Monday 30 August 2021

COVID-19 World Review, Part 3: Transmission

This is the third of my review reports on COVID-19 worldwide as at August 20th 2021. This report will cover transmission of the virus. I’ll be looking not at cumulative cases or daily new cases, but at the (weekly) rate of growth of new cases. I’ll also be looking at the R-rate, which is defined as the average number of new infections resulting from one infection.

Conceptually, there are some differences between weekly case growth and R-rate. Firstly, weekly case growth is based on measurements of new cases (I calculate the weekly case growth as the weekly percentage growth in new cases counts which are already weekly averaged). R-rate, on the other hand, is based on using an equation-based model to estimate infections. The methodology for calculating the R-rates supplied by Our World in Data is documented at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3581633.

For the R-rate and weekly case growth to correlate well requires not only that the model is accurate, but also that the proportion of infections which are not detected as cases is constant. The graphs for individual countries in the first two papers show that the two statistics do correlate quite well, at least since the first three or four months of the epidemic; and thus, suggest that these assumptions are valid.

Europe

Here are the weekly case growth and R-rate graphs for the four regions into which I have divided Europe:








The weekly case growths seem to be a lot more volatile than the R-rates, particularly in smaller countries. Weekly case growths are also seriously perturbed whenever there is an adjustment made to the case data of a particular country. That said, when the R-rates are in their mid-range of approximately 0.5 to 2, they do seem to correlate well with the weekly case growths. Because of this, for the rest of the world I’ll only show the R-rate graphs.

One notable feature of the above R-rate graphs occurs where several countries have suffered sudden, sharp rises in their R-rates at much the same time. A good example is the precipitate increase in many of the Europe 14 countries starting around the middle of June 2021. The sudden and almost parallel rises in many countries suggest that this may have been down to the spread of the, relatively new and more transmissible, delta variant. In the previous few weeks, the UK and Portugal had been above the others in the general level of R-rate; and they were less affected by the subsequent spike, suggesting that they may have received the delta variant earlier than others.

In Eastern Europe (North) there was a similar spike, a couple of weeks later. Russia, Moldova and Belarus were the three countries above the others in the previous few weeks; but all three had sudden rises in R-rate during June. Might the new variant, in this area, actually have spread from east to west? In the southern part of Eastern Europe, virtually every country has been affected by a spike in R-rate in late June or July, though Cyprus had it first by several weeks. Malta also had a big spike during June.

Americas

In North America, there are strong increases in R-rate in the USA starting in June, and in Canada a little less than a month later. That dark brown line at the bottom, Nicaragua, doesn’t look right to me.

South America does not show any precipitate increases, except for Suriname which, having a population of only 600,000, is always likely to show up as more volatile than larger countries.

In the West Indies, those recent spikes in Antigua and Barbuda (mid blue) and Saint Kitts and Nevis (dark brown) are reminiscent of the spikes in Europe. And the recent R-rate figures for Dominica (light blue) and Grenada (dark blue) don’t look right to me, given that Dominica has had a big spike in cases starting in early August, and Grenada (a country of only 100,000 or so) a smaller one.

Middle East and North Africa

Not much to see recently in the northern Middle East, except for the recent strong rise in the Lebanon.

In the southern Middle East, there has been a big rise in Israel starting in June, and another a few weeks later in Palestine. Delta variant, perhaps?

In North Africa, there were big recent rises in Libya, Morocco and Egypt. Could the delta variant be spreading on the North African coast too?

Sub-Saharan Africa

In Central Africa, the R-rates are generally lower than in other places, and are frequently below 1. They are also very variable. It looks as if the virus has difficulty spreading in central African conditions; perhaps because of the heat and humidity? And there is no co-ordination between the R-rates for the different countries, possibly reflecting a relatively low amount of international travel.


There have been recent strong and sudden rises in the R-rates in Uganda, Rwanda, Malawi and Ethiopia. But the first three, at least, look a bit early to have been down to the delta variant.

The four recent peaks are in Zambia, Zimbabwe, Mozambique and Mauritius. There was an earlier peak, starting from a much lower base, in Eswatini too. The straight, steep rises are reminiscent of some of the European R-rate graphs. Except for Mauritius, I would guess this could be the signature of the spread of the beta variant (originally found in South Africa), which like the delta variant is highly transmissible, into these countries. It’s possible that some of the rises in East Africa might be due to this, too.

