Consistently, tenants who have not moved have occupied the same space for an average of 21 quarters, compared with just over 13 quarters for those who have. Beyond this descriptive difference, a logistic regression model shows that each additional quarter of occupancy increases the probability of staying, with an almost zero likelihood that this result is due to chance¹, even when controlling for building class.

If building class does not alter the result, then building quality does not meaningfully affect tenant retention.

According to the data, the share of tenants that have not moved since occupying their space within the period analyzed is virtually the same across office classes—79% in Class A assets, 80% in A+ and 83% in B—suggesting that the decision to stay is driven less by building quality than by the operational, financial and logistical costs of replacing the occupied space.

Taken together, the analysis indicates that the office market does not operate as a system of constant replacement, but rather as one of prolonged occupancy, where the friction to change space outweighs differences in asset quality.

This behavior does not remain at the individual level; it carries through to overall market performance, as limited tenant turnover ultimately conditions absorption and vacancy, leading to gradual adjustments.

In that context, with contained turnover and a more staggered delivery of new inventory in recent years, net absorption has grown faster than gross absorption—11% versus 9% compound annually over the past six years—putting gradual downward pressure on the vacancy rate. While this pattern is not new, it has been amplified by recent market conditions, particularly after the pandemic, which slowed both supply and demand before a more gradual recovery.

This perspective—in which retention outweighs replacement in shaping market dynamics—redefines how competition works: it is no longer just about attracting tenants, but about understanding what keeps them. Thus, value lies not only in offering better buildings, but in reducing the explicit and implicit costs of occupying them.

To explore this dynamic further and access detailed market data, explore SiiLA Market Analytics or contact us at contacto@siila.com.mx.

***

¹ Estimated in R using a logistic regression model (binomial glm), where the dependent variable is a binary indicator of whether a tenant remains in the same space (has not moved), and the main explanatory variable is time in occupancy (in quarters), incorporating building class as a control variable. The significance of the time coefficient was evaluated using a z-statistic and found to be highly significant (p < 2e-16). It does not describe individual cases or causal relationships.