Coronavirus and Cities

There’s a meme going around the American discourse saying that the Covid-19 outbreak is proving that dense cities are bad. Most of this is bullshit from politicians, like Andrew Cuomo. But now there’s serious research on the subject, by a team at Marron led by the excellent Solly Angel. Solly’s paper looks at confirmed infection rates in American metropolitan areas as of late March and finds a significant correlation with density, but no significant correlation between deaths and density. In this post, I’m going to look at Germany. Here, big or dense cities are not disproportionately affected by the virus.

Why Germany?

Germany has pretty reliable data on infections because testing is fairly widespread, so far covering 1.6% of the population. Moreover, testing is this high throughout the country, whereas in the US, there are vast differences in testing as well as in other aspects of response by state, e.g. New York has tested 2% of state population, Louisiana 1.9%, Florida 0.8%, California and Texas 0.4%.

I also have granular data on infection rates in Germany, thanks to Zeit. The data I’m using is synchronic rather than diachronic, i.e. I’m looking at current infection rates rather than growth. Growth rates aren’t the same everywhere – in particular, they’re lower in North Rhine-Westphalia, which was the epicenter of the German outbreak weeks ago, than in southern Germany – but they’re low enough that I don’t think the situation will change in short order.

Size and density

Within Germany, there aren’t huge gradients in density between cities. More central neighborhoods have taller buildings than less central ones and higher ratios of building to courtyard, but there are no huge differences in residential built form the way there are between American cities.

For example, look at densities by neighborhood in Berlin, Hamburg, Munich, Frankfurt, Cologne, Stuttgart. There aren’t big differences in the pattern: the densest inner neighborhoods have about 15,000 people per square kilometer, and density falls to 3,000-5,000 in outer neighborhoods. Hamburg has a few areas with no residents, since they include the city’s immense port. Stuttgart’s densest districts are in the 5,000-6,000/km^2 range, but that’s because the districts are not very granular and the dense ring of inner-city neighborhoods just outside the commercial center is not congruent to district boundaries.

The upshot is that the big question about density and the risk of epidemics cannot be answered by comparing German cities to one another, but only to the surrounding rural areas. So the real question should be, are the major German cities more afflicted by the virus than the rest of the country?

Infection rates by city

As of the end of 2020-04-09, Zeit reports 118,215 confirmed coronavirus cases, which is 14.2 per 10,000 people. The six states of former East Germany, counting the entirety of Berlin and not just East Berlin, total only 12,873 cases, or 7.9 per 10,000. The Robert Koch Institute’s definitive numbers are slightly lower, but are also slightly outdated, as states sometimes take 1-2 days to report new cases. Going by Zeit data, we have the following infection rates by major city:

City Population Cases Cases/10,000
Berlin 3,644,826 4,357 12
Hamburg 1,841,179 3,518 19.1
Munich 1,471,508 4,123 28
Hanover* 1,157,624 1,389 12
Cologne 1,085,664 1,947 17.9
Frankfurt 753,056 730 9.7
Stuttgart 634,830 1,056 16.6
Dusseldorf 619,294 737 11.9
Leipzig 587,857 451 7.7
Dortmund 587,010 507 8.6
Essen 583,109 578 9.9
Bremen 569,352 425 7.5
Dresden 554,649 476 8.6
Nuremberg 518,365 733 14.1
Duisburg 498,590 525 10.5

*Zeit reports Hanover data for the entire region; the city itself only has 538,000 people

The sum total of the fifteen largest cities in Germany, with 15.1 million people, is 21,552 cases, which is 14.3 cases per 10,000 people. This is the same as in the rest of the country to within measurement error of total population, let alone to within measurement error of Covid-19 cases.

State patterns

Bavaria and Baden-Württemberg both have high confirmed case counts, averaging 23.6 and 21.7 per 10,000 people respectively. Munich’s rate is somewhat higher than the Bavarian average, but its suburbs are on a par with the city, as are some entirely rural areas all over the state. Oddly, the second and third largest cities in the state, Nuremberg and Augsburg, have lower rates – though both Fürth and the rural areas around Nuremberg and Fürth have very high rates as well.

The pattern around Stuttgart is perhaps similar to that around Nuremberg. The city’s infection rate is not much higher than the national average, but the infection rates in counties and cities around it are: Esslingen (24.8/10,000), Reutlingen (29.3), Tübingen (47.9), Böblingen (28.4), Ludwigsburg (22.9).

NRW’s rate is 13.9/10,000, i.e. essentially the same as the national average. The worst is in areas right on the Belgian border, like Heinsberg. Cologne has a noticeably higher rate, but Dusseldorf has a lower rate, and the cities of the Ruhr area a yet lower one. Don’t let the fact that these cities only have around 600,000 people each fool you – they’re major city centers, with the density and transportation network to boot. Dortmund alone has three independent subway-surface trunks, meeting in a Soviet triangle; total public transportation ridership in Dortmund across all modes is 130 million per year. Essen has two subway-surface trunks, one technically light rail and one technically a streetcar tunnel; total ridership there and in Mülheim, population 170,000, is 140 million per year.

What’s going on in Frankfurt?

There is some correlation between wealth and a high infection rate, since Bavaria and Baden-Württemberg have high rates of confirmed cases and the East German states have low ones. However, Frankfurt’s rate is fairly low as well, as are the rates of surrounding suburbs like Offenbach and Darmstadt. Frankfurt is not as rich as Munich, but like Hamburg and Stuttgart, it is fairly close, all three metro regions surpassing Ile-de-France and roughly matching London per Eurostat’s per capita market income net of rent and interest table.

In particular, it is unlikely that the greater international connections of rich cities like Munich explain why they have higher rates. Frankfurt Airport is the primary international hub in Germany, with many passengers standing in line at the terminal and coughing on other people. It would have been the easiest for imported infections to arise there rather than in the Rhineland, and yet it doesn’t have a major cluster.

Frankfurt also has extensive O&D business travel; Wikipedia puts it third after Berlin and Munich, but Frankfurt’s visitors are most likely disproportionately business travelers rather than tourists. This is important, since February and March are low season for tourism, whereas business travelers are if anything more likely to be going to Frankfurt during low season because during the summer high season they go on vacation in more interesting places.

So, is urban density more vulnerable to infectious diseases?

Probably not. Rural Germany has some areas with Korean levels of confirmed cases per capita, and some where 1% of the population and counting has tested positive. Overall, there isn’t much of an urban-rural difference – the 15 largest cities in Germany collectively have the same rate as the rest of the country, and moreover, where there are notable state-level patterns, they also hold for the states’ big cities. If Munich’s high infection rate is caused by its high rate of U- and S-Bahn usage, then the suburbs should have lower infection rates (they’re more auto-oriented) and the rest of Bavaria should be much lower; in reality, nearly the entirety of Bavaria has high rates.

The highest density in the developed world does not exist in Germany. German neighborhoods top at 15,000/km^2, with individual sections scratching 20,000; Paris tops at 40,000 in the 11th Arrondissement, New York scratches 50,000 on the Upper East Side, and Hong Kong has entire districts in the 50s. So the “density doesn’t matter” null hypothesis, while amply supported on German data, requires some extrapolation for the handful of world cities with the highest density.

Nonetheless, these are not huge caveats. German data is pretty reliable in the density range for which it exists; if cities today had the infection rates they did before modern plumbing, when a noticeable fraction of a city’s population might die in a single epidemic, it would be noticeable today. But there is no mass death, nor are urban hospitals here collapsing under the strain. On both the level of a basic sanity check and that of looking at the data, cities do not appear to be vulnerable to disease.

What does this mean?

There is no need to redesign the world to be less urban or dense in the wake of the coronavirus. Nor is there any need to let go of collective public transportation. The Rhine-Ruhr and Frankfurt are not Tokyo or Hong Kong in their public transportation usage, or even Paris or Berlin, but they have extensive urban and regional connections by train. And yet, the Heinsberg disaster zone and the high infection rate of Cologne have not been exported to the Ruhr, nor is southern Hesse particularly affected by German standards.

The virus has exposed serious issues with cleanliness. But even given Germany’s current levels of urban cleanliness, those issues are not enough to turn Berlin, Frankfurt, Hanover, or the Ruhr cities into hotspots. There is no danger to public health coming from urbanization, density, development, or public transportation. Cities should keep investing in all four in order to reduce the costs of transportation and environmental damage, even if the occasional failed politician blames the virus on density to deflect attention from his own incompetence.


  1. Herbert

    It’d probably be a spurious correlation because data isn’t granular enough to be meaningful, but could there be a correlation between air pollution and Covid19 cases?

      • michaelrjames

        There is lots of anticipation about the effects of the onset of the northern summer but it is mostly speculation. Pollution would probably have its major effect (in cities) via weather (ambient temperatures, humidity, sunlight).

  2. michaelrjames

    Nice. You should have submitted it to CityLab or the like. Worth a bigger audience. There will be lots of similar studies eventually but possibly only Germany really is amenable this early in the course of this virus.