Here, the recent peaks come from Liberia, Sierra Leone, Senegal and Nigeria. Could there be some more transmissible variant of the virus spreading between these countries? If so, it seems surprising that I can’t see anything untoward for other countries in the area, such as Ghana or Cote d’Ivoire.

Rest of Asia

In East Asia, one thing to note is that enormous R-rate in South Korea (light blue) at the beginning of the epidemic. What happened there? Was that a variant that no-one else got? Another thing to note is the three recent peaks in Taiwan, South Korea and Japan. Taiwan and Japan’s peaks were probably down to the delta variant; could that have been so in South Korea too?

The recent rise in Azerbaijan (light blue) might possibly be an effect of the delta variant?

I wonder whether India, Nepal and the Maldives in mid-June, and Sri Lanka in early July, may have received a more transmissible variant – such as the delta?

The big, sudden recent spikes in Laos and Singapore may well be due to the delta variant.

Oh, and why was the R-rate in Singapore so low as far back as the end of February 2020? As far as I can see, no-one else in the world except China had such a low R-rate that early. If that’s right, they must have had an earlier wave! But their cases continued to grow through February and into March. So, the correlation between R-rate and weekly case growth was broken at that stage of the epidemic in Singapore.

Australasia and Oceania

Australia seems to have had as many separate R-rate peaks as anyone in the world. And New Zealand, after the initial outbreak, seems to have kept its R-rate consistently as low as anyone outside Central Africa. (Up till August 20th 2021, of course).

Some examples

I decided to pick some countries I haven’t looked at in detail before, which show different behaviours in epidemic profile from their neighbours. I chose: Germany (lowest cases in its area). Russia (a big country, where the virus may well spread differently from its neighbours). Cyprus (an island with high cases, out of sync with the rest of its group). Nicaragua (can I really believe how few cases and deaths they have had?). Pakistan (unexpectedly low cases for its area). Saudi Arabia (ditto). South Korea and Australia (both having multiple peaks of different sizes). And New Zealand, to compare with Australia.

Germany



German lockdowns have tended to be quite high over the course of the epidemic; even after several successive unlocks, the stringency is now 62%. But interestingly, the Germans have at no point had a nationwide face covering mandate. The characteristic signature of the delta variant is visible at the right, but it hasn’t (yet) produced a huge surge in cases, probably because it started from an unusually low R-rate in mid-June of only just above 0.5.

Russia


The Russian epidemic is, necessarily, on a large scale, and it tends to move more slowly than in other places. Nevertheless, there have been substantial bumps in the R-rate in September/October 2020 and in early June 2021. The latter is probably due to the delta variant. There is some doubt over the accuracy of the Russian statistics; it’s said in some quarters that both cases and deaths have been under-estimated. If true, that could prove to be an advantage from this stage of the epidemic onwards!

For much of the time, Russian lockdowns have been relatively light; particularly since many of them have been regional only. The recent surge in cases has prompted a reversal, but the lockdowns are still less than in many countries further west. It’s worth noting that the Russians have kept on moving between quarantining international travellers, banning some of them, and from March to July 2020, closing the borders entirely.

Cyprus


Cyprus has had several big peaks in the R-rate, two of them being in midsummer. It held cases down until October 2020, at which point cases rose towards the first of three peaks. The most recent peak has the signature of the delta variant, and suggests that Cyprus was one of the earliest countries to get it. The R-rate is now below 1, and Cyprus looks to be in better shape than most other countries in its region. Lockdowns have been gradually relaxed, and the only mandates remaining are screening of international arrivals, a medium restriction on gatherings, and face coverings required when with others.

Nicaragua


When I first looked at Nicaragua’s figures, I didn’t believe them. I still don’t believe them. The reported cases per million are extremely low for the region, only 0.16% of the population; and only 1.9% of these have died of COVID. Cases and deaths are only being reported weekly, and the recent deaths per case ratio, if I could believe it, would be outstanding compared with its neighbours, at 0.28%. The reported death counts each week have been exactly 1 for the last seven weeks in a row. The lockdowns have been consistently among the lowest in the world. The R-rates since the new year don’t make sense at all. In short, no lo creo.

