This paper examines the impact of South Africa's state old age pension on the labour supply of working age members of pension-receiving households. A range of identification strategies are employed in an attempt to make full use of recent labour force survey data. Results suggest fairly modest, negative impacts on participation and employment on average, but a pronounced, positive migration impact. As such, they marry the results of Bertrand, Mullainathan and Miller (2003) and Posel, Fairburn and Lund (2006), using more recent (and nationally-representative) data. Concluding sections consider implications, and question whether gender-differentiated effects are grounds for rejecting the income-pooling hypothesis.