Fraud Theory Explained: The Fraud Triangle, the Diamond, the Scale

Fraud theory is the body of explanation that tries to answer a question audit procedures cannot: not how a fraud was concealed, but why an otherwise law-abiding person committed it. The models are used to design controls, to assess risk at the planning stage of an audit, and to decide where in an organisation the exposure actually sits.

Four models do most of the work, and they are not competing accounts of the same thing. Three explain the individual offender; the fourth explains the environment that produces unusual numbers of them. Treating them as rivals is the commonest mistake, because the interesting cases need both levels.

The fraud triangle

The original model came from a study of imprisoned embezzlers and asked what they had in common. The answer was three conditions, all present together.

Pressure. A financial problem the person regards as non-shareable — one they cannot admit to the people whose opinion matters. The problem need not be poverty. Gambling losses, a business failure concealed from a spouse, an addiction, or the cost of maintaining a standard of living are all classic, precisely because they cannot be disclosed. In a corporate setting the equivalent is an earnings target or a covenant that will be breached.

Opportunity. A perceived chance to resolve the problem in a way the offender believes will not be detected. This is the only limb an organisation controls directly, which is why it receives almost all the practical attention.

Rationalisation. A way of describing the act that leaves the person's self-image intact. Borrowing, not stealing, taking what is owed, keeping the company alive through a bad quarter, everybody does it. The rationalisation precedes the act; it is part of the decision, not an excuse constructed afterwards.

The model's strength is that it explains why fraud is committed by people with no criminal history. Its weakness is that pressure and rationalisation are internal states an organisation cannot observe, and that it says nothing about capability.

The fraud diamond

The diamond adds a fourth element: capability.

The argument is that the three original conditions can all be present without a fraud occurring, because the person lacks the position, the technical knowledge, the confidence or the capacity to manage the deception over time. Opportunity describes a gap in the system; capability describes whether this particular person can exploit it.

The addition has practical consequences. It directs attention to the individuals whose role, seniority and technical understanding make them capable of both committing and concealing a material fraud, which is a much smaller group than the population with theoretical access. It also explains why long-running frauds are so often committed by people with a reputation for competence: sustaining a deception across reporting periods requires the ability to answer questions convincingly, and that is a skill.

It is also why the traditional control of enforced holiday works. A fraud requiring continuous management by one capable person fails when that person is absent for two consecutive weeks.

The fraud scale

The scale replaces rationalisation with personal integrity, and treats the three factors as continuous, not binary.

The claim is that pressure, opportunity and integrity trade off against each other. High pressure and weak controls may still produce no fraud where personal integrity is high; low pressure and strong controls may still produce fraud where integrity is very low. Fraud becomes likely when the combination crosses a threshold, not when three boxes are ticked.

The practical advantage is that integrity is observable in a way rationalisation is not. It shows in how a person deals with small ambiguities: expense claims, minor rule-bending, how they describe previous employers. It is also the limb that connects individual explanation to organisational culture, because integrity is influenced by what an organisation visibly tolerates.

The scale is generally regarded as a better fit for financial statement fraud, where the pressure is an external performance expectation rather than a private financial problem, and the offender frequently gains nothing personally in cash.

Anomie and the American Dream

The fourth model is sociological rather than psychological, and it operates at the level of the society or the organisation instead of the person.

The argument begins from a mismatch. Where a culture places overwhelming emphasis on a single measure of success — monetary attainment — and simultaneously fails to make legitimate means of reaching it available to everyone, it generates pressure towards whatever means are available. Deviance is then not a failure of the culture but a predictable product of it.

Applied to organisations, the model explains a pattern the individual theories struggle with: concentrations of fraud in particular firms, functions and periods. Where the only respected measure is the number, where the means of hitting it are unequally distributed between units, and where the institutions that would restrain the pursuit are weak relative to the reward system, misconduct becomes normal, not exceptional. Nobody in such an organisation needs an unusual pressure or an unusual rationalisation, because the environment supplies both as standard.

This is also where the model earns its practical value. The individual theories suggest screening, segregation of duties and supervision. The cultural theory suggests examining the incentive structure itself, and asking what the organisation actually rewards as opposed to what its values statement says. An audit that reviews controls without reading the bonus plan has looked at only one half of the exposure.

What the models share and where they diverge

All four accept that fraud is a decision and not a compulsion, that it is usually committed by people without prior convictions, and that the offender's own account of the act will not be theft.

They diverge on where the leverage lies. The triangle and the diamond point at the individual and the control environment. The scale points at selection and tone. The anomie explanation points at the reward system and at the relative strength of the institutions that are supposed to restrain it.

The divergence matters because it produces different recommendations from the same facts. A fraud discovered in a sales division can be read as one dishonest manager exploiting a control gap, or as the predictable result of a commission scheme that made the target unreachable by legitimate means. The first reading produces a dismissal and a new authorisation limit. The second produces a redesigned incentive. Organisations overwhelmingly prefer the first, which is one reason the same frauds recur.

How the theory is used in practice

Three applications recur, and each maps onto a different model.

Risk assessment. Auditing standards require the engagement team to discuss the susceptibility of the financial statements to material misstatement due to fraud, and to presume a risk in revenue recognition and in management override of controls. Both presumptions are triangle reasoning: revenue is where the pressure concentrates, and management override is where opportunity and capability coincide.

Control design. Segregation of duties, authorisation limits, mandatory leave, job rotation and independent reconciliation all attack opportunity, because it is the only limb an employer can reduce directly. Capability reasoning refines this by concentrating the strongest controls on the roles that could actually sustain a concealment.

Detection and interview. Rationalisation is the most useful limb in an interview, because an offender who will not admit theft will frequently explain a borrowing, a temporary transfer, or a correction they intended to reverse. The admission of the mechanism arrives before any admission of dishonesty.

Limitations worth stating

The models are frameworks for structuring an explanation; none of them predicts individual behaviour.

They are also derived largely from cases that were detected, which is a biased sample by construction. Frauds that were never found are, by definition, absent from the studies that produced the theory, and there is no reason to assume they share the characteristics of the ones that failed.

A further limitation is that the models describe the decision to begin a fraud far better than they describe its continuation. Most material losses come not from a single act but from a small initial taking that had to be concealed, which required a second act, and then a third. At that point the original pressure has been replaced by the pressure of the concealment itself, and the offender is no longer solving the problem that started it. That is why the amounts in long-running cases grow in a pattern nobody planned, and why an offender's account of the first act frequently sounds proportionate while the total does not.

And two of the four limbs of the original model are unobservable from outside. A risk assessment that claims to have evaluated pressure and rationalisation across a workforce has claimed more than the method supports. What can be evaluated is opportunity, capability, the incentive structure and the observable tolerance of small breaches, which is why practical work concentrates there.

How it is examined

The typical question describes a fraud and asks for an explanation using the theory, sometimes with a recommendation.

Identify the limbs on the facts explicitly rather than reciting the model: what the non-shareable pressure was, what specific control gap supplied the opportunity, what the offender's own account of the act would have been, and what position or knowledge made them capable of sustaining it. A script that names the three elements without attaching them to the facts scores poorly.

Then take the second step that separates a good answer. Ask whether the case is better explained at the individual or the organisational level, and say why. Where the pressure was a reporting target, where several people were implicated, or where the conduct persisted across periods, the cultural explanation carries more of the weight and the recommendation should address the incentive structure rather than only the controls.