Pakistan


Pakistan’s recorded cases count is very low compared with countries around it (other than Afghanistan); only 0.5% of the population. Lockdowns have been consistently high, with an average stringency of 63% throughout the epidemic, and a current stringency of 67%; but, because Pakistan is a big country, many lockdowns have only been regional, and some regions may have got off more lightly. The last week of July 2021, which saw a big surge in cases yet only a small increase in the R-rate, suggests that a lot of the population are still in the “susceptibles” category. This is corroborated by the vaccination levels, with only 5.5% of the population fully vaccinated. So, Pakistan still has a long way to go.

Saudi Arabia


Saudi Arabia, like Pakistan, has low cases per million compared with most of its neighbours. Lockdowns have been high, particularly during the first wave. It would seem that this is what got the Saudis to their relatively low cases per million. There has been a very gradual series of unlocks, but the average stringency throughout the epidemic is close to Pakistan’s at 62%, and the current stringency is 54%. The Saudis are better off than Pakistan, particularly since they have managed to fully vaccinate 33% of the population. But they, too, still have a long way to go.

South Korea


South Korea’s total cases are 0.46% of the population, way below European levels (about a tenth of Germany’s). They locked down early – closing schools in February 2020. And they locked down hard – mandating stay-at-home and travel restrictions, and closing some workplaces, in the third week of March. Lockdowns have averaged 54%, and are now at 51%. But the delta variant seems to have caught up with them in late June and early July 2021, and their cases are now at their highest point so far, 35 new cases per million per day – though still low by European standards. They have a long way to go to reach herd immunity! And only about 20% of the population are fully vaccinated, so they have a long way to go on that score too.

Australia


Australia’s cases per million are even lower than South Korea’s, at 0.17% of the population. And like South Korea, the delta variant has recently caught up with them, and propelled new cases to previously unseen heights. But unlike South Korea, the R-rate doesn’t seem to want to come down again; despite recent lockdowns. Like the USA, Australia is a multi-state country, so the stringency (averaging 58%) will often overstate the actual lockdown severity. But Australia’s biggest lockdown is a national one: the borders have been closed since March 20th 2020. It looks very much as though they may now be finding themselves in a Uruguay situation, of a good start followed by a big spurt of cases. Their vaccination levels are similar to South Korea’s, so the vaccines aren’t going to save them – not in the short term, at least. Their early successes, I think, may turn out to have been a shot in their own foot.

New Zealand


The New Zealanders closed their borders on the same day as Australia did, and have kept them closed since. If it wasn’t for that uptick at the right-hand end of the cases graph, this would look like the perfect strategy – wouldn’t it? Lock down hard to get rid of the first wave, then keep lockdown levels generally low (averaging 35%), while zapping the virus occasionally with targeted lockdowns to keep cases down. Unfortunately, that uptick looks very much like the arrival of the Dreaded Delta. As at the time of writing (August 29th), new cases had gone up by a factor of 5 inside a week. Vaccination levels are similar to Australia’s; so now, they’re both together in the same pickle. It will be interesting to see how the New Zealanders play things from here.

To sum up

The weekly growth in new cases, and the reproduction rate (R-rate) estimated by use of an equation, correlate well with each other, at least after the first few months of the epidemic.

The “signature” of the delta variant is even clearer in the R-rate than it is in the daily new case counts. Particularly in Europe, many countries have seen sudden increases in the R-rate, moving successively from country to country. These now seem to be extending to more and more parts of the world. There are similar signatures in some countries in Southern and East Africa, which may well be evidence of spread of the beta or “South African” variant. And there is something transmissible spreading among some of the countries in West Africa, too.

As to the countries I picked for further detail: Australia and New Zealand appear to have had their entire strategies against the virus thrown into confusion by the arrival of the delta variant. South Korea is in a similar situation, if maybe a bit less serious. Saudi Arabia and Pakistan have achieved their relative success so far by hard lockdowns; but there isn’t any evidence the delta variant has reached them yet. Ouch!

