Property financial modelling in Morocco: building models that survive scrutiny

A model is not an accounting exercise. It is an instrument of discussion: it makes a chain of reasoning visible, therefore contestable, therefore improvable. A model nobody can contest is not a strong model — it is an opaque one.

1. Why modelling separates assertion from demonstration

The professional does not content himself with saying that a scheme is good. He opens a file, shows his assumptions, explains where each figure came from, and accepts that they will be attacked one by one. That posture is only available to somebody whose file can withstand it, which is why the discipline is worth acquiring deliberately rather than by accretion.

It is sometimes said that specialised software has made the spreadsheet obsolete. Field experience says the opposite. A specialised tool produces a result, but it encloses the reasoning in a black box; when someone asks why the vacancy allowance is set where it is, or why the major-repairs budget is phased that way, you need to be able to answer by pointing at a cell. Three properties keep the spreadsheet where it is, and nothing has replaced them: the transparency of the formula, auditability — anyone can walk back up the calculation chain — and portability, since the file passes as it stands to a co-investor, an adviser or a valuer, without a licence and without a proprietary format.

In Morocco this matters more than elsewhere, for a reason worth stating plainly to anyone arriving from a mature market: market data is fragmented and rarely public. You work on constructed assumptions far more often than on long series available at a click. In that context the quality of a model does not come from its mathematical sophistication but from the traceability of its assumptions — where this reference rent came from, which comparable transactions support this exit price, what margin of uncertainty is being carried. A model that documents its sources beats a brilliant model that will not say where it is speaking from.

2. Architecture: assumptions, calculations, output

The first competency taught in a serious course is not a function. It is a discipline of structure. A property model divides into three strictly separated blocks, ideally on distinct sheets.

• Assumptions — every input, grouped and annotated: areas, reference rents, programme and timing, unit costs, financing terms, holding period. No assumption lives anywhere else.
• Calculations — the engine, containing only formulas that point back at the assumptions. It is the part you do not touch when you test a scenario.
• Output — the summary written for the reader: cash flow table, indicators, charts, the range of results. It calculates nothing; it displays.

From that separation follow a few rules that are simple and not negotiable. One source per assumption: if a reference rent appears on three tabs, it will eventually hold three different values. Never a hard-coded number inside a formula: a figure typed into the middle of a calculation is a hidden assumption, and hidden assumptions are the leading cause of wrong models. A constant colour convention — one colour for inputs, another for formulas, another for links between sheets — lets any reviewer see in three seconds what he is allowed to change.

Two further elements distinguish a professional model from an improvised file. First the control sheet: a handful of coherence tests that must return zero or true at all times — uses equal resources, the sum of the annual flows equals the total, no cash balance is incoherent, no cell displays an error. Then versioning: a model lives, gets corrected, gets extended. Without a naming convention and a change log you will eventually present an investor with a version that is not the one you think it is.

3. The toolkit that genuinely matters

The good news is that the number of tools genuinely required is small: the overwhelming majority of property operations can be modelled with about a dozen, properly mastered. The bad news is that mastering them means knowing their traps as well as their syntax.

Logic, lookup, aggregation

• Conditional logic and clean error handling — on condition that error suppression is not used to hide a real structural problem. A trapped error that nobody investigates is a bug with better manners.
• Positional-independent lookup — retrieving a value from a reference table without depending on the order of the columns, which is what breaks a model six months after it was written.
• Multi-criteria aggregation — summing by unit, by works package, by period.
• Dynamic ranges — powerful, but volatile and awkward to audit. Use sparingly, and never where a structured table does the same job.

The financial functions

• Discounting and internal rate of return over regular periods — with care over the treatment of the initial outflow, which is the classic source of a wrong answer that looks plausible.
• Internal rate of return over dated, irregular flows — indispensable the moment cash lands on real dates: equity drawdowns, staged sales, construction payments.
• Loan amortisation — building a repayment schedule, separating principal from interest, and measuring debt service period by period.

