Currency Risk for Foreign Buyers in San Miguel de Allende

17.4601

USD/MXN spot, 23 July 2026

European Central Bank reference rate. A widely circulated figure of 17.51 for the same day was wrong.

19.6%

Rise in the dollar cost of a fixed peso bill since the start of 2025

Arithmetic from the implied start-2025 rate to the 23 July 2026 spot. It applies to renovation, labour, predial and closing costs, not to asking prices.

18.330 vs 19.212

IRS annual average MXN per USD, 2024 and 2025

The averages moved the opposite way to spot. The two measures are different and are routinely conflated.

None

Sources measuring the currency effect on San Miguel sales volume

The purchasing-power arithmetic is sound. The causal link to transaction counts is inference, not data.

Every appreciation claim published about San Miguel de Allende is nominal, and most are nominal in pesos. Not one of the brokerages, aggregators or forecasters producing them restates the figure in the currency the reader actually holds, net of the exchange rate over the same window. That omission is not an oversight of arithmetic. It is what happens when the publisher earns a commission on the transaction and the honest restatement is unflattering.

The arithmetic is not difficult. It is set out below in full, along with the parts of it that cannot be done because the underlying rates are not in any record we can source.

The measurement trap: spot rates and annual averages are different things

Two series describe the peso and they moved in opposite directions over the same period. Both are correct. Almost every published discussion of the peso in a property context confuses them.

Read the spot column and the peso strengthened dramatically. Read the average column and the peso was 4.8 percent weaker on average in 2025 than in 2024. There is no contradiction: 2024 began with a strong peso and ended with a weak one, so its average is low, and 2025 began from that weak close and worked back down, so its average is high. The direction of the annual averages is the opposite of the direction of the spot move, for the same two years.

The practical rule is that a spot rate answers what a transfer converts at, and an annual average answers what a tax authority will use to translate a year of income. Neither answers the other’s question. Any page that reports a strengthening peso from spot and a weakening peso from averages in adjacent sentences does not understand its own data.

Even within the spot series the size of the 2025 move is reported inconsistently. One outlet describes the peso as appreciating close to 14 percent that year, another as nearly 16 percent. Part of that gap is genuine disagreement about the start and end points, and part is that the same move produces two different numbers depending on which currency sits in the numerator. Working back from the year-end close near 18.00, the published figures imply a starting rate somewhere between roughly 20.5 and 20.9 pesos per dollar; from an implied 20.9 to 18.00, the dollar lost about 13.9 percent of its peso value and the peso gained about 16.1 percent against the dollar. Both describe one event. We publish the disagreement rather than pick a winner.

Where the exchange rate actually bites

San Miguel de Allende is an unusual case, and the reason most generic Mexican currency advice fails here. Academic work on the city finds that prices are quoted predominantly in dollars rather than pesos, and the market is described by its own participants as almost exclusively cash. So the asset behaves as a dollar asset sitting inside a peso jurisdiction.

One line in that table is in dollars. Everything the state charges, and everything a local trade charges, is in pesos. A stronger peso therefore cannot make a San Miguel house cheaper. It can only make the second column more expensive.

The dollar buyer’s arithmetic

From the implied start-2025 rate of 20.88 to 17.4601, a dollar buys 16.4 percent fewer pesos. Inverted, that is the number that matters: a fixed peso obligation costs 19.6 percent more dollars than it did at the start of 2025.

Worked against statutory figures, on a property with a recorded value of MXN 10,000,000:

The tax bill did not rise. The statute is unchanged and San Miguel has now held predial rates flat for four consecutive years. What changed is the dollar cost of paying it, by about a fifth, without any legislature voting on anything.

The renovation line is where this becomes material rather than interesting. Predial is a few hundred dollars either way. A renovation, which in a market of older central stock is frequently the larger half of the project, moved by $18,762 on a 2 million peso budget. A buyer who priced the work at 2024 rates and is spending in 2026 is short by roughly that proportion. The same mechanism applies to any dollar income funding peso living costs: a US$3,000 monthly transfer that once converted to about 60,000 pesos yielded roughly 51,000 at the rate prevailing in late January 2026.