    • Alon Levy

      Italy has pretty good testing nowadays too, I just haven’t seen municipality-level data, only province-level data. In the Italian data (link)you can’t even tell where the big cities are – the Province of Milan is the one immediately to the west of the darkest-color cluster. Belgian data (link), same thing – testing isn’t as extensive as here or in Italy but there’s enough of it that if it’s uniform nationwide we can draw conclusions, and one of these conclusions is that the infection rates in the big cities are mostly in line with the national average.

      • Yom Sen

        In Italy I read that the first known case (or first death) was in Codogno a small town in rural Lombardia between Lodi, Cremona and Piacenza and it happens the provinces with the highest rate of cases per capita are Lodi, Cremona and Piacenza (over 1%).
        Looking at the confirmed cases per province, we can see that the virus spread all over Italy proportionally to the distance from Codogno. Exceptions with more hit areas like Bergamo or Adriatic coast (Rimini, Pesaro) or less hit areas like Veneto have nothing to do with density. Densest provinces like Naples, Rome, Florence, Genoa have very similar rates than their regions

        For Paris and petite couronne, I saw this map that you might find interesting, you can see confirmed cases per capita per commune/arrondissement of residence. The poorest parts of the city seem to be the most hit:


        • michaelrjames

          For Paris and petite couronne, I saw this map that you might find interesting, you can see confirmed cases per capita per commune/arrondissement of residence. The poorest parts of the city seem to be the most hit:

          The biggest points on that graphic are all within intramuros Paris. It is a bit peculiar that it is concentrated in the 18, 19, (20, 12) and 13th arrondissements (clockwise from northernmost). Is it possibly the highest number of residents of Chinese (returning from China or East Asia generally)?
          The first wave of infections were the regions bordering Germany, Italy and Spain. Alsace was the hotspot for France until Paris eventually overtook it.

          • Alon Levy

            It’s not Chinese residents; in New York there are racial statistics and Asians are less infected than the other three major racial groups. Nor are the 18th, 19th, and 20th particularly Asian? They’re just poor.

            And Alsace is the hotspot not because it borders Germany – at the time, Baden-Württemberg across the border wasn’t a big cluster and even now the Württemberg parts of the state have much higher infection rates than the Baden parts. Rather, it’s because of that Evangelical meeting.

          • Yom Sen

            Yes, the 13th is indeed the main chinese neighbourhood in Paris and Belleville (between the 19th and the 20th) has also a strong chinese community.
            The biggest points were on the eastern part of Paris intra-muros the first week, but in the 2nd and 3rd weeks you can see that the virus spreaded more in Val-de-Marne and Seine-Saint-Denis and not much in the western part of Paris.
            I think the first hotspot with community spreading has been in Oise around Crépy-en-Valois, 50km north of Paris and Paris itself had cases very soon. End of Feb, 36 of France’s 100 confirmed cases were in Oise. Mulhouse hotspot was stronger in March due to a large evangelical meeting that happened there end of February and Haut-Rhin has still by far the highest number of deaths per capita (over 700 per 1 million).

        • Alon Levy

          In New York, too, the working class is more affected than the middle class; you can see it in the racial statistics, with blacks and Hispanics having around twice the infection rate of whites. (Asians, who in the city proper tend to be poor as well as the Asian middle class suburbanizes, have slightly lower rates than whites, which I ascribe to greater use of masks and earlier social distancing upon reading the news from China.) I believe it comes from two places: first, higher crowding levels among poorer people, and second, essential workers are somewhat more likely to be working-class, e.g. sanitation and retail workers.

          • Onux

            I know the two words are not exactly synonymous, but isn’t it a contradiction to say that “density” doesn’t affect infection rates, but poorer people have higher infection rates because of “crowding”?

          • Alon Levy

            It’s not a contradiction, because density (people per km^2 of land) and crowding (people per m^2 of apartment) aren’t the same.

            That said, New York has very high density but, by non-US standards, not high crowding. Like other rich American cities, its middle class often lives in large apartments with housemates, which is unknown in rich Continental European cities (the Berliners I know who live with housemates are working-class or at the low end of freelancer income), but I don’t quite think this is it either. The US just had a really poor response to the crisis even by European standards, turning New York into a disaster zone.

          • RossB

            Yeah, white collar jobs can often be done at home. By the time some cities closed their bars and restaurants, white collar workers had been sitting at home for days. Not all white collar workers are well paid, but compared to those at a typical restaurant or bar, they are.

            I don’t think you can ignore the initial information about this virus, either. From the get go, it was clear that it hit older people harder. Those that are well off, and older, were therefore way more likely to stay at home, even if it meant using vacation time.

            The only countering factor is the upper and middle class is much more likely to travel by airplane, an obvious vector for the disease.

          • adirondacker12800

            It might be more interesting to consider that poor people are more likely to do essential work. I realize that baristas in fancy coffee shops are invisible to upper middle class symbol manipulating twats but poor people have work that involves interacting frequently with strangers they tend to call “customers”. Lots of them for “tips” that you are aware of. That if they are left in cash they themselves have to handle. And some of them work places that might be described as “labs” where they are working right now because what they do is essential not something they do when they aren’t pounding a keyboard having incestuous email exchanges with colleagues.
            Lower middle class people called flight attendants do a lot more flying than the general population, how are they doing? The head of the union appears on the talking heads show quite frequently these days and she seems more focused on that all of her members are out of work because the TSA and airports are reporting a 96% drop in passenger numbers.

            I know someone who is blue collar middle class union member. A cop for the Port Authority of New York and New Jersey Police.


            He says he saw someone get off a plane wearing “exam” gloves. I don’t have any, don’t have any travel plans other than perhaps going to a supermarket with 20 aisles sometime next week and I have clean drinking water and soap so I wouldn’t do that but if you want to wear exam gloves until you get home, okay. She stuck a fingertip between her teeth, in Newark Airport, pulled them off and threw them away. … okay. There is a technique to taking them off, more than one actually and none of them involve teeth.. But she though she was doing something good and he has a bachelors in something amorphous that was described in the catalog as Criminal Justice.

          • Onux

            Fair enough on density vs crowding.

            How can you say that the US has had a response that is poor by European standards? Per the current US per capita fatality rate is 57/million. In Europe:
            Spain 344/M
            Italy 312/M
            Belgium 260/M
            France 202/M
            Netherlands 147/M
            UK 132/M
            Switzerland 116/M
            Sweden 86/M

            Cases per capita are higher in all of these countries too, except for the UK and Netherlands, and the % of confirmed cases that result in death is higher for all of them as well. At this point, the US is even reporting more tests per capita than Italy and Switzerland.

            Other countries in Europe are doing better, particularly Germany, but if you add all of Western Europe (Portugal to Sweden, Ireland to Austria) and compare those 406M people to the 331M in the US then rates per 1M are:
            Infection 1867 Eur / 1519 US
            Death 168 Eur / 56 US
            Tests 10236 Eur / 7670 US

            In Europe 9% of reported cases are ending in death, in the US 3.7%.

            The US response is poor by East Asian standards, but Western Europe is poor by US standards.

          • michaelrjames

            In a country the size of the US those comparisons aren’t very meaningful.

            On Friday, New York State reported 92,384 cases and 5,663 deaths in New York City
            UK: 73,758 cases, 8,958 deaths
            If you normalise on population (v. approx.) the NYC to UK, then NYC = 677,482 cases, 41,528 deaths.

            This may not be such an outrageous comparison because both are at similar stage, and UK is predicted to be worst in Europe.

          • Matthew Hutton

            Trump and the Republicans are I’m sure trying to suppress the deaths.

          • RossB

            In some ways, it is easier to think of each individual U. S. state as being, well, a state (it’s own individual country). That is because the federal U. S. response to this pandemic was basically nonexistent. It has been up to each individual state to respond. The same could be said for the European Union. There are larger cultural differences and more independence within European countries, but just like the U. S., a lot of people move between the countries quite easily (and not just via the airports).

            Thus it is probably more meaningful to look at the individual states and compare them to European countries. In Europe, both Andorra and San Marino got hit with the virus, and they are towards the top in terms of deaths per million. But the numbers are so small (double digits) that it is possible the same thing happened in some U. S. territories (e. g. Guam). If you throw out those areas, these are the states and countries with deaths over 100 per million:

            New York State: 440
            Spain: 352
            Italy: 322
            Belgium: 289
            New Jersey: 246
            France: 212
            Connecticut: 175
            Louisiana: 173
            Netherlands: 154
            UK: 145
            Michigan: 140
            Switzerland: 120

            It is worth noting that other than the small islands, all of the worst hit areas in terms of deaths per million residents are listed above. This mix of European countries and the U. S. continues as you go down the list, until you get to Iran, which has 52 deaths per million. It is also worth noting that Canada has 17 deaths per million (bad compared to the world, but not as bad as most states, and a lot of Europe).

            So, obviously, both the U. S. and Europe have responded very poorly to this crisis. I’m surprised that the European response was so poor, although there are plenty of countries that have responded well. I’m not sure why that would be. In contrast, I can understand why various states struggled. U. S. states are not used to solving this problem, any more than they would be expected to create their own army. That responsibility falls to the federal government, and has for a very long time. Or at least, it was expected to.

