Blog

Construction tolerances

📐 Article9 min read

What this page covers The function of the mechanism, the ordinary 2 per cent limit, the graduated regime introduced by the Salva Casa decree for works predating 24 May 2024, the anti avoidance rule on the reference area, execution tolerances, a worked calculation, the limits of applicability and the professional's certification.

Not everything departing from the approved design is a breach. Tolleranze costruttive, construction tolerances, identify a band of departures which, given their limited extent, do not constitute a building violation. On existing stock this is the mechanism deciding whether a survey closes with a certification or opens a regularisation procedure.

The ordinary limit

Paragraph 1 of article 34-bis of the code provides that failure to observe the height, separation distances, volume, footprint and any other parameter of individual units does not constitute a building violation where it is contained within 2 per cent of the measurements provided in the consent. The wording sets a percentage below which a departure is not a non conformity.

Three elements of that wording repay attention. Each of them decides the outcome of the calculation.

The reference is to the consent, not to the design as filed or to the expected as built condition: the percentage is calculated on the measurements the consent provides. The basis of calculation is therefore documentary and not surveyed.

The unit of reference is the individual property unit, not the building as a whole. The check must be run unit by unit.

The list of parameters is open, because the rules refer to any other parameter after identifying height, separation distances, volume and footprint. The closing words extend the tolerance to parameters not named.

The graduated regime for earlier works

Decree Law 69 of 2024 inserted paragraph 1-bis into article 34-bis, which for departures relating to works carried out by 24 May 2024 replaces the uniform 2 per cent limit with a progression inversely proportional to the net floor area of the unit. The new regime applies only to works predating the date given.

Net floor area of the unit Departure limit
Above 500 square metres 2 per cent
Between 300 and 500 square metres 3 per cent
Between 100 and 300 square metres 4 per cent
Below 100 square metres 5 per cent
Below 60 square metres 6 per cent

The rationale for the progression is that the same percentage produces very different effects on a thirty square metre unit and on a thousand square metre one, and that on small units a uniform 2 per cent limit yielded departures of a few centimetres. The sliding scale corrects that disproportion.

The last two rows should be read as bands rather than overlapping thresholds: an eighty square metre unit falls under 5 per cent, a fifty square metre unit under 6 per cent. Each unit falls within one band only, set by its floor area.

The date of 24 May 2024 is decisive and must be documented. This is why dating departures in the report on lawful status is not a detail: it determines which of the two regimes applies.

The anti avoidance rule on the reference area

Paragraph 1-ter of the same article provides that the net floor area relevant for applying the thresholds is determined by reference to the original consent that authorised the works, disregarding any subsequent subdivision. The rule holds even where the unit was later subdivided.

The purpose is expressly anti avoidance: to prevent a unit being subdivided solely in order to fall within a wider tolerance band. Without it, subdivision would become a planning device.

The practical effect is that the calculation must be carried out on the extent approved by the original consent, and that a unit of eighty square metres today deriving from the subdivision of an original unit of four hundred remains in the 3 per cent band and not in the 5 per cent one. A unit that is small today may therefore fall within a narrower band.

Execution tolerances

Article 34-bis also governs a second category, distinct from the first and often confused with it. The difference lies in the test, qualitative rather than percentage based.

Tolleranze esecutive, execution tolerances, concern irregularities of an executional or geometric nature, without substantive significance, which may emerge during approved works. They are not tied to mathematical percentages but to the minimal and immaterial nature of the departure from the consent.

The tests delimiting their scope typically concern not affecting the external appearance of the building, the load bearing structure, the dimensioning or the non derogable separation distances. None of them is expressed as a percentage.

Here too the 2024 decree introduced a dual regime, with ordinary rules and rules referring to works predating 24 May 2024 which widen the range of situations while conditioning them on the limits set for the ordinary regime. Which rules apply depends on the date of the works.

A worked calculation

Applying the bands is clearest on a case. Take a unit with a net floor area of 90 square metres, built in 2015, whose consent provided for an internal height of 2.70 metres, where a survey returns 2.60 metres.

Step Value
Measurement in the consent 2.70 m
Measurement surveyed 2.60 m
Absolute departure 0.10 m
Percentage departure 3.70 %
Band applicable, work predating 24 May 2024 and area below 100 m² 5 %
Outcome Within tolerance

The same departure on the same unit, had the work been carried out after 24 May 2024, would fall under the ordinary 2 per cent limit and would constitute non conformity. And if the 90 square metre unit derived from the subdivision of an original 400 square metre unit, the applicable band would be 3 per cent and the outcome would likewise be negative.

Three variables therefore determine the outcome on a single departure: its extent, the date of the work and the net floor area referred to the original consent. None of the three can be deduced from the survey alone.

A limit of applicability to know

Administrative case law has clarified a boundary that practice tends to ignore. It concerns properties legitimated by an amnesty.

Tolerances do not apply to extensions carried out on a property that has already been the subject of a condono, an amnesty, because the situation covered by article 34-bis arises only where design measurements set by a previously issued building consent have not been observed, and not following retrospective consent. An amnesty follows its own rules and does not combine.

The principle has a general reach worth spelling out. The mechanism presupposes a consent fixing design measurements from which the departure is measured. Where that reference is absent, or where the legitimation derives from an instrument of a different nature, the percentage mechanism has no base on which to operate.

The professional's certification

Tolerances do not require a procedure: they require a certification. It is the qualified professional who declares that the departures surveyed fall within the applicable limits, and that declaration allows the property to fall within the sphere of lawful status.

The minimum content of the certification covers identification of the reference consent, the measurement the consent provides for each parameter concerned, the measurement surveyed, the departure in absolute and percentage terms, the net floor area assumed for the applicable band and the date the work was carried out. The measurement surveyed and the date must also be stated.

Any element omitted makes the certification unverifiable. The net floor area and the date in particular are not formalities: they are the two figures determining which limit applies.

The most frequent errors

Applying the paragraph 1-bis bands to works carried out after 24 May 2024 is the first, and attributes to the property a tolerance it is not entitled to. The property is credited with a tolerance it does not have.

Calculating the net floor area on the current condition rather than on the original consent is the second, and contravenes the anti avoidance rule. The anti avoidance rule is thereby breached.

Confusing construction tolerances with execution tolerances is the third, and leads to percentages being applied to departures governed by a qualitative test, or the reverse. Percentages end up applied to departures governed by a qualitative test.

Assuming the mechanism applies to properties whose legitimation derives from an amnesty is the fourth, and case law has already excluded it. Case law has already excluded it.

Certifying without documenting the measurement provided, the measurement surveyed and the date is the fifth, and produces a document without evidential value. The document ends up with no evidential value.

Note: the legislative references in this page relate to Italy and are current as at the date of publication. Several regions have received the 2024 amendments through their own provisions and case law is still developing: the text in force, the regional legislation and the municipal building regulation should be checked before any operational use.

Frequently asked questions

2 per cent of the measurements provided in the consent, for each property unit.

No. They concern departures relating to works carried out by 24 May 2024, on a progression inversely proportional to net floor area.

On the net floor area referred to the original consent, disregarding subsequent subdivisions.

No, a certification by a qualified professional is provided for, and it should be drawn up with the figures that make it verifiable.

Building consents in Italy: permesso di costruire SCIA and CILA