ENESFRAR
Request a diagnostic
Project Margin Architect PMA
ENESFRAR
Request a diagnostic Menu Menu

Risk & investment consulting Cross-border infrastructure

Margin by design, not by chance.

Request a diagnostic Read the method

We are called in when a project's own documents stop agreeing with each other. We locate the disagreement, price it, and rebuild the case the money is lent against.

Four questions a credit committee cannot answer from the file

  1. 01

    Is the capex in this file the real capex?

  2. 02

    What is unsecured, and how much of it?

  3. 03

    Is the revenue contracted, or assumed?

  4. 04

    What breaks first, and at what point?

01Leakage

Margin is rarely lost. It is given away — one document at a time.

Every engagement begins from the same observation: the failure is almost never inside a single file. It sits in the space between files, where nobody is responsible for the comparison.

Severity

The feasibility study, the business plan and the financial model quote three different capital costs. Financing is sized on the smallest of them, and the gap surfaces on site, when the cheapest money is already spent.

Root cause

No single owner of the consolidated cost base. Each function prepares its own document, at its own time, for its own purpose — and the reconciliation is nobody's job.

Severity

Money moves to the contractor before the bank guarantee is in force. From that moment recovery is a negotiation rather than a right, and the lender's position is unsecured in fact if not on paper.

Root cause

The payment schedule is driven by the contractor's cash-flow needs, not by the lender's security terms. The guarantee is left for later.

Severity

In the contract folder there is no ship-or-pay, no volume commitment, no counterparty. The entire debt service rests on an assumption written in a spreadsheet.

Root cause

A letter of intent is given the weight of a contract. Commercial assumptions are never stress-tested, because nobody owns the downside case.

Severity

Site fill and piled foundations, seismic and drainage works, rolling stock and container flat wagons, grid connection, IT infrastructure. None of it is in the model, all of it is unavoidable on site, and each item is large enough to move the payback.

Root cause

Finance teams build the financial model; engineers write the engineering documents. The two sets are never read side by side.

Severity

The arithmetic is correct; the project it describes does not exist.

Root cause

Optimism is entered as an input, not tested as an assumption. The model is built to justify the investment, not to challenge it.

02Root causes

Why this keeps happening.

We have read files from banks, funds, developers and state corporations. The sectors differ. The symptoms differ. The causes repeat.

Six reasons the file does not hold

Siloed documentation
A file is assembled, not written: consultants, engineers, the sponsor's finance team and lawyers each contribute a part and are accountable only for that part. The divergence comes from the division of labour, not from intent.
Optimism as input
Volumes, prices and timing are each set at their most favourable value. One by one, the assumptions can be defended; together, they add up to a project without a single bad year. The sponsor's approval process rarely questions the combination.
Security treated as a formality
Payments are approved by the project team; guarantees are handled by lawyers and the back office. No control makes the first conditional on the second, so an advance can go out on a signed contract alone. The lender learns of the exposure once it already exists.
Revenue assumed, not contracted
Before an asset is built, customers are reluctant to commit to volumes. The model still needs revenue, so expressions of interest are turned into throughput — and the distance between interest and obligation is never priced.
Engineering costs invisible to finance
The model is often frozen before the surveys are finished. What the engineers find later — ground conditions, seismic requirements, the real cost of a grid connection — arrives as technical reports in technical language, and never makes its way back into the numbers.
No independent seat at the table
The contractor controls the schedule, the budget and the information. The lender sees the file, not the site. The divergence reaches the lender last, and is paid for at a premium.
03What we find

Not every file has every problem. But the same patterns recur.

Across infrastructure, logistics, construction and cross-border projects, these are the gaps we record most often.

Ten findings

  • Capital cost quoted in three documents, reconciled in none.
  • Advance payments with no bank guarantee behind them.
  • Throughput modelled as revenue, with no contracted counterparty.
  • Engineering works costed in the surveys but missing from the financial model.
  • Year one at nameplate capacity, ignoring ramp-up and seasonality.
  • Grid connection still unconfirmed, yet already built into the construction schedule.
  • Equity share stated in the business plan, not committed in the corporate filings.
  • Contingency set below the cost of known unknowns.
  • Non-binding letters of intent standing in for contracts.
  • A consolidated cost base that belongs to no one.

Each of these has a price. We measure it — and we rebuild the case the money is lent against.

04Instrument

We do not audit documents. We audit the joints between them.

Interlock diagnostics is the core of the Margin Management Framework. Every parameter that moves the number is traced across every document that states it — and the disagreement is measured, not noted.