            It is worth noting that all data at this point needs to be taken with a grain of salt. The most reliable number is deaths, and yet it is clear that many deaths were attributed to other causes. We probably won’t have a good idea of how this spread, or good estimates for the total number killed by it for quite some time. Thus it is possible that other countries are performing as poorly as the U. S. and Europe.

          • Alon Levy

            The problem with using deaths is that they’re a lagging indicator, and the US got hit later than Europe.

            And the problem with a state-by-state analysis is that federalism is a policy choice. France, Italy, Spain etc. decided against federalism at all. Germany decided to be federal, an this delayed the initial response, but after a week of foot-dragging an emergency plan was enacted to coordinate the response and prevent states from sniping at each other. The US decided to be federal, make states compete with one another, steal PPE shipments, and complain about opposition-party governors. As I said on Twitter, there are no red states and blue states, only failed states.

          • adirondacker12800

            Donald Trump is not the U.S. He decided that it was too much bother to do his job.

          • Onux

            MichaelRJames, it is a very meaningful comparison when people make statements like “The US had a really poor response” and lump the entire country together; hence the comparison to an equally large grouping of people in form of Western Europe.

            RossB, I am ok with your approach of comparing US states to other countries, but only looking at the top of the list is misleading. 40 of 50 US states are below the death rate of Iran, but only a handful of European nations. Setting aside Wyoming with no reported deaths, there is even a US state with a death rate per capita lower than South Korea. You mentioned Canada as being better than most US states, but in fact 26 have a rate equal or higher and 24 lower, so Canada is right at the median for US state performance. States doing better than Canada include very populous ones like California and Texas.

            Alon, almost the exact same facts can be used to challenge your statement. Do the failed states stealing each other’s medical supplies include Utah, Minnesota or Florida? Because they, among many others, have rates of infection and death lower than Germany with its delayed centralized response, and far lower than countries whose response was centralized from the start such as France, Italy, etc. Just as you cannot say density is an enabling factor when some of the densist cities on earth have a low infection rate (c.f. Tokyo, Seoul, Taipai) you cannot say that federalism is a problem when polities in a federal system are equaling the performance of some of the very best in a more centralized system (N. Carolina and Finland have an equal death rate) but the federal system as a whole (the US) is seeing far fewer deaths than the collection of centralized entities (the EU).

            Deaths are a lagging indicator, but they are also the most reliable since you cannot have an asymptomatic death that goes unnoticed, and also the most meaningful since cases that recover are just someone being sick versus a life permanently lost, and since death rate presumably reflects both infection rate and quality of care.

          • RossB

            @Alon — Deaths are a lagging indicator, but they are the best we’ve got, by a big margin. Testing numbers are less reliable, as they rely on people actually testing. That has been so poor, and so inconsistent in the United States as to be meaningless.

          • RossB

            >> And the problem with a state-by-state analysis is that federalism is a policy choice.

            Everything is a policy choice. My point is that no one is looking at European Union virus numbers, so why should we look at total U. S. numbers? There is no reason why the E. U. couldn’t have decided to handle this crisis the same way they handle fiscal or security matters — with an organization that unites the countries. It is quite likely it would have been better (it couldn’t have been much worse). But they didn’t, just like the United States.

            Likewise, the states could have allowed counties to do their own thing. But they didn’t. It was states in the U. S., and countries in Europe.

            As it turns out, the E. U. and the U. S. are remarkably similar in size; the biggest difference is that there are twice as many states as E. U. member countries. Germany is twice as big as California. Luxemburg is the size of D. C.

            The only reason to compare a country like Germany to the entire United States would be if the entire U. S. had a single policy and Germany had a single policy. That clearly wasn’t the case with this crisis.

          • fjod

            RossB – I think this misunderstands the structure of the EU, which is a confederation with quite strictly limited powers over what it can legislate on, and even stricter limitation of powers on how that legislation can be implemented. The EU doesn’t have the power to, for example, enforce lockdowns, shut schools, close borders, test residents (though it can buy tests and has done so), build or expand healthcare capacity, or most other things that could form a response to the coronavirus crisis. So yes, there is a very good reason why the EU couldn’t handle coronavirus like it handles some security matters (not fiscal policy, that’s the Eurozone) – this would have required a new treaty to give the EU that power, and possibly multiple national referendums on this treaty. That was obviously not possible on this timescale.

          • michaelrjames

            I want to second what fjod wrote.
            Especially in healthcare the EU has almost no involvement and thus very little power in that domain. (It probably doesn’t want any because it is the biggest single budget item for nations, and one of the most politically charged.)
            The perception, by many Europeans and non-Europeans, that the EU is some gigantic all-powerful and all-controlling monster bureaucracy is simply untrue. It’s budget is approx. one percent of GDP, and its “giant bureaucracy” is a fraction of the size of, say, London or Paris local government.
            The EU is a lightning rod for all kinds of criticism, often by national politicians who know it is quite untrue.

          • michaelrjames

            Alon: “Asians, who in the city proper tend to be poor as well as the Asian middle class suburbanizes, have slightly lower rates than whites, which I ascribe to greater use of masks and earlier social distancing upon reading the news from China.”

            Fairly dramatic confirmation:

            A CHINESE BRIGHT SPOT, IN ITALY — The medieval city of Prato, near Florence, has the largest concentration of Chinese residents in Europe. Defying the fears of many, the city’s Chinese community has been credited with helping to protect Prato from the outbreak. The city has just 479 cases of the coronavirus — fewer than anywhere else in Tuscany. Not a single member of the Chinese community in Prato has tested positive for the virus, according to officials, because of a scrupulous adherence to the rules and a fastidious attention to hygiene. “The Chinese community set a good example, creating a virtuous circle,” said Matteo Biffoni, Prato’s mayor. “The seriousness of the virus was understood in Prato, perhaps before anywhere else in Italy, and thanks to their example, we all got into line.”

      • Marco Chitti

        The italian case explains pretty well why establishing a correlation between density and covid19 spread is tricky at least.
        Initial cases of community spred were recorded in Codogno and Vo’ Euganeo in mid-February, Those are small villages of few thousends surrounded by fields. Codogno is in Lombardy, but in the less densely populated province of Lodi, that like Cremona and Crema had large number of infected/population, even if they are far less dense of the less touched Varese province.

        The worst outbreak happended in Bergamo province, but in particular in Seriana valley, a strech of low density, rich industrial towns and villages just north-east of Bergamo itself. It is now believed that the reason of such a strong outbreak happening there was the failure of local authorities to lockdown the area early as they did in Codogno and Vo’, because they feared the economic consequences of closing down factories (this was 3 weeks before Italy general lockdown, when nobody in the west was considering coronavirus other than a “Chinese/Asia thing”). Another factor that draw up the contagion numbers was the fact of not separating properly suspected covid19 cases from other people in hospitals, and sending the elderly showing mild symptoms back to isolation in retirees homes. Those places has become a major source of spreading among the most vulnerable population, killing thousend. No one of this has anything to do with density itself, or even crowding in public places, but more with the existence of specific “hot spots” for contagion (hospitals, retirement homes, etc.) and/or the work of so-called superspreaders (like family doctors, for example).

        Early spread pattern seems much more related to chance (someone got infected abroad and visited many crowded places/people when back) and/or the response from authorities. In Lombardy we now know that it was pretty bad, compared to Veneto and the epidemic evoluted differently there, even if both regions are pretty dense in terms of popultaion/km2.

        So, I can understand that we use “density” as a proxy of “higher chances people interact with eachother so closely that they are more exposed to contagion”. But nobody lives in crowded tenements or mietskasernen like a century ago, Household size is even lower in dense cities than in suburbs (people tends to live alone/couple and move out when they get kids also in Italy), so transmission within family, that is a big factor of spread, can be worse in suburbs than in central cities, for example.

        • RossB

          The study is interesting, especially the part about total deaths not being tied to density, but exposure is. That is scary. It suggests that this will spread through the countryside unless various areas react better than New York City. That is a very low bar to clear, but I fear many will not clear it.

          But it is easy to poke holes in the study. For example they have yet to look at how each city has responded to the crisis. I can imagine that in other countries, it didn’t matter that much. In the U. S., it was everything. Essentially, the President said “every man for himself”. Some responded reasonably well (e. g. folks on the West Coast) while some screwed the pooch (New York, New York).

          It is especially unfortunate that they limited their study to the U. S., given the wide range of response to the virus. This greatly limits the value of the data, especially since New York City responded so poorly, and has a very large amount of the density and virus cases (as well as deaths) in the United States. Since the study seems to ignore governmental response, you might as well gather from a larger set of cases. Italy and Spain, for example, would definitely fall into the same range of responses as the U. S., and you would have more meaningful data.

          To be clear, quantifying the response will be difficult. Looking at “stay at home” orders is bound to result in misinformation. By the time Washington State issued a stay at home order, most people were staying at home. Schools, restaurants, bars, concerts and plays were closed. Even hair salons were closed (I’m letting my freak flag fly). The stay at home order was basically a formality. Given the exponential nature of an epidemic, this makes a huge difference. Acting early, but not thoroughly is much better than acting a few days later, but closing … I don’t know … a few more construction projects? I honestly think that the lives of very few people in Washington State changed once the issue was ordered, yet it is used as important metric, thus showing the difficulties of measuring the response.

