Mr. Gerry Fitzpatrick:

I think the history has shown that perhaps it didn't cause the ... identification of provision as earlier as the new standard would.

I think at the time ... I think ... and as I've referenced earlier, there have been huge concerns in relation to the kind of, the discretionary impact that previous provisioning had which would actually hide ... could cause people to make provisions in a good time and then wouldn't see the performance of the bank. So I thought it was a good standard in relation to the comparability factor. I think the danger would be that anybody would think that a new standard would actually be able to predict the future. So the new standard ... and unfortunately it's not yet actually EU endorsed so we don't actually know for 2018 whether we'll be able to use that standard. It is an expected loss model and therefore books more of the provisions. I guess, as an auditor, the losses have already been incurred before you're measuring it for provisions purposes, so that the lending criteria are long made before you're actually measuring for impairment. Impairment happens down the road. But an expected loss model certainly will allow some greater prediction of impairment, albeit with significant huge judgments. And if you look at the house price index I referenced earlier, even an expected loss model will only use the date and the time, and won't necessarily say, well, that ... what could predictably happen in the future in relation to asset values. It will use current asset values.