Traced across

  • Feasibility study
  • Business plan
  • Financial model
  • General contract

The check

Whether the capital cost reconciles across every document that states it, line by line, including works priced in one file and omitted from another.

Red flag

A divergence of two times or more between documents is not a clerical error. It is a governance failure, and it has a price.

Found on a live engagement

Reconstructed capex diverged from the stated figure by a factor of 2.5. Real project cost was understated by 30–50%.

Traced across

  • General contract
  • Bank guarantee
  • Payment schedule
  • Corporate filings

The check

What share of the advance is covered by a bank guarantee, when the guarantee takes effect, and what happens to it if the contractor is replaced.

Red flag

Any material advance released against no instrument. The lender discovers this after the payment, not before.

Found on a live engagement

83% of the advance carried no bank guarantee of any kind.

Traced across

  • Business plan
  • Ship-or-pay contracts
  • Financial model
  • Letters of intent

The check

Whether the volume the model earns from is committed by a counterparty, and whether that counterparty can perform.

Red flag

Debt service modelled on throughput that exists only as an intention. Letters of intent counted as contracts.

Found on a live engagement

No guaranteed throughput was contracted. The revenue line rested entirely on projected volumes.

Traced across

  • Construction schedule
  • Design brief
  • Survey reports
  • Grid connection permits

The check

Whether the ramp-up curve, the works schedule and the permits that gate them describe the same project on the same dates.

Red flag

Full design capacity assumed from the first day of operation, with no ramp-up and no seasonality.

Found on a live engagement

The model took the terminal to full capacity immediately, against a schedule that had not yet secured grid connection.

Traced across

  • Financial model
  • Design brief
  • Engineering surveys
  • Technical conditions

The check

Which physically necessary works appear in the engineering documents but never reach the financial model.

Red flag

Ground conditions in the survey report that no line in the model pays for — waterlogging, seismicity, drainage, piling.

Found on a live engagement

IT infrastructure, rolling stock and container flat wagons, grid connection, site fill and seismic works were absent from the model.

05Framework

Five stages. One deliverable a credit committee can act on.

The Margin Management Framework is our own method, held as a trade secret. It merges the discipline of audit, the arithmetic of risk management and the output of strategy work into a single sequence.

Five principles it runs on

  1. 01Systemic view — the documentation is read as one system, not as separate files.
  2. 02Interlock diagnostics — the finding lives between documents, not inside them.
  3. 03Quantified risk — every risk carries a probability and a number.
  4. 04Three scenarios — base, stress and upside, each costed to DSCR, IRR, NPV and payback.
  5. 05Margin as the objective — the target is a better project, not a longer report.
01

Documentary reconciliationDiagnostics

The full pack is assembled — feasibility study, business plan, financial model with DSCR and stress tests, cash-flow budget, general contract with annexes, design brief, engineering surveys, grid connection conditions, corporate filings. Key parameters are then reconciled across all of them: capital cost, dates, specification, equity share, committed throughput, advance and security terms. Every divergence is recorded, including the small ones.

Deliverable —Discrepancy table and a red-flag register with a first severity rating.

02

Field studyOn site

For projects under construction or in pre-construction. The site is inspected: actual readiness, condition of utilities, ground conditions — waterlogging, flooding, seismicity. Interviews with the technical client, the contractor, the designer, local authorities and utility operators. Access to quarries, the state of approach roads and rail.

Deliverable —Site inspection report with photographic record; design documentation confirmed or refuted.

03

Financial stress testDeep dive

The model is checked for technical faults — broken formulas, inconsistent sheets — then rebuilt where it needs to be. DSCR, IRR, NPV and payback are recalculated against the reconstructed cost base. Stress cases are run: throughput down 20%, interest up 5 points, capex up 15%. The question is whether cover holds above 1.0 or the project moves into default territory.

Deliverable —Verified model, the real capex range, and a measured statement of financial headroom.

04

Risk and margin matrixRating

Risks are classified — financial, technical, commercial, legal, organisational, regulatory — and each is rated for probability and impact. Those that are both likely and critical are separated out as requiring an immediate decision. Three scenarios are then built: base, upside with a specialist operator and long-term volume contracts, and downside with the current risks left in place.

Deliverable —Rated risk matrix and three costed scenarios.

05

RoadmapDecision

A step-by-step plan out of the current position, grouped by urgency: immediate measures, one to three months, three to six months. It names what must be fixed in the documentation — capex reconciled, equity share raised, ship-or-pay concluded, guarantees obtained, surveys completed, grid conditions secured — and sets out the alternative configuration where one exists.

