This paper examines the socio-economic consequences of teenage motherhood for a cohort of British women born in 1970. We apply a number of different methodologies on the same dataset, including OLS, a propensity score matching estimator, and an instrumental variables estimator, using miscarriages as an instrument. We bound the biases introduced through IV due to non-randomness, and misreporting of the instrument. Our results are sensitive to the methodologies used. Taking only observed characteristics into account, the effects of teenage motherhood appear large and negative. The pathways are through bigger family size, and negative labour market outcomes for the mother and her partner, and are mitigated by transfers from the state through the British benefit system. Our IV estimates show that almost all these effects are reduced to zero once unobserved heterogeneity is taken into account. However our IV bounds show that biases introduced by non-randomness and misreporting of our instrument could be responsible for all of this apparent reduction in effects.