How to Check Valuation Calculations Properly

A valuation can be analytically sound and still contain a figure that does not reconcile. A market value may have been updated in the executive summary but not in the valuation rationale. A floor area may be stated in square metres on one page and used as square feet in a rate calculation on another. The purpose of a final review is to check valuation calculations as a connected whole, before the report leaves your desk.

This is not about distrusting the valuer’s judgement. It is about recognising how reports are produced under time pressure. Figures are revised, comparables are added late, assumptions change after a call with the client, and content is carried between sections. The more moving parts there are, the more useful it is to test whether the arithmetic, inputs and narrative still tell the same story.

Start with the figure the lender will rely on

Begin with the reported market value, market rent where applicable, and any special assumption value. These are the figures most likely to be read first and repeated elsewhere in the report. They should agree precisely across the instruction summary, valuation conclusion, executive summary, valuation certificate, security section and any lender-facing form.

A difference is not always a mistake. A report may properly distinguish between market value subject to a special assumption and unrestricted market value. The point is that each figure must be clearly labelled and used consistently. A £50,000 difference between the valuation conclusion and a summary table will invite a question, even where the underlying valuation is well reasoned.

Then work backwards. If the property is valued at £2,400,000 on a rate of £4,000 per sq m, the adopted area should be 600 sq m. If the report instead states 580 sq m, establish whether the adopted rate, area basis or capital value has changed. Do not assume it is a rounding issue until the calculation proves it.

Check valuation calculations from the inputs upwards

A useful review follows the calculation chain rather than reading the report only from page one to the end. For each valuation approach, identify the source data, the adopted input, the calculation and the reported output.

Areas and rate applications

Area errors are often small in appearance and material in effect. Check that the area used in each calculation matches the stated basis of measurement and the unit used by the comparable evidence. A rate per sq ft cannot be applied to an area in sq m without conversion. Nor should a gross internal area be compared directly with a net internal area without a clear adjustment and explanation.

This matters particularly where a property has mixed accommodation. A ground-floor retail unit, ancillary basement and upper parts may each have different area treatments and rates. The total may reconcile arithmetically while the component values do not reflect the adopted approach.

Where rounded figures are used, calculate from the unrounded working where possible. A reported area of 1,250 sq ft and a rate of £400 per sq ft might reasonably produce a rounded figure of £500,000. But several individually rounded components can create a final difference that needs explaining, rather than simply being carried into the conclusion.

Rent, yield and capital value

Investment valuations require the same discipline across rent, term and yield. Confirm the passing rent, estimated rental value, market rent and adopted income are not being used interchangeably. If the valuation capitalises £180,000 per annum at 6.00%, the initial capital indication is £3,000,000 before allowing for purchaser’s costs, rent-free periods, voids, non-recoverable costs or other adjustments. The report should make clear which of these have been reflected.

Yields need context as well as arithmetic. A stated net initial yield may not reconcile to the purchase price and passing rent if costs have been omitted, or if the rent is expressed net of a deduction elsewhere in the report. This is not necessarily wrong, but it should be traceable. The reader should not have to infer whether the stated yield is equivalent, net initial or reversionary.

Check lease dates carefully. One transposed digit can turn an unexpired term of 8.4 years into 18.4 years in the narrative, affecting the reader’s understanding of covenant risk and investment appeal even if the valuation calculation itself uses the correct date. The lease commencement, expiry, break dates, rent review dates and any incentives should agree wherever they appear.

Development and residual calculations

Residual appraisals deserve a slower review because a modest input change can have a disproportionate effect on land value. Check the gross development value, unit numbers, sales rates, construction costs, professional fees, finance assumptions, contingency, marketing period and profit basis. Then test that the residual output follows from those inputs.

The appropriate level of checking depends on the instruction. A straightforward residential investment report does not need a development-appraisal review. Where a residual valuation is material, however, it is worth testing the logic behind the model as well as the totals exported into the report. A stated 18-month programme alongside finance calculated over 12 months is a useful example of a contradiction that can be missed in a page-by-page read.

Reconcile the comparables, not just the adopted figure

Comparable evidence is where the calculation and the valuation judgement meet. Each comparable should reconcile internally: price or rent, area, analysed rate, transaction date, tenure and any adjustment stated in the commentary. If a comparable sale price is £1,050,000 and the reported area is 500 sq m, the analysed rate should be £2,100 per sq m, subject to the stated treatment of costs and any other adjustment.

It is equally important to compare the evidence table with the narrative. If the report says the adopted rate sits below the best evidence due to inferior specification, the adopted rate should actually sit below it. If the narrative describes a comparable as letting evidence but the table records a sale, that distinction may affect the weight a lender gives the analysis.

Not every comparable will require the same adjustment. A transaction with a short unexpired term may be highly relevant to investment sentiment but less directly comparable to a long-let asset. The review should test whether the numbers support the explanation, not force every item into a uniform format.

Read the report for contradictions

Arithmetic is only part of the task. A good final review reads the report as one document. The property address, tenure, inspection date, valuation date, market conditions, planning position and basis of value should remain consistent from instruction to conclusion.

This is where repeated figures cause difficulty. A previous draft may refer to a marketing period of six months, while the final report says three. An EWS1 disclosure might appear in the risk section but be absent from the executive summary required by the instruction. The market value can still be correct, yet the report may not meet the client’s stated criteria.

Build a review around the points that are easy to overlook when concentrating on the valuation itself: report-wide figure consistency, required disclosures, lender-specific fields, minimum comparable evidence and conflicts between narrative and tables. Red Book compliance is not a matter of finding a single phrase. It depends on whether the report, assumptions, reasoning and disclosures are appropriate to the instruction.

Use an independent second pass

A manual second review remains valuable, particularly where the reviewer understands the asset and instruction. But a second person may still follow the same assumptions, skim familiar wording or focus on the conclusion rather than every repeated figure. Confirmation bias is a normal feature of technical work, not a failure of professional care.

This is where a structured audit can help. WriteUp reviews a draft valuation report in one to two minutes and flags figures that do not reconcile, contradictions across sections and missing criteria such as EWS1 disclosures or comparable evidence thresholds. It is private, encrypted processing, with no report data retained or used for model training. The surveyor decides whether each finding is relevant and remains responsible for the final report.

The value is not in handing judgement to software. It is in directing attention to the few lines that deserve another look before submission, especially when the report has been revised several times or is part of a high-volume instruction.

Before issuing, ask a simple question: could another experienced valuer trace every reported value back through the adopted inputs, evidence and assumptions without having to guess? If the answer is yes, the report is far more likely to stand up under scrutiny.