Nicaragua’s figures are not believable. Cyprus, right now, looks as close as I’ve seen to having dealt with the virus both successfully and with at least a modicum of concern for people. Russia is… Russia. Enigmatic, and maybe dealing with the situation better than some think. And Germany is… Germany. I suspect they’re in a similar situation to the UK: not out of the woods yet, but with some open country apparently visible in the distance.

Friday 27 August 2021

COVID-19 World Review, Part 2: Cases

This is the second of my review reports on COVID-19 worldwide as at August 20th 2021. This report will cover cases. I’ll show graphs of cumulative case counts and daily new case counts for countries in different regions of the world. To enable inter-country comparisons, I’ll use cases per million population (or per hundred of the population) throughout.

Europe

Again, I’ll start in the core of Western Europe:


There is a factor of 2.6 between the total cases in the top country (Luxembourg – 12% of the population) and the bottom (Germany – about 4.6%). Many of the countries seem to have had peaks and troughs of daily new cases at much the same times; while, at other times, one country seems to come out above the rest. And several times, a pattern has been followed where new cases in several countries have suddenly increased one after another.

The first wave, in March 2020, peaked higher in Luxembourg and Spain than anywhere else, followed a month later by peaks in Ireland and France. This was followed by a period of longer and lower peaks, in Sweden, then Luxembourg again, then Spain again. Then, starting in October, daily cases reached new heights; first in Belgium, then proceeding through Switzerland, Luxembourg again, and France. Followed, in the new year, by peaks in the UK, Ireland, Portugal and Spain again. There was a lesser series of peaks in April, affecting mainly Sweden, France and the Netherlands. After a quiet period in the first half of June 2021, the UK’s cases started to climb, ushering in in a new series of peaks in July, mainly in the Netherlands, Spain and Portugal, and also rising cases in France, Ireland and Switzerland.

This tends to suggest that, when a new and more transmissible variant gets loose in Europe, it tends to hit several countries in quick succession. This is characteristic of an area with (relatively) free movement across borders. This happened first with the initial strain, then with the one that hit Belgium in October (which may or may not have been the “beta” variant). Whether the peaks in April were caused by the same variant or a different one is not clear. But the very steep climbs in new cases in early July in several countries are one of the signatures of the “delta” variant, which is the primary variant now in the UK and many other countries.


In the northern part of Eastern Europe, Czechia is way ahead, with almost 16% of the population having become cases. The remaining countries range from Lithuania and Estonia just above 10%, down to Belarus at 5% and Russia at 4.5%. Russia (dark brown, relatively smooth line) is always likely to be an outlier in this group, as due to the country’s large size the virus is likely to spread slower there than further west.

Eastern Europeans did not suffer from the first wave as badly as those further west. But the virus would not be denied! The peaks in daily new cases from October onwards were in Czechia first (with a smaller peak in Slovakia), then Lithuania, then Czechia again (twice), then Estonia, and later Hungary, Poland, Lithuania and Latvia. By June, things seemed to be fairly well under control – except in Russia, which as above is an outlier. The sudden spurts in cases in Lithuania and Estonia during July suggest that they may have been among the first in the area to receive the delta variant.


Here, it is Montenegro which has the most cases per head of population, at 17.3%. (It is unfortunate that Montenegro is one of the countries which is not reporting any stringency information, which may make it difficult to work out why their cases are so high.) The other countries range from Slovenia and Cyprus, both just above 12%, down to Greece at 5.3% and Albania at 4.8%.

Apart from Montenegro, big peaks in daily cases per million have been experienced in Slovenia, Serbia and Croatia. The two recent large peaks come from Cyprus. And the recent upturns in daily new cases in virtually all the countries except Cyprus suggest that the delta variant may well have found its way to this part of the world too.


Andorra and San Marino, small communities sandwiched between France and Spain and within Italy respectively, have the highest cases as percentage of population here, at 19.4% and 15.4% respectively. Iceland, Norway and Finland are at the other end with just 2.9%, 2.7% and 2.2%. The Vatican (black line, horizontal in the first graph since October) has 3.3%. Liechtenstein and Monaco, both landlocked, and Malta, an island, are all close to 8%.