        • michaelrjames

          Oops, I posted this on the wrong article:

          Sweden is interesting in that it represents a genuine experiment. The thing is that even if a clown mentions it, the herd immunity approach is not without foundation. Apparently Sweden’s was (is?) deliberate policy, however what I cannot tell is whether they have enacted any protection strategy for the most susceptible (elderly, esp. nursing homes). Looking at their rising toll, I presume not, which indeed makes it totally irresponsible. Rare to have old people in Sweden living with their younger offspring, which appears one of the big problems in italy (and Spain?). Yesterday I read the first report with some (incomplete) data on nursing home deaths and, as long suspected, it is shocking in many places. Some homes have lost one third to half their patients, and sometimes workers have spread it around as they work in multiple homes. Officialdom seems to have done nothing much, but I don’t know how anyone could be unaware of the massive risk. How has Germany avoided this (or has it?). Or Australia for that matter. Such places are the equivalent of those cruise liners, only worse, and one third of Australia’s deaths come from just one ship, Ruby Princess.

          In Spain, the army has reported finding dead and abandoned people in their beds after it was drafted in to help disinfect care centres. Care homes in the Madrid region alone have reported the deaths of 4,260 residents who were diagnosed with coronavirus or had associated symptoms since 8 March, the regional government said on Wednesday.
          In France almost a third of all coronavirus deaths have been of residents in care homes. According to the latest figures released on Tuesday a total of 3,237 people have died in care homes. In Paris alone there were 172 deaths and over 2,300 homes have reported at least one case of Covid-19.
          In Italy …. 3,859 people have died in care homes across the country operated by the RSA organisation since 1 February …
          In Germany there have been reports of deaths in homes totalling hundreds across the country. In the worst case so far, 29 out of 160 residents at a care home in the northern city of Wolfsburg died after 74 residents became infected. Prosecutors are now investigating the home on charges of death through negligence.
          In Canada, health authorities have been grappling with coronavirus cases in long-term care homes across the country. At one retirement home in Bobcaygeon, Ontario, 29 of its 65 residents have died after contracting the virus.

          Also, re the argument about density versus crowding, both cruise ships and nursing homes demonstrate this. They are crowded, ie. the overwhelming factor is proximity, even if some of it is via the air system, same thing really just a particular mechanism.
          Likewise with Italian villages, one shouldn’t consider them low density. They’re not twee sprawled housing estates! More important is the behaviour which produces crowding and proximity. The mass was continued to be performed all over Italy long after warnings were loud and clear. And those not pious were sitting in crowded bars for half the day!

          • michaelrjames

            I should say that in Australia they do have protocols (mandatory I think) to deal with certain high-risk events. When such an event happens the home sends out a text or phones the regular visitors to advise them that the home is not allowing any for a given time, ie. effectively lockdown. The problem with this virus is that asymptomatic young people (ie. the care workers) can spread it around for a week or longer without anyone being aware. They should have gone into this mode as soon as the problem appeared in China, and maybe here they did and that’s why we hear nothing about it. I don’t know and the media haven’t been reporting about it (but as also noted everywhere, the media is a mere threadbare shell of its former self–if it’s not on social media, with free pics, they’re probably not going to cover it).

  3. adirondacker12800

    dense cities are bad. Most of this is bullshit from politicians, like Andrew Cuomo.
    If you had actually bothered to check the date on it and read it, it doesn’t say that. It says ( the stay at home order went into effect at 8 PM on March 22 and we did not close the playgrounds) that people were not observing it in playgrounds so he was closing them. Or you had nine days to figure out this and have proven you can’t use them responsibly. I’m closing them.
    Nothing about the city as a whole, just that things were too dense in playgrounds. Nine complete days and two fractions on either end of that after the order went into effect. The parks are still open. He didn’t mention anything else being closed either. There may have been a separate announcement about any other but from that announcement alone it was just the playgrounds. A basketball court is a basketball court all across the country and they have the same density when they are being used by the same amount of people. Or a tennis court. Handball courts are very rare outside of New York but they are the same where they do exist. Bocce courts..

    • Benjamin Turon

      I have listen in or watched Gov. Andrew Cuomo’s daily briefings, and while he said density was an issue in containing the pandemic, he praised it overall, and said that density overall was a good thing, including making New York City a great economic machine, and would help the city comeback. He said this the morning right after Mr. Levy attacked him in a blog post for being anti-density. One may disagree with Mr. Cuomo on many things, but I don’t see him as being anti-NYC. Many Upstaters loath him for being pro-NYC (common complaint) and he has been attack for wanting to “seize” ventilators from Upstate, for Downstate hospitals. Surprisingly, a few of my red hat gun-touting, ammo stock-piling friends and relatives, think that the governor is doing overall a good job. As for social distancing, its obviously harder in a city, up here I can get in my car and drive to a bike path, that if I see a dozen people in an hour, is very crowded. In Manhattan if everyone goes to Central Park, it looks like Woodstock.

      • Benjamin Turon

        For some people this seems an issue of “semantics”, of folks who opposed the use of the word “density” in the mandates for “social distancing”, with critics preferring the words “crowded” or “proximity” instead. This is ridiculous, since for most people they mean the same thing, I could see my sister saying: “Manhattan is very dense, crowded place, with everyone in close proximity to everyone else”. So no more calling the crowds at Penn Station or Yankee Stadium a dense! No more dense throngs of humanity! This is a typical fight you see now days, a war of and over words, instead of meaningful debate and action.

        Op-Ed: Dear Gov. Cuomo, The Problem Is Crowding, Not ‘Density’!

        Cuomo: Social density controls may be slowing hospitalization rate, data suggests

        • Benjamin Turon

          On density, in that press briefing a week or so ago, Gov. Cuomo talked about how the density of New York City — he has pointed out numerous times that he is from Queens — made thinks difficult, but also help by giving people many neighbors willing to help out, more social services within easier reach, including healthcare and hospitals. Yes, he has use the word “density” in terms of “social distancing” but I have not heard him say NYC is too dense overall, that their are too many high rise buildings, and we should all live in suburbia. He has not said that. If so, Mr. Levy should provide a quote.

          Now have others said this — perhaps on Fox News or the Reason Foundation — no doubt, but its not news that there are anti-city folks who need no excuse to attack urbanism.

          • adirondacker12800

            Since nobody in the U.S. is doing antibody testing to a great extent who knows? There is this thing called Google. Type the right thing in, it can be a wonder. And very very quaint, the New York Times, is a wealth of information.

          • Benjamin Turon

            Mr Levy, I don’t understand your point. Adirondacker and me are not arguing against you main point — that we shouldn’t abandon urbanism and public transit — but against your opening statement:

            “There’s a meme going around the American discourse saying that the Covid-19 outbreak is proving that dense cities are bad. Most of this is bullshit from politicians, like Andrew Cuomo.”

            First, that is not true about Mr. Cuomo, it is a canard, a mistruth, a falsehood, a lie. I have just in the past few days heard him say the exact opposite of what you clam he is saying. I have heard him praise the density of New York City. Just because you hate Mr. Cuomo doesn’t mean you can make him your bogeyman for “the war against density”. Criticize him all you want for what he does and says — for his airport trains that go the wrong way — but you can cut words from whole-cloth and place them in his mouth. That is just bullshit, and expected better from you, since you seem to want a fact based discussion.

            Look, if you think a high density of children on a playground or joggers in the street is a great idea during a pandemic, then just say so. If you think shutting down and a economy and mandating social distancing is wrong, the just say so. If you think all the Broadway theaters should be open and packed then just say so. If you think everyone should go to Sunday mass, then just say so. If you think subway cars full of infected and sick people will slow the spread, then just say so.

            Second, your blog is just about the first I have heard that there is a “American discourse” on cities and public transit needing to be abandon to combat pandemics. Mostly I have read and heard that we should have a better public health service, like South Korea, Taiwan, Hong Kong, Singapore, and Germany. Yes, while NYC has gotten a lot of focus, there also has been a few stories on rural places like Albany, Georgia have been hit hard. There also has been a bit of panic by Upstate NY politicians and the media that if rural areas have a big outbreak, it will be really bad because of fewer resources. But then again, admittedly I’m not watching Fox News.

            What I have heard about changing the way we live after this pandemic has been more nuance and informed. A better public health system and universal healthcare have been on the top of the list. Reducing air pollution as been put forth. Using new antibacterial and/or easy to clean materials in building and vehicle designs. Having a larger cleaning staff. Reducing office density back to were it was a decade ago, before corporations started piling people into rows of bench desks and hot desking. Building more parks and a better pedestrian environment. There has been some worry about more people working from home hurting commercial office space, but that hurts broadly both downtown and suburbia, and suburban office parks could be hit worse because they are on less valuable land or be converted to other uses, like residential.

