In a March 11 post, co-blogger Scott Sumner went Jeopardy style, with:
The answer: One week.
And the question? “What’s the difference between Italy and the rest of the Western world?”
Scott made the point that we are only a week behind Italy in the cumulative number of cases of Covid-19. He turns out to be roughly right. But fortunately, the most important comparison is the number of deaths and on that, we are doing way better than Italy did a week earlier.
On March 12, Italy’s total number of cases was 15,113. On March 19, the United States had 13,789. Pretty close.
But now, let’s look at fatalities. On March 12, Italy had 1,116 deaths. On March 19, the United States had 207 deaths. That’s over 5 times as many. So already I think we can say that we in the United States are strongly outperforming Italy on the main measure that matters.
READER COMMENTS
Alan Goldhammer
Mar 20 2020 at 7:04pm
David – I hope this holds!!! The critical test will be what happens in New York City given the density of population. Lots of elderly people as well. It’s perhaps the most similar to the dense regions in Lombardy. The next 2-3 weeks will tell us if this holds.
David Henderson
Mar 20 2020 at 7:22pm
Good point.
Mark Bahner
Mar 20 2020 at 11:42pm
Hi,
There are many, many differences between the U.S. and Italy. (In fairness to Scott Sumner, I had to see the different trend between the U.S. and Italy to realize the many, many differences. So my quarterbacking on Monday is much better than Sunday. :-))
The differences between the U.S. and Italy include: 1) we have a significantly younger population, 2) we have more square footage per person in our residences and buildings, 3) we are less densely populated, 4) we have customs involving less social contact (e.g., no cheek kissing), 5) we have fewer situations with extended families living in one residence, 6) we are signficantly more wealthy per capita, and 7) we have more capita (even though we are less densely populated) so that we have much more total money. (If you’ve got a disease running around, it’s better to have 330 million people making $60,000 a year than it is even if the other country has $80 million people making the same $60,000 per year. That’s just like in a war, where it’s better to have a larger population than a smaller population, even if both populations have the same median income.)
Mark Z
Mar 21 2020 at 1:27am
Developed countries with comparably old populations seem to be doing much better than Italy, and with dense populations as well (Japan, South Korea, Germany). In the US mortality rates in NY are pretty low as well.
The other cultural factors you mention might help explain the difference, but a factor of 5 is a huge difference that requires quite an explanation; Italy really stands out and I don’t find the explanations that have been suggested in recent weeks fully satisfying.
Thomas Knapp
Mar 21 2020 at 6:09am
Two other differences between the US and Italy: Italians are twice as likely to smoke as Americans, and Italy’s air quality is worse than America’s. Both of which are negative factors for good lung function.
Mark Bahner
Mar 21 2020 at 8:23am
Based on this website, it looks like Italians smoke much more than people in the U.S., but not “twice as much” (unless you somehow now about amount of cigarettes smoked, versus simply percentage of people who do smoke). The U.S. is listed at 17.25% smokers overall, versus Italians at 24.00% overall.
And if we’re considering smoking, it looks like Germans (30% overall), Japanese (22%) and South Koreans (27%) are all big smokers, too. The Germans and South Koreans are even bigger smokers (in prevalence) than the Italians.
Air quality seems like a possibility…though I don’t know much about the air quality in South Korea.
One thing I’ll cling to after Mark Z. destroyed all my other possibilities is something about building construction. Even there, the Japanese, Germans, and South Koreans have smaller residences, just like the Italians. Sooooo…much of the buildings in Japan, Germany, and South Korea are probably more modern than Italy. That’s just guessing that many more buildings were destroyed in WWII in Germany and Japan than in Italy…and that South Korea’s buildings are mostly post-Korean war growth in South Korea.
So other than the Italians’ habit of close personal social contact, and possibly something about buildings, and maybe something about air quality, I’m still pondering.
The difference between Italy and Germany seems particularly puzzling, in light of the Germans being even bigger smokers than the Japanese.
