[Goose]

I also believe the death count is artificially low because the German health system generally reports the proximate cause of death (i.e. pneumonia) even when the patient is COVID-19 positive.

I've seen that argument made several places, but I haven't seen any evidence that this is the case.  In fact, analyses of excess mortality in Germany (mortality above and beyond what would have been expected, which would catch COVID-19 related deaths that were attributed to something else) don't suggest there is a large number of deaths reported as being caused by something else that were really called by COVID-19.  Here are a couple of articles:

BC, I have located a couple of references to this subject. However, I can't find the one that I saw early on about how deaths were reported. The references I have are to the fact that if you died without a COVID-19 test having been performed, you wouldn't be reported as a COVID case no matter what. Even if the symptoms matched COVID-19, without a positive test, it wasn't COVID-19. That probably mattered some in the early going (March early April) but not much in the last two months or so when testing was very plentiful. One of the links is:
https://www.businessinsider.com/why-germ...ate-2020-4
I have a couple of similar links, but you get the idea.

The link I can't find is to an article that specifically said that in Germany cause of death reports are listed as the proximate cause, not the underlying cause. The example given was Pneumonia as a cause of death even though the patient was COVID-19 positive. The government put out a directive to stop doing that, as I guess that was the standard of practice in Germany. I am sure I read this early on in the game, but I can't locate it now.

By this point in time, both these effects probably have washed out, to whatever unknown extent they mattered early on. Germany, like the rest of Europe, has pretty well flattened their curve and can start opening up with some hope of a predictable outcome. We in the USA certainly have not.

https://www.dw.com/en/coronavirus-german...a-53537500

Germany saw a slight increase in the average number of deaths in April, the Federal Statistics Office reported on Friday. April's death rate was around 5% more than the average number of deaths in April between 2016 and 2019....

"Excess mortality in Germany is low compared to other European countries," the statisticians reported. "Italy reports 49% more deaths in March 2020 than in the years 2015 to 2019, on average [...] Sweden reports twice as many deaths in the agglomeration area of Stockholm for weeks 14 to 16 [mid-April] as, on average, the five years before."

https://www.ft.com/content/a26fbf7e-48f8...5839e06441

Through May 24, the FT identified 9200 excess deaths in Germany.  Through May 24, Germany had reported 8,371 COVID-19 deaths, which is 90% of the excess deaths, which is within the error of the method of identifying expected mortality based on prior years at the same time.  Moreover, the gap is lower than it is in other European countries.  Also, excess deaths would include deaths due to other reasons (for example heart attacks) that were impacted by delay in seeking care either because of lockdown or fear of contracting COVID-19 in a hospital.

But it would also include fewer deaths due to the lack of activity during the lockdown, such as traffic accidents, industrial accidents etc., would it not? Certainly traffic accidents would be significant. This could be normalized out by looking at each cause of death and compensating accordingly by historical data on each cause. Do they do that? It also would include fewer deaths due to other infectious diseases that are suppressed by the lockdown, like flu. If they don't "correct" the numbers, it would artificially reduce the excess deaths, so this would be important.

I conclude that German reports of COVID-19 deaths are a reasonably accurate reflection of the true number of deaths that occurred in Germany due to COVID-19, and that the idea that Germany was systematically mis-coding cause of death is either apocryphal or does not significant impact the numbers.

BC

I would go with not significant. Even a 25% undercount isn't going to change the basic idea that they got better results. The articles I have read attribute this to better testing, better contact tracing, better ICU availability (although it is not clear that this ever actually mattered), and better compliance. Seems reasonable to me. Many articles throw in references to lucky timing both in initial cases and in a significant delay before the full onset, which helped as well.

[Goose]

I also believe the death count is artificially low because the German health system generally reports the proximate cause of death (i.e. pneumonia) even when the patient is COVID-19 positive.

I've seen that argument made several places, but I haven't seen any evidence that this is the case.  In fact, analyses of excess mortality in Germany (mortality above and beyond what would have been expected, which would catch COVID-19 related deaths that were attributed to something else) don't suggest there is a large number of deaths reported as being caused by something else that were really called by COVID-19.  Here are a couple of articles:

BC, I have located a couple of references to this subject. However, I can't find the one that I saw early on about how deaths were reported. The references I have are to the fact that if you died without a COVID-19 test having been performed, you wouldn't be reported as a COVID case no matter what. Even if the symptoms matched COVID-19, without a positive test, it wasn't COVID-19. That probably mattered some in the early going (March early April) but not much in the last two months or so when testing was very plentiful. One of the links is:
https://www.businessinsider.com/why-germ...ate-2020-4
I have a couple of similar links, but you get the idea.

The link I can't find is to an article that specifically said that in Germany cause of death reports are listed as the proximate cause, not the underlying cause. The example given was Pneumonia as a cause of death even though the patient was COVID-19 positive. The government put out a directive to stop doing that, as I guess that was the standard of practice in Germany. I am sure I read this early on in the game, but I can't locate it now.

By this point in time, both these effects probably have washed out, to whatever unknown extent they mattered early on. Germany, like the rest of Europe, has pretty well flattened their curve and can start opening up with some hope of a predictable outcome. We in the USA certainly have not.

https://www.dw.com/en/coronavirus-german...a-53537500

Germany saw a slight increase in the average number of deaths in April, the Federal Statistics Office reported on Friday. April's death rate was around 5% more than the average number of deaths in April between 2016 and 2019....

"Excess mortality in Germany is low compared to other European countries," the statisticians reported. "Italy reports 49% more deaths in March 2020 than in the years 2015 to 2019, on average [...] Sweden reports twice as many deaths in the agglomeration area of Stockholm for weeks 14 to 16 [mid-April] as, on average, the five years before."

https://www.ft.com/content/a26fbf7e-48f8...5839e06441

Through May 24, the FT identified 9200 excess deaths in Germany.  Through May 24, Germany had reported 8,371 COVID-19 deaths, which is 90% of the excess deaths, which is within the error of the method of identifying expected mortality based on prior years at the same time.  Moreover, the gap is lower than it is in other European countries.  Also, excess deaths would include deaths due to other reasons (for example heart attacks) that were impacted by delay in seeking care either because of lockdown or fear of contracting COVID-19 in a hospital.

But it would also include fewer deaths due to the lack of activity during the lockdown, such as traffic accidents, industrial accidents etc., would it not? Certainly traffic accidents would be significant. This could be normalized out by looking at each cause of death and compensating accordingly by historical data on each cause. Do they do that? It also would include fewer deaths due to other infectious diseases that are suppressed by the lockdown, like flu. If they don't "correct" the numbers, it would artificially reduce the excess deaths, so this would be important.

I conclude that German reports of COVID-19 deaths are a reasonably accurate reflection of the true number of deaths that occurred in Germany due to COVID-19, and that the idea that Germany was systematically mis-coding cause of death is either apocryphal or does not significant impact the numbers.

BC

I would go with not significant. Even a 25% undercount isn't going to change the basic idea that they got better results. The articles I have read attribute this to better testing, better contact tracing, better ICU availability (although it is not clear that this ever actually mattered), and better compliance. Seems reasonable to me. Many articles throw in references to lucky timing both in initial cases and in a significant delay before the full onset, which helped as well.