Analysis by Keith Rankin.
On 7 March 2023 I published a summary table of death tallies in a wide range of countries, comparing 2019‑2022 with 2015‑2018. (Covid19 Pandemic-era Facts: Irrefutable, Inconvenient, Important, Evening Report.)
I let the data stand on its own, largely without interpretation. The most important findings to explain are the reasons why Sweden weathered the Covid19 pandemic so well, and why East Asian countries fared so poorly. These facts run counter to the mainstream narrative; a narrative which has presumed that the truth is the precise opposite of what the recently available data shows. East Asian countries relied very heavily on government mandates and the compulsory wearing of facemasks for extended time periods. Sweden, on the other hand, came to be known for adopting one of the least interventionist public health policies during this early-2020s’ pandemic.
This is the main issue that needs to be explained. The demographic data, of course throws up other issues as well – including the high death tolls in the USA and the remainder of the Americas, and the lower toll in Eastern Europe than might have been expected given earlier health data. Demographic imbalances may be contributing to countries’ different experiences; imbalances relating to diverse and changing birth rates and economic migration, in addition to life expectancy issues.
Interpretation
To answer a question such as the main query posed here, we need data, at least one hypothesis, and at least one counterfactual. We also need a contestable academic environment, whereby multiple interpretations can be freely posited and reasonably argued.
We have an important set of data in my report in Evening Report.
The hypothesis that I posit is that pandemic-related mortality in general has been lowest in societies which have good balances with respect to pathogenic exposure and hygiene, that imbalances lead to reduced levels of general immunity and/or raised levels of morbidity, and that societies with high levels of general morbidity will have more excess deaths during an event such as a pandemic; indeed, during any catastrophic event.
It is possible to have too little (as well as too much) exposure to environmental pathogens. In these situations, there are two types of risk – which statisticians, prosaically, call Type 1 and Type 2 – and the reduction of one type of risk in itself raises the other type of risk. The recently oft-said phrase ‘an abundance of caution’ is an example of attempts to reduce Type 1 risk.
In teaching statistics, it is commonplace to use a criminal courtroom setting to explain these risks. A ‘risk-averse’ approach (ie an ‘abundance of caution approach’) is to acquit an accused person if there is any doubt whatsoever about the person’s guilt of the crime in question. The expression ‘beyond reasonable doubt’ expresses balance; the expression ‘beyond all doubt’ expresses an abundance of caution.
It is easy to see that, minimising the risk of an innocent person being convicted also increases the risk of a guilty person not being convicted (and thereby being ‘free’ to commit further crimes); this is the ‘Type 2’ risk. Reduced risk to the defendant means an enhanced risk to society. ‘An abundance of caution’ simultaneously means ‘a scarcity of caution’; more caution with respect to pathogen exposure means less caution re general immunity deficiency. (The quality of the evidence – eg the data – minimises both types of risk; it also minimises the quality of interpretative reasoning in relation to that evidence.) In pandemics, a good practical compromise is to adopt ‘Type 1’ caution for a brief period of acute danger associated with an unknown threat, and to as soon as possible to revert to a normal ‘balanced caution’ approach.
The most straightforward counterfactuals are purely demographic. And comparative. This is why we need much better, and more comprehensive, demographic information. Demography is the statistical analysis of births, deaths, and migrations. The most important demographic variable is a person’s ‘age’. While race/ethnicity/ancestry, sex/gender, and (to a lesser extent) religion are also demographic attributes, ‘age’ is more important to understanding outcomes (noting that the most important demographic ‘outcome’ is death). Age is the most important predictor of a person’s likelihood of dying; after that, it is socio-economic and lifestyle attributes such as income, housing, education, happiness (leading a meaningful life) and access to healthcare services which determine the likelihood of both mature and premature death.
In the context of Covid19 pandemic mortality, the counterfactual is what levels of mortality would have occurred had there been no Covid19 and hence no pandemic. The usual ways to establish such a pandemic counterfactual is to evaluate and project normal patterns of mortality in the previous few years; if necessary making comparative-country adjustments for any abnormal events in those prior years. And then to use those normal data to predict an ‘alternative present’; in essence, this process of forecasting the immediate past is a valuable use of forecasting techniques.
The next process is, if possible, to compare your (affected) place (eg country) with some other place (or places) which were unaffected (or lesser affected) by the phenomenon you are seeking to evaluate. Some countries are better comparators than others. While Australia – with its many cultural and economic similarities – is the most widely used comparator for New Zealand, Scandinavian countries are also widely used.
With a pandemic, no country is unaffected. But countries pursuing different public health policies become useful comparators; they help to answer the question as to what would have happened had one country followed another country’s policy. Thus, Sweden’s experience can be built into a counterfactual for New Zealand, because Sweden’s policies were different in both substance and in style. Australia is less useful because its pandemic public policies were very similar to New Zealand’s.
