Using the 28-day count is silly for England and Wales where there is good death-certificate data. The first five weeks of the first wave did have undercounting by at least 2000 COVID-19 deaths per week but the guidance was clarified (not changed) to say that one didn't need a Coroner to diagnose COVID-19 as a cause of death and from then, the death certificate data has been pretty accurate.
The ONS has weekly data going back to wk31,1999 for all cause deaths in its excess winter mortality reports. So using that, I produced the following:
Year running wk31-wk31 (which separates the two waves)
Baseline for excess - not the usual 5-yr average for the week, but the median for that week number from 1999-2021 (where those data are available - so median from 2000-2020 for wks 19-30) - for comparing 2020 and 2021, this might have an offset but its the same for both. And as the death numbers haven't changed very much in a normal week from 2000-2019 the baseline wouldn't be very different anyway.
The data is presented as below, with also COVID-19 on the death certificate as either raw or running to total to week number
|Weekly Raw deaths | Weekly Excess deaths|
|Cumulative deaths to week| Cumulative Excess deaths to week|
[IMGW=640]https://pbs.twimg.com/media/E2FSqnYWUAo7eVs?format=png&name=large[/IMGW]
The reduction in excess deaths below the COVID-19 deaths in the second wave is almost certainly due to COVID-19 mitigation measures (like lockdowns and improved hygiene) reducing other causes of death - for example Norovirus outbreaks and flu have been almost completely suppressed this last year.
By my previous counting, the first wave had about 70,000 deaths - about 20,000 more than those recorded on the death certificates for the first wave.
Which makes both waves very similar, but the second wave slightly worse.