Thank you, Mr. Chairman.
I think that the problem relates to the state of advancement of the science of modelling rather than to the data as such.
We have a way of looking at statistics in order to assess the impact of the change. The way we do it is to say that if the tariff on a product being sold in a market is moved from its current level down to zero, let's assume that the price in the country falls by the same amount; what will the effect be on the quantity of the product shipped from Canada?
There are two key questions in that analysis. One is whether the price is going to fall by as much as the tariff has fallen; the other is whether the elasticity--that's the nature of the response of the consumer in that country--is accurately specified in the model. Then when you aggregate all these things across a wide range of products, are you sure the supply impact in Canada is going to be accurate?
My colleague Dan Ciuriak may get into a bit more detail in a couple of seconds, but if you were, for example, to imagine that a country were to lower its tariff on Canadian pork, would the Canadian market be able to produce as much pork as that country would buy? Would we be taking product market share away from another country also selling in that market? Would we be taking market share away from domestic producers? The real problem in modelling is not whether we have 20% of the share of the market now or 17%; it is what's going to happen when the price of the Canadian product falls. Have we got our estimate of what the new price is going to be accurately in our model? We do the modelling, but we don't make our decisions on whether to negotiate a trade agreement exclusively on the basis of the model's predictions, because we know there are limitations in that.