As well, looking at that, drug prices are of course moving targets. We have probably the best dataset available within the public sector right now with respect to the composition of drugs and drug consumption. That is changing dynamically over time.
Going back to a point I made earlier, which my colleagues will likely strangle me for saying, we produce these forecasts not because we're going to be right. We produce these forecasts to set up a planning environment for parliamentarians. Conceptually, based upon this paper, are we comfortable that this is a balanced estimate of the potential savings? Absolutely. In order to achieve those savings, are there many open questions with respect to the implementation? Absolutely. Will that implementation directly affect the actual savings that could potentially be realized by the federal government? Most certainly.