Mark Bahner
Mar 21 2020 at 8:26am
Oops, forgot to include the smoking website:
Smoking rates by country
P.S. What a world this is, that I can find that wonderful website in a one-minute search. Before the Internet, think of how long it might take me to find a website like that! How would I even do it…what research library would even have something like that?
Dylan
Mar 21 2020 at 8:38am
Germany however, has an even higher rate of smoking than Italy, highest in Western Europe I think with the exception of Austria.
Glen Raphael
Mar 22 2020 at 1:34am
A factor of 5 almost perfectly reflects the difference in population. Italy’s number of cases is about 5 times larger per capita so their hospitals are far more likely to be overburdened than ours.
Mark Bahner
Mar 21 2020 at 8:06am
Excellent points about Japan, South Korea, and Germany.
Yes, I think the cultural factor stays, even for Japan, South Korea, and Germany. But I agree, a factor of 5 difference explained by cultural factors (closer personal contact) seems surprisingly large.
Good points! If you find anything that seems like strong evidence explaining Italy to you, I’d be interested to read it.
Mark Z
Mar 22 2020 at 4:31pm
Thanks. A possible explanation I just came across is how death are reported in Italy. It’s mentioned in a Telegraph article (https://www.telegraph.co.uk/global-health/science-and-disease/have-many-coronavirus-patients-died-italy/); basically, Italians are far more generous in attributing death to the virus; if you die and you are infected, you are counted as a COVID-19 death. But: “On re-evaluation by the National Institute of Health, only 12 per cent of death certificates have shown a direct causality from coronavirus, while 88 per cent of patients who have died have at least one pre-morbidity – many had two or three.”
Almost 90% of deaths in Italy are people over 70, and they’re disproportionately sick for other reasons too. So you have many patients who’s life expectancy is maybe only a few more years. I’d ask, what are the odds of them dying over a given ~4 week period? It’s possible that a sizable fraction of these deaths are merely coincident with, not caused by, the virus.
Another suspect: nosocomial infections. Within-hospital transmission is apparently a huge problem there. I’m not sure why it’d be worse there than elsewhere, but possible more cramped hospitals, worse hygiene, etc. could really exacerbate it, especially since there’s already such a high concentration of sick old people in hospitals. A relevant article: https://www.statnews.com/2020/03/21/coronavirus-plea-from-italy-treat-patients-at-home/
Matthias Görgens
Mar 21 2020 at 1:05am
The main measure that matters is quality adjusted life years (QALY) lost.
Of course, death removes all your expected remaining life years. Italy has a lot of old people. They are more likely to die to the disease, but each death has less impact than the death of a young person.
QALY also counts time lost to acute disease and the impact of long term impairments.
In the widest sense, QALY also counts lower quality of life from economi
Matthias Görgens
Mar 21 2020 at 1:08am
(Hit submit too early..)
QALY also counts lower quality of life from economic disruptions. But that’s a bit more controversial to measure. Given that quality adjustments for eg blindness or losing a limb are already pretty subjective.
Interestingly, I remember an estimate that the Coronavirus has resulted in a net of something like -40,000 deaths so far. That’s right, fewer people died.
That calculation mostly relied on vastly fewer premature deaths from air pollution, because the Chinese economy shut down. That wasn’t a QALY calculation.
Mark Bahner
Mar 21 2020 at 8:36am
Not only does COVID-19 kill old people at a much higher rate, it is particularly hard on old people with co-morbidities (e.g. COPD, heart disease). That’s not surprising, of course. But what it says is that those old people actually had fewer expected years of life than would be indicated by their age on an actuarial life table:
U.S. Social Security actuarial life table
I hate to give another spoiler–and probably other people have done similar things–but the U.S. response to COVID-19 has been absolutely insane, in terms of over-reaction.
Matthias Görgens
Mar 21 2020 at 12:26pm
They are over reacting and under reacting.
Mostly they are also doing lots of silly things like banning independent testing.
