The UK's sero-surveillance data is out, & it shld be quite big news. It finds that at least 17.5% had CV (at least in the sense that some may have had it without antibodies) in London by mid-April. Seems quite plausible London's now close to herd immunity. https://www.gov.uk/government/publications/national-covid-19-surveillance-reports/sero-surveillance-of-covid-19

Only a week or so back there was a lot of discussion of serology data out of Spain suggesting only 5% had had it there, with speculation our figure might be similar. Actually it appears the UK figure is more like 10% antibodies (potentially higher exposure or natural immunity).

If 17.5% in London hv had CV & 7,864 died of it (as per the ONS figures), that wld be a (differential-attack-rate) IFR of 0.5%. That now appears to constitute the upper limit. (It'll be lower than that if people can get it without producing antibodies, as appears to be the case.)