Referral and job performance: evidence from the Ghana colonial army

Using data compiled from army archives, we test whether the referral system in use in the British colonial army in Ghana served to improve the unobserved quality of new recruits. We find that it did not. If anything, referred recruits were more likely to desert and be dismissed as inefficient or unfit. We find instead evidence of referee opportunism. Army recruiters may have been aware of this problem by insisting that referred recruits have better observed characteristics.