A machine learning model is being used to detect fraudulent transactions in a banking system. Explain why recall might be a more important metric than precision in this context. Justify your answer with a real$-$world example of the consequences of false negatives versus false positives.