Deliverable —Final report with the roadmap, and the presentation for the board or credit committee.

06Record
Engagement record Cross-border container terminal Anonymised at the client's request

A file that had been financed on the smallest of its three capex figures.

A rail-and-river container terminal on the Russia–China border, at pre-investment stage. The documentation had passed internal review. Reconciled against itself, it did not hold.

Capex reconstructed against capex stated: a 2.5× divergence.

Across the feasibility study, business plan and financial model.

Real project cost understated by 30–50%.

Once omitted engineering and infrastructure works were priced in.

83% of the advance carried no bank guarantee.

Released against the contract alone.

No contracted guarantee of throughput.

Revenue and debt service modelled on projected volume.

Alternative configuration: capital cost reduced by 70%.

Pilot-scale build, phased capacity, restructured contracts.

Payback shortened from 16 years to 5–7.

On the alternative configuration, base scenario.

Payback, before and after restructuring

As presented16 years
Alternative configuration5–7 years

Figures as stated in the engagement's final report. Client and project identifiers withheld. No outcome is offered here as a forecast for another project.

07Practice

Four practices. The audit is the front door, not the product.

Classic audit hands over a report and leaves. We stay: the diagnostic is priced as a diagnostic, and the rest of the work is priced against what the project keeps.

The front door. Three depths, chosen by how much of the file is in question.

  • Health checkFive to seven core documents read against each other. Returns a three-to-five page report with the red flags: capex divergences, advance exposure, missing volume guarantees, absent cost lines.
  • Full due diligenceFifteen or more documents, a site study, the financial model stress-tested, contracts reviewed in law. Returns fifty to a hundred pages with a rated risk matrix and three costed scenarios.
  • Construction and technical auditEstimates, schedules and design decisions checked; contractor risk assessed. Most often commissioned by lenders before or during a construction drawdown.

The centre of the business model. We sit inside the project as an independent project office, with decision rights on the questions that actually move the number: replacing a contractor, increasing a budget, signing a strategic contract. Capital cost, schedule and quality are controlled from that seat; negotiation with partners is run from it.

The project strategy written as one document that holds: objectives, phases, key indicators, funding sources, risks and named owners. Alongside it, structuring of the financing, and the building or verification of business plans and financial models.

Where a project's economics depend on how goods and money cross a border, the fix is often structural rather than financial. We design the ecosystem — and, through AMP LLC, we operate it.

  • Financial Architect ecosystemThree product lines: deferral of import VAT for up to ninety days, deferral of logistics payment alongside specialist operators, and turnkey leasing of warehouse infrastructure.
  • Guarantee and settlement designA settlement structure that can be integrated with existing payment systems, so that deferral does not simply move the risk to another party.
  • Strategic partnersSourcing and bringing in specialist operators and strategic partners, with the analytical and presentation material the negotiation requires.

Foreign-trade operations — AMP LLC

  • Design of inbound freight routes into Russia
  • Documentation packs and customs clearance at speed
  • Audit of accompanying and permit documentation
  • Phytosanitary control and permit paperwork
  • Customs audit to minimise clearance and post-clearance risk
  • Representation in customs disputes, including administrative proceedings
  • HS-code risk audit, with a written opinion and remedies
  • Advance classification rulings
  • Restoration of tariff preferences
  • Foreign-trade agent service
  • Review against low-risk-operator criteria
  • Customs-value risk control on the client's nomenclature
  • Full administration of import contracts and payment for delivered goods
08Engagement

We are paid for what the project keeps.

A fixed fee buys the diagnosis. Everything after that is tied to a number both sides can verify — cost taken out, or contracts signed.

01

Diagnostics

A health check or a full due diligence. You get the red flags, the reconstructed cost base and a clear statement of what is not yet financeable.

Fixed fee, agreed against scope.

02

External project office

We take a seat inside the project with decision rights on contractor, budget and strategic contracts, and control cost, schedule and quality from it.

Monthly retainer.

03

Success fee

A share of the cost saving both sides have verified, or of the contracts concluded with operators and strategic partners.

Share of verified result.

If the cost base does not fall and no long-term agreements are signed, no success fee is due.

Fees are set by agreement against the scope and effort a specific file requires. We publish no price list, because no two files cost the same to read properly.
09Request

Send us the file that worries you.

A health check reads five to seven core documents and comes back as a three-to-five page red-flag report. Start there; decide the rest afterwards.

Please enter your name.
Please enter your organisation.
Please enter a valid email address.
Tell us briefly what you need looked at.
Please confirm before sending.
Or write to us directly.

Request a diagnostic

Margin by design, not by chance.