Homicide escalation did not alter Mexico’s long fertility decline
Since 2006, Mexico has lived through a severe rise in homicidal violence tied to national policy shifts and international drug trafficking pressures. The new analysis asks a deceptively simple question with major population-health consequences: did that violence change how many children women had, or how many they wanted to have?
The answer is restrained. Across 2,443 municipalities and two decades of population data, fertility trends continued their long descent with remarkable continuity, even as homicide rates climbed. The study found only very small negative associations, and no overall effect of homicides on fertility during the period examined. Birth rates fell, but they did not appear to bend sharply in response to homicide shocks.
The same pattern held when the analysis moved from municipal registries to women’s stated intentions. Among 6,341 women in the Mexican Family Life Survey, higher homicide rates were not associated with shifts in fertility desires. Education, age, and parity did not change that null result.
Municipal fertility measures stayed on their prior trajectory
Investigators used administrative data on births, deaths, and homicides for municipalities across Mexico between 2000 and 2020. They examined the total fertility rate, the crude birth rate, and birth counts, then applied fixed-effects models and a staggered difference-in-differences estimator to estimate the effect of homicidal violence over time. This design matters because it helps separate local violence exposure from slow-moving municipal differences that can otherwise distort fertility comparisons.
The result was not a dramatic fertility response to violence, but a slight negative association that never rose to a meaningful population-wide shift. In practical demographic terms, that means a municipality experiencing more homicides did not show a clear, statistically important departure from the broader national pattern of fertility decline. The Mexican fertility transition kept moving in the same direction it had already been traveling.
Violence can reshape daily life, but fertility decisions sit at the intersection of biology, partnership stability, economic expectations, and access to care. A spike in homicide may increase fear, disrupt routines, and strain household planning, yet those pressures did not translate into a measurable aggregate reversal in births across the country.
Fertility desires among women remained stable despite local violence
The second arm of the analysis examined whether violence altered what women wanted for their families. Using random-intercept and fixed-effects models in the Mexican Family Life Survey, the study linked municipality-level homicide rates to fertility desires reported by 6,341 women between 2002 and 2012.
Here too, the signal was absent. Higher homicide rates showed no association with fertility desires, and the finding remained consistent across education, age, and parity. Women with different reproductive histories and social positions did not diverge in their stated preferences as local violence increased.
A single sentence captures the core clinical-demographic message: homicidal violence did not meaningfully move fertility intentions.
That absence of association may reflect the fact that fertility desire is shaped by more durable forces than short-term security shocks. Preferences about childbearing often evolve from long-run family norms, partnership status, economic resources, and existing reproductive goals. Even in a setting of sustained violence, those deeper determinants may have outweighed the immediate influence of homicide exposure.
Why the fertility response may have been muted
The study does not claim to identify a biologic mechanism, but the demographic logic is straightforward. Fertility outcomes are not only a reaction to danger; they are also the product of accumulated social change. Mexico entered the homicide surge amid an ongoing fertility decline, so the violence arrived on top of a trajectory already moving downward for reasons that likely included education, urbanization, contraceptive access, and shifting family expectations.
When a structural trend is already strong, an additional shock may not leave a visible imprint in national fertility indicators. Municipal birth counts can absorb many pressures at once, from migration and union stability to labor markets and access to reproductive services. The net effect can be surprisingly flat even when local insecurity is severe.
The authors’ use of both administrative and survey-based data strengthens the interpretation. Birth records answer one question: did fertility change? Survey responses answer a second: did women’s desired family size or childbearing intentions change? Both pointed in the same direction, which makes a hidden, widespread fertility response less likely.
What the null finding means for population health research
This work adds an important counterweight to a literature that often focuses on how violence affects mortality, displacement, mental health, and social functioning. Fertility is a central demographic behavior, and it is reasonable to expect violence to disturb it. Yet broad municipal data from Mexico show continuity rather than disruption.
The absence of a strong effect should not be misread as evidence that violence was harmless. Homicide can still affect women and families through trauma, insecurity, economic instability, and constrained mobility. The point is narrower: those harms did not translate into a detectable population-level shift in fertility rates or fertility desires over the study period.
For demographic researchers, the finding is a reminder that not every large social shock reshapes every downstream behavior in the same way. For policy makers, it suggests that fertility trends in Mexico during these years were driven more by preexisting demographic change than by the homicide surge itself.
A deeper unanswered issue remains embedded in the data structure rather than the results: local violence may have affected specific neighborhoods, subgroups, or time windows in ways that municipal averages cannot fully reveal. The study’s aggregate estimates are strong enough to describe the country-level pattern, but finer-grained effects could still have been missed if they were small, highly localized, or short-lived.