Right. So on a comp in particular, the frequency decline there is we think a very temporary phenomenon and therefore as we stated, we have taken a very close look at our loss picks and we have taken into consideration the unknowns of COVID-19. We’ve weighed that against our – the reduction that we're seeing in frequency when we looked at the accident year and we say in totality – and there's other factors as well, but in totality, there isn't severity in comp that is creating any concern, but it is a matter of frequency down. And then as I went into detail in my opening comments, there is some impact from COVID-19. So we put it all together and we say we're confident in our accident year loss pick. And I might just take the time to say, that is why we know a lot of our competitors, our peers might be putting up bulk reserves or they all are going out with headlines on their cat, the cat number that they're putting up. And we just don't see that as necessary. First off, we don't write international – and I'm moving a little outside of the comp question, but we don't write any international business other than Canada. We don't have political risk exposure, trade credit exposure, entertainment and insurance exposure. So far, we haven't seen any surety losses. From a commercial property standpoint, we have an immaterial amount of policies out there without a virus exclusion. Our loss on our travel accident business is immaterial after reinsurance is applied. So far, GL losses have been immaterial. We're not a main street retail insurer. And you put all those things together and there just is no reason for us to put up a specific COVID-19 number. The only thing that we certainly will have is the work comp losses. And that's where we took a hard look at our accident year loss picks and we're very comfortable with it. And we have no intention of reducing it prematurely. So we will hold that. And if things change with respect to COVID-19, if they change for the worse, we would actually look at maybe increasing it. But right now, given all those reasons, the loss-sensitive nature and the mild proportion of the losses, we're very comfortable with where we're at. And then on auto, the other big one. There, yes, frequency is down. But again, we think it's temporary and we need the reduction we're seeing in frequency to offset the increase we're seeing in severity. So therefore, there too, we would say that we're comfortable, confident in our loss pick for auto. We're not looking at decreasing it or increasing it because it's a different dynamic on that line. And that is that the reduction in severity is helping, along with the rate increases to offset the things we're seeing with severity.