Prices and Appreciation in San Miguel de Allende

$568,829

Average resale sale price, 2025

Down from $650,462 in 2024. An average, not a median, and reported by a single brokerage.

29.6 months

Average inventory, Jan-May 2026

Against 17.6 months in the same period of 2025. The same source calls 15 to 18 months normal here.

860

Active listings, May 2026

Up 45 percent on 593 a year earlier. The count fell to 745 in June.

None

Published medians from closed sales

No registry, association or MLS feed publishes one, for any year.

San Miguel de Allende has no public transaction registry. The Guanajuato Registro Público de la Propiedad publishes no compraventa statistics. AMPI San Miguel de Allende, the local realtor association, publishes listings and an agent directory and nothing else; we checked its site directly on 23 July 2026. There is a shared local MLS, but it is not publicly accessible and issues no aggregate reports.

The practical consequence is that essentially every hard number about San Miguel prices, including the ones below, traces to a single brokerage reporting on its own view of that MLS. It is the best data that exists, and we use it, but nothing independent can verify it. Treat that as a permanent caveat attached to every figure on this page.

The headline: 2026 has cooled

The clearest signal is supply, and it is not ambiguous.

Two notes on that table. First, the same brokerage describes 12 or more months of inventory as ordinary for San Miguel and 15 to 18 months as normal, on the argument that this is an almost entirely cash market whose absorption rate never resembles a mortgage-driven one. By that benchmark 29.6 months is roughly double normal, not merely elevated. In March 2026 the firm said inventory in several price ranges was as high as it had been since the period following the 2008 financial crisis, though it published no band-by-band numbers.

Second, the June fall from 860 to 745 is not absorption. The source attributes it to withdrawals, expirations and sellers choosing other strategies. It also states the decline was 81 properties, which is wrong: 860 minus 745 is 115. We publish the arithmetic error because the source has not corrected it, and because it bears on how much weight any single monthly figure can carry.

What a home actually sold for

Applying the reported 24.68 percent increase to 404 closings implies roughly 504 sales in 2025. That figure is arithmetic, not reporting. The source never stated a 2025 count, and any page presenting 504 as a published number is misrepresenting it.

The two changes for 2025 are the interesting pair: units up 24.68 percent, dollars up 9.03 percent. More transactions, less money. The average sale price fell $81,633, or 12.6 percent, and sales above $2 million fell from 19 to 12. The brokerage attributes roughly half the decline to the loss of those luxury transactions and the rest to elevated inventory.

That attribution is worth taking seriously, because it is also the reason the average is a weak statistic here. A market closing four to five hundred homes a year, in which a dozen sales above $2 million occur, has a mean that a handful of $3 million to $5 million transactions can move by tens of thousands of dollars. The 2025 average did not fall because typical houses lost 12.6 percent of their value. It fell because the mix of what sold changed. A 12.6 percent drop in the average is not a 12.6 percent drop in prices, and should never be published as one.

The correct fix would be a median from closed sales. No source publishes one, for any year. That is the single largest hole in this market’s public record.

Price per square metre: every published figure is an asking price

There is no price per square metre derived from closed sales in San Miguel de Allende. Not one. Every figure available is either scraped from listing portals or crowdsourced, and they do not agree.

The city-centre crowdsourced figure sits 74 percent above the citywide asking median published three days earlier. Part of that is a genuine centre premium and part is that the two are not measuring the same thing, which is precisely the problem: the published figures cannot be reconciled because none of them documents a method. Numbeo’s sample is nine people. Publishing a two-decimal-place peso figure off nine contributors is a presentation choice, not a measurement.

The neighbourhood-level table below is the only one of its kind we could find. It is one aggregator’s listing-portal scrape with no disclosed sample sizes, and we reproduce it labelled rather than endorsed.

Three things to notice. Ojo de Agua and Los Frailes are absent from it entirely, so no price per square metre of any kind is published for either. The same publisher gave San Antonio as a point figure of MXN 39,600/m² on 2 April 2026 and as a range of 32,000 to 55,000 ten weeks later, without reconciling the two. And the same publisher’s land table puts Centro land at MXN 18,000/m², well below its own figure for built Centro property, which is not impossible but is unexplained.

With inventory at 29.6 months, asking prices sit above clearing prices by a margin nobody has measured. Mexico has no published sale-to-list database; the 92 to 97 percent ratio in circulation is inferred, and the publisher carrying it says so itself.

The 2026 turn, month by month

The January average of $649,316 is the only 2026 price point published by anyone. It sits 14 percent above the 2025 full-year average, and it covers a single peak-snowbird month in a market closing on the order of 40 homes a month, where one large sale moves the mean. It is not a 2026 run rate. No monthly average or median has been published for any month since.

March is the pattern in miniature: more closings, materially less money. The firm itself described the market as showing conflicting signals.

The closing count nobody can settle

The 2026 year-to-date volume figure is unresolved, and we will not publish one.

Realty San Miguel’s May 2026 post reports 197 closings through May 2026 against 190 in the same period of 2025. Its June 2026 post reports 186 resales through June 2026 against 220 through June 2025. A year-to-date count cannot fall from 197 at five months to 186 at six, and the 2025 comparative cannot rise from 190 to 220 for an overlapping period. Either one post carries revised or differently defined data, or one is wrong. The source does not acknowledge the discrepancy.

Taken at face value the June figure is a 15.5 percent year-on-year decline in closings. Taken at face value the May figure is a 3.7 percent increase. Those are different markets. Until the firm reconciles them, the honest position is that the 2026 transaction count for San Miguel de Allende is not publicly known.