AMT
Mar 21 2020 at 1:12am
Given the flat out lies and utter incompetence of our government (which anyone leaning libertarian ought to agree on, given the minimum expected incomptence in a system with such lacking incentives), it seems clear that American cases are VASTLY understated:
https://www.stltoday.com/news/local/one-st-louis-doctor-s-ordeal-i-am-so-frustrated/article_afdab0dd-de1e-54eb-b7c8-a573a311e894.amp.html?__twitter_impression=true&fbclid=IwAR1IdjMEddQkNgYjZ3iat6fGPosm_Sf0N_bbn9NDFqWCQjsRYB4kro_A8I4
Do you believe there is no shortage of tests? Well, sure, if you close your eyes, plug your ears and sing “lalalalala” you can maybe pretend nothing bad is happening.
Do we have any evidence or reason to believe that the extensive lying and deliberate obfuscation of the truth is anywhere near as widespread in Italy?
On a Monday, Trump said Corona was no big deal and just fake news. On a Wednesday he closed travel from Europe. O a Friday he declared it a national emergency. I don’t that we can just take his word for it that “everyone who needs a test can get one.” We have heard from too many disinterested sources that there is a masive lack of testing. So acting like American numbers can be compared to others is extraordinarily optimistic.
Trevor W Adcock
Mar 21 2020 at 2:58am
In this scenario if America’s number of cases is even higher, but deaths are the same, then that means we are doing even better!
Unless you think we are counting a lot of deaths from COVID-19 as deaths from the flu, but I haven’t heard anything to suggest that is the same.
Maybe our health care system implodes soon if you’re right, but that we haven’t already even with under counting suggests America isn’t doing nearly as bad as Italy.
Mark Z
Mar 21 2020 at 3:22am
The lack of tests should generally bias the observed mortality rate upward though, not downward. If Italy is testing far more extensively than the US, then that only heightens the oddness of their abnormally high observed mortality rate.
AMT
Mar 21 2020 at 10:27am
Did you actually read the article? Two people died, very likely of Corona virus, but they will be counted as “pneumonia,” because we cannot confirm with testing.
Alan Reynolds
Mar 21 2020 at 11:07am
I agree that the number of U.S. cases are vastly underestimated, which is why I have argued since March 2 that the case death rates have been vastly overestimated. If you only test the very old and very sick (e.g., residents of a Kirkland WA nursing home), the death rate is sure to look very high even if it was mainly from, say, lung cancer or COPD.
We are still testing only those with a fever and symptoms, but about 90% of those tests are negative. The wider the net of testing becomes, the more milder cases will be added to the denominator thus pushing down the ratio of deaths to “confirmed” infections. New York State quickly found over 7000 cases by aggressive testing but the death rate there at last count had dropped 0.05%. It will likely fall further.
Alan Goldhammer
Mar 21 2020 at 11:33am
One of the difficulties right now is the limitation of RT-PCR testing. Most major medical schools have implemented their own testing protocols so they can get real time data on patients entering (my daughter is at UCSF Children’s Hospital). This type of testing is critical for patient isolation and care. I don’t know whether these groups are reporting to CDC and in any event it’s a very small number. Commerical testing labs take a 2-3 days to turn around results as specimens need to be shipped to where the testing is taking place.
What is badly needed is a serological test for COVID-19 antibodies. These are much easier tests to do and can be run as part of normal blood panels. It can be expanded quicker and to asymptomatic individuals to assess past exposure/infection. In addition to keeping up with the clinical trial and drug/vaccine development data, I’m now looking at reports for antibody testing of those who have been infected. This morning I saw maybe a half dozen pre-prints from China that point to the development of a blood test over there. I don’t know what the state of this research is in Europe as I’ve not seen any papers.
Brent Buckner
Mar 21 2020 at 10:01am
Not to be alarmist, BUT: *if* the only material difference were total number of ICU beds available then a few additional days of exponential growth in the US would put the US also over capacity and fatality rates would converge.
Scott Sumner
Mar 21 2020 at 1:48pm
I agree, but notice that Spain already has more deaths than Italy did on March 12. It will take the US a bit longer to get there, but we are not all that far behind, assuming exponential growth.
