San Miguel de Allende's Luxury Segment Above $2 Million

12

Resale sales above $2M in 2025

Down from 19 in 2024. One brokerage's MLS view; no public registry exists.

~$34,700

Mean shift from seven fewer $3M sales

Our arithmetic. The full average-price decline was $81,633.

29.6 months

Average inventory, Jan-May 2026

Up from 17.6 a year earlier. No breakdown by price band is published.

0

Luxury figures published for 2026

The only luxury release of 2026 contains no numbers at all.

Twelve resale properties closed above $2 million USD in San Miguel de Allende in 2025. Nineteen did in 2024. That is the entire luxury dataset for this city, and it is the reason most commentary about the top of the market is wrong before it starts.

The segment above roughly $1.5 to $2 million does behave differently from the rest of San Miguel, but not for the reasons usually given. It is not that wealthy buyers are more patient or that trophy property is recession-proof. It is that twelve transactions a year is too small a sample to produce a trend, and every published statement about the segment rests on it.

The only published boundary is $2 million

There is no $1.5 million band in any San Miguel dataset. The single threshold anyone reports is $2,000,000, and it is reported by one firm. Realty San Miguel publishes monthly updates drawn from its own MLS view, and those updates are the origin of every hard number in this market.

San Miguel de Allende has no public transaction registry. The Guanajuato Registro Público de la Propiedad publishes no compraventa statistics, and the local AMPI chapter, checked directly on 23 July 2026, publishes listings and an agent directory and nothing else. Realty San Miguel’s own updates note that AMPI San Miguel has been navigating an MLS system transition and that not all agencies participate in the same MLS, so even the source figures are a partial count of unknown coverage.

So when this page says twelve sales, it means twelve sales visible in one brokerage’s MLS view, unaudited, with a commission motive attached. We use it because it is the best that exists. It is not a census.

Twelve sales is not a market

Two of those cells need flagging. The 2025 closing count was never published: the source reports 404 closings for 2024 and a 24.68 percent increase for 2025, which implies roughly 504, but never states the number. Wherever that figure appears as a count, including in other publishers’ work, it has been derived and presented as reported. The share percentages are our own arithmetic on that derived base. The 2023 count was 425, so 2024 was already down 4.9 percent before the luxury segment contracted.

A drop from 19 to 12 is a 37 percent decline. It is also a difference of seven houses. In a market closing roughly 40 homes a month, seven transactions cannot be distinguished from randomness with any published data. One estate settlement, one developer completing a project, or two buyers closing in January rather than December would move the number by as much.

The arithmetic of the average-price fall

The average resale price fell $81,633 in 2025, from $650,462 to $568,829, a decline of 12.6 percent. Realty San Miguel attributes roughly half of that to fewer luxury transactions and the rest to elevated inventory. The half claim is close to right, and it is worth showing the working rather than asserting it, because the effect is usually overstated in both directions.

Start with a consistency check. Multiply 404 closings by the 2024 average of $650,462 for about $262.8 million of dollar volume, then apply the reported +9.03 percent dollar change and +24.68 percent unit change. The resulting 2025 average is $568,806 against a published $568,829. The three percentages are internally consistent with a 404 base to within about $25, which is the strongest reason to believe the numbers are at least arithmetically real.

Now the luxury effect. The seven missing sales did not remove seven transactions from the count, because unit volume rose. They were replaced by ordinary sales. So the question is what seven $3 million sales add over seven ordinary ones near $500,000, spread across roughly 504 transactions.

At the $3 million assumption the luxury contraction accounts for about two-fifths of the decline, not the whole of it. For seven sales to explain the entire $81,633 they would have to have averaged close to $6.4 million each, which is well above the $2 million to $5 million band the source describes them as occupying.

This is derived arithmetic on a derived denominator. Every figure in that table is ours, not published. The honest summary is that the luxury contraction is a substantial minority of the average-price fall, plausibly a third to three-fifths, and the remainder is mix shift across the middle of the market that no published dataset is granular enough to isolate. Nobody publishes a median from closed sales in San Miguel, for any year, which is exactly the statistic that would settle it.

Days on market at the top: the number does not exist

The most-quoted figure about luxury property in San Miguel is that it sits for 180 to 365 days. We traced it. It comes from a content aggregator citing a Realty San Miguel market update, and that update contains no days-on-market figure at all. The citywide figure of roughly 150 days, from the same publisher, is attributed in part to the INEGI 2020 census, which cannot produce a days-on-market statistic under any methodology.

