In an investment committee session, two entirely different files can arrive back to back: an automation upgrade on an existing production line, with a known customer, contracted demand and mature technology, followed by a proposal to develop a facility in a new geography where no offtake commitment yet exists. Both are presented in the same template, with the same rows populated, and both are judged by whether they clear or fail the same threshold. The rate itself is rarely discussed in the room; what is discussed is whether the project reaches it. Asked where the number comes from, the answer offered is usually institutional rather than analytical — this is the company's target return, and it has been for years.
This behaviour appears in almost every organisation that allocates capital, and its cause is not negligence. A single threshold simplifies internal debate, establishes what looks like a fair contest between competing business units, and removes the burden of conducting a separate risk negotiation for every file that comes forward. Yet in fixing the rate, an organisation quietly fixes what the rate measures as well: an average cost of capital, which by construction fits no project that is not itself average, and which becomes progressively less descriptive as the range of projects widens.
The mechanism has a name — discount-rate inconsistency — describing the separation between the risk borne by a cash flow and the rate used to bring that flow back to present value. In principle a discount rate prices the systematic risk specific to the flow being valued, meaning the question is not what return this company aspires to earn but what return an investor purchasing this particular stream would demand. A corporate hurdle rate answers the first question and is then deployed against the second. Where the projects in a portfolio carry comparable risk, the two answers coincide in practice; as the risk distribution widens, they diverge, and because the rate itself never moves, the divergence leaves no visible trace.
The inconsistency takes several recurring forms, each entering from a different direction. The first is a mismatch of inflation treatment, in which a nominal cash flow is discounted at a real rate or the reverse, so that inflation is counted twice or not at all — on long-lived assets that error alone can shift present value by an order of magnitude. The second is discounting levered free cash flow at a weighted average cost of capital, which allows the benefit of debt to enter both the numerator and the denominator and makes the project appear more attractive than its economics support. The third is currency mismatch, where a flow denominated in local currency is assessed against a rate reflecting the parent's funding currency. The fourth, and by some distance the most common, is applying one rate to a mature replacement investment and to an early-stage development option that has not yet entered permitting.
None of these tendencies begins as an error. A fixed threshold rate is a carrier of institutional discipline: it is a defensive line built against the pressure exerted by business unit heads to lower the bar for their own proposals, and it performs that function genuinely well. The difficulty is that the defensive line stays where it was placed after the conditions that justified its placement have changed. For a company operating in one line of business, in one geography, with broadly similar assets, one rate is a defensible simplification; as the portfolio diversifies and geographic risk decouples from technology risk, the same rate ceases to enforce discipline and begins to manufacture a systematic selection bias instead.
The institutional cost of that bias runs in two directions, and each enters the accounts through a different line. A single rate held high screens out low-risk, low-return projects — maintenance capital, efficiency improvements, capacity expansions underwritten by contracted demand — even though the predictability of their cash flows argues for evaluation at a materially lower rate; failing the threshold, they are declined. The trace left by those declined investments is invisible in the first year and surfaces several years later in the maintenance expense line, in unplanned downtime and in rework cost. Simultaneously, projects carrying high risk and high expected return clear the threshold comfortably and are approved, their modelled returns systematically overstated because they have been discounted at a rate containing no premium for the risk actually assumed.
The composite effect of the two errors is a portfolio drift that no one recorded as a decision: the risk profile of the company migrates upward through the mechanics of the allocation filter rather than through any strategic deliberation. That migration finds an echo on the financing side, where a lender observes a balance sheet in which the proportion of contracted cash flow is falling and the proportion of development-stage assets is rising, and calibrates covenants and security requirements accordingly. On the valuation side, the multiple prices an increase in risk that has never been articulated; even where the company appears to be meeting its stated return target, the volatility attaching to that return has risen, and the valuation gap does not close.
In a sale or capital-raising process, this structure surfaces at the diligence table with some speed. Reading investment committee decisions backwards, the reviewing party tends to ask not whether the projects were profitable but by what measure they were selected; and where the measure proves identical across every file, the finding concerns the capital allocation process rather than any individual project. A finding of that character is typically reflected not as a direct line item against price but as a condition precedent, a broadened representation and warranty package, or an earn-out structure tied to management continuity. Where a founder or long-tenured finance director has been adjusting the rate by instinct from project to project — as most mature organisations do informally — that knowledge sits in an individual's judgement rather than in the institutional record, and the buyer prices it as founder dependency.
The structural intervention lies less in identifying the correct rate than in building an architecture that records how the rate was chosen, and it has three separable components. The first replaces the single threshold with a matrix of rates keyed to risk class, so that replacement and efficiency capital, capacity investment underwritten by contracted demand, new-geography or new-technology investment, and early-stage development options are each assessed within their own band, with the spacing between bands reasoned and written down. The second is a consistency check applied to every file: whether the cash flow is nominal or real, levered or unlevered, denominated in which currency, stated pre- or post-tax, and whether the rate sits on the same plane as the flow on each of those questions. The third is an explicit choice about where risk is placed: diversifiable, project-specific risks are carried in cash-flow scenarios and probability weights rather than in a premium loaded onto the rate, because a premium buried in a rate cannot be debated while an assumption placed in a scenario can.
The intervention BEIREK operates across capital-intensive project portfolios is constructed around that third component. For every investment file, and before the decision point rather than after it, a single-page rationale record is produced showing which risk class the applied rate belongs to, on which plane the cash flow has been built, and which risks have been placed in the rate against which have been placed in the scenarios; the record is kept at the moment of proposal, not at the moment of approval, since a rationale written after the fact documents the defence of a decision rather than the decision itself. These records are then read at portfolio level and on a periodic rhythm rather than file by file, so that the number of proposals generated by each risk class, the classes being screened out systematically, and the direction in which the weighted risk profile has moved across quarters all appear in the same table.
A second effect of that rhythm shows up in negotiation. Once the rationale for a rate is written, what is defended in front of a lender or an incoming investor is no longer a target return but a risk classification; and because a classification is a discussable object, the counterparty's objection attacks a single assumption rather than the entire model. Founder dependency diminishes through the same mechanism, since knowledge of how rates are selected, held in the institutional record rather than in one person's working memory, attaches to the process rather than to any individual's tenure.
The genuine fragility in capital allocation is not that the wrong projects are approved but that no one knows which projects were eliminated before they were ever discussed, and that knowledge can only be recovered where the reasoning behind the threshold has been written down. Asked where an organisation's risk profile has moved over the past three years, it is worth considering whether the answer should be sought not in the strategy document but in the accumulated sum of its discount rate choices.