Scenarios and volume

• One- and two-variable data tables — the fastest way to produce a sensitivity matrix that fits on one readable page.
• Back-solving — finding the input that drives an indicator to a chosen threshold; for instance the acquisition price at which the NPV falls to zero.
• Constrained optimisation — when a single back-solve is no longer enough because several constraints bind at once.
• Structured tables, pivot summaries and a repeatable import routine — the moment you handle schedules of units, rent rolls or exports of comparables, these replace hours of manual handling and remove a large share of copy-paste error.

4. The two reference models

Almost all property modelling reduces to two families. Knowing how to build both, and knowing which one applies, covers the bulk of professional situations.

The income asset cascade

Here you follow an already-built, let property through time. The cascade is always the same and its order is not negotiable: start from potential gross income — the asset fully let at market rent — deduct vacancy and bad debt to reach effective income, then the non-recoverable operating expenses to arrive at net operating income. NOI is the pivot: capitalised, it produces a value, and it is the first thing a purchaser looks at.

Below NOI come the items that depend on the investor rather than on the asset: capital expenditure — major repairs, renewal, reinstatement — and debt service, giving cash flow to equity. At the end of the horizon you model an exit value, generally by capitalising a normalised final-year NOI net of disposal costs. The whole set of flows is then discounted at a rate reflecting the risk of the project. Any charges arising under the rules in force sit outside this cascade and should be confirmed with your own advisers. One teaching point carries more weight than the rest: the exit capitalisation rate and the discount rate are assumptions, not truths. They must be displayed, justified and tested — never copied across from a previous model because it was to hand.

The developer appraisal and its residual reading

The developer appraisal models a construction-and-sale operation. Its lines are conventional: forecast sales revenue derived from the price schedule and the absorption rate; land cost; construction cost; professional fees for design, engineering and inspection; finance costs arising from the mismatch between outflows and receipts; marketing costs; a contingency line; and the margin the operator requires.

The professional interest of the model lies chiefly in reading it backwards. In the residual method you do not compute the margin from the price of the land: you start from achievable sales revenue, subtract all costs, fees and the required margin, and what remains is what the land can support. That is the reasoning of the developer negotiating a site, and equally that of the valuer who has to value development land. Its fragility is well known and must be stated whenever the result is presented: the residual is a difference between large aggregates, so a modest movement in sales price or construction cost shifts the land value substantially. All the more reason never to present it as a single figure — a point we develop in our note on valuing developer land by the residual method.

5. Reading the indicators without telling yourself stories

A model always produces several indicators, and each says a different thing. The IRRexpresses an internal rate of return: intuitive, but it ignores the size of the operation, becomes very unstable over short horizons — a few weeks' movement on one flow can make it jump — and it rests on an implicit assumption that intermediate flows are reinvested at the same rate, which is rarely realistic. The NPV expresses value creation as an amount, but it depends entirely on the discount rate adopted: change the rate and you change the conclusion, which is why displaying and justifying it is compulsory rather than optional. The cash-on-cash return measures what the equity actually committed earns, without discounting. The payback period says how fast the stake comes back, and ignores everything that happens afterwards.

None of them stands alone. A useful course teaches, above all, how to spot the case where one of them is lying: a flattering IRR on a short, small operation; a high NPV obtained with an accommodating discount rate; an attractive cash-on-cash concealing a deferred capital expenditure requirement. Any value handled in a training room is purely illustrativefor the same reason: there is no such thing as a “good IRR” or a “normal capitalisation rate” applicable without discussion to a market, a city or an asset class.

Sensitivity analysis: the real deliverable

The output of a good model is not a figure. It is a range and a set of break points. Concretely: a two-variable table crossing the two most structuring assumptions — often sales price or market rent on one axis, construction cost or exit yield on the other; three coherent scenarios, prudent, central and favourable, built by changing a complete set of assumptions rather than a single cell; and the identification of the break-even points — from what fall in price does the operation stop creating value, what minimum occupancy still covers debt service. That is the information a decision is made on, not the third decimal of an IRR. The same logic applies to the scenario tables inside a valuation report, as set out in our note on reading DCF sensitivity in a valuation report.