No days-on-market figure derived from MLS or closed-sale records exists for this city at any price point, and none exists at neighbourhood level. What circulates is an estimate wearing a citation that does not support it, and no buyer should price a negotiation off it.

What can be measured instead

Months of inventory is the one supply statistic reported with a method attached, and in mid-2026 it is stark.

Three caveats belong with that table. First, these benchmarks are not comparable to US markets, because San Miguel is almost exclusively a cash market and absorption never mirrors mortgage-driven demand. No source quantifies the cash share; “almost exclusively” is the strongest published statement available.

Second, the June decline is not absorption. The firm attributes it to withdrawals, expirations and sellers choosing alternative strategies, and describes the fall from 860 to 745 as “a reduction of 81 properties”, which is wrong by 34 units and was never corrected.

Third, and most relevant here, the March 2026 update states that in several price ranges months of inventory are as high as they have been since the period following the 2008 financial crisis, but publishes no per-band figures, and no inventory breakdown by price band exists anywhere in the public record. That luxury inventory specifically is at post-2008 highs is plausible. It is not verifiable.

Negotiating discounts, and why nobody can tell you the number

The sale-to-asking ratio quoted for San Miguel is 92 to 97 percent, implying a discount of 3 to 8 percent. Its own publisher rates it medium confidence and states that Mexico has no published sale-to-list database, which makes it inferred rather than measured. We record it as a claim, not a benchmark.

The one forward-looking segment statement available is a forecast: over the twelve months from July 2026, well-positioned homes at plus 4 to 8 percent, average overpriced listings flat to plus 3 percent, stale luxury homes needing cuts of 5 to 10 percent. That is an aggregator’s model, not a record of transactions, and it is quoted in pesos.

What a buyer at this level actually has is a base rate: twelve sales above $2 million in a full year, against citywide inventory of 745 to 860 homes running at roughly twice normal. The scarcity of buyers at the top is the negotiating position. It does not need a discount percentage to be real.

Who buys at this level

Nobody publishes a buyer-nationality split for San Miguel de Allende. Not one source gives percentages for American, Canadian, European and Mexican buyers. The only official nationality data is the 2020 INEGI census of residents, which records 1,100 US arrivals over five years and an implausibly low 15 Canadians, and which measures residency rather than purchases.

What is known structurally is more useful than a demographic guess. San Miguel prices are quoted and paid predominantly in dollars, a finding from peer-reviewed research on the city rather than from brokerage material, and the market is almost entirely cash. Both facts mean the top of this market tracks North American sentiment and liquidity, not Mexican housing fundamentals. Mexico’s official house price index, which showed 8.7 percent national appreciation in Q1 2026, covers only mortgage-financed homes and is structurally blind to this segment.

The same academic work records an average property price of about $520,000 in 2017, and historic-architecture homes in the centre quoted at around $3,000,000 just before the pandemic. Asking prices above $2 million in Centro are not a recent invention.

What buyers at this level get wrong

Treating “the luxury market” as a market. It is twelve sales. Any statement about its direction, including those on this page, rests on a sample that would not support a conclusion in any other context.

Reading brand marketing as data. The only luxury-segment release published for San Miguel in 2026 came from Sotheby’s affiliates and described Q1 2026 as 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. Zero figures. It sits alongside the same quarter’s market data showing March closings up 13 percent while closed dollar volume fell 26 percent. It is a marketing announcement, not evidence of luxury market strength in either direction.

Buying branded product without resale evidence. The Rosewood Residences are quoted at $1.8 million to $2.8 million, with one 6,125 square foot, four-bedroom unit cited as having sold at $2.2 million and five residences described as unsold. The only page carrying those figures is undated brokerage marketing, so the “five remaining” count could be years stale. There is no published price per square foot, no HOA or amenity fee schedule, and no record of any resale. Whether a Rosewood Residence has ever traded on the secondary market is unknown, and until one does there is no appreciation evidence for branded product in this city.

Confusing Rosewood with Otomi. They are frequently grouped as one prestige area and are not the same place. The Rosewood Residences sit at the Rosewood hotel near Parque Juárez, inside the walkable core. Otomi is a gated residential and equestrian development outside the urban core, reached by cobbled road. Its six listings span $219,500 to $1,712,600, a range that wide because it includes raw land. No price per square metre, lot-size norm, HOA schedule or confirmed drive time is published for Otomi.

