Menu
G7 mortality rates span a wide range of outcomes, with Japan and the US at the extremes. Their COVID and excess death rates diverge by a factor of 6x and 2x respectively. With the latest data, this post examines why these gaps are so large.
The club of the Group of Seven countries has Canada, France, Germany, Italy, Japan, the United Kingdom and the United States as its members. These are some of the world’s richest countries. But they probably couldn’t be more different in terms of how the pandemic has affected their health outcomes.
The chart below traces the evolution of the cumulative COVID-19 death rate over the full course of the pandemic. This expresses officially reported deaths with COVID-19 as the underlying cause of death relative to the size of the population. The rate is normalized per 100,000 people.
A few patterns pop out: the effect of the different waves (note how the curves steepen) and the large range in outcomes at any point in time. The gap is enormous: they differ by a factor of 6. We also notice that the evolution of the Japanese fatality rate has been far smoother, avoiding the spikes that all other countries of the rich-country club have seen.
Let us repeat the above exercise with excess deaths. These capture the gap between the total number of deaths that occur for any reason and the number that would be expected under normal circumstances. Given the massive undercounting of the death toll attributed both directly and indirectly to COVID-19, excess deaths provides a glimpse of the true and total impact of the pandemic.
The chart below shows how the cumulative excess death rate has evolved over the course of the pandemic. These rates are based on excess death estimates produced by The Economist and we take the mid-points of their excess death model (see here for more on the model and its results).
The differences across countries remain stark. Again our two countries stand out (even though Italy ranks first and Canada last now) and cumulative excess death rates vary by a factor of 2. This range is a little more compressed than the previous one, in part because Japanese excess deaths have started to rise recently on the back of omicron. Broadly speaking, however, its curve remains relatively steady (thank the vaccines and other factors, as we will discuss later).
For completeness, the table below presents the data as of the latest date. We show both numbers and the rates. Note how the American numbers for absolute deaths stand out. That is entirely predictable given the size of its population and keeping all other factors constant. But not all other factors are constant as can be observed in the relative numbers, where we see – as in the charts above – that the cumulative rates are much higher than elsewhere.
We review a number of factors that typically have a large effect on country performance: (1) infection control, (2) vaccination and treatment, (3) general health profile of the population and (4) demography. Let us discuss these in turn.
The Japanese have a long and well-established tradition of wearing masks, which has proven to be effective in mitigating the spread. There has never been an official mask mandate in the country and at least half of the population is reported to be planning on wearing masks for the foreseeable future. The government distributed masks to the population early on and launched during spring 2020 the “Avoid the three Cs” public awareness campaign (relating to closed spaces, crowded places and close-contact settings). For the most part of the pandemic, the country maintained a policy of strict border control, which has undoubtedly dampened infections.
With individual behavior conducive to infection control, the chart below suggests there has been less of a need to impose stringent government responses. The index of government response stringency shows a pretty low and stable level, especially compared to the American experience that was marked by more intrusive interventions during the early stage of the pandemic.
The low mortality rates are remarkable because of the high Japanese urbanization rate (the share of population that lives in urban areas). High urbanization is accompanied by high population density (the country was the 12th most dense country in the world as of 2021 among countries with a population larger than 7.5 million people). All else equal, these factors will have facilitated the spread of the virus. Yet, as the mortality stats show, other factors will have mitigated or offset the mortality impact of high urbanization and population density.
The above measures on infection control gave the Japanese population some welcome breathing space, which was particularly useful and necessary given the slow start of the country’s vaccination campaign (see chart below). But as soon as vaccines became available, primary vaccination rates rose quickly and the share of the vaccinated population with at least 1 dose and also the share of fully vaccinated people (fully as per the primary protocol) caught up within a few months and overtook American rates. The experience with boosters followed exactly the same pattern of an initially slow take-off followed by rapid catch-up and an eventual overtaking.
The American experience was different. Benefiting from the typical home manufacturing bias in vaccine coverage, American rates rose quickly initially. But the rate of progress decelerated prematurely, with levels settling at comparatively low rates. Note also the big discrepancy between the at-least-1-dose and full-vaccination curves, where American rates continue to be subject to a large gap. The booster picture seems to be an exaggeration of the primary picture: quick initial progress, but a considerable slowdown afterwards.
On treatments (such as Paxlovid), we also note that the up-take has been subpar. As Eric Topol notes in this post (dated October 8, 2022): “Paxlovid is only being given to less than 25% of the eligible and less people age 80+ are getting Paxolovid than those age 45. The reasons that doctors are not prescribing it – worried about interactions for a 5-day course and rebound – are not substantiated.”
Another key determining factor of COVID mortality is the general health profile of the population and the prevalence of pre-existing conditions. In this respect, what is likely to have helped enormously in the Japanese context are the low prevalence of obesity, the low intake of red meat (especially saturated fatty acids) and the high intakes of fish (specifically n-3 polyunsaturated fatty acids), plant foods (such as soybean) and non-sugary beverages (such as green tea). The American patterns are different with the population being more prone to the typical underlying conditions that pose especially high risk for severe COVID: obesity, diabetes, hypertension and heart failure.
The differences in the general health of the population are captured well in the summary statistic of the chart below: life expectancy at birth. The Japanese population has by far the highest life expectancy at birth in this list of rich countries and in fact more broadly in the world, whereas the American population comes last in this group of countries.
One could point to other factors that may have played a role. But let us conclude with a note on demography. Demography has played an oversized role in this pandemic given the age-discriminating nature of COVID. There are many aspects to demography but the single factor that has mattered the most in this context is age structure.
The chart below shows the share of the 60+ and 70+ cohorts in the total population of each country.
As we can see, the American elderly share is the lowest in this group whereas the Japanese one is the highest. The latter was to be expected since the country has the the second-oldest population in the world (after Monaco). But this is also precisely what makes Japan’s pandemic performance so remarkable.
Among G7 countries, the mortality gap between American and Japanese mortality is remarkably large. Many explanatory factors present themselves, where some work to enlarge gap and others to reduce it.
Conversely, age structure, population density, urbanization and vaccine coverage during the early stages of the global vaccination campaign should have worked the opposite way, even though the net effect clearly is one of the very sizable gap indeed.