The Canadian buyer: two legs, and we can price only one

A Canadian buying in San Miguel de Allende is exposed twice. The house is quoted, negotiated and settled in US dollars, so the first conversion is Canadian dollars into US dollars. Everything after closing is in pesos, so the second is US dollars into pesos. The legs multiply rather than add.

We can price the second leg exactly, because it is the 19.6 percent above. We cannot price the first, because our sourced record contains no Canadian dollar series against either currency, and a rate lifted off a converter and presented as analysis is precisely what this publication exists not to do.

What can be shown is the shape of the compounding. If the Canadian dollar had fallen 10 percent against the US dollar over the same window, the Canadian’s peso costs would have risen not by 19.6 percent but by 31.5 percent, because 1.196 multiplied by 1.10 is 1.315. That 10 percent is an assumed input chosen to illustrate the multiplication. It is not a claim about what the Canadian dollar did, and it should not be quoted as one.

The consequence is structural rather than numerical. A Canadian modelling only the purchase price has modelled the leg that carries one conversion, and ignored the leg that carries two. Nobody publishes the second calculation, and the FX brokers who rank for these searches sell conversion services rather than model the triangle.

The European buyer: three currencies, two conversions

The European position is the Canadian one with an additional gap in the record. Capital is in euros. The asset is priced in dollars. The obligations are in pesos. Neither euro leg appears in any source we hold, so no euro figure is published here at all, including the euro value of the residency thresholds.

The structural point survives the missing data. A European buyer’s return is the property’s dollar performance, multiplied by the euro against the dollar, with a separate peso exposure running continuously against carrying costs and landing again at sale, where the tax is computed on peso values. Three of those four terms are unobserved in the published record for this market. A forecast quoted in pesos and read in euros has passed through two conversions nobody has measured.

Restating the published forecasts in dollars

This is the calculation the incumbents do not run on their own numbers.

The forecasts in circulation for San Miguel de Allende are stated in pesos. One aggregator gives full-year 2026 scenarios of 5 to 7 percent base case, 3 to 5 percent downside and 8 to 10 percent upside for the best walkable homes. The sourced consensus for the peso at end-2026 clusters at 18.75 to 19.00, weaker than the 17.4601 prevailing on 23 July 2026. Converting a peso-denominated gain at a weaker peso gives this:

Every scenario except the top of the upside case converts to a dollar loss. That is the entire content of this page in one table, and it is why a commission-earning publisher does not print it.

Two cautions apply, and they cut against overreading the table as much as against ignoring it. The property forecasts are aggregator estimates rather than measured statistics, from a publisher whose figures have shifted between revisions of the same page. The exchange-rate consensus is also a forecast. Multiplying two forecasts produces a third forecast, not evidence.

The backward-looking restatement is more awkward still for the incumbent record. The same aggregator reports San Miguel asking prices about 9.5 percent higher in nominal pesos than a year earlier. Because the peso strengthened across most of that window, restating a peso asking-price rise into dollars makes it larger, not smaller. Meanwhile the only closed-sale series anyone publishes shows the average dollar sale price falling 12.6 percent in 2025. Converting the asking-price claim into dollars therefore widens the gap between asks and closes rather than narrowing it. We cannot put a precise figure on the restatement because our sourced record contains no mid-2025 spot rate, only the calendar-2025 endpoints and the 2026 series. The direction is unambiguous even where the magnitude is not.

The exit: a tax computed in pesos

The deed records a peso value. Mexican income tax on the sale is computed in pesos. A foreign owner thinking in dollars can therefore generate a taxable gain in a year the dollar price did not move.

Buying at US$500,000 with the peso at 17.4601 records about MXN 8,730,050. Selling at the same US$500,000 with the peso at 19.00 records MXN 9,500,000. That is a peso rise of about 8.8 percent on a dollar gain of zero. Note the direction: it is a weakening peso after purchase, not a strengthening one, that manufactures the phantom gain, and the sourced consensus points to a weaker peso by the end of 2026 than the rate available today.

