Compared to the trends on January 2020, has Australia or Sweden lost more wellbeing in 2020? And which has seen the greater damage to expected future wellbeing years for after 2020? The Table below summarizes the answers to this.
For the first calculation, let us only count the main elements going into the wellbeing of these two countries in 2020: the experienced wellbeing of the population in 2020 and the excess deaths in 2020. Lots of the other things we normally look at in these covid-calculations, such as changes to GDP, will show up in the anticipated effects for after 2020.
|Lost Wellbeing years in Sweden and Australia due to covid and lockdowns|
|in 2020||beyond 2020|
|Future debt repayment||7773333||886340|
|Loss per million||20077||6288||306026||86667|
First, what was the wellbeing drop in Australia? Well, an ANU-sponsored longitudinal panel found a drop from 6.9 to 6.5 in their life-satisfaction poll from January to April, a huge decline that is similar to the drop in the UK. This panel is based on over 3,000 individuals, which is why this drop is strongly statistically significant.
What about the rest of 2020 in Australia though? The same ANU team, headed by Professor Biddle said at the end of 2020 about the lockdowns in Australia that they were “a massive hit to happiness, experienced by Australians from all walks of life”. They document the increased inequality and how things are particularly bad among the young and the vulnerable in general. Their key graph for the whole of 2020 is above.
This graph aggregates the changes in Australia in the various states, though the ANU team found one can clearly see that life satisfaction plummets in every state during a lockdown but recovers very quickly afterwards, so the top graph is essentially directly related to the average severity of lockdowns in Australia. Now, in terms of actual numbers, we should note that relative to December 2019-Jan 2020, (6.95 on average), Australia from March onwards had an average of around 6.8: the country was right back at the end of 2020 when there almost no lockdowns anywhere, but experiencing various degrees of suffering at other periods.
Now, how bad is an average loss of 0.15 for 9 months, ie a loss of 0.1 for the whole of 2020? A loss of 0.1 means a loss for the 25.4 million Australians of 2.54 million units of wellbeing (called WELLBYs). Since a wellbeing level of 2 is when individuals are, roughly speaking, indifferent between continued living or not, and a regular wellbeing year is spent at level 7, a regular year of life in Australia is worth 5 units of wellbeing. So the 2.54 million lost units are equivalent to 510,000 lost years of wellbeing in Australia in 2020.
In that 510,000 wellbeing years is all the misery caused by lockdowns and social distancing: all the cancelled weddings, all the additional cases of depression, all the extra anxiety, all the worry about unemployment, all the loneliness, all the abused women who could not escape, all the people who were not helped at hospitals, all the people whose IVF treatments were cancelled, etc. These numbers are also born out by other Australian surveys, such as on the strongly worsened mental health situation of Australia’s youth due to barred access to school, evidenced in official reports. The fact that the media does not talk about these lockdowns-victims much doesn’t mean this suffering does not exist.
One can note, btw, that the data just released by the World Happiness Report 2021 fits in perfectly with this picture in that the drop reported there for Australia since 2019 of about 0.1 refers to data gathered in February to mid-March 2020, so largely before the drop but just nipping its first beginnings, hence the drop you expect you see for that period.
And wellbeing in Sweden? Unfortunately for Sweden, there is no neat large panel gathered during the whole year like there is for Australia to base the numbers on, but we do have some comparable data. What we know for Sweden is that, unlike in strong lockdown-countries like the UK or Victoria in Australia, there was no drop in the main wellbeing indicator (life-satisfaction) in the first weeks of the panic in Sweden: Kivi et al. (2020) found among the group one expects to see a drop (the 65-75 year olds) no change in their life satisfaction in early April relative to the previous year, just as, by the way, found in places like the Netherlands which in that period had very light lockdowns.
For the rest of the year, the data for Sweden is patchy. An imperial college panel covering the period after the initial lockdowns(so not the early change) found no significant drop in life satisfaction over time in Sweden for the rest of 2020, but that is essentially because that panel is extremely small (50-100 observations per two-week period). Yet, if we aggregate that data over three month periods, we see that there was a drop in the last 6 months of 2020 relative to the April-June period of 2020. Not a significant drop, but the drop was concentrated in the last 3 months which was when the Swedish government was over-riding its health authority with more restrictions, so exactly the period you do expect to see a drop. So on the balance of probabilities, Sweden probably did have a slight drop in 2020, concentrated in the last 3 months, amounting to a drop over the whole year of 0.03 (this is very close btw, to the insignificant small drop found for the 2020 Gallup data for Sweden underlying the World Happiness Report).
Does this make sense if you look at the policies?. Sweden did not close the schools, which is the thing that has been found to cause massive wellbeing losses among the young and their parents in other countries. Sweden also did not prevent the generations from mixing or friends and workers socialising, which are the big elements in the drop elsewhere. So the lack of a big drop fits with what we know does the damage to wellbeing in strong lockdown places. But the government in Sweden did become more difficult about travel and socialising at the end of 2020. If we go with the data we have available then one should say wellbeing remained constant in Sweden for most of 2020 but probably did drop in the last few months.
Then the excess deaths in Australia and Sweden. The notion of excess deaths depends on what one counts as expected deaths, which is not so easy to say and on which reasonable people can disagree. Yet, the simplest thing to do is to just look at how many deaths there were in 2020 more than the average of 2015-2019. The Oxford group looking into this found that, at the time of recording, there were 112 less deaths in Australia in 2020. In comparison, in Sweden there were 5983 more deaths in 2020. So here we see the benefits of Australia having avoided a large glut of covid deaths, and perhaps also less traffic deaths, less flu deaths, and other sources of death. Sweden too will probably have avoided some traffic deaths and of course many who died of covid would have died in 2020 anyway, reducing the Swedish measured excess deaths to well below their reported number of covid-deaths.
