Researchers cut South Africa's hiring gap with a one-page reference letter

South Africa's unemployment rate has stayed among the world's highest for decades, and the usual policy toolkit — skills training, wage subsidies, job-search assistance — treats the job-seeker as the thing to fix: not skilled enough, not searching hard enough, not subsidized enough to be worth the risk of hiring.

Economists Martin Abel, Rulof Burger and Patrizio Piraino targeted a different half of the transaction: employers screening large numbers of applicants from a labor market where almost no one has a verifiable work history couldn't tell a reliable worker from an unreliable one, and defaulted to rejecting the whole pool rather than gambling on any one candidate.

The researchers ran two linked experiments: a resume audit sending otherwise-identical applications with and without a one-page reference letter from a previous employer, and a field experiment encouraging real job-seekers to obtain and attach one. The letter contained no new skill, credential, or subsidy — just a former employer's plain assessment of the person's reliability, made visible to a screener who previously had nothing to go on.

Gave job-seekers a one-page reference letter from a former employer, making existing reliability visible to screeners who had nothing to go on, at essentially no cost.

The letter carried information only a past employer could supply, which résumés couldn't convey in a market with no verifiable work history.

It changed nothing about the candidate's actual skill, only what the screener could see, so it worked because the failure was informational.

The effect was strongest for women, where the information gap was widest, showing it fixed the exact invisibility problem.

It cost almost nothing, so it scaled without training or subsidies.

When buyers can't tell good from bad, find the cheapest credible signal that only a genuine past counterparty can supply, and make it standard.

Reference letters raised callback rates by more than 60%. Among job-seekers encouraged to obtain and use one, employment rates rose 49% — driven overwhelmingly by women, for whom the information gap had been widest — closing a meaningful share of the sample's gender employment gap with no training, subsidy, or change to the applicants themselves.