The graph of daily cases per million doesn’t really show much. The two biggest peaks are both from the Vatican (with a total of 27 cases and no deaths), and the remaining peaks are dominated by Andorra and San Marino. There is, however, a suggestion of a delta variant signature in early July 2021 in Andorra, Malta and Monaco, and a little later in Iceland.

Americas


In North and Central America, the USA tops the list with about 11.4% of the population having become cases. Panama and Costa Rica also have cases per million comparable to European levels. All the other countries (even Canada!) are below 4%, with El Salvador coming in at 1.4%, and Nicaragua at a hard-to-believe 0.16%.

As far as daily new cases are concerned, the graph shows many of the same features as the European ones, with significant peaks during the Northern Hemisphere winter, and the characteristic shape of the delta variant hitting the USA in about the third week of June.

In South America, Argentina and Uruguay lead in cases as a percentage of population, with just over 11%. Other cases per million levels, down to Bolivia at 4.2%, are comparable with Europe, and generally higher than in North and Central America – a little bit unexpected. Venezuela is a low outlier, with only 1.1%.

The case of Uruguay is salient. Having been bottom of the list in December, less than six months later they had risen to the top, and are now in second place. Uruguay is also the highest vaccinator in South America; although its full vaccination rate didn’t really start to pick up until June. By which time, the damage had been done. This goes to show that, however well a country does against the virus early on, if there’s still a significant population of susceptibles, it will get among them in the end!

The daily cases per million in different countries look to be more independent than in Europe; perhaps because there is less international travel between them.

Cases per million in the West Indies are lower than in Europe and South America, and comparable with Central America.

The “Uruguay syndrome” of doing well against the virus to start with, then suffering a major and sudden spurt in cases, has happened in Cuba, Dominica, Saint Lucia and Saint Kitts and Nevis recently. It also happened earlier in 2021 in Saint Vincent, Saint Lucia, Trinidad and Tobago and Saint Kitts and Nevis; and in the Bahamas earlier still. I would guess that this pattern may have something to do with a country being relatively isolated from other countries around it. Also, the most recent outbreaks (except possibly Cuba) may be due to the import of the delta variant.

Middle East and North Africa

These countries divide into two groups: low-case and higher-case. The low-case countries are Syria, Afghanistan and Pakistan, in all of which less than 0.5% of the population have become cases. Two of these, of course, have political problems at the moment. The higher-case countries have cases in the 4% to 9% range. There is an obvious multiple-wave pattern in most of the countries.

Lebanon (light blue), Turkey (dark brown) and Iran (grey) are concerning. You can control cases, up to a point; you can lock down, and so recover from a spurt. But if susceptibles are there, the virus still has them in its sights. Particularly if it is a relatively new variant.

Many of these countries have cases per million comparable with Europe. In Bahrain, 16% of the population have become cases; and in Israel, an early vaccinator, 11.3% and rising fast. There are two outliers on the low side: Saudi Arabia at just 1.6%, and Yemen at a mere 0.25%. But Yemen, too, has its political problems; COVID is not their highest priority.

Again, there are multiple waves in many of the countries. And the peaks are tending to get higher as time goes on. Not good. Though Bahrain, the UAE and Kuwait seem to have escaped high peaks, instead having a more sustained daily case load.


Here, the countries again seem to divide into two groups. Tunisia, Libya and Morocco tend to follow the multi-wave pattern of the rest of the Middle East, with total cases in the range 2 to 5% of the population. The others have low total cases, all below 0.7% of the population.

Sub-Saharan Africa

As case numbers here are generally low, I’ll show only the total cases per million graphs.




The Seychelles is the high outlier here, with 19.7% of its population having become cases. Cape Verde, Botswana, Namibia, South Africa and Eswatini are all in the 3% to 6% range. The multi-wave pattern is in evidence, too. The other countries are all around or below 1%.

Rest of Asia


Case counts are low here, with the Philippines leading the way at about 1.6% of the population. Taiwan had almost zero cases before May 2021, and China – if we can believe their figures – has had all but zero cases per million ever since the end of the first wave in about May 2020. The multi-wave pattern is apparent in Hong Kong, South Korea and Japan, at least. And it looks as if Japan, and possibly the Philippines too, have received the delta variant.


In north-east Asia, the multi-wave pattern is with us again, and Georgia is leading the pack with more than 12% of the population having become cases. Cases in Uzbekistan and Tajikistan are minimal; 0.44% and 0.17% of the population respectively. The remainder are in the 2% to 6% range.

The timing of the recent rise in cases in Georgia suggests that it may be due to the delta variant, and it could very well be contributing in Azerbaijan, and maybe Kazakhstan, too.

The Maldivian islanders seem to be having a European style epidemic, with more than 14% of the population having become cases. None of the others have gone much above 2%, and Bhutan is the low outlier with only 0.33%.

In South-East Asia, cases are low, not more than 1.5%; except for Malaysia, where about 4.5% of the population have become cases.

Australasia and Oceania

The only country here with a significant case count yet is Fiji, with almost 5% of the population having become cases since June 2021. Again, showing that once the virus does get in, it is difficult to contain it, even if you have previously been successful against it.

Top and Bottom 20

Here are the lists of the top and bottom 20 countries in cases per million.



The top 20 all have more than 10% of their populations having been diagnosed as cases. But many of these countries have small populations; and the virus seems to spread more easily in small countries than in larger ones. Serbia, Andorra, Montenegro, San Marino, the Maldives and Luxembourg all have populations of less than a million. Bahrain and Cyprus have less than two million, and Slovenia and Lithuania not far above that number. Czechia, Sweden, Israel and Serbia are all near the 10 million mark. Only the Netherlands, the USA and Argentina among these top 20 have populations significantly over 10 million.

The bottom 20, on the other hand, are a mixed bag. There are remote island countries. There are countries in Africa, where the virus seems to spread relatively slowly; and many of them have relatively little interaction with their neighbours. There are countries which closed their borders hard and early, like Taiwan and New Zealand. There’s Yemen, which is a war zone. And then, there’s China.

So, which countries will I follow up this time round? Czechia is a must; as are Bahrain, the Netherlands, the USA and Uruguay. And from the bottom 20, Taiwan and China are the obvious choices; similar ethnicities, but very different politics.

The current “hot spots”

But before I do individual country follow-ups, I’ll show the top 20 countries in order of current level of daily cases per million (weekly averaged):


They look to be well spread out around the world! It’s worth noting that Israel, the Seychelles and the UK were among the countries I looked into in detail in the vaccinations paper. Thus showing, that having a high level of vaccinations isn’t the automatic panacea that some seem to think.

The bottom 20 are also worth showing, if only for the enormous difference in the scales of the two graphs, and the position occupied by China:

Some examples

Czechia



The most significant things on these two graphs, I think, are on the right-hand side. Cases have been pretty much flat-lining since the middle of June, at just a few hundred new cases per day – orders of magnitude below the values of over 12,000 at the peaks. And there’s the sudden drop-off in the R-rate and weekly case growth at the end of June, following an equally sudden rise, which may have been due to the delta variant getting loose.

This looks much like what happened in Belgium at almost exactly the same time. The difference is that the Czechs, unlike the Belgians, chose to tighten lockdown from 35% to 37%. On July 23rd they closed schools in some regions, and tightened their national face mask mandate to “Required when with others.” This has had no visible effect, as the drop in R-rate and weekly case growth had already happened by the time the lockdown came into force; and the cases have continued to flatline up to the date of writing (August 26th).

Bahrain


In Bahrain, the cases have been less heavily concentrated in large peaks than in most other countries; which may help to explain the high total cases per million. Lockdowns seem to have been constantly and tightly micro-managed. While the current stringency is only 54%, the measures are quite harsh, including mandatory closure of all schools and some workplaces, and face coverings required everywhere outside the home.

The right-hand end of the cases graph looks very much like Czechia’s. The current trend in the R-rate looks interesting; if that continues to go down, the Bahrainis could be in good shape. But it isn’t clear from these figures whether or not they have had the delta variant yet.

The Netherlands


As in Belgium, there was a big spike in cases, starting at the end of June; probably the delta variant making its mark. This coincided with a major unlock, which was partially reversed on July 10th by closing some workplaces. There was then a spectacular drop in new cases. But given the example of Czechia, I wonder if the Dutch might be close enough to herd immunity that this drop might have happened without the closures? New cases are now (August 26th) going up, but barely so. We’ll have to see when the R-rate next peaks.