On the other hand, the breakdown in the functioning of much of the health system in Australia during local lockdowns will have meant more cancer deaths and deaths from other preventable health problems. The balance seems to be 112 less deaths in total. Note that these numbers are relatively small. The 6000 extra deaths in Sweden in 2020 means a loss of around 3000 wellbeing years in 2020 (because people on average died in the middle of the year and thus lost 6 months in 2020), whilst the 112 less deaths in Australia meant a gain of 56 wellbeing years.
So Australia in 2020 lost 510,000 wellbeing years, but did gain 56 wellbeing years at the same time because of those 112 less excess deaths. Sweden lost 3000 wellbeing years because of their excess 6000 deaths and 62,400 lost wellbeing years from the probable drop at the end of 2020. This comes with a very simple conclusion: 2020 was a relatively horrible year for Australia. Per one million citizens, Australians lost 3 times more wellbeing than Sweden. The fact that the media chooses to focus on the minuscule gains and not the enormous losses is essentially a form of dismissal of the misery of those who have lost. As if they don’t exist.
So Australia compared to Sweden has had a terrible 2020. But, you might wonder, how about losses yet to come after 2020 because of the things that happened in 2020. How about the lost years of life of those who died in 2020? And how about government debt, employment, and all that?
Well, the key statistic there is that Sweden has already absorbed the small peak of unemployment mid-2020 and has ended the year with 3% less GDP at the end of 2019, expected to rebound fully by the end of 2021. The extra government debt in 2020 was 6% of GDP, but the country had stopped running a deficit by the end of 2020. Now, that extra debt really matters because it means reductions in future government expenditure in order to pay back this debt. Equivalently, you might say that money could have been spent in 2020 to fund things that would have increased future wellbeing, such as via setting up mental health programs or improving the health service. As a rule of thumb, based on the finding that governments buy a wellbeing year for around 30,000 Euro (5 WELLBY which are generated by 6000 euro of government spending on average), 1% GDP debt in Sweden (which has 10 million inhabitants) costs Sweden 0.15 million wellbeing years in terms of the effect of those lower future government expenditure, so the extra debt will cost Sweden 886,340 wellbeing years.
How is Australia looking in this regard? Well, the surging iron ore prices have helped keep the 2020 GDP drop to around the same 3% in Australia in 2020, though it has to be said that the expected growth for 2020 was higher than in Sweden, so the drop is relatively bigger. The more relevant difference is the increase in debt: already 10% of GDP has been added to government debt in 2020, but unlike in Sweden, that debt is still climbing fast because of the closed borders and the extremely costly recurrent lockdowns. So the RBA expects the debt to climb at least another 10%. Paying back that debt will involve huge future wellbeing costs, as we are already starting to see with the announced cut-backs.
Per person, Australia and Sweden have essentially the same GDP. So the 10% additional debt in Australia of 2020 will cost Australia 3.9 million wellbeing years. The projected further debt of another 10% will cost another 3.9 million wellbeing years. That is 7.88 million lost wellbeing years in Australia in total.
The 6,000 extra deaths in Sweden of 2020 will going forward mean around 2.5 years of lost life each after 2020, meaning 15,000 lost wellbeing years in the future from those higher excess deaths. By the same token the 112 less excess deaths in Australia mean 288 gained wellbeing years to come.
Summarising just these basic effects, per 1 million inhabitants, in 2020 Sweden lost 6288 wellbeing years, made up of their higher excess deaths and a probable drop at the end of 2020. Australia lost 20,077 wellbeing years per million citizens in 2020, due to the misery of its much harsher intermittent lockdowns. Australians thus experienced a wellbeing drop of 3 times the Swedish wellbeing drop in 2020.
In terms of losses going forward, Sweden’s additional debt and higher excess deaths mean a loss of about 89 thousand wellbeing years per million Swedes. In contrast, the much higher and continuing government debt in Australia means a loss of about 306 thousand wellbeing years per million Australians. Australians thus should expect to lose 3.5 times as much wellbeing going forward as the average Swede.
I can mention that if I add disrupted IVF services, loss of child education, sectoral disruption, and loss of basic freedoms into this calculation, that Australia’s losses loom much worse still than Sweden since all these were much worse affected in Australia than Sweden. So you should see the calculations above as rigged to look good for Australia. I intend to do another post with an extended comparison when some of the key data on other effects, like the covid baby bust and the extent of unemployment in Australia, have become clear.
The conclusion is inescapable: in terms of the wellbeing of the population, which counts everything that makes life worthwhile, 2020 was a much worse year for Australia than Sweden, with the events of 2020 also blighting the future much more for Australia than for Sweden. The fact that much of its media and commentariat do not seem to realise this, tells you how abandoned the many victims of lockdown policies in Australia must feel.
[This post was updated on 17-18/04 after a commenter convinced me it was more reasonable to say Sweden did see a drop in wellbeing in the last 3 months of 2020 that should be reflected in the numbers. That of course changed the ratio of losses between Sweden and Australia because that is almost entirely related to the wellbeing drop among the general population in 2020.
I also took the opportunity to make the definition of a wellbeing year equal to 5 WELLBY for both countries (previously I used a lower number for Australia because Sweden is happier, but that just leads to confusion). For the GDP baseline I used 2019 GDP figures.]