As Jean-Denis mentioned earlier, there's not much information on this type of expense, so the best we had was a study in 2002 carried out by the State of Michigan. The study concluded that on new construction projects, 50% of the cost was materials and 50% was labour, as opposed to historic rehabilitation, where the labour part could go up to 70% of the total cost, while the materials stay at 30%.
We simply took the average cost based on the survey of household spending done by Statistics Canada, and we took half of the total amount that was spent and said that was for materials. Suppose that amount represents only 30%, rather than 50%, what would be the remaining 70%? That's how we came up with the factor of 67%.
It may seem a bit high, which is why in appendix B we provide multiple different scenarios. Obviously the cost of the credit is going to change relative to the different assumptions we make, so in appendices B and C, we present different hypotheses to show what the range of the credit could be if the cost changed.