Well, again, I think that's an empirical question. I think we do have some research from the U.S. that does show the response rate goes down, as Mr. Drummond indicated. I'm also not too concerned about the decrease in the response rate.
The U.S. Census Bureau actually came to the conclusion that if they expended more resources, they could still get the same reliable data on a voluntary basis just as they did on a mandatory basis.
So if it's a question of data reliability, the U.S. Census Bureau certainly came to a different conclusion.