The explanation is structural. Beyond a rent sensitivity (or elasticity) that changes as supply and demand evolve², Mexico’s industrial market operates with five- to ten-year contracts denominated in pesos or dollars, pre-established escalation clauses, and relocation costs that can easily exceed one million dollars for a manufacturer or logistics operator. In addition, much of the sector is dominated by REITs, institutional funds, and major developers whose financial value depends directly on the rents they report.

In that context, rents tend to rise more rapidly when vacancies are low because, in markets with limited available space, even small changes in supply quickly shift negotiating power toward landlords. By contrast, when vacancies increase, maintaining pricing—even at the cost of tolerating more available space—also means protecting valuations. As a result, adjustments tend to appear first outside visible rent levels—through additional free rent periods, property improvements, and greater contractual flexibility—and only afterward, when keeping available square footage becomes more expensive than recognizing the adjustment, do they begin to show up in pricing.

Rigidity, however, is not uniform. Results vary by market, suggesting that it also responds to structural differences between markets: institutional depth, land constraints, manufacturing profile, logistics pressure, border proximity, and integration into global supply chains.

Guanajuato, for example, is the only market with an individually statistically significant elasticity: each additional percentage point of vacancy rate is associated with a decline of more than 5% in real rent. Saltillo shows an even greater magnitude within the panel, though with less statistical robustness, possibly due to a smaller sample size and higher relative market volatility. Tijuana and Mexico City also show relatively elevated sensitivities compared to the rest of the markets.

At the opposite end are Monterrey, Ciudad Juárez, Reynosa, and Querétaro, where the simple relationship between vacancy and rent is much weaker. Monterrey, in particular, appears to operate under a different logic. With the country’s largest industrial inventory and one of the highest quarterly absorption levels nationwide, that market shows a much greater capacity to absorb new supply before transferring direct pressure to rents.

The difference is not limited to geography either. Asset class adds another layer to the adjustment.

In Class A industrial properties, sensitivity to a lagged vacancy rate—that is, increases in available space during previous quarters— is greater than in Class B assets, while the direction of adjustment remains the same in both segments. In other words, asset quality does not change how the market adjusts pricing, but rather how quickly and how strongly that adjustment occurs. Part of that difference stems from the fact that newer facilities tend to attract companies with greater negotiating power, more comparable alternatives, and greater ease of relocating within the same market.

That difference in speed also helps explain current market tensions. 

Between 2024—when nearly eight million new square meters were added to the market, the largest volume in the series since 2019—and 2025—when net absorption began moderating from the peak of the nearshoring cycle—a gap began to emerge that is still not fully reflected in rents. If that trend were to continue consistently and new inventory kept growing faster than absorption, pressure would likely first appear in incentives, leasing timelines, and commercial flexibility, before gradually—and with a lag—transferring into pricing.

For more information on Mexico’s industrial real estate market, visit SiiLA Market Analytics or contact us at contacto@siila.com.mx.

***

¹ Estimated contemporaneous coefficients range between -1.0 and -1.1 in log-linear real rent specifications. In these models, coefficients can be interpreted approximately as percentage changes in rent resulting from an absolute one-percentage-point increase in the vacancy rate. Lagged coefficients for the same variable range between -10.7 and -12.2, depending on the econometric specification, which is where the reference to an effect up to ten times greater in subsequent quarters originates.

² The central structure of the results—limited contemporaneous effect, significant lags and positive pressure from lagged net absorption on rents, with coefficients ranging from 9.4 to 13.1, versus negative pressure of similar magnitude from lagged new inventory—remains consistent when excluding the pandemic period, separating asset classes and individually removing markets and regions from the panel. The main coefficients maintain direction and significance under fixed-effects specifications, and Driscoll-Kraay standard errors, which correct for serial autocorrelation, heteroskedasticity, and cross-market dependence. The Hausman test favors fixed effects over random effects. The analysis also finds no conclusive statistical evidence of a specific vacancy threshold at which pricing power automatically collapses.