Three bids arrive at an evaluation session, the technical scoring has been completed, and the winner is typically separated from the runner-up by a narrow margin; nothing about the session itself is irregular. Read across the last three award cycles in the same category, however, the same supplier finishes ahead each time by a similarly narrow margin. Alongside this, certain clauses of the technical specification — a particular connection standard, a particular warranty term, a particular reference-project size — align with one manufacturer's catalogue characteristics at an unusual level of precision. The team that drafted those clauses did so, in most cases, not out of bad faith but by copying the prior year's specification, which had itself descended from the year before. The process thereby settles into a closed loop that no single person established through a single decision, and that everyone sustains through small omissions.
A second observation surface is the distribution of order values. Wherever the threshold sits at which approval authority passes to a committee, order values tend to cluster just beneath it; three separate orders placed with the same supplier within the same month, if consolidated, would have reached the size requiring committee review. A third surface is the change order. The work was awarded at the lowest bid, yet within six months of contract signature, additional line items, scope clarifications, and site-condition revisions have carried the contract value materially above the original bid differential. A fourth surface is the frequency of the emergency-purchase exception, which tends to concentrate around specific suppliers and specific requesters. None of these four patterns constitutes a conclusion on its own; appearing together, they form a strong signal that the process is optimising relationships rather than price.
This pattern carries the name procurement fraud — the steering of a purchasing process so as to maximise the interest of the individual holding authority within it rather than the interest of the institution — and its mechanics operate not at a single moment but across six distinct nodes: specification drafting, supplier prequalification, bid evaluation, approval of post-award variations, delivery and acceptance inspection, and invoice authorisation. Each node legitimately contains discretion, since none can be fully mechanised; technical suitability requires interpretation, site conditions genuinely change, and an urgent requirement may genuinely be urgent. Manipulation arises not from the existence of discretion but from the same individual, or the same reporting line, holding authority over more than two of these nodes. Where authority converges at two nodes, a relationship forms; where it converges at three, scrutiny disappears altogether.
The mechanism remains incompletely described until one sees why the tendency is functional under certain conditions. Discretion exists precisely to reduce transaction cost: competitively tendering every line item costs more than the price advantage obtainable on low-value purchases; a long-standing supplier relationship generates advantages that appear nowhere in the contract, such as delivery flexibility and payment terms; the emergency exception keeps a production line from stopping. The difficulty lies not in the shortcut itself but in the shortcut persisting after the condition that legitimised it has dissolved. Once purchasing volume rises by an order of magnitude, once the supplier base broadens, or once a position goes unrotated for years, the mechanism that lowered transaction cost slowly becomes a rent channel — and no one observes the moment of conversion, because each step is only a modest extension of the one preceding it.
The behavioural layer enters here, moving the matter out of the register of morality and into the register of design. Reciprocity obligation — the disposition to return a small gesture — normalises gradually within supplier relationships: first a meal, later a trade-fair visit, later a training programme, each stage only one step removed from the last, so that at no stage does anyone experience the sensation of having crossed a threshold. To this is added the perceived indispensability accumulating around a person who remains in the same position for years: only that individual knows the supplier, the price history, the technical detail, and this informational monopoly raises the institutional cost of asking questions. Where the purchasing manager's performance is measured largely through savings ratios, the system directly rewards behaviour that suppresses the contract price and recovers it through variations.
The first layer of institutional cost is not, contrary to expectation, the amount diverted. Direct loss, once identified, generally appears one-off and bounded; the substantive cost accumulates as a unit-price differential spread across years. Each contract closing above market produces a small deviation in isolation, yet repeated across three years within the same category it surfaces as a quiet drift in gross margin — and because that drift carries no dedicated line in the financial statements, it is attributed to raw material prices, currency movement, or volume effects. Its balance-sheet correlate is usually hidden not in the inventory line itself but in the supplier concentration ratio and the count of single-sourced items; the fact that an overwhelming share of purchasing within a category flows to one supplier without a written framework agreement carries far more information than any accounting entry.
The second layer surfaces once the company enters a sale, partnership, or financing process. When the buy-side diligence team matches the supplier list against commercial registry and shareholding records and finds a related-party connection, or when it computes the ratio of change orders to contract value, the finding is priced not merely as a compliance issue but as a question about the repeatability of margin. The practical consequence follows a predictable sequence: an increased escrow percentage, a specific indemnity carved out for the identified exposure, an expanded representation and warranty package, renewal of supplier contracts made a condition precedent to closing, and deduction of the quantified margin deviation from normalised EBITDA. Because the last adjustment operates through the multiple, the aggregate effect can reach several times the direct loss.
The third layer surfaces in financed projects and is harder-edged. Where loan drawdowns are tied to progress certification and cost verification, inflated purchase values do more than raise project cost; they distort the cost-to-complete calculation and pull forward the equity contribution call, and covenant tests are sensitive to the timing of that call. Added to this is the effect on the insurance and security side: where prequalification has been steered in contractor selection, a gap opens between the coverage of the performance bond and the actual risk profile of the work, a gap that becomes visible only once delay or defect materialises. At that point the institution carries the direct cost and the loss of contractual protection simultaneously.
This tendency must be managed through institutional architecture rather than individual will, and the intervention separates into four components. The first is authority segregation: the roles that draft the specification, evaluate the bid, accept delivery, and approve the invoice are distributed across distinct reporting lines, and the technical suitability decision does not converge with the commercial decision under one signature. The second is threshold architecture: approval thresholds are not merely defined but accompanied by monthly monitoring of clustering beneath them, with split orders placed to the same supplier within a single period automatically consolidated for review. The third is control over supplier master data, whereby bank account changes, address changes, and new supplier registrations require approval independent of the requester. The fourth is rotation combined with mandatory uninterrupted leave; even in relationships that require continuity, allowing one person to manage a category without interruption does not preserve institutional memory but monopolises it.
BEIREK's intervention in this area begins not with drafting a compliance policy but with relocating the decision points within the purchasing process. The authority matrix is constructed by decision type rather than by expenditure line — specification drafting, prequalification, evaluation, change order, acceptance, and payment addressed separately — and technical specification drafting is removed from the line that manages the commercial relationship with the supplier and placed within project engineering. The decision record is maintained at the moment of recommendation rather than at the moment of approval: why a supplier is being proposed, why particular alternatives were eliminated, and which technical constraint generated which clause are all written before the outcome is known. Taken alone, this is the single most effective mechanism for making after-the-fact rationalisation structurally difficult.
On the operating-rhythm side, a monthly exception report is run whose subject is not the purchases made but the purchases departing from the rule — emergency exceptions, tenders closing with a single bid, orders clustering beneath thresholds, change orders exceeding a defined proportion of contract value, and items priced outside the standing schedule. In financed projects, independent cost verification is added: physical measurement preceding progress certification, direct-to-supplier payment structures for critical equipment, and alignment of the payment stream with the loan drawdown calendar. What these mechanisms share is that they function without accusing anyone and without assuming anything about intent; a deviation report is not an allegation but a ranking that indicates where a question ought to be asked.
An organisation's purchasing discipline rests not on an assumption about the honesty of its people but on a number — how many points within the process a single individual controls — and once that number exceeds two, whatever protection remains accumulates in personal choice rather than in written policy. The question that properly reaches an investment committee or a board is therefore not whether an irregularity has occurred in the past. It is this: were an irregularity to occur today, in which reported line, within how many months, and independently of whom would the institution be able to see it?