It can be debated what model should be used and governments don't always have consistent methods of baselining. With different models you can compute some excess mortality in Sweden but those have their own flaws.
But it's worth calling out here that the reason you have to dive into the details of what precise baseline model you use is because the number of deaths was so very tiny on the scale of a country. A simple graph showing the absolute numbers puts it in perspective:
It looks initially dramatic but that's only due to the massive drop in 2019. In reality 2020 had the same number of deaths per pop as in 2015! Nothing special happened in 2015.
Given that 2019 was an abnormally low year, 2020 would have been a bit higher than normal anyway even with no pandemic. But we're still looking at numbers pretty much within the realm of normal variance. If somehow you'd never heard of COVID, then someone had shown you that graph and asked "did something catastrophic happen in this decade" you'd probably have said no. Although 2020 is a bit higher, it's on the order of a few thousands of people, not something you could have noticed in a country of millions short of saturation-level media coverage.
That's why it's hard to say but important to hear: these levels of death simply do not matter. How do we know? Because some countries have had high and persistent levels of excess death since the end of the pandemic, but as COVID isn't the explanation hardly anyone cares. There's just no link between levels of death and severity of reaction, in our society.
Many European countries have an ageing population, and thus the number of yearly deaths increase year to year. For example Finland in 2010–2019 had about 0.8% increase in yearly deaths every year. If the deaths followed this model and thus there was absolutely no excess mortality (and no random variation), the simple method you cite would compare 2020, 2021 and 2022 to the average of 2015–2019, and would estimate 2.4%, 3.2% and 4.0% excess mortality for years 2020, 2021, 2022 for Finland.
Whereas Sweden doesn't have an ageing population structure, due to lots of immigration in the past 20 years. And Sweden didn't have an increasing trend in yearly deaths in 2010–2019.
This simple method overestimates the excess mortality for many countries, if they have an ageing population structure. Then you end up comparing overestimated values to Sweden's value that it not overestimated.
You can project a baseline like that and get a higher number for 2020, but 2019 is still abnormally low and those people who survived 2019 had to die at some point.
I'm happy to grant a different baseline though. It goes from being a really small number to a really small number. Remember that everything that happened was justified by claiming a 3.something% IFR and 100% infection rate in a single giant wave. That never came close to happening and the places that used very light touch approaches had no different outcomes to places that went full totalitarian. That's the important thing here.
It can be debated what model should be used and governments don't always have consistent methods of baselining. With different models you can compute some excess mortality in Sweden but those have their own flaws.
But it's worth calling out here that the reason you have to dive into the details of what precise baseline model you use is because the number of deaths was so very tiny on the scale of a country. A simple graph showing the absolute numbers puts it in perspective:
https://www.statista.com/statistics/525353/sweden-number-of-...
Alternatively, a graph of deaths per 100k people shows how surprisingly normal 2020 was in Sweden:
https://i0.wp.com/swprs.org/wp-content/uploads/2023/04/swede...
It looks initially dramatic but that's only due to the massive drop in 2019. In reality 2020 had the same number of deaths per pop as in 2015! Nothing special happened in 2015.
Given that 2019 was an abnormally low year, 2020 would have been a bit higher than normal anyway even with no pandemic. But we're still looking at numbers pretty much within the realm of normal variance. If somehow you'd never heard of COVID, then someone had shown you that graph and asked "did something catastrophic happen in this decade" you'd probably have said no. Although 2020 is a bit higher, it's on the order of a few thousands of people, not something you could have noticed in a country of millions short of saturation-level media coverage.
That's why it's hard to say but important to hear: these levels of death simply do not matter. How do we know? Because some countries have had high and persistent levels of excess death since the end of the pandemic, but as COVID isn't the explanation hardly anyone cares. There's just no link between levels of death and severity of reaction, in our society.
reply