Assuming price bands map to neighbourhoods. Los Frailes listings span $236,000 to $3,539,300, from two-bedroom units to hacienda estates. Ojo de Agua, the area most often described as San Miguel’s high-end hillside, is absent from every quantitative price table we could find, and the range in circulation is unattributed brokerage narrative we will not publish as data. It carried thirteen active listings in one brokerage’s colonia directory. At that count a single sale is not a comparable set.

Assuming San Miguel is insulated from Guanajuato. In February 2026, following the death of cartel leader El Mencho, roughly 77 attacks were reported across about 23 Guanajuato municipalities within 24 hours. San Miguel was among them: an arson incident on 22 February, with businesses and universities suspending activities, and a body showing signs of violence found in the municipality on 25 February. Contemporaneous brokerage commentary described no unusual incidents and minimal cancellations. That characterisation is not supportable.

The cost arithmetic changes at the top

San Miguel’s acquisition tax is progressive, not the flat 2 percent repeated across most buyer guides. The 2 percent is only the first bracket. Above MXN 1.5 million the tariff is a fixed MXN 39,125 plus 4 percent of the excess, so the effective rate climbs toward 4 percent.

At the ECB reference rate of 17.4601 MXN per USD on 23 July 2026, MXN 30 million is about $1.72 million, so a $2 million purchase sits above the top row at an effective rate just under 4 percent. ISAI is assessed on the highest of registered fiscal value, transaction value or a certified appraisal under a year old, so it cannot be managed down by under-declaring alone.

Predial is charged on cadastral value, not market value, and the ratio between the two could not be sourced anywhere. That is the most important missing number in the carrying-cost picture, and it matters most where market value is largest. Urban land with buildings on less than 5 percent of its area is taxed under the vacant-land table at 0.438 to 0.705 percent rather than the built rate of roughly 0.35 percent, which is the obvious exposure for large-lot estate property.

On exit, a non-resident seller faces 25 percent of gross proceeds with no deductions whatsoever, or may elect 35 percent of the computed gain. The election wins whenever the gain is under roughly 71 percent of the sale price, and contrary to widespread guidance no Mexican legal representative is required to make it when the sale is by public deed, which in practice is every sale. The principal-residence exemption caps exempt sale price, not gain, at 700,000 UDIs, MXN 6,166,187 at the 10 July 2026 UDI value or about $353,000, and requires Mexican tax residency, so at $2 million it is close to irrelevant. One further trap for long-held property: the building component of cost basis is depreciated 3 percent a year to a 20 percent floor, and where the deed does not separate land from construction, land is deemed 20 percent of cost, which in San Miguel inflates the taxable gain.

2026 so far, and how thin it is

There is no luxury data for 2026. None. The only segment release contains no numbers, and the citywide reporting has published no average or median sale price for any month after January 2026, when it stood at $649,316 on a base of roughly forty homes.

The citywide 2026 volume figure is also unresolved. The May update reports 197 closings through May against 190 in 2025; the June update reports 186 through June against 220. A year-to-date count cannot fall from 197 in five months to 186 in six, and the prior-year comparative shifts too. Neither post acknowledges the discrepancy, so no 2026 volume figure should be treated as settled and we are not publishing one.

What can be said about the top of the market in 2026 is this. Inventory is running at roughly twice the level the reporting brokerage calls normal. The March data shows more closings at materially lower aggregate value. The only luxury communication of the year contains no evidence.

Currency is the fourth factor, and it is routinely mangled. Because asking prices here are quoted in dollars, a stronger peso does not cut headline prices; it raises the dollar cost of peso-denominated items, which at this end means renovation, labour, predial and closing costs. The ECB reference rate was 17.4601 on 23 July 2026. The IRS yearly averages were 18.330 for 2024 and 19.212 for 2025. Those are averages across a whole year, not spot levels, and the two are frequently quoted interchangeably. They should not be blended.

What is not knowable

The gaps that bear directly on this segment: no public transaction registry and no independent check on any figure here; no published 2025 closing count; no median sale price from closings in any year; no days-on-market data at any price point; no measured sale-to-list ratio; no inventory breakdown by price band; no 2026 luxury figures of any kind; no resale record or fee schedule for branded residences; no quantitative price data for Ojo de Agua; no cadastral-to-market value ratio; no buyer nationality mix; no measured appreciation by neighbourhood.

A publisher with a commission would fill those with estimates. We would rather name them, because at twelve sales a year the difference between a measured figure and a confident one is the whole of the analysis.

All figures USD unless noted. Past performance does not indicate future results.