Appreciation: what can and cannot be said

No repeat-sales index, no hedonic index and no MLS time series exists for San Miguel de Allende. There is therefore no measured appreciation rate for this city, at any level of geography.

What exists instead:

• Closed-sale averages, which fell 12.6 percent in 2025 and are a mix statistic, not a price index.
• Asking-price estimates, which one aggregator puts about 9.5 percent higher in nominal pesos than a year earlier and 17 to 20 percent higher over two years. These are asks, not sales, and in pesos, not dollars.
• Modelled neighbourhood appreciation of 8 to 13 percent from the same aggregator, which lists Realty San Miguel among its own sources and so had access to the closed-sale data pointing the other way.
• The national SHF index, which recorded 8.7 percent annual growth in Q1 2026, with a national median appraisal value of MXN 1,331,000 and an average of MXN 2,024,337. It covers only homes bought with a mortgage, which structurally excludes almost all of San Miguel’s market, and its national median is a fraction of what foreign buyers pay here. SHF publishes no San Miguel series. The nearest metro datapoint is León at 8.2 percent, which is a poor proxy and not the same market. A Guanajuato state figure circulates but we could not verify it against a fetchable page, so we do not publish it.

Aggregator forecasts for the twelve months from July 2026 put well-positioned homes at 4 to 8 percent, average and overpriced listings at flat to 3 percent, and stale luxury as needing cuts of 5 to 10 percent. Full-year scenarios from the same publisher run 5 to 7 percent base, 3 to 5 percent downside, 8 to 10 percent upside. These are forecasts, not statistics, and they are stated in pesos. A 5 to 7 percent peso gain against a peso that has strengthened against the dollar is roughly flat to negative for a dollar buyer.

For the luxury tier specifically, the only 2026 release we found is a Sotheby’s announcement describing one of the busiest first quarters in company history. It contains no sales counts, no dollar volume, no prices, no days on market and no inventory. It is a marketing statement and cannot be cited as evidence of anything.

The currency, which is where dollar buyers actually lose

The peso strengthened sharply through 2025 to close the year near 18.00 per dollar — published estimates of the full-year move disagree, at close to 14 percent against nearly 16 percent — and the rate stood at 17.46 on 23 July 2026 on the ECB reference rate. That is a large move and it is frequently mangled, so two distinctions matter.

Spot rates and annual averages are different measures. The IRS yearly average was 18.330 pesos per dollar for 2024 and 19.212 for 2025. Those describe the average level across each year, not the start-to-end move, and quoting one as if it were the other produces nonsense.

The second distinction is what the move actually touches. San Miguel is overwhelmingly a cash, dollar-referenced market, so a stronger peso does not lower dollar asking prices. It raises the real dollar cost of everything denominated in pesos: construction and renovation, domestic labour, predial, notary fees and transfer taxes. A US$3,000 monthly transfer that once converted to about 60,000 pesos yields roughly 51,000 today. Budget renovation work at today’s rate, not the rate you remember.

No source measures how much of the 2026 slowdown the currency explains. The purchasing-power arithmetic is solid; the causal link to San Miguel transaction volumes is inference.

What the asking price leaves out

The headline price is not the cost. Two figures here come from primary law rather than from brokerages, which makes them among the most reliable numbers on this page.

Acquisition tax (ISAI) in San Miguel de Allende is not the flat 2 percent widely repeated online. The 2026 municipal Ley de Ingresos sets a progressive tariff topping out at MXN 39,125 plus 4 percent of value above MXN 1.5 million. Applied to the statutory table:

Those effective rates are arithmetic applied to the published tariff, which is deterministic. ISAI is assessed on the highest of registered fiscal value, transaction value or a certified appraisal, so it cannot be reduced by declaring a low price alone.

Total buyer closing costs are less settled. Mexico News Daily puts San Miguel at 5 to 6 percent against roughly 8 percent in Querétaro; Mexperience gives a general Mexican range of 5 to 10 percent; brokerage and publisher estimates for San Miguel cluster between 4 and 9 percent. None of these is derived from recorded transactions.

Annual predial runs on marginal rates of 0.234 to 0.354 percent for urban built property, applied to fiscal value rather than market value, with a 15 percent discount for paying in full in January. The ratio of fiscal to market value in San Miguel is not published anywhere, which means the statutory rate cannot be converted into a percentage of what you paid. That gap is worth stating plainly rather than papering over with an estimate.

What would have to exist for this page to be shorter

A median sale price from closed transactions. A price per square metre from closed transactions. A days-on-market series from the MLS. A published year-to-date closing count that does not contradict itself. An inventory breakdown by price band. A buyer nationality mix. A cash-versus-financed split. A fiscal-to-market value ratio. Any of these from a body without a commission motive.

None of them exists. What exists is one brokerage’s monthly posts, a handful of aggregators recycling listing-portal scrapes, and a national index that excludes this market by construction.

One further reason to hold the single-source data at arm’s length: in February 2026 the same brokerage reported no unusual incidents in San Miguel following cartel leader El Mencho’s death, and minimal travel cancellations. Mexican press reported an arson incident in San Miguel de Allende on 22 February with businesses and universities suspending activities, and a body showing signs of violence found in the municipality on 25 February, within a statewide reprisal wave that reached roughly 23 Guanajuato municipalities. San Miguel is not insulated from Guanajuato. A source that reported that month as uneventful is the same source supplying every price figure above.

Use the numbers. Know what they are.

All prices in USD unless stated. Peso figures are converted at rates noted in the text. Past changes in average sale price are not appreciation and do not indicate future performance.