BTW, I also said I did not expect the growth rates to continue, so while I think they’ll rise sharply in the short run, I also believe they will level off as social distancing kicks in.
Your post does point to something I’ve been discussing over at Econlog. The differences in death rates are extremely large, probably too large to be explained by any one factor. Thus compare Germany and Spain. It can’t just be demographics, or the length of time the virus has been in each country, Spain and Germany don’t differ that much. Perhaps Spain has many more unreported cases than Germany.
David Henderson
Mar 21 2020 at 2:43pm
You write:
Sure, but the point is that we’re not a week behind. I’m trying to communicate to people who might not have read your original post as carefully as I did.
In fact, I bet we’re more than two weeks behind.
Scott Sumner
Mar 21 2020 at 5:55pm
Agreed.
Mark S Barbieri
Mar 21 2020 at 1:53pm
Given the large variance in testing rates and protocols on who gets tested, is it meaningful to compare the rates of those who tested positive vs those who died from the disease? High death rates per positive test may just show that they are only testing the sickest of the infected people.
David Henderson
Mar 21 2020 at 2:41pm
You write:
It’s meaningful, but it is true that rates of testing may drive this.
Fortunately, it doesn’t affect my point. I wrote about deaths not about rates. I think we can be fairly sure that they’re measuring deaths correctly. So my point stands.
AMT
Mar 21 2020 at 5:00pm
I disagree. I think it’s fair to say that generally speaking if you have no shortage of tests available, you will test all the more severe cases and many of the marginal or mild cases as well. If you have a severe shortage of tests available, then you are much more likely to not only miss some marginal cases, but also serious and fatal cases, given the somewhat arbitrary rationing that takes place. See the article I linked to above.
We aren’t just avoiding testing the mild cases that don’t result in deaths. So, I think you could very easily end up understating the death rate. To know how much, you’d have to get an idea of how common it is where people are refused to be given tests and end up dying. If it was only a slight shortage, I’d put the risk of failing to test moderate or more severe cases as very low, but it appears we are suffering very acute shortages of tests, given how extreme the rationing is. Again, see the article I linked to above.
I’ll add that given Trump claims there is an abundance of tests, that makes it all the more certain there is a massive shortage…
AMT
Apr 5 2020 at 9:05pm
Not even close:
https://www.msn.com/en-us/news/us/coronavirus-death-toll-americans-are-almost-certainly-dying-of-covid-19-but-being-left-out-of-the-official-count/ar-BB12bSIa?ocid=spartanntp
Gordon
Mar 21 2020 at 5:00pm
Italy has a population of 60 million which means 15,000 cases has more of an impact on the medical system there than the US. And a week ago, the hospitals in Italy were so overwhelmed that critically ill people ended up dying at home. If US cases were on the order of 75,000 and we had a lower fatality rate, that would be meaningful.
AT
Mar 22 2020 at 5:14am
If all the older people and those with health problems who are at risk from the virus isolate themselves, so that only young and healthy people get infected, then you could have a much lower death rate.
Possibly some countries have been better at doing this.
In Italy, things happened very quickly, so people had less time to find out about the virus and prepare for it.
Eric Garris
Mar 22 2020 at 11:58am
I think there is a possible and simple explanation about the high rate of COVID-19 in Italy, France, and Spain. These are heavily Catholic nations, where many residents attend Mass weekly. It involves lining up to have a priest place a wafer on their tongue. The priest does not clean his hand between each person. In addition, attendees often drink from the same chalice during the Mass. This is followed by the attendees each shaking hands with the priest as they exit the church.
