Social science has come under a firestorm of criticism following the discovery that many findings cannot be replicated. This paper argues that increasing the use of out-of-sample tests would improve the reliability of findings in all fields, including “complex” fields such as macroeconomics....