The confusion is expensive, and it is everywhere. A business plan states an intention: here is the project, here is what it should return. A bankable feasibility does something else — it gives a credit committee grounds to decline. That is a difference in kind, not in length.
A banker is not looking to be convinced. He is looking for the points where the structure breaks: the month cash goes negative, the absorption assumption that does not hold, the cost line nobody itemised. A file that hides those points does not reassure him — it worries him.
The one-question test. Take any figure in your file, at random. Can you trace it back to its source in under a minute — a quote, a statement, a standard, a dated comparable? If the answer is no for even one figure, the file is not bankable. It is merely handsome.
In the Dominican Republic, a construction loan goes before a committee applying a precise framework: the rules of the Superintendencia de Bancos and the prudential ratios derived from Basel III. It is not negotiable, and it does not vary by institution.
In practice, four families of checks:
Most files that reach us address the first family and ignore the other three. That is precisely the reverse of what decides the outcome.
A bankable file is not signed by its author. It is signed by six distinct responsibilities, each binding the person who carries it:
A single signature, however prestigious, does not replace the six. A bank knows this, and looks first at who did not sign.
The financial statements in a feasibility are worth exactly what the method behind them is worth. We apply the criteria of the international standards on auditing (ISA) across five dimensions, in two passes — the second re-reads what the first corrected, which is the only way to catch errors introduced by the corrections themselves.
The result comes as three separate annexes, and that separation matters:
Merging the three into one document is the reflex that gives an amateur file away. A bank wants to read the attestation on its own.
A real estate project is not a financial object. It is a legal, technical, planning, tax, commercial and environmental object, and any one of those dimensions can sink it alone.
Our internal standard covers seventeen disciplines, each audited through six layers of control, with a threshold of 95 out of 100 per discipline. Below it, the section is regenerated — not annotated, redone.
That threshold is not decorative. A discipline sitting at 80% in a bankable file is a one-in-five chance the credit committee lands precisely on the gap.
Why 95 and not 100. The perfect file does not exist: what exists are files whose limits are known, written down and owned. The threshold leaves a margin — but a documented one, not a hidden one.
Every file carries a serial code and a SHA-256 fingerprint. Anyone can verify that the document received is exactly the one produced, byte for byte.
This answers a very concrete and rarely named risk: assembly. A file is not a binder into which pages from elsewhere get slipped. One project's sections are not interchangeable with another's, however adjacent, however shared the developer. The fingerprint makes assembly detectable.
And one rule governs all the others: zero invented figures. A missing data point stays missing and is declared as such. A plausible but unsourced figure is more dangerous than an empty cell, because it prompts no question at all.
A credible structure rests on three visible sources, and on how they interlock over time:
The bank looks at the commitment ratio before anything else. A lightly exposed developer transfers his risk to the lender, and the committee sees it immediately.
To this is added a disbursements and receipts ledger. Not an annual summary table: a monthly flow, with a realistic absorption assumption, CAPEX itemised line by line, and a sensitivity analysis. That is where a project's truth is read — in the tightest month, not in the average.
We do not start a construction phase before reaching 52% pre-sales on that phase. Every project is phased, with a schedule per phase and a Go / No-Go decision between each.
The threshold protects three parties at once. The buyer, whose funds do not disappear into a site with no demand. The bank, whose loan rests on proven commercialisation. The developer, who does not commit fixed costs to an unsold phase.
It is a real commercial constraint: it slows launches down. It is also why projects structured this way get financed, and the others do not.
Dominican electronic invoicing (e-CF, Directorate General of Internal Revenue) is not a last-minute administrative detail. A project whose invoicing chain is non-compliant from the outset creates a regularisation liability the bank provisions for — and which therefore weighs on the structure. We build it into the structure, not after it.
If you are an investor: ask for the feasibility file before you ask for the price. A developer who cannot produce one is telling you something important about what follows.
If you are a developer: the file is not the entry cost of financing, it is the financing. A sound project badly documented gets declined just as reliably as a weak one.
If you are a banker or trustee: our files are built to be contradicted. That is the point. Every assumption is traceable to its source, and every limit is written out rather than left to be discovered.
The projects we develop in the Dominican Republic — in San Cristóbal, Bávaro, Las Terrenas and Santo Domingo — all go through this framework. Without exception, and without a lighter version.
A feasibility study answers "is the project viable?". A bankable feasibility answers "can a bank lend against this document?". The second contains the first, plus regulatory compliance, ratios set against the lender's thresholds, covenants, conditions precedent, external audit and binding signatures.
The timeline depends on the availability of source data, not on the writing. A project with topographic surveys, builder quotes and title deeds in hand moves fast. A project missing those items does not move — and no amount of writing fills that gap.
The reference file is in Spanish, the language of Dominican institutions. We produce French and English versions for foreign investors, but the Spanish version is the one filed and the one that governs.
The section is regenerated, not annotated. A footnote does not fix an insufficient analysis: it flags it while leaving it in place, which is the worst of both worlds.
Because a published price has to be honoured, and an honoured price presupposes a validated, dated schedule tied to specific units. Until that schedule is settled for a project, we publish no prices. An indicative figure that moves is worse than no figure at all.
The audit follows the criteria of the international standards on auditing and produces three separate annexes — report, management letter, attestation. The binding signatures cover six distinct responsibilities, including that of a certified public accountant independent of the developer.
A question about structuring your project or about your bank's requirements? Our teams answer directly.
Talk to an expert Email us