Abstract:
Are estimates typically closer to the true parameter value when those estimates are published in highly-ranked economics journals? Using 14,387 published estimates from 24 large literatures, we find that, within literatures, the mean and variance of parameter estimates have little or no correlation with journal rank. Therefore, regardless of what the true parameter value is that a literature is attempting to estimate, it cannot be that estimates in higher-ranked journals are on average noticeably closer to it. We discuss possible explanations and implications.
