The paper by James G. MacKinnon discusses the application of bootstrap inference in econometrics, highlighting its advantages and limitations. Bootstrap inference allows economists to base statistical inferences on simulated distributions rather than asymptotic theory, leveraging the significant increase in computer performance over the past two decades. The author reviews Monte Carlo tests, various types of bootstrap tests, and bootstrap confidence intervals. While bootstrapping often yields more accurate inferences, it is not always reliable, especially in models with serial correlation, heteroskedasticity, or simultaneous equations. The paper provides detailed examples and simulations to illustrate the performance of different bootstrap methods under various conditions, emphasizing the importance of careful application and the need for caution in certain scenarios.The paper by James G. MacKinnon discusses the application of bootstrap inference in econometrics, highlighting its advantages and limitations. Bootstrap inference allows economists to base statistical inferences on simulated distributions rather than asymptotic theory, leveraging the significant increase in computer performance over the past two decades. The author reviews Monte Carlo tests, various types of bootstrap tests, and bootstrap confidence intervals. While bootstrapping often yields more accurate inferences, it is not always reliable, especially in models with serial correlation, heteroskedasticity, or simultaneous equations. The paper provides detailed examples and simulations to illustrate the performance of different bootstrap methods under various conditions, emphasizing the importance of careful application and the need for caution in certain scenarios.