What we have the clearest view on are the changes that were made in 2016 that affected insured mortgages. It was the same kind of change that we saw put in place in January of this year, with major part being that households looking for a new mortgage loan would have to pass a stress test of increased interest rates so you could tell that they could withstand that. One of the indicators we looked at, which had been giving us cause for concern, was the share of mortgages that had a loan-to-income value greater than 450%. If you're leveraged greater than 4.5 times your income, you're less resilient if interest rates increase or you lose income because you're working fewer hours.
We saw this ratio, which was around 18% at the time of the measure, fall to well under 10%. I think it's probably now between six per cent to eight per cent. That change in the composition of debt, in the new mortgages that are written, means that over time the quality of the debt out there should continue to improve. We expect a similar effect from the most recent changes, which, as you know, are related to uninsured mortgages. It's too early to have those data, but we expect to get them as the year progresses and to be able to follow them over the coming months.