I think you can be confident that we look at the micro-data, the segregated data, when we try to understand what the interest rate effect is, because the distribution of who is indebted and what kind of assets they have really matters. It matters to the macroeconomic outlook. We've tried to do that in our model changes.
What we also look at in the context of the financial system review are scenarios where it's not just interest rate increases that are happening; it's also perhaps a context in which instead of interest rate increases you end up with a spike in unemployment and things like that. It's there that you see other reactions such as defaults and arrears that would have stronger macroeconomic impacts. According to our modelling, the kinds of interest rate increases we look at in a cycle like this one don't have enough of a reaction in arrears and bankruptcies to create a macroeconomic issue.
That's not to say we don't take it into account, but it's just not big enough. It's really in the tail event, a recessionary situation, or something like that where you would see a remarkable aggregate increase in arrears. That doesn't mean it's not difficult for individuals who are facing that situation. It's just that when you're targeting inflation, it's something that we take into account, but it's not driving the results.