This paper studies the relationship between the age of enrolment in Peru’s conditional cash transfer programme, JUNTOS, and the foundational cognitive skills of a sample of children aged between 5 and 12 years old. Using a difference-in-differences approach and exploiting within-household variation, we show that younger siblings in recipient households display significantly higher levels of inhibitory control than their older counterparts (0.11 standard deviations), having benefited from the programme for the first time at a relatively earlier age. In high-income countries, this behavioural trait has been linked to later-life outcomes such as job success, physical health, and even reduced risk of criminality. Conversely, we find little evidence that enrolment age is associated with long-term memory, working memory, or implicit learning. Employing a threshold estimator, we show that relative gains in inhibitory control are most clearly defined where a child benefits from the programme before they reach 80 months of age (6.7 years). In an extension to our main results, we then conduct mediation analysis, demonstrating that a small but meaningful proportion of this benefit (6.5%) operates through changes in the probability of the child’s timely entry into primary school.