The methodology of a lot of the studies is a bit dodgy. The ONS study adjusts by 'region', which is ludicrous considering that there can be very many deaths in say Luton, and very few in North Norfolk, which are in the same region (East of England).
The IFS study adjusts by coronavirus test count, which is even more ridiculous.
There is also a large systematic bias in that people over, say, 90, are hundreds of times more likely to die of covid-19 or indeed anything else, than people of say, 40, and there are many many thousands of extra deaths not encompassed by official covid-19 death stats in the 70+ age group, which will be disproportionately white, whereas there are no extra deaths at all up to about age 55 or 60, a group which will be disproportionately non-white.
In addition, it seems that old white people are more likely to be in care homes (to which some people trot out tropes about ethnic populations caring more about their elderly populations, which ignores the fact that the white population of the UK is only 5% higher than in 1939, whereas the populations of many developing countries are 10 times higher, hence there are vast numbers of young people to care for the very small number of elderly people in these countries - so ethnic minority populations are structurally much better equipped to care for their own elderly, since their population is a pyramid, whereas the white population of the UK has a baby boom bulge that will continue to outnumber younger generations for quite some time). Only around 30% of the old white people in care homes are being counted in the official death stats as having died from covid-19.
To properly test for ethnicity, you would need to compare on quite a low level, ideally at a Lower Level Super Output Area level, so that white populations are compared with BME Populations living in the same areas.
It's also very much worth noting that the Bangladeshi community is incredibly insular in that 25% live in one borough (Tower Hamlets), whereas the Chinese population is very spread out across the whole country. The Pakistani community is much more insular than the Indian community, which is more insular than the Chinese community.
Some white communities are very insular, but whereas the black and Asian communities are concentrated in major cities, which have airports, public transport, etc. places like Hull, Blackpool, etc., which are very white, clearly lack such links and would be expected to experience covid-19 much, much later than say Birmingham, or inner London. I am not aware of any instances of ethnic minority populations living in large numbers in places which are relatively disconnected (transport etc.) from the rest of Britain in the way that places like Hull or Blackpool might be.
The actual numbers of people being discussed are also small.
In addition, the risk of death from covid-19 is massively overwhelmingly and conclusively linked to age - a 10 year age difference absolutely will increase the risk of death by 3x or more, but it is not proven that a black person and a white person with the same job/background have any difference in risk.
In other words, just because the average black person is poorer than average, it doesn't follow that every black person is poorer than average, and it doesn't really tell you anything personally.
Old people, of whatever race and sex, should worry about covid-19. Young people of whatever race and sex. The gradations within that, whether you are man (higher risk) or or a woman, or whatever else are just noise in this context.