On July 1, the S&P/BMV IPC closed at 67,248, up 0.42% on the day and back above the 67,000 line it had surrendered twenty-four hours earlier. The peso firmed to around 17.5 per US dollar. Measured against what investors had braced for, a tariff wall and a broken treaty, this passed for a good day. The index had already recovered 3.2% from its late-June low near 65,142. The verdict from the trading desks was close to unanimous: less harsh than feared.
We think the relief is the error.
The United States did not tear up the USMCA on July 1. It declined to renew it. That distinction carries the entire story, and the market has read it backwards. The treaty stays in force. What ended was the feature that made it worth anything to the single activity Mexico’s equity case is built on: committing capital for a decade at a time. By July 7 the index had already handed back most of the bounce, closing at 66,675, down 1.17% on the day and still 6.9% below its 52-week high of 71,601. The relief was real. It was also thin.
What actually happened
Article 34.7.4 required the three governments to meet on July 1, 2026 and decide whether to extend the agreement’s 16-year term. Mexico said yes. Canada said yes. US Trade Representative Jamieson Greer said the United States “did not agree to renew the USMCA in its current form.”
One refusal was enough to change the machinery. The agreement now faces a joint review every year through July 1, 2036, the date it sunsets if the three parties never agree to extend. Between now and then nothing mechanical changes: the same rules of origin, the same zero tariff on compliant goods, no disruption to a trade relationship that carried roughly USD 545bn of Mexican exports into the United States last year. The three governments can still sign a fresh 16-year extension at any point before 2036, and doing so requires only written agreement, not a reopened negotiation. The catch is that they now have to reach for it every July, across the table from a US administration that has already shown it treats the calendar as leverage.
The original design asked North America to make one decision every six years. The new one asks for a decision every twelve months, nine times over, before the clock runs out. There is a second erosion hiding inside the mechanics. Washington is now running a separate bilateral track with Mexico, with a negotiating round set for late July, while Canada, whose trade minister has prioritized the fights over steel, aluminum, autos, and lumber, had not yet begun substantive talks. A trilateral treaty is quietly being managed as two bilateral relationships.
The relief is priced on the wrong number
Markets rallied because the tariff stayed benign. Compliant goods still cross the border at zero. That was the number the desks were watching, and on that number they were right.
It was the wrong number to watch.
Nearshoring, the thesis underneath three years of “Mexico wins the trade war” research, is not a tariff bet. It is a capital-expenditure bet. A manufacturer moving a plant from Asia to Nuevo León or the Bajío is underwriting a facility across ten to fifteen years of output. That arithmetic only closes if the rules hold across those ten to fifteen years. A treaty that can, at least in theory, be unwound at any one of nine consecutive annual reviews does not supply that. It supplies a countdown.
So “the deal survived” is not the reassurance it sounds like. Survival on a twelve-month leash is worse for long-horizon investment than a clean multi-year extension would have been, even one signed under duress. A tariff is a cost, and a cost can be absorbed, passed through, or engineered around. Open-ended uncertainty cannot be put into a model, and capital that cannot be modeled does not get deployed. As CSIS argued in a recent study of Mexico’s investment gap, when the same rules can be reinterpreted over time, firms cannot reliably model costs or returns, and money that could fund expansion sits idle as a hedge instead. Each annual review becomes its own forum for pressure, negotiation, and conditionality.
The damage is already in the data
This is not a projection. It is already sitting in the 2025 investment numbers, and the market walked straight past it.
The headline reads well. Total FDI rose about 16% over the past year to roughly USD 44bn, close to a record. One layer down, the figure inverts. Most of that gain was profit from existing operations reinvested in place, not new money entering the country. The greenfield line, meaning new projects and new plants, the literal physical footprint of nearshoring, fell 50% in announced value and 16% in the number of projects. New commitments announced under the nearshoring label dropped 78% year over year in the first quarter of 2026.
Widen the lens to total fixed investment and the weakness is not confined to foreigners. It fell roughly 10% in 2025, with public investment down more than 26% and private investment off about 2%. The greenfield damage concentrated exactly where the annual reviews will aim: autos, electronics, and machinery, the sectors whose rules of origin sit at the top of Washington’s list. Companies are not waiting for a tariff to land. They are pricing the uncertainty already, and they are pricing it by not building.
Record FDI and new investment are not the same thing. The money that would build the next decade of nearshoring is precisely the money now pulling back.
What the reviews will actually fight about
These reviews will not be ceremonial. Washington has already named its targets: tighter automotive rules of origin, Chinese content and transshipment routed through Mexico, labor enforcement under the Rapid Response Mechanism, and the running disputes over steel, aluminum, and agriculture. A US-Mexico bilateral round is set for the week of July 20 in Mexico City to fix Washington’s priorities for this cycle.
Autos are the pressure point, and Mexico’s exposure there is not marginal. The country now supplies 46.25% of every auto part the United States imports. Step back to the whole economy and the concentration is starker: about 82% of Mexican exports go to one customer, the United States, and foreign trade runs near 75% of GDP.
There is no diversifying out of that inside a decade of annual reviews. The concentration is what gives the new structure its teeth. Every July becomes a negotiation in which Mexico brings one dominant buyer to the table, and that buyer writes the agenda.
This lands on top of a growth problem we flagged in “Trapped at 6.50%.” Banxico is already boxed between sticky services inflation and a stalling economy. A decade of annual trade uncertainty keeps a risk premium embedded in the peso and in long MBono yields, which makes every rate cut more expensive to justify and leaves the economy with less monetary cushion exactly as investment softens.
What we are watching
Our base case is still that the USMCA gets extended before 2036. Mexico wants it, Canada wants it, and the US manufacturers who depend on integrated supply chains want it. The extension mechanism needs only signatures, not a fresh treaty. We are not forecasting a lapse.
We are calling the relief rally early. The market repriced the tariff risk, which genuinely improved, and ignored the certainty risk, which got materially worse. Those are two different trades, and only one of them cleared on July 1.
Two things are worth watching. In the near term, the week of July 20 bilateral: the tone and the length of Washington’s demand list will show whether this first annual cycle is procedural or confrontational. Structurally, watch the greenfield FDI line rather than the headline FDI number. If new-project value keeps sliding through 2026, the nearshoring thesis is not on pause. It is being repriced in real time, one plant that never breaks ground at a time.
For equities, the read-through is selective, not blanket-bearish. The exporters whose entire multiple rests on the nearshoring story, industrial FIBRAs, auto-parts suppliers, the border-logistics names, carry a discount they are mostly not paying yet. The domestic-demand names, the ones that sell to Mexican consumers rather than to a trade agreement, look relatively better than they have in three years.
That is the trade the relief rally missed.