The classic errors

• The illegible model — endless formulas, unnamed tabs, no reading logic. Nobody can verify it, so nobody can trust it.
• Hidden assumptions — values typed into the middle of a calculation, coefficients with no comment, arbitrations left undocumented.
• Uncontrolled circularity — the textbook case of finance costs that depend on the funding requirement, which depends on the finance costs. Recognise it, treat it cleanly, and document the choice made.
• No coherence checks — without a control sheet, a range error goes unnoticed all the way to the investment committee.
• False precision — displaying a result to the dirham from approximate assumptions buys a confidence the model has not earned.
• Copying an existing model — reusing a file without understanding its assumptions means inheriting somebody else's choices, usually made for a different asset in a different context.

6. Training: content, prerequisites and audiences

A course in property modelling has to be built around files, not around slides. The natural progression follows the profession: structure a clean model (assumptions, calculations, output, control sheet); build an income asset cascade through to cash flow and exit value; build a developer appraisal and then turn it round into the residual method; attach a loan schedule; and finally produce the layer that makes the difference — sensitivity tables, scenarios, break points, and a summary note legible to a non-financial decision-maker.

• Real prerequisites — moving around a workbook, writing a simple formula, understanding relative and absolute references. Nothing beyond that: the financial content is taught in the room.
• Agents and negotiators who want to defend a price with figures rather than with sales arguments.
• Developers and client bodies arbitrating land acquisitions, who need to know what a site can support before they bid on it.
• Asset managers and family offices tracking a let portfolio and producing reporting somebody else will rely on.
• Analysts and valuers who receive business plans and need to audit them rather than absorb them.

This is a different exercise from writing the business plan itself, where the work is the defensibility of the underlying assumptions rather than the construction of the file — a subject treated separately in our note on building property business plan assumptions that hold.

7. The ReaConsult approach, and what anchors a model

What distinguishes ReaConsult Academy is where its trainers come from: they are valuation practitioners. Since 2019 the practice has completed more than 5,000 valuations across 6 cities, with RICS-certified experts and a rating of 4.9/5 across 47 reviews. That changes the teaching on one essential point: a model is worth no more than its value assumptions, and those assumptions come from the ground — comparables, observed rents, costs actually incurred. An independent valuation remains the best way to anchor the entry or exit value of a model on something other than intuition. It is a private valuation, produced to inform a decision or an arm's-length negotiation.

On formats, ReaConsult Academy offers an in-person day in Casablanca at 1,500 MAD incl. tax, video sessions at 150 EUR, and certifying pathways between 15,000 and 17,500 MAD for those aiming at a complete competence in valuation. The 21 sessions already delivered to professionals have shaped a simple format: you open an empty workbook at the start of the session and leave with a model that runs. For companies, funding may run through the OFPPT Special Training Contracts (CSF), on review of the file and under the conditions in force — a matter to be framed with the HR function in advance, not at the point of registration, as set out in our guide to funding property training through the CSF.

8. Instructing the work

Where a model needs an anchor a spreadsheet cannot supply — an entry price, an exit value, a market rent — our reports are prepared by RICS-certified experts and comply with Red Book standards: named assumptions, cited sources, a stated methodology, and conclusions built to be argued with point by point rather than defended in a block. Fees start at 3,500 MAD excl. tax for standard assets, with a firm quote within 24 hours and delivery in 5 to 8 days. ReaConsult has been advising owners, investors and institutional clients since 2019, with more than 5,000 valuations completed, offices in 6 cities and a rating of 4.9/5 across 47 reviews.

Note:this article sets out a modelling methodology and the training routes associated with it. No IRR, capitalisation rate, discount rate, salary, placement rate or course duration is quoted, because none is established as a general figure: those points are rendered qualitatively for that reason, and any value used in a training room is purely illustrative. No qualification is presented as mandatory. Training funding through the OFPPT Special Training Contracts is granted on request and subject to review of the file, under the conditions in force. Any charges arising under the rules applicable to an operation are a separate matter to be confirmed with your own advisers. A private valuation informs a decision and an arm's-length negotiation. To instruct us, see our contact page or the property blog.