The Hidden Effects of Algorithmic Recommendations

Abstract

Algorithms provide human decision-makers with data-driven predictions, but they can also provide explicit recommendations. I demonstrate that algorithmic recommendations have significant independent effects on human decisions. I leverage a natural experiment in which algorithmic recommendations were given to bail judges in some cases but not others. Lenient recommendations increased lenient bail decisions by 30-40% for marginal cases. The results are consistent with algorithmic recommendations changing the cost of certain decisions. In this way, algorithms can affect human decisions through preferences as well as predictions.