How the figures are computed

Every number on a property page comes from a stated formula over public data. This page gives the formulas in words, so an owner, an assessor or a council can check them. Nothing here is a judgement about any assessment.

What the site holds

The assessment roll of every Quebec municipality, as published by the ministry of municipal affairs: for each property, its land value, building value, total value, building area, lot area, year built, use, category, position, the neighbourhood unit the assessor placed it in, the date its value describes, and the previous roll's value. Rolls back to 2014, keyed on the matricule. The global tax rate of each municipality from the ministry's financial reports, and the ministry's market factor per municipality. The site holds no sale prices and nothing about the inside of any building.

Which homes are compared

Comparison is always inside one municipality, because every property of a category in a municipality pays the same rate, so the rate cancels and what is left describes tax burden. Across a boundary that stops being true.

Peers are drawn first from the neighbourhood unit the assessor assigned to the property on the roll. The assessor defines a unit as properties close to each other, with homogeneous characteristics, in a similar environment; it is the group the assessor analyses and compares within. Inside the unit, peers must be of the same category, with a building area within 25%, a lot area within 35% and a year built within 15 years, and where the roll records them, the same physical link and the same number of dwellings. The nearest forty are kept. When a unit yields fewer than eight such peers, the same criteria are applied over a widening area of the municipality instead, and the page says which of the two was used.

The repriced comparison

Totals are not compared, because a larger house would read as a higher one. Instead each peer's rates are applied to this property's own dimensions:

implied = (peer building value ÷ peer building area) × this building area
        + (peer land value ÷ peer land area) × this lot area

That gives, for every peer, what the roll would say about a house of this size at that peer's rates. The median of those is the value the property is measured against, and the gap is the property's value minus that median. A peer missing a land or building part is left out, never estimated.

The band

The implied values are split into quarters. A property inside the middle half is about in line. Above the third quartile it is above; below the first, below. Beyond a quartile by one and a half times the interquartile range it is clearly above or clearly below: the standard box-plot rule for an outlier. "Above" will always describe about a quarter of homes, because any ranking does; it is not by itself a reason to file.

The dollar figure

a year = (value − median implied value) × municipal rate ÷ 100

The rate is the municipality's global rate, blended across categories, adjusted to the roll in force; the page names the kind of rate it used. A difference under $60, or under 5% of the bill, is reported as about right: two assessments that agree, not a grievance.

The three checks

1. Against its peers' rates. The band above.
2. Against its own past. The property's change since the 2022 roll, minus the median change of its peers over the same rolls (the peers' tails clipped at the 5th and 95th percentiles). A gap of more than ten points fails. A failing gap on a house that sits in line with its peers today reads as a catch-up from an earlier under-assessment, and the page says so rather than calling it a case.
3. The land. This lot's land value per square metre against the median of its peers'. Land in one sector is usually assessed at a common rate, so a rate more than 15% above the peers' is reported.

A review is suggested only when the property is clearly above its peers' implied values, or its land fails, and the yearly difference is material. It is never suggested to a property below its peers: a review can raise a value.

The stake

stake = yearly difference × years left on the roll, counting the current year

A roll is in force three years. No review fee is subtracted, because the site does not hold the fee schedule, and a number built on a guessed fee would be worse than none.

How the comparison behaves across the province

On 30 October 2026 the comparison was run on 150 randomly chosen residences in each of the 1,132 municipalities, 165,151 homes in all. Among the 63,343 that had enough peers for a verdict:

A verdict exists for 38% of sampled homes when every municipality counts equally, because most municipalities are small and many homes there have no comparable within the bands; in cities the figure is above 90%. The median municipality's spread among similar homes (interquartile range over median of the implied values) is 20%, with the tenth and ninetieth percentiles at 11% and 30%. These shares are what a reader's own result should be read against: "above" describes a fifth of homes everywhere, and "clearly above" about one in thirty.

What the flags predicted, checked against past rolls

Every record carries its value on the previous roll. On 30 October 2026 the band was computed on previous-roll values for 27,092 randomly sampled residences across every municipality, among the same assessor-unit peers, and compared with what happened to each home at the current roll relative to those peers.

Three things follow. The flag is not noise: of homes clearly above on the previous roll, 66% are still clearly above and 31% are above today; only 0.3% of homes that were in line became outliers. Assessors do pull outliers back, but gently: the median clearly-above home moved about one point toward its peers, the average about five. And two thirds of outliers stayed outliers, which means that for most flagged homes the gap is durable, consistent with a difference the roll sees and this site cannot, such as quality, condition or a view. That is why the site asks the owner about finish before drafting anything, and why a flag alone is a point to raise, not a case.

This checks whether the flag predicts the assessor's own later corrections. It does not check against sale prices, which the site does not hold, and it was computed on total values at both dates because the previous roll's land and building split is not published.

Corrections

If a figure on this site is wrong, write to contact@monrole.ca and say which property or municipality and what you expected. Every report is checked against the data and the method above. A confirmed error is corrected on the site and listed here with the date, what was wrong and what changed. Any figure quoted to the press is recomputed before it is repeated, and a correction is sent to the outlet that carried it.

No corrections listed yet for the method described on this page, which took effect on 30 October 2026. Earlier figures compared total values and used fixed percentile cuts; the changes and the reasons are described above.

What none of this says

It compares assessments with assessments. It does not say what a property would sell for, and it does not say any assessment is wrong. It cannot see finish, condition, renovations or a view, which the assessor may have graded and which are legitimate reasons for a difference. A neighbourhood of mixed housing shows a wider spread than a uniform one for that reason alone. Where a figure cannot be computed honestly, the page says so and shows nothing.

Questions about the method: contact@monrole.ca.