To go to the question in terms of whether we should rethink trade data and the way we do it, the administrative data we get from Customs is by far the best data we're going to get.
In the early 1990s, for international merchandise trade, we had a project actually called the alternate data sources project. We looked at surveying importers, surveying exporters, or talking to carriers. The conclusion we derived out of that was that we would never get the detail that we have in terms of the province of origin or even the commodity level detail. It just wouldn't be feasible.
With the Customs data, we have this rich data set. It has its problems, but it's better than anything we could ever collect in terms of detail and information. So that gets at the first part of your question.
On the other part, in terms of maybe improving it, we've been working and negotiating quite a bit with our American colleagues. One of the things we'd like to try to get access to is what is called U.S. in-transit documentation. We've been trying for a number of years, and we've succeeded for very small windows. That's what enables us to talk about under-coverage and put some parameters on the size of it. We can then match that transport documentation with export declarations.
We're doing that now in a marine mode. For the ports of Halifax, Montreal, and Vancouver, we're trying to match the shipping manifests, if you will, with the export declarations, to try to get some sense of what implication or what result comes out of these measures that our colleagues at CBSA have implemented, in terms of improving the quality of the data and reducing the under-coverage.