home  /  proving the scheme
department of proving the scheme

undefined

The schemes that are hardest to prove are the ones the victim’s accounting records cannot see. Proving those is a different exercise from reading a ledger, and it ends in a courtroom where the constraint is not what the analysis found but what the witness may say about it.

begin here

What did you find, and how did you find it?

Start a conversation with the Forensic Concierge, already scoped to proving the scheme. Select a subject area to prompt it, or describe the matter directly.

Forensic Conciergeproving the scheme · orientation, not a finding of fraud
Tell me roughly what is alleged, where the money is said to have moved, and what records exist. I will help you see which scheme family it belongs to, what evidence trail that family leaves, and where proof usually breaks down. I will not tell you whether fraud occurred, whether a particular person did it, or what your claim is worth.

Skimming takes cash before it is recorded, so no entry exists to find and no reconciliation will surface it — which is why a balanced ledger answers nothing, and why proof has to be assembled from operational data and from the other side of the transaction. Cash larceny takes the same money after recording and leaves a difference anyone can see. Corruption is the extreme case: in a kickback or bid-rigging arrangement the money often never touches the victim’s accounts at all, which is why the Justice Department’s own list of what proves collusion is almost entirely non-accounting. Financial statement fraud is the inverse — the trail sits entirely inside the books, in entries, reserves and period-end adjustments. Outside the books, inside them, or nowhere at all: that single question sets the record set, the method and the proof strategy for everything that follows.

specialization areas

The three problems between an accusation and a factfinder

What the scheme was, what the records can honestly be made to show, and what the witness is permitted to say about it — in that order, because each one constrains the next.

methodology

How this department investigates.

The work that turns an allegation into something a court can receive.

Scheme classificationPlacing the allegation on the taxonomy first, because the branch decides which records could possibly show it.
Off-book reconstructionWhere no entry exists, building from independent operational data and counterparty records rather than the ledger.
Ledger and adjustment analysisGeneral ledger, subledgers and ERP data, and the adjustments that never became formal journal entries.
Indirect methods of proofNet worth, expenditures and bank deposits, used where the direct record is missing, inadequate or withheld.
Summary constructionSchedules that tie line by line to produced source documents rather than to the analyst’s working papers.
Opinion scopingConclusions drafted in the vocabulary of conduct rather than the vocabulary of the statute.
common questions

Proving the scheme — common questions

If the expert cannot say it was fraud, what exactly is being proved?

Conduct, mechanics and elements — which is nearly everything short of the verdict. AICPA Statement on Standards for Forensic Services No. 1 ¶10 reserves the ultimate conclusion of fraud to the trier of fact and prohibits a member performing forensic services from opining on it, while expressly permitting ‘expert opinions relating to whether evidence is consistent with certain elements of fraud or other laws based on objective evaluation’. The ACFE Code §V.B.2 draws the same line for certified fraud examiners: ‘No opinion shall be expressed regarding the legal guilt or innocence of any person or party.’ The ACFE’s own guidance then makes the scope clear — an examiner may conclude that a person misappropriated cash, misrepresented a transaction or concealed funds, and may conclude that each element of a fraud statute is satisfied. The finding survives; only the label is withheld.

Why does the type of scheme change the investigation so much?

Because the scheme determines which records could ever have recorded it. An on-book scheme leaves a discrepancy inside the accounting system, so the work is reconciliation, journal entry testing and source-document examination. An off-book scheme leaves nothing, so the work moves to independent operational data and to the counterparty. A corruption scheme frequently leaves nothing in the victim entity at all, so the record set becomes bid documents, travel and expense reports, telephone records and diaries. Choosing the wrong record set is the most expensive error available early in a matter, because preservation notices and third-party subpoenas are aimed at the wrong custodians while the window is open.

Our books reconcile. Does that not answer the allegation?

Not against an off-book allegation, and this is the most common misreading in the field. The ACFE defines skimming as a scheme in which an incoming payment is stolen before it is recorded on the organization’s books and records. Money that was never recorded produces no entry to flag and no difference to reconcile, so the accounting system will balance perfectly whether or not the scheme ran. Cash larceny — the same theft after recording — does leave a reconciling difference. In Occupational Fraud 2024 the two appeared in 10% of cases each, with medians of $43,000 and $50,000: close in the numbers and completely different in provability, which is why the first question in a cash matter is when the money was taken relative to when it was recorded.

Where does this stop and damages begin?

By question, not by profession. Forensic accountants routinely quantify losses — the AICPA’s CFF and NACVA’s MAFF bodies of knowledge both include damages and lost profits work, and the same practitioner very often does both halves. The division here is editorial and subject-matter: this Institute covers operations on the historical record, including what moved and when. Our Economic Damages Institute covers measurement against a but-for state — lost profits, apportionment, present value, event studies. FRE 703 is what lets one matter support both: the damages expert relies on the reconstructed factual record as the predicate for a model, provided experts in the field would reasonably rely on that kind of data.

What is the most common way this evidence fails in court?

Not usually a Daubert loss. Three quieter routes account for most of it. The first is a disclosure defect: a Rule 26(a)(2)(B) report that states conclusions without the how and why, met with self-executing exclusion under Rule 37(c)(1). The second is the lay-versus-expert trap — a tracing witness offered as a lay summary witness who, on the Tenth Circuit’s reasoning in United States v. Joseph, 108 F.4th 1273 (10th Cir. 2024), had to choose among competing methodologies and therefore should have been disclosed as an expert. The third is phrasing: an opinion delivered in the statute’s own vocabulary, which is what the Second Circuit set aside in United States v. Scop, 846 F.2d 135 (2d Cir. 1988).

Work out which records could show it before you ask for any.

Describe the allegation and what has been preserved. The Institute will help you see which record set the case will actually be built from.

forensic conciergeorientation · not a finding of fraud
Tell me roughly what is alleged, where the money is said to have moved, and what records exist. I will help you see which scheme family it belongs to, what evidence trail that family leaves, and where proof usually breaks down. I will not tell you whether fraud occurred, whether a particular person did it, or what your claim is worth.