Class Doesn’t Distinguish Companies, But Its Office Absorption Size Does

In Mexico City, for example, Class A+ buildings accounted for nearly eight out of 10 of the largest absorptions, but six out of 10 of the smallest. The opposite occurred in Class B buildings, which accounted for 16% of the smallest transactions and less than 4% of the largest. The same pattern, although with varying intensity, was observed in Guadalajara and Monterrey.

The gap also persists after accounting for where and when each absorption occurred. After statistically controlling for differences across markets and years, the estimated size of an absorption in a Class A+ building was about 30% larger than in a Class A building. In comparison, in a Class B building it was nearly 18% smaller¹.

This relationship does not mean class determines how much space a company takes, since that decision depends on both the occupier’s needs—such as headcount, work model, or budget—and the building’s characteristics and the space it can offer at that time. By grouping some of these characteristics under a single category, the classification can distinguish among building types, but not identify which characteristic drives the result.

That uncertainty does not prevent a practical consequence from emerging, since an increase in large absorptions does not necessarily benefit all available inventory equally. At the same time, a higher frequency of small transactions does not mean that all buildings compete for them on equal terms.

For that reason, looking only at how many square meters were absorbed can obscure an important part of the market, since the same absorption volume can consist of transactions of different sizes and be distributed differently across building classes.

That distribution also adds nuance to how competition is understood, since a Class B building and a Class A+ building do not necessarily compete for the same demand, but can instead be part of an inventory that meets different needs.

Find more data on office absorptions, properties and markets in Mexico with SiiLA Market Analytics or email us at contacto@siila.com.mx.

***

¹ The analysis covers 4,729 absorptions recorded by SiiLA between 2019 and the second quarter of 2026 in Mexico City, Guadalajara and Monterrey. Suites occupied by the same tenant in the same building and quarter were consolidated into a single transaction, while unidentified tenants were kept as separate records to avoid unsupported groupings. The dataset was audited for exact duplicates, inconsistencies in class or submarket within the same transaction, and variations in tenant names that could affect consolidation; no cases were found that altered the analytical universe. A linear regression was estimated using the logarithm of each absorption’s size, with Class A as the reference category and controls for market and year. Standard errors were clustered by property to account for possible correlation among transactions within the same building (619 properties). The association remained for A+ (30.0%; 95% CI 8.0%–56.4%; p=0.005) and B (−17.5%; 95% CI −29.7%–−3.2%; p=0.018). As sensitivity tests, the analysis was repeated excluding unidentified tenants and removing the largest 1% of transactions in each market, without changing the direction of the result. The distribution by class among the largest and smallest transactions was also examined using 20%, 10%, and 5% cutoffs, with the overall pattern holding. Finally, interactions between class and market were tested, and no statistical evidence was found that the relationship differed across the three cities (p=0.53). The main model had an R² of 0.07, meaning that class, market, and year explain only part of the observed variation in absorption size.