We apply an unsupervised machine learning algorithm to revisit legislative lags of U.S. tax reforms and show that at least two lags have been longer than previously identified. Our approach offers an alternative way to approximate U.S. tax foresight, given that the relationship between tax...