That's a very good question. I'm going to tread carefully so I don't veer into policy.
Maybe I can pile onto the model that I described for transportation. In that instance, we found issues that are very similar to what we're discussing here today on the state of energy statistics. In the transportation model—again, it's very early days in rolling this out—we started working among the federal partners involved with transportation information to first identify where the information lies and exactly who was holding what, and then we worked to eliminate duplicative inefficiencies. We—