So, a simple counterfactual for New Zealand would be to project 2015‑2018 mortality data into the 2019‑2022 period. The documented excess of deaths compared to that counterfactual represents an estimate of New Zealand’s ‘quantity of life’ pandemic outcome. Then, repeating the exercise for Sweden yields a comparable quantity of life outcome for that country. The country with the smaller percentage excess of deaths probably pursued the better set of policies, noting that two quite different policy approaches could yield similar outcomes.
A well-reasoned counterfactual is an essential part of any interpretation of historical facts. In a scientific process, for which the reasoned use of counterfactuals is an example, a counterfactual is commonly called a ‘control’. It was widely noted (eg in the book The Herd) in 2020 and 2021 that Sweden potentially contributed substantially to the scientific understanding of the Covid19 pandemic, by providing demographers and epidemiologists with a control. So far, however, I have seen little evidence that Sweden’s value as a control – as an important policy counterfactual – has been well-utilised.
The quality of demographic information throughout the world is rather poor. While New Zealand is better than most countries, getting good information about the ages of the population (and where people of different ages live) is difficult. Indeed, until a few years ago, demographic information about immigrants and emigrants was collected somewhat casually as an accessory to tourist data. (Indeed, re yesterday’s ‘compulsory’ population census, the government’s target was only ninety percent compliance; meaning it is regarded as acceptable to ignore the 500,000 ‘harder to reach’ people in this country.) Travellers were assessed as immigrants – rather than visitors – based on their stated intentions on arrival, and not on the actual outcomes of their travel. Many of the people who die in New Zealand are not born in New Zealand, and vice versa.
This makes it hard to create good projections of what mortality in any country, let alone the world, would have been from 2020 in the absence of the Covid19 pandemic. Nevertheless, indications are that, in the absence of the pandemic, New Zealand would have had a higher increase in mortality (maybe a four percent increase from 2015‑2018 to 2019‑2022) than most other countries in the world would have had. This finding relates in particular to New Zealand’s particular pattern of aging, noting some substantial variations in birth rates in the years from 1930 to 1960; and, also variations in the age distribution of older foreign-born New Zealanders.
There are other indications, based on the use of the United States as a comparator country, that some of the increases in morbidity occurring in the USA before 2020 were also occurring in New Zealand; for example, reasons around income inequality, housing, and mental health. The pandemic in USA was substantially more severe than in Sweden, with New Zealand falling in between.
Evaluation of the Hypothesis
In New Zealand, an excessive emphasis on hygiene – including the mandated wearing of facemasks in many public settings – most likely contributed to a loss of general immunity to infectious diseases. New Zealand, by in large, followed the East Asian public health policy model.
This loss of general immunity was countered by a comprehensive, though belated, Covid19 vaccination programme. Vaccination immunity almost certainly contributed to low excess mortality in the period from October 2021 to March 2022. But specific immunity (whether arising from infection or vaccination) to coronavirus diseases – which include around ten percent of ‘common colds’ – and influenzas has always been known to be short-lasting. So general immunity that arises from lifestyle factors remains an important protector of life; general immunity is enhanced by balanced diets (avoiding excesses of foods that create morbidity, such as alcohol, sugar, some fats, and salt) and some ongoing exposure to a range of less-dangerous pathogens.
People living in West European urban environments probably have had closer than most other people to ideal levels of balanced nutrition and general immunity. So these countries have generally had the least pandemic mortality, and (if my hypothesis is correct) probably have the best outlook for the next few years with respect to deaths arising from respiratory infections. People living in Eastern Europe, especially in the European Union, seem to have regained high levels of general immunity, though they bore a high cost in 2021; and membership of the European Union gives them lifestyle options not available in many other countries.
My hypothesis, if correct, suggests that excess deaths will continue in East Asia for another year or so, due to compromised general immunity arising from excessive hygiene. And, in the Americas, increased morbidity seems to be a growing socio-economic problem, making those populations particularly vulnerable to respiratory pandemics. Both of these regions will experience increased levels of morbidity arising from the after-effects of Covid19 infections.
Africa and South Asia are hard to evaluate due to lack of data. But, for Africa at least, indications are that Covid19 excess deaths have been less than in the Americas, and maybe even comparable to Western Europe. Lifestyle morbidity remains less in Africa as a whole than in the Americas. And general immunity levels in Africa have always been high; it is a continent widely associated with ongoing pathogenic exposures. The critical factor for Africa in the coming years will be nutrition. South Asia most likely the same, though with complications arising from substantial air and water pollution.
New Zealand and Australia? Harder to predict, based on my hypothesis, because both countries contain elements of the East Asian, American and Western European experiences. I just hope that New Zealanders are able to get their pre-winter boosters in time. There is every reason to anticipate a dangerous new outbreak of Covid19 in the early winter, much as occurred in Western Europe in the three months to mid-January 2023.
To get to the truth we need reasoned argument, scientific argument. The pandemic has touched on our lives sufficiently to deserve mainstream media attention be given to contestable analysis of its impact; and to question politicised narratives of the form ‘the science says this’ when in fact science is a contestable and argumentative process.
————-
Keith Rankin (keith at rankin dot nz), trained as an economic historian, is a retired lecturer in Economics and Statistics. He lives in Auckland, New Zealand.
