Major projects around the world have long been notorious for cost overruns and schedule delays. In response to the cost overruns, the U.S. Department of Transportation established a task force and took a range of steps to strengthen the stewardship, management, and oversight of major projects at the beginning of the 21st century. Through a classical test of hypotheses, this study first investigates the prevalence and magnitude of the cost overruns of the major infrastructure projects in the U.S. before and after the reform. The before and after comparison finds that the projects experiencing cost overrun have been reduced from 77% to 45% following the reform. Results revealed that the U.S. Department of Transportation has developed measurable outcomes but their success may be limited. To address the cost overruns, conversely, the U.K. adopted Kahneman’s Nobel Prize-winning theory to challenge biases in human judgment and mandated reference class forecasting for major projects at the beginning of the 21st century. Through before-and-after and with-and-without comparisons of major infrastructure projects, this empirical study then examines the practical relevance of reference class forecasting for infrastructure investments. A before-and-after comparison reveals that the average cost overrun declined from 50% to 5% following the introduction of reference class forecasting. A with-and-without comparison also demonstrates that the U.K. surpassed its targeted probability of completing projects within budget by 12% using reference class forecasting, whereas the U.S. underperformed by 23%. Thus, reference class forecasting has engendered notable improvements in estimation in the U.K. Although schedule delays not only result in significant financial implications to project sponsors but also lead to substantial social costs to the public, scant attention has been dedicated to schedule delays. Therefore, this study further investigates the frequency, magnitude, and characteristics of schedule delays occurring during the construction of major infrastructure projects built in the U.S. and the U.K. between 1999 and 2018. A multivariate regression analysis confirms that length of implementation phase and project type plays the most important role in project schedule performance. During extensive efforts were made to project controls in the U.S. and the U.K. over the last two decades, this study also finds, project performance has not improved at all from a schedule perspective. In order to address the issue, this study suggests reference class forecasting to develop more realistic and reliable schedule estimates. This empirical study demonstrates the benefits of supplementing or replacing the current forecasting method. The findings can be used to reduce substantial financial risks for the government as well as social and economic welfare losses for society.