USA


In the USA, the weekly case growth spiked hugely at the end of June, like Czechia, the Netherlands and Belgium. Very likely, this was the delta variant. It’s not entirely obvious on the cases graph, but the week-on-week rate of increase of new cases has dropped significantly since then, though it is still above zero. Other things being equal, the next few weeks should see a flattening out of the new cases curve, hopefully followed by another fall.

Lockdowns in the USA are not easy to track, as almost all the measures are at the individual state level, and large states even have their own regional lockdowns. Moreover, the national stringency tends to overstate the stringency levels in individual states. But the general lockdown trend since June and before has been slowly downward.

Uruguay


I recounted earlier the Uruguayan story, of initial success (due to relative isolation?) followed by two big peaks. Things seem to be better again now. Cases are still (just) going downwards. The R-rate is still below 1, but going up. The Uruguayans are unlocking, but the lockdown stringency is still above 50%. The borders are closed, there is a medium restriction on gatherings, and face coverings are required in some places; but the rest of the mandatory measures are gone. It will be interesting to see what happens when they re-open the borders.

Taiwan


The Taiwanese story looks at first sight similar to Uruguay’s. Almost no cases at all, helped by early closure of the borders; until a sudden surge in the middle of May 2021. But unlike Uruguay, this surge only caused 0.7% of the population to become cases.

Lockdowns were generally low, until the surge. The borders are now closed again, some schools and workplaces are closed, and face coverings are required when with others. The Taiwanese don’t seem to be bothering too much with vaccinations: 39% of the population have had one jab, while only 3% have had two. And there’s no evidence from what I see that the delta variant has found Taiwan yet. It looks to me as if the epidemic in Taiwan still has some way to run.

China


China. Ah, China. Do I believe that epidemic profile? I’m inclined not to, even though there is a certain resemblance to Taiwan’s profile. I’m dubious about the enormous claimed number of vaccinations, too. And I’m not sure whether, as suggested by events in the Seychelles – if not also, according to reports, Bahrain, Chile, Indonesia and Mongolia – the Chinese vaccines might not be as good as they claim them to be.

But the Chinese lockdown figures have been generally high, which I suppose might be believable. As in the USA, the national lockdown figure tends to overstate the lockdown levels in individual provinces. But even provincial stringencies have not often been below 50%. The Chinese have micro-managed lockdowns in a way that makes the Bahrainis look like amateurs; in August 2021 alone, there have been 70 different changes in lockdown measures in 26 provinces (out of 27!) And over the course of the whole epidemic, there have been 1,060 such changes.

To keep on managing these lockdowns, and to keep them going for such a long period, when none of the outbreaks (since the initial one) have looked at all serious by the standards of other countries, gives me furiously to think. What is their actual strategy for dealing with the thing in the long term? And what might the Chinese know about this virus that we don’t?

To sum up

I’ve learned a lot from this exercise.

First, initial success in keeping cases down is, to use a phrase from the finance industry, “no guarantee of future performance.” Even absent new variants, if there is a pool of susceptible people anywhere, the virus will eventually find them. And even if you manage to stop it the first time, it will try again. The Uruguayans, among others, have found that out the hard way.

Second, the cases per million to date are generally highest in Europe, next highest in South America, a little lower in North and Central America and the Middle East, and lowest in Asia and Africa. Small countries, particularly islands, may get a lot more cases per million than their locations suggest. But there are also a lot of countries, in which the penetration of the virus into the population in percentage terms is low or even very low.

Third, the delta variant seems to be present in most of Europe, in the USA and perhaps in the West Indies. Also in Japan, and perhaps in the Philippines and a few other countries, particularly in the Middle East.

Fourth, three European countries I have looked at in the first two papers (Belgium, Netherlands, Czechia) appear to be approaching a kind of equilibrium, in which new cases remain low while the level of lockdown is far lower than would have allowed this earlier in the epidemic. How far this equilibrium is from full herd immunity, and how far the herd immunity target might go up if a new variant gets loose, remain to be seen.

Fifth, the Chinese are not to be trusted. Not their data, nor their vaccines; and, very probably, not their intentions.