Whether the phantom gain costs anything depends on which method applies:

• The non-resident default, Article 160 of the Ley del Impuesto sobre la Renta, is 25 percent of gross proceeds with no deductions whatsoever. Currency is irrelevant to it, and so is whether the sale made money at all. A sale at a dollar loss still produces this bill.
• The elective method, also Article 160, applies the top Article 152 rate of 35 percent to the computed gain instead. Here the peso arithmetic is decisive. The election does not require a Mexican legal representative where the sale is by public deed, contrary to widespread guidance, because the statute disapplies that requirement in the fourth paragraph.

Three basis mechanics compound the effect and none of them is about currency. The construction element must be depreciated 3 percent for each year of ownership, floored at 20 percent of initial cost, whatever the building’s condition. Where land and construction are not separated in the deed, land is deemed 20 percent of cost, which is usually unfavourable in San Miguel, where land is a high share of value. Only capitalised improvements are deductible, never maintenance, and only with proper facturas. Indexation of basis is by Mexican inflation, which has no relationship to the dollar.

The principal-residence exemption is a peso number too, and its dollar value therefore drifts with the exchange rate. Seven hundred thousand UDIs came to MXN 6,166,187.30 at the UDI published for 10 July 2026, which converts to roughly US$353,000 at 17.4601. It caps the exempt sale price rather than the gain, it belongs to Mexican tax residents rather than to foreign nationals as such, and the UDI is inflation-indexed and republished twice monthly, so it must be recomputed at the actual sale date.

What holds the peso up, and what could move it

The main support under the strong peso is the interest-rate differential. Banxico’s policy rate stood at 7.0 percent against the Federal Reserve’s 3.75 percent as of 6 March 2026, a gap of 3.25 percentage points. That differential has been compressing: Banxico has cut from 11.25 percent to 7.00 percent, and a narrowing spread is the principal argument for peso weakness later in 2026. Remittances to Mexico fell to US$61.8 billion in 2025 from US$64.7 billion in 2024, a 4.5 percent decline and a peso-negative factor.

Volatility in 2026 has been low. The rate ran between 17.1061 on 23 February and 18.1490 on 30 March, averaging 17.4452 over six months, a band of roughly 6 percent. Low volatility removes timing risk for a buyer converting a lump sum, and locks in the stronger peso for everyone paying peso costs from dollar income.

We publish no forecast of our own. The consensus figures quoted above are other people’s.

The currency line item at closing

One item on a San Miguel closing statement exists purely because of the exchange rate. A local brokerage’s buyer FAQ describes an exchange-rate fluctuation reserve of US$1,000 to US$2,000 on the final transfer, mostly refundable to the buyer. It is not fixed by any statute, the page carries no date, and the source has a commission motive, so treat the range as market colour rather than a rule. It is included because it is a real cash requirement at closing that no statutory analysis captures, and because its existence is itself the point: the transaction settles in dollars and the obligations discharge in pesos, and somebody has to absorb the gap between the two on the day.

What no source measures

• The effect of the exchange rate on San Miguel transaction volumes. The arithmetic above is sound. The causal link from it to the 2026 slowdown is inference. No source measures the relationship, and the brokerage reporting the slowdown attributes it chiefly to conditions in the US resale market.
• Any Canadian dollar or euro series against either the dollar or the peso, for this market or generally, in the record we hold. No CAD or EUR figure appears on this page for that reason.
• A mid-2025 spot rate, without which the aggregator’s year-on-year peso asking-price claim cannot be restated in dollars with a number attached.
• Retail conversion spreads and wire costs actually paid by buyers here. Every rate on this page is a reference or interbank rate. Nobody receives one.
• The ratio of cadastral to market value in San Miguel, without which the statutory predial rates cannot be turned into a percentage of what a buyer paid, in any currency.
• A San Miguel price index of any kind. There is no repeat-sales index, no hedonic index and no MLS time series for this city, so there is no measured appreciation rate to restate into a foreign currency in the first place. Everything above operates on other people’s estimates.

The last of those is worth sitting with. This page restates figures that are themselves weak: asking-price scrapes, aggregator forecasts, and closed-sale averages from a single brokerage’s view of a non-public MLS. Applying accurate currency arithmetic to an inexact price series produces a currency-adjusted inexact price series. It is still better than the alternative, which is quoting a peso number to a dollar reader and letting them assume the two are the same.

This is published research, not legal or tax advice. Your notario is the legal authority on your transaction. We have no financial interest in whether it closes.