In Southern Italy an outbreak was linked to this activity during a religious retreat: https://cruxnow.com/church-in-europe/2020/03/coronavirus-outbreak-in-southern-italy-linked-to-neocatechumenate-retreat/
There is currently a debate among leaders of the Eastern Orthodox churches about the practice of using a shared spoon to deliver communion. “We believe that no virus or disease can be transmitted through communion,” said Metropolitan Ilarion, of the Moscow Patriarchate, on Rossia-24 on March 7: https://www.rferl.org/a/coronavirus-vs-the-church-orthodox-traditionalists-stand-behind-the-holy-spoon/30492749.html
Warren Platts
Mar 22 2020 at 2:18pm
The main difference is that on February 1st, at the behest of the Chinese Communist Party (who produced a couple of slick videos of the event), they held a “Hug a Chinese Day” event in northern Italy. In that region there are at least 300,000 Mainland Chinese citizens living there mainly working in the garment industry sweatshops there. Not to mention there were planeloads of people from Wuhan returning from their New Years holiday because the Italian government was afraid of offending China.
The result was that the epidemic in Italy, instead of being set off with a slow fuse, it was set off with a detonator. The first cases started showing up the week of February 16. The first recorded death was February 21. As a result, the Italian health system was caught totally flatfooted. The reason the death rate is so high there is because they have been forced into a triage situation where there is one respirator for each 3 people who need one.
Consider that on February 1, in China, the “official” number of cases was 14,380. The number of “official” number of deaths was 304. (I put “official” in scare quotes because any information coming out of the CCP is highly suspect. Consider that between December and February, the number of cell phone accounts in China declined by 21 MILLION. If you know anything about China, you know you cannot survive there without a cell phone.)
Therefore, the Chinese authorities by that time knew full well what was going on. They knew full well that hugging and kissing random strangers is the diametrical opposite of what people should have been doing. Consequently, it is really, really hard to believe that the “Hug a Chinese Day” event was not an intentional biological warfare attack.
That is the difference between Italy and the United States. And don’t kid yourselves. Our day is coming, starting right now. Everybody is going to contract the virus. We are only trying to slow the spread so we don’t wind up like Italy where a lot of unnecessary death is going on because they don’t have the resources to handle it. My advice is to try and get as healthy as you can. Eat healthy, get exercise, take your vitamins, and if you are a smoker, now is a good time to quit…
Warren Platts
Mar 22 2020 at 2:25pm
And in case you are curious, here are the videos I was referring to. The first was made by Chinese filmmaker Hai Chi who lives in Milan now. The second I think was made by CGTN–China Global Television Network–one of many propaganda arms of the CCP. (Coincidentally, I am sure, the only reporter turned away from the recent White House briefings for having a fever was a CGTN reporter.)
https://www.youtube.com/watch?v=JEjDMRENuy4
https://www.youtube.com/watch?v=mNMdg4morQs
marc alberts
Mar 23 2020 at 9:00am
“But fortunately, the most important comparison is the number of deaths and on that, we are doing way better than Italy did a week earlier.”
That actually isn’t the most important comparison, unless we’re just looking at the scorecard. From an economic perspective, it is not the most critical number. The death rates don’t kill the hospital capacity, which is the real crisis. The most important number, from a macro view, is the number of ICU cases and the number of critical ICU cases requiring a ventilator. Those are the ones that will tell us if we’re in more or less of a crisis than Italy. A bed taken up with COVID-19 is a bed that isn’t available for a car accident victim, a heart attack patient, etc., so the societal overflow effect is compounded by that more than death rates. And when you measure death rates, raw numbers aren’t really a good indicator. It needs to be deaths/recovered cases. For anyone not dead or recovered, you still have an outcome in doubt where you can’t ascribe it automatically to survival, which is what most reports seem to be doing.
David Henderson
Mar 24 2020 at 11:45am
Good point.
Ashley
Mar 27 2020 at 5:51pm
Another way to look at this is that the average time from catching the disease to death is 9 days, and that each of the deaths today was a potential ‘detected case’ 9 days ago.
Another way to look at the data from the same date range in your original post:
USA:
March 19th – 207 deaths
8 days earlier… 994 Detected Cases
Italy:
March 07 – 233 deaths
9 days earlier – 655 Detected Cases
There is still a gap, but maybe not a 5X gap.
TapeOp
Apr 8 2020 at 5:49pm
Well, this post didn’t age well.
Comments are closed.