This article evaluates the computational reproducibility of nearly 500 articles published in Management Science before and after the journal introduced its Data and Code Disclosure policy in 2019. Using assessments by more than 700 reproducibility reviewers, the study shows that reproducibility improved substantially under the new policy, especially when data access and technical requirements did not prevent reviewers from running the materials. At the same time, the results highlight persistent barriers, including inaccessible datasets, missing or incomplete code, code errors, weak documentation, and software or hardware constraints. The article provides evidence that journal disclosure policies can materially improve reproducibility, while also identifying areas where such policies can be strengthened.