This paper investigates how the permanent departure of the father from the household affects children's school enrolment and work participation in rural Colombia. Our results show that departure of the father decreases children's school enrolment by around 4 percentage points, and increases child labour by 3 percentage points. After using household fixed effects to deal with time-invariant unobserved heterogeneity, and providing evidence suggesting strongly that estimates are not biased by time varying unobserved heterogeneity, we also exploit an interesting feature of our setting, a conditional cash transfer programme in place, and show that it counteracts the adverse effects. This, and other pieces of evidence we give, strongly suggests that the channel through which departure affects children is through reducing income. It also highlights the important safety net role played by such welfare programmes, in particular for very disadvantaged households, who are unlikely to find formal or informal ways of insuring themselves against such vagaries.