlarge loan proceeds or front-load draws against work that has not yet occurred.

Phantom progress

A claim of completed work that did not happen, supported by staged photos, recycled photos, or fabricated inspector sign-off.

Front-loaded draw

A draw request that pulls a disproportionate share of loan proceeds early in the project, often masking thin equity or covering unrelated cash needs.

Document fraud

Doctored invoice

An invoice with edited line items, totals, dates, or vendor information — typically produced in a PDF editor to support a larger draw than the underlying work justifies.

Synthetic bank statement

A bank statement assembled or generated — increasingly by AI tools — that mimics a legitimate institution's format but reflects transactions that did not occur.

PDF tampering

Modification of a PDF after issuance, detectable through object-stream analysis, font substitution, incremental update history, or mismatched producer metadata.

Cross-document inconsistency

A contradiction between two documents in the same loan file — for example, a payoff letter whose balance does not reconcile to the title commitment or borrower's bank statements.

AI-generated document

A statement, letter, or form produced by a generative model rather than the named issuer. Detection relies on glyph, layout, and metadata signals plus issuer verification.

Identity fraud

Synthetic identity

An identity composed of real and fabricated elements — often a real SSN paired with a fabricated name and history — used to obtain credit without a single real person behind the file.

Straw borrower

A nominal borrower whose name appears on the loan but who does not control the property, the funds, or the project. Frequently used to obscure the real principal.

Entity verification

Confirming that a borrowing LLC, corporation, or trust exists in good standing, that its registered agents are real, and that the signer has authority to bind the entity.

Beneficial owner

The natural person who ultimately owns or controls a borrowing entity. Required disclosure under FinCEN's Corporate Transparency Act for most private lending borrowers.

Device intelligence

Signals derived from the device used to submit an application, draw, or capture — including device fingerprint, emulator detection, and reuse across unrelated loan files.

Photo & video

EXIF metadata

Metadata embedded in a photo by the capturing device: timestamp, GPS coordinates, camera model, and edit history. Frequently stripped, edited, or fabricated in fraudulent submissions.

Geolocation validation

Confirming that a submitted photo or video was captured at the subject property, using a combination of EXIF GPS, network signals, and reverse-image scene matching.

Recapture detection

Detecting that an image was photographed from a screen or printout rather than captured live — typically through moiré patterns, screen reflections, and color-gamut shifts.

Liveness capture

A capture flow that proves media was produced in real time by a present human and device, rather than uploaded from a camera roll, AI tool, or third party.

Scene match

A comparison between a submitted photo and prior photos of the same property to confirm or deny that the location matches and progress is consistent.

Deepfake collateral

Generative-AI imagery or video depicting property, construction progress, or inspections that never occurred, presented as authentic evidence to a lender.

Underwriting

Business-purpose loan

A loan whose proceeds are used primarily for a business purpose, including most private-lender rehab, ground-up, and bridge loans against 1–4 unit and small commercial real estate.

Loan-to-cost (LTC)

The loan amount divided by total project cost, including acquisition and budgeted rehab. A key sizing constraint for construction and rehab lending.

After-repair value (ARV)

The expected market value of a property once the budgeted scope of work is complete. ARV inflation is a common driver of overstated loan proceeds.

Loan-to-ARV

The loan amount divided by after-repair value. Used to size construction and rehab loans against expected stabilized value rather than current as-is value.

Pass, review, or fail

A three-state verdict applied to a draw, document, or borrower submission. Pass releases the workflow; review routes to a human; fail blocks the action and logs the reason.

Confidence score

A numeric or banded estimate of how strongly evidence supports a verdict. Reported alongside every automated decision so reviewers can calibrate human attention.

Operations

Audit packet

A bundled, time-stamped record of every signal, document, photo, and decision tied to a loan event — produced so a third party can reconstruct what the lender knew and when.

Audit trail

An immutable sequence of events — captures, verdicts, overrides, releases — attached to a loan, borrower, or draw, used for examiner review and post-loss investigation.

LOS integration

A connection between a verification platform and a lender's loan origination system, typically via API or webhook, so verdicts and evidence flow back into the system of record.

Loss event

A confirmed instance where a lender suffered or would have suffered loss due to fraud, used to calibrate fraud-rate denominators in this index.

Fraud consortium

A multi-lender data network that shares anonymized signals about repeat bad actors, recycled collateral, and known fraud patterns, while preserving each lender's confidentiality.

Frequently asked questions

What is draw fraud in private real estate lending?

Draw fraud is any attempt to obtain construction or rehab loan proceeds for work that has not been performed, has been overstated, or is supported by fabricated, recycled, or staged evidence. The most common forms are recycled photos, doctored invoices, phantom progress claims, and front-loaded draws against thin actual work.

How is a synthetic identity different from a stolen identity?

A stolen identity reuses a real person's full credentials without their consent. A synthetic identity stitches together real and fabricated elements — often a real SSN paired with a fabricated name, address, and credit history — so there is no single real person behind the file. Synthetics are harder to detect because no victim ever reports the fraud.

What does a 'pass, review, or fail' verdict mean?

It is a three-state outcome applied to a draw, document, or borrower submission. Pass releases the workflow automatically, review routes the item to a human with the supporting evidence, and fail blocks the action and logs the reason. Every verdict is paired with a confidence score and an audit trail so reviewers can calibrate attention.

Why does EXIF metadata matter for photo verification?

EXIF metadata records when, where, and on what device a photo was captured. In private lending, EXIF is one of the first signals checked to confirm a photo was taken at the subject property at the time the borrower claims. EXIF is also frequently stripped, edited, or fabricated, which itself is a strong fraud signal.

Can I cite these definitions in reporting or training materials?

Yes. The glossary is free to cite and free to reproduce with attribution to 'Glossary of Private Lending Fraud Terms, Zack Farris, zackfarris.com/glossary'. For interviews, on-record quotes, or custom definitions, email press@zackfarris.com.