            Overall I agree that “Cities should keep investing in all four in order to reduce the costs of transportation and environmental damage” but to state that Gov. Cuomo is a “failed politician” blaming “the virus on density to deflect attention from his own incompetence” is a lie, just like how President Trump makes up lies about his opponents. Mistruthing undermines a otherwise well built argument. I thought our side aimed to be better than that — well I guess human nature is human nature — you should be ashamed of yourself.

          • Benjamin Turon

            Governor Andrew Cuomo in his daily press briefing this Easter Sunday mention the density of New York City again, in answering a question on the preparedness of rural Upstate counties, specifically Genesee County. The governor stated that New York City had been it by the first wave because of international travel, while being the most densely populated city (he statistically compared NYC to LA) in the country made social distancing harder, when people have to ride the subway.

            However, he continued that the pandemic was spreading and increasing amounts of cases would pop up in Upstate NY, triggering the state to send in resources to contain these outbreaks. He pointed out the biggest cluster of COVID-19 cases to date was not in New York City, but suburban New Rochelle. He also pointed out that the virus until recently had been quickly spreading in the two Long Island counties. He said that his mandating of social distancing, closing of schools and non-essential businesses was made for the WHOLE state, because every part of the state was at risk. He said he was glad they canceled St. Patrick Day parades, because those mass social events going forward could have made things worse.

            He did NOT state that population density (or public transit and airline travel) was a problem long-term. You can criticize Mr. Cuomo all you want for the specifics of his plans and management of megaprojects and transit agencies, but if he really hates public transit and population density, why has the 2nd Ave Subway, Penn Station, LaGuardia Air Train, MTA Reform, Upstate Rail Service feature so prominently in his agenda? I have heard him many times in in-depth interviews by WMAC Northeast Public Radio out of Albany NY, and he has often and in detailed talk about (even when not asked) about the MTA, Penn Station, and the Gateway Program. I live Upstate, and many believe — even in the opinion press — that Mr. Cuomo loves New York City and hates Upstate.

            Governor Andrew M. Cuomo: Coronavirus

            I thought this recent CNN article was pretty good, pointing out density as an issue for the spread and wide spread testing* for the virus giving statistically Metro NYC a higher number of confirmed cases; yet the article continues by pointing out that New York City was closer to “maturity in the outbreak compare to other places in the USA when widespread social distancing was enacted. This factor, combined with a poorer population with widespread pre-existing conditions and lack of access to good healthcare, are likely primary factors as what was according to their figures at CNN, a higher death rate than other places.

            CNN — Why New York has been hit so hard by coronavirus

            “How about population density? Covid-19 spreads most easily when people are packed together — in churches or cruise ships, in outdoor events like concerts or Mardi Gras, or perhaps in small apartments with multiple roommates or large families. New York’s population density, about 27,000 people per square mile, is easily the highest in the country, though it’s nowhere near the top for cities across the globe. Many cities — in Asia, for example — have a density of nearly 40,000 people per square mile… The density explanation may explain some of the difference, but not all of it. New York’s densest borough is Manhattan, while relatively sprawling Queens sits at fourth out of five. But Queens has more than twice the cases and twice the rate of cases as Manhattan.” — From CNN Article

          • michaelrjames

            Sorry, but Cuomo won’t be running for president, at least in 2020.
            Unless there is the fantasy scenario where the US totally loses control over this thing and Joe contracts covid-19, and some of the other gerontocracy who run the US … then it could indeed become a true turning point. But then why not turn to another New Yorker like AOC? If we’re into fantasies …

          • adirondacker12800

            No he won’t and answered his brother’s questions about that with “No” recently. That single word, not more elaborately. On Chris Cuomo’s regular hour long program on CNN. While Chris Cuomo was still recovering from his apparent infection with Covid-19, isolated in his basement. I think it is unlikely ever because Andrew Cuomo is 62.
            While the myth is that anyone can grow up to President of the United States there are some eligibility requirements. AOC doesn’t have one of them. Yet.

          • Herbert

            There’ll be tremendous pressure on AOC to run in 2024. And the “establishment” will try to stop her…

          • michaelrjames

            It took adirondacker to take literally a joke based on a fantasy.
            But I think someone should start a new birther movement. AOC has faked her birth certificate (probably to win that scholarship to BU) and she is in fact eligible!

          • adirondacker12800

            She was born in the Bronx. She has spent more than 14 years in the U.S. but she’s not 35. I don’t think there is any argument that she will actually be 35 on inauguration day 2021. She is too young. I’m almost sure everybody agrees she is too young.

          • michaelrjames

            I’m almost sure everybody agrees she is too young.

            That is ambiguous.
            If you meant that everyone knows she is too young under the current rules of eligibility, then yes.
            But whether she is too young to be in such a role, that is at least debatable. Mostly I am not a fan of the cult of youth etc (and as a boomer have a certain … reticence … about Millennials-genZ) and generally would like the politicians who rule over us to have well-rounded world experience etc. The trouble today, especially–dare I say–in the Anglosphere, politicians are increasingly machine-men (and women), ie. they often have done nothing else their entire lives except within the party-political embrace. They begin while undergrads as political gofers, or do an internship with a politician, progress to junior political staffers and progress up the ladder. Further, in the American system, the longer they are in this system the more likely they have become beholden to it and all the usual vested interests. The “wisdom” is of incrementalism and the art of the doable which is almost entirely self-serving nonsense. Thus, youth becomes almost the only protection (hah, like against sars-co-v2).

            If you get Biden, save a miracle of happenstance (which feasibly could come from this virus but probably not), then who doubts that it will be BAU? Indeed Biden is selling himself as the BAU-candidate! Funny enough there is more likelihood of meaningful change under a second-term Trump but of course that will be because he has destroyed the kingdom and brought chaos and made revolution more likely. And if you get Biden then the Veep will probably be Julia Louis-Dreyfus, err.. Kamal Harris or similar, and in turn that person will most likely be the next prez. Thus, bleak logic suggests that talk of waiting patiently for 2024 by the likes of the younger hopefuls (several in this years contenders all of whom were dismissed early) is futile. In 4 to 8+ years, will AOC still be so eager for change or will she have been beaten into conformity? That eligibility rule on age is there for these reasons, to select not wisdom but conformity.

            I was being facetious when I recommended AOC, but also wistful. Your country needs serious change of direction.

          • adirondacker12800

            35 is not ambiguous, especially these days with accurate calendars and clocks. That isn’t going to change without a constituitional amendment and good luck with that.

  4. Nicolas Centa

    The recommendations to practice social distancing, including working remote, taking public transportation as little as possible, and avoid crowded places, do not seem to favor density though.

    It is clear here in Ibaraki that it is far easier to get food and toilet paper without meeting anyone than it is in Tokyo. And it has been the case for any catastrophe before COVID-19 as well.

    If people notice they can work remote and get entertainment online while still being so afraid of viruses as to practice social distancing, it may become worrisome for density proponents.

    • Matthew Hutton

      Partial remote working also opens up the possibility of more housing further out of big cities being viable for jobs. I still can’t see people wanting to drive into big cities due to congestion.

      • michaelrjames

        I’d predict the opposite. Sure it shows some work can be done at home. But most people will come to appreciate more than they would have thought, having a workplace to go to.

          • SB

            Yes, there is decrease in productivity but there is trade off between productivity and costs.
            Company don’t have to rent/own property and paid workers less if employees WFH in areas with lower cost of living.
            For tech trade-off from decreased productivity is not worth it.
            Industries that doesn’t necessarily need high productivity, WFH might make business sense.
            Not sure for academics. Distance learning was already popular and it may popular after the pandemic.
            Obviously primary and secondary education shouldn’t be distance learning.

          • Alon Levy

            Distance learning was always unpopular among students and professors. University administrators like it and at one American university responded to faculty complaints about increased workload during the virus with “we can hire you some consultants so that you can improve your time management skills.” But also central admin at universities is generally loathed by both the faculty and the students, and at Columbia even by department-level administrators.

          • Onux

            Is their productivity taking a hit because of working at home per se, or because of other factors inherent in the current situation, i.e.trying to work while also taking care of children not in school, or the 95% of office work suffering because they can’t get 5% needed time in a lab?

          • Nicolas Centa

            Now you really start sounding conservative about remote work.

            People don’t like change, especially when it is happening fast and being imposed on them, so of course now they don’t think remote work is great.

            As far as tech work is concerned, I am also a hobbyist free software developer, and I think that none of the Linux, Python or any of the major projects that dominate the server and development landscape are being primarily developed in offices and coordinated in meeting rooms.

            I think this pandemic is scary because if it is too long it has the potential to really reorganize a large part of the economy to function during social distancing, and because of the importance of remote working, the fact nobody can go on vacations, and reduced non-digital non-essential consumption overall, CO2 emissions will drop as well.

            Public transportation has no place in this because it has no potential to solve the infection issue (just cleaning and having larger trains is not enough!), whereas the Internet, remote working and having your groceries delivered has.

          • Eric

            I’m a tech worker, and less productive because I have a young child whose child care facility is closed. But without a child, I think I’d be equally productive. One of my close friends is a tech worker who has worked 100% remotely for half a year now (unrelated to corona). Another of my close friends is a professor who for years has only gone into campus once a week. In general I think a large majority of tech workers would be equally productive working from home. In the long term there would be a hit to productivity due to not actually knowing and forming human connections with coworkers, but this could be remedied with 1 or 2 days a week of work on-site, 5 days would not be necessary.

    • adirondacker12800

      If you are working from home being able to pop out the door and take a short walk or even a short drive to many places to… get lunch?… has a premium. If it’s a 15 minute drive, you can’t.

      • SB

        You don’t have to leave home to get lunch?
        Even if one is terrible cook, one can just make sandwiches, eat prepared meals or order delivery.

        • adirondacker12800

          You don’t have to go out for lunch if you work in a glittering enormous office building either but people do. You can bring it, go to the cafeteria if there is one and have a better chance of having delivery available than if you are out in cul-de-sac exurbia. But they still go out.

  5. Eric

    It seems the main correlation here is that Bavaria has more cases, presumably because it has more travel to northern Italy.

    The other thing that needs to be looking is the exact timeline of social distancing or lockdown. In the US, California had roughly a week’s headstart in this compared to New York, and predictably a couple weeks later the situation had gone from California having a few more detected cases to New York having ten times as many cases.

    • Alon Levy

      In Germany the timeline of social distancing is more compressed – there was variation in state-level responses, but it wasn’t as big as in the US, e.g. Berlin canceled events maybe a week behind when it would have if Müller were better at his job.

      I don’t think it is really about the Italian connection? Bavaria has the highest infection rate in Germany right now, but a month ago this was not the case, but rather the epicenter of German infections was western NRW.

      I guess it could be about the ski resorts – I’m told Southern Germans go to these more often than Northern Germans?

      • Herbert

        Most Bavarian school kids go on at least one school organized skiing trip during their school life…

        • michaelrjames

          Which may be (looks like) it could be irrelevant. ie. the mysterious absence of children as vectors, which is another big difference to seasonal influenza.

          • RossB

            >> the mysterious absence of children as vectors

            Say what? I’ve seen no evidence to suggest that children aren’t vectors.

          • michaelrjames

            That’s because you’re RossB and not RossA (sorry, the virus made me do it, sir; it’s cabin-fever humour!)

            It is true that absence of evidence is not evidence of absence …. but… in all other influenzas and colds, children are major vectors. That’s why there has been this knee-jerk reaction of school closure almost everywhere (except Singapore & SK, except recently) but if children were (major) vectors then the effect would have been super-obvious like John Snow’s Soho drinking well. It is one of the abiding mysteries of this virus. The experts don’t know if children mostly avoid infection (probably not), or fight off the infection at a very early stage so they never produce enough free virus to be vectors.

            Professor Peter Collignon is an infectious diseases physician and microbiologist at Canberra Hospital and professor of medical school at Australian National University:
            There is a lot we still don’t know about Covid-19. What we do know is that children appear to very rarely have serious disease and complications, compared to those in the older age groups – especially their grandparents.
            The data from a range of countries shows that children rarely, and in many countries never, have died from this infection. Children appear to get infected at a much lower rate than those who are older, although confirmation of that will need to await the rollout of large-scale antibody testing, which tells you that you have had an infection in the past.
            Although there is still some uncertainty of infection rates in children, there is no evidence that children are important in transmitting the disease. Children can get infected and children with infections – usually acquired from older family members – have been at school, but there have been no documented outbreaks in those schools.
            Marylouise McLaws, epidemiologist and member of the World Health Organisation (WHO) Health Emergencies Program Experts Advisory Panel for Infection Prevention and Control Preparedness, Readiness and Response to Covid-19:
            Kids are unusual in that they represent across the world, less than 2% of all cases. Scientists are starting to look at whether or not it’s due to a receptor site that the virus likes and that perhaps the children aren’t expressing it. Or that their immune system somehow protects their lungs from becoming overly inflamed.

            When Collignon says “there have been no documented outbreaks in those schools” the obvious problem is that most children, if infected, are asymptomatic so you’d only know this if whole schools had been systematically screened and I am not aware of any such studies. Studies on small numbers of children in hospitals etc. are revealing other things but are not very informative re epidemiolgical questions.
            Incidentally the antibody studies (which reveal prior infection) are not so simple because the current serology tests apparently are quite insensitive, giving unacceptably high false negatives, and of course if the hypothesis is that children have a mild reaction one might have to search for lower levels. Korea has reported that in one study they found 40 people who had recovered from infection have later shown virus again–at this stage they don’t know if it is true re-infection, or latent virus in their own body resurfacing. The point is that no one knows how persistent the antibody response is, and immunity to earlier coronaviruses has been short-term (lost after a year). It is also not clear if these people are infectious which would be particularly bad news, eg. in relaxing lockdowns. FWIW, I think they won’t be infectious but will be like children (because if not, we’d have seen this effect more overtly everywhere).

            School closures have little impact on spread of coronavirus – study
            UCL say small benefits should be weighed against profound economic and social costs
            Sally Weale, 7 Apr 2020
            The UCL-led study concludes that the evidence to support the closure of schools to combat Covid-19 is “very weak”, and statistics from influenza outbreaks suggest school closures “could have relatively small effects on a virus with Covid-19’s high transmissibility and apparent low clinical effect on schoolchildren”.
            The research team reviewed 16 studies of recent outbreaks of other coronaviruses, including the 2003 Sars epidemic in mainland China, Hong Kong and Singapore, and found that school closures did not help control the epidemic.
            “We know from previous studies that school closures are likely to have the greatest effect if the virus has low transmissibility and attack rates are higher in children. This is the opposite of Covid-19,” said the study’s lead author, Prof Russell Viner, of UCL Great Ormond Street Institute of Child Health.

          • RossB

            They have largely traced the disease by symptoms, and as it turns out, children rarely have symptoms. Yet it is clear that children can get the disease. Doesn’t it stand to reason that not only are children vectors, but they are the worst kind of vectors (asymptomatic ones). As one scientist put it — it is hard to think of a better place to spread the disease than a school. (I can think of one, Mardi Gras. Not many people were exhibiting symptoms there, either. There was not a single case reported until well after the festivities).

            The decision to close schools early in Washington State may have saved thousands of lives, while the decision to keep them open for a few more day likely resulted in the opposite. Oh well, I’m sure Timmy feels so much better having spent a couple extra days in class while half the city is shut down and everyone is crying about the dead relatives. Keep in mind, when Washington State shut down the schools, it was only supposed to be for a couple weeks, and it seemed a bit extreme at the time. After those two weeks, it was obvious that starting them up would be very stupid, and much of the state was shut down.

            I guess we will all know once we have widespread antibody testing. But don’t be surprised if it turns out that there are huge numbers of children who had it, and obviously spread it in New York.

          • michaelrjames

            They have largely traced the disease by symptoms

            Not in Singapore or S. Korea.

            Doesn’t it stand to reason that not only are children vectors, but they are the worst kind of vectors

            Evidence, mamm, nuttin’ but the evidence.
            And so far, the evidence doesn’t support it. Your statement is simply assuming this virus follows the rules of influenza and others. But it doesn’t and is peculiar (or different) in many ways. The most important information to obtain is the antibody tests but they haven’t been done yet, partly because the tests are lousy. Of course if most children are seronegative that doesn’t answer the question of whether children are infected at the same rate as adults, though it will suggest the virus doesn’t take hold in children.
            OTOH if lots of children are seropositive it won’t prove what you assume, that children have spread it. I believe the working hypothesis of the experts is that children are infected but for whatever reasons (poor replication of the virus, quick or natural immune suppression) they don’t produce enough virus for either symptoms or transmission. (They may not produce enough virus for an antibody response if there are other reasons for their resistance.)
            Too many open questions but the pattern does not conform to the influenza pattern of schools as hotspots of infection and transmission. Your own examples, Mardi Gras and related (church meetings, sports events), shows how such hotspots can be easily identified retroactively (contact tracing). Nothing has traced back to a school. It’s almost the first thing epidemiologists look for.

            The decision to close schools early in Washington State may have saved thousands of lives

            Seriously, stop it! Speculating way beyond any data or evidence to support such hysteria. Closing of schools may be justified on the basis of precautionary action but there is zero evidence for it. The reason it is important not to come to the wrong conclusion is the role schools will play in getting society back to some level of normal functioning.
            Far more important than closing schools is lockdown of old-age nursing homes, and failure to do that really has cost thousands of lives.

          • adirondacker12800

            Evidence, mamm, nuttin’ but the evidence.
            It would be nice to see some that hints at having everyone wear a rag over their face helps. I’m sure you have been digging deep for that.

          • michaelrjames

            It would be nice to see some that hints at having everyone wear a rag over their face helps.

            Below is from one of the top science publishers (“Science”). Admittedly, some of it is inferential rather than pure data-based, but such a slow-building consensus by scientists has prima facie validity. I don’t see how the west can hope to get back to normality without masks in public, especially where proximity is unavoidable (and really this means anywhere in the public domain). Estimates are that some 80% of those infected are asymptomatic and thus not easily detectable (and most in the west, esp. USA, currently are incapable of contact tracing).

            Would everyone wearing face masks help us slow the pandemic?
            By Kelly Servick, 28 Mar 2020

            But the greatest benefit of masking the masses, Cowling and others argue, likely comes not from shielding the mouths of the healthy but from covering the mouths of people already infected. People who feel ill aren’t supposed to go out at all, but initial evidence suggests people without symptoms may also transmit the coronavirus without knowing they’re infected. Data from contact-tracing efforts—in which researchers monitor the health of people who recently interacted with someone confirmed to have an infection—suggest nearly half of SARS-CoV-2 transmissions occur before the infected person shows symptoms. And some seem to contract and clear the virus without ever feeling sick. “If I knew who was asymptomatic and presymptomatic [for COVID-19], I’d … triage the face masks to those individuals,” Monto says. Unfortunately, he adds, “We don’t know who these are.”

            Masks can protect the wearers but only high-quality masks (N95 etc) and only if fitted and worn correctly. Hence the panic over PPE shortage inside hospitals or amongst paramedics in those ambulances etc. But out in the more public world, the more important point concerns the asymptomatic or pre-symptomatic spreaders:

            Below is a summary of papers in the very top science publisher (Nature):
            (The author of this piece is not a Guardian journo but “a distinguished research scientist at the University of San Francisco and the co-founder of Masks 4 All”. Yes, having worked at UCSF myself, and as it happens on the same floor as the Bishop & Varmus lab, I do have a certain bias towards these conclusions.)

            To help stop coronavirus, everyone should be wearing face masks. The science is clear Even people without symptoms can infect other just by speaking but a simple cloth covering can stop us spreading harmful droplets
            Jeremy Howard, 4 Apr 2020
            You might walk into stores over the next few days and sicken dozens without knowing it. Some might die. Others will think they are dying before they recover. That’s the worry I have after reading a paper by Roman Wölfel and colleagues, published this week in Nature. It shows that people are most infectious in the first week after catching Covid-19. During that time they often show no or few symptoms.
            In other words, Covid-19 moves like a silent assassin, with unwitting accomplices. Maybe you’ll be one of them. The best way to ensure that you’re not: wear a mask, and keep your distance from others. Don’t wear an N95 respirator, the type in desperately short supply in hospitals, which is designed to keep doctors safe even when doing potentially dangerous medical procedures. But almost any kind of simple cloth covering over your mouth, such as a home-made mask, or even a bandanna, can stop the assassin in its tracks.
            The Wölfel paper explains we must focus our efforts on stopping the spread of droplets. This is because the virus is primarily transmitted through tiny droplets of saliva ejected when we speak. You can’t see them, but they are there. We also know that these droplets can go significantly further than the 6ft which is widely cited as a safe distance.
            Research supported by Nobel prize-winning virologist Harold Varmus tells us that placing a layer of cloth in front of a person’s face stops 99% of the droplets.
            So, the science is clear. We do not know when we are sick. If we are sick, then when we speak we are projecting virus-laden droplets into the air. Wearing a simple cloth mask stops those droplets in their tracks. “I’m not going to wear a surgical mask, because clinicians need those,” said Dr Harvey Fineberg, chair of the National Academy of Sciences’ standing committee on emerging infectious diseases and 21st century health threats. “But I have a nice western-style bandanna I might wear. Or I have a balaclava. I have some pretty nice options.” Fineberg led a committee of experts that has just released an expert consultation explaining that the virus can spread through talking, or even breathing.
            Prof David Heymann CBE, a World Health Organization (WHO) adviser, said, “I think that wearing a mask is equally effective or more effective than distancing.”
            In a paper published in Nature on Friday, a five-year study from the University of Hong Kong and the University of Maryland has found that a simple non-fitted mask blocked 100% of coronavirus droplets and aerosol. There’s a vast chasm between what the science is showing and what many countries are doing. Masks may be the most important weapon in our war on the virus. But we’re not even using it.
            ….. One huge challenge is that mask wearing only really pays off when most people do it. A Food and Drug Administration analysis of the flu estimates that if 50% of the population uses a mask, virus transmission is reduced by half. If 80% of the population uses a mask, the virus is “essentially eliminated”.

            A clutch of Nobel prize winners, and papers across both Science and Nature, really what more do you need, adirondacker?
            Incidentally, Harold Varmus won his Nobel (with Mike Bishop) for discovery of the oncogene (v-sarc) in the retrovirus Rous sarcoma virus and linking to the equivalent genes (proto-oncogenes, in this case c-sarc) in mammalian genomes, opening up the whole field of oncogenetics. He did this at UCSF as post-doc to Mike Bishop, but Varmus was a New Yorker and after his stints as director of the National Cancer Institute (within NIH) and then director of the National Institutes of Health, he returned home to be director of the Sloan Kettering Cancer Centre in NYC. If you want to learn about a genuine American (and New Yorker) giant, read his autobio The Art and Pollitics of Science (2009). Superb. Worth mentioning on this blog because “Harold E. Varmus is an avid bicyclist and an Advisory Committee member of Transportation Alternatives the New York City-based advocacy group for pedestrians and cyclists. He is also a runner, rower, and fisherman.”

          • adirondacker12800

            I’m aware of the recommendations. Which is why I’ve been fooling around with knit fabrics that stretch in all directions, knits, and are something I have. I’m a old fashioned kinda guy, I separate my laundry into different colors and wash the whites in hot water with bleach. I have have a whole laundry basket full of tighy whitey’s and t-shirts to rummage through before I move onto the lights, the reds and the darks. This one has something that will work with t-shirt fabric and doesn’t require any sewing.

            Use white cotton it can be boiled or bleached. Soak it a bit first to loosen any snot it caught and rinse the snot out. You need bleach in the house, if you want to bleach instead of boiling it, after being used. Read the label, you are likely using too much bleach. Ironing it with steam, slowly, would probably be good enough. A lot of you have only heard of bleach and irons. It could probably be washed in a machine but that might not be a good idea depending on the fabric and the machine. The colored stuff could probably be boiled briefly if it’s cotton. The color might change if you use bleach. I lean towards heat instead of bleach, bleach can leave a faint odor.

            People with reputations to protect are coming up with those numbers from someplace. Probably stuff their coequals in Asia pointed out in studies done there. It will remind me to not touch my face, if I decide to wear one, which is something. If I decide to go out, I have that luxury. Something not tangential perhaps maybe pertinent if you read it in warm dry air that smells good. I hear that, with a cup of tea can make it even better. People noticed that other people who only drink tea get sick less often. The probably though it had something to do with the medicinal properties of the herb or perhaps boiling the water drove off a bad smell. And that drinking only beer had a similar effect. Perhaps boiling the grain made some sort of beneficial elixir?

          • michaelrjames

            It’s taken a couple of months but some scientific evidence and analysis about children and this virus.
            Note that this concerns infection rates. Additional factors will be how much infected children transmit the virus to others. I’m betting that will be lower too.

            [Nature email, 23 June 2020]
            Evidence that children are less susceptible to infection
            Children and the elderly are less likely than adults under the age of 65 to show evidence of past SARS-CoV-2 infection, according to a survey of people in Geneva, Switzerland. Only one out of the 123 tested children aged 5–9 came back positive, although 21 of them lived with someone who had COVID-19 antibodies. A separate study in Guangzhou, China found that people under the age of 20 had a 5.2% risk of being infected by a member of their household, compared with a 14.8% risk for people aged 20–59.
            Reference The Lancet paper & Lancet Infectious Diseases paper

      • fischerhias

        For Bavarians it is very common to go skiing for a week or a few days at the end of february / first half of march. Combine this with the situation in Southern Tirol and the catastrophic handling of the situation in Ischgl and you’ll understand the situation in Bavaria. So we got this situation with realtively few cases througout march and an extreme low death-rate (people who go skiing are normal not part of the risk-group) at the beginning. But now both cases and death rate exploded.

        • Herbert

          You’d expect a certain North-South cleavage then first because of distance in the physical sense, and second because of linguistic, cultural and so on differences. Lutheran Franconians are less likely to be big fans of Austro-Bavarian Catholic regions…

          • fischerhias

            You can find a certain North-South-divide. The counties with the most covid-19-cases lie near the Austrian border (Rosenheim, Miesbach, Traunstein) or round Munich (like Furstenfeldbruck, Erding). Counties with a lot of wealthy people, who like to go hiking or skiing in the alps. The only exception is Tirschenreuth in the east-northern part of bavaria.

  6. fjod

    Nice analysis. It doesn’t seem to fit the data in France or England, however – I think both these countries show an effect of density between the major cities and everywhere else.

    In France, the Paris region has more deaths than elsewhere, but there’s minimal correlation between density and deaths in the other départements*, perhaps because none of them are particularly dense. The Mulhouse area also clearly has far more deaths, but excluding it doesn’t reveal a correlation.

    In England, there is a similar pattern. Case rates are highest in London, next highest in Birmingham, Manchester, Liverpool and Newcastle urban areas – all the large (pop >1m), dense (>2000/km²) urban areas – and then lower in the rest of the country. The case rate in the rest of England might have a small effect of density, but again it doesn’t jump out at me on a graph. Here are the relevant graphs:

    So what I’d concluded was that big cities are magnets for coronavirus, at least in France and England. It’s interesting that Germany (as well as Italy and the Netherlands) doesn’t show this pattern; I wonder what the explanation for this is.

    * with the possible exception of Lyon; I can’t really tell the significance just by eyeballing

    • Nicolas Centa

      And it is necessary to look at the mechanisms not only the figures.

      If you live in a tower and have to take transportation to do at least some of the purchases you need, you will meet people in the elevator and on the bus or train, who may infect you.

      If you live in a house and go shopping by car it is obvious that the number of people that can infect you will be lower.

      But that is only if you take precautions. Otherwise even in villages people will gather in cafes and get easily infected all at once by the few that got the virus while traveling elsewhere.

      I think if nobody is careful it is very easy for the virus to spread even without density because a lot of people are moving and meeting. However once social distancing becomes the norm, it is far easier without density.

      • RossB

        But what about all the things that enable that relatively safe activity? Someone has to repair the car. Someone has to take the money, to pay for the gas. Someone has to pump the gas, and if the driver pumps, then you better hope they are wearing gloves (that they then throwaway).

        My point is that it gets really complicated, and it yet it is clear that the most important thing is behavior. Extremely densely populated areas (Taiwan, South Korea, Hong Kong, Singapore) had really good behavior — model behavior, really — and they came out of it unscathed. Very low density areas (Michigan, Georgia) as well as medium density areas (northern Italy) and high density areas (New York City) have struggled, to say the least.

        So it is really the opposite. If you have good behavior, it really doesn’t matter where you are. But if you don’t, you are probably better off in someplace like Michigan instead of New York (although you will still have problems).

        By “behavior” I don’t mean just actions taken by ordinary individuals, I mean actions taken by the state (which largely guide the former and are probably more important).

    • Eric

      Big cities are magnets for coronavirus because they are magnets for travel. The virus starts spreading in cities, there are a lot of cases in cities before there are many rural cases, then the entire country goes to a lockdown at once. So the rural cases stay low. Not because it inherently spreads less in rural areas though.

  7. MucklaSSR (@mucklasuntermf1)

    Offenbach Nordend is 20,000 per km². With 52 cases on 130,000 inhabitants Offenbach is one of the least affected cities in Germany despite being one quite dense. Anyway calling Offenbach a suburb is quite misleading its Frankfurt’s biggest neighborhood. Half of Frankfurt’s official neighborhoods are located farther away from the city center than Offenbach.

    • Alon Levy

      I was thinking of Offenbach Landkreis, but yes, Offenbach Stadt is also very low-infection by German standards. (I emphasize “German standards” because Seoul’s infection rate is 0.6 per 10,000 and even disaster-zone-by-Korean-standards Daegu is only 27.9, i.e. same as Munich.)

  8. Benjamin Turon

    On cities and density in general: obviously it’s a problem – and has been since the first ziggurats were built in ancient Mesopotamia. The history of the city is in part the history of engineering, technology, and public services that overcome the challenges of density to allow for successful urban societies, often with ever greater levels of population and density.

    This includes water and sewer systems, mass transit, public police and fire services, fire and building codes, and public health services. It is these innovations and investments that are the reason why modern urban centers are not ravage by epidemics of typhoid or yellow fever. Its also why modern cities don’t burn down repeatedly – outside major natural disasters including earthquakes and wildfires – like London, New York, Chicago, and Baltimore have in the past.

    Yes, if in the 19th Century you lived in a tiny New England village, then wells and outhouses are likely to work without major health issues. You also likely could walk to work, the local general store, post office, or school. For a city like Boston, things were different, and society needed to build infrastructure and support public services to make urban neighborhoods safe and livable.

    We also know that cities that took decisive and effective steps to curb the 1917-18 Spanish Flu in the USA, like St. Louis and even New York City, had a lower death toll and faster economic recovery then those that dithered, like Philadelphia. Right now, it seems that places like California, Washington State, and Ohio have done better than Metro New York. The idea that the density of New York City makes for a bad result from a pandemic, is being tested in Louisiana, and likely to be tested in largely suburban Florida.

    Besides, in modern society, most life is basically urban life, from the high rises of Manhattan to the suburbs of Long Island. You might live in a single-family house and drive your personal vehicle, but your still entering places of great density and activity, from workplaces to retail stores to schools to Sunday church services. Right now, as COIVID 19 cases are falling in New York City, they are reportedly still on the rise in the suburbs and secondary cities of Long Island, Connecticut, the Hudson Valley, and New Jersey.

    And in our modern global society of trains, planes, and automobiles; it only takes one infected visitor from to spread a pandemic to a whole community, as seen in Albany, Georgia. And as seen in Albany, a smaller community can be quickly overwhelmed, with a disaster that in terms of the percentage of people sick and killed, equaling what is seen in larger cities like Wuhan, China.

    So yes, density is a challenge, but one that can be successfully overcome. The history of the city proves that, and this seems to be what has occurred in Germany during the COVID-19 pandemic, as well as South Korea, Taiwan, Hong Kong, and Singapore. Good public health infrastructure and administration in very well-run urban centers, successfully (so far) confronted the pandemic. Places more poorly led with poorer public heath infrastructure – like the United States – have fared worse, even with plenty of warning of what was to come from first China, then Italy.

    Overall, I doubt that the current Pandemic Emergency is going to affect the increasing urbanization of the world, including the growth of dense megacities. Hopefully, it will inspire or force through necessity, the building up of better cities. Including in the United States.

    • Luke

      I will say, anecdotally, that at work–in a hospital in conservative-for-New-England Manchester, NH–whenever COVID-19 comes up and NYC with it, the first thing out of everyone’s mouths is “density”. And this is among a group of people that know better, that cleanliness is what actually matters. They think this because they all live in suburban neighborhoods, and look at the skyscrapers of Manhattan and the skyscraping infection rates, and say, “ah, there is the causation”. These are not people I would largely regard as stupid, but unless you are primed to think critically, and open to the possibility that cities are, in fact, better-equiped to deal with this sort of situation because of the density of social and communal services, and not worse of because density is “foreign”–and foreign is at least strange, if not out-and-out bad, just like the coronavirus is bad–then people have and are and will make that connection, wrong though it is.

      • Benjamin Turon

        I have seen the enemy – and it is airline miles…

        What makes New York City vulnerable — and Gov. Andrew Cuomo in one of his recent press briefings stated this — is travel and social interaction, not density of population per se. You see this with the large outbreaks in winter ski resort areas, density is low overall, but high in hotels, restaurants, and lodges, with a big overturn of people from afar coming and going. An infectious disease like the flu will spread slower dense city of couch potatoes and basement dwellers, as opposed to bar hopping hard partiers in the rural countryside.

        New York City is the same, just on a larger scale. Now — as Mr. Cuomo said — of course international and domestic travel is not bad but very, very good for New York City, yet it did placed the city towards the front line of having a major outbreak first in the USA. This was compounded by the failure of the authorities to do what some nations did in East Asia to contain the spread of the pandemic through international air travelers.

        Despite the Trump Administration’s travel banned on China in January, hundreds of thousands continue to fly into American airports from China afterwards, with only the most preliminary of medical checks, at best. There was no testing for the virus, and no enforcement of quarantine. With that said, it is reported that most cases of COVID-19 in New York City can be traced genetically to Europe, and not China.

        The New York Times has reported after the first confirmed case of COVID-19 enter NYC from Iran through Kennedy Airport, the CDC failed to do tracing of the people that infected traveler came in contact with on the plane, at JFK, and after. This despite, the official announcement by state officials that it would be done by the federal authorities.

        And the first big confirmed outbreaks in Metro New York occurred not in the city, but at a synagogue in Westchester and a birthday soirée in Connecticut. For the Connecticut outbreak, the local health authorities did do infectious disease tracing, but gave up when several of the attendees to the house party attended several large major social events attended by hundreds of people. The man who is believed to be the “super spreader” at the soirée flew into New York, attended the birthday, and then soon flew out to South Africa, falling visibly ill only on that flight to Africa, were he was eventually tested and confirmed to have the virus.

        While it is the poor who will be hit the hardest, the pandemic in Metro New York seems to have been spread broadly initially by jet-setting high society. More evidence of this is the early closure, sickness, and one death of a prominent reporter at CBS Broadcast HQ in New York. In my own family I knew from the start that I was at the greatest risk of spreading the virus to my family, because my sister works a office job from home, my parents are retired and sell stuff on Etsy, yet I work in a major hotel in Saratoga Springs. Clearly the social density and churn at the hotel put me at greater risk of contracting the virus. And the risk was the same for people working and living in the Grand Tetons or Disney World.

        Its not population density that spread the virus around the globe and within nations, but rapid and widespread travel by jet aircraft, high speed trains, and superhighways. The Spanish Flue had steamships and trains, while the Black Death had sailing ships and horses. COVID-19 had Boeing and Airbus. Look, the one nation not hit so far according to WHO and seemingly confirmed the lack of mass graves (unlike Iran) seen in satellite images is the hermit kingdom of North Korea. Gee? I wonder why? They must not get out much, or have many visitors.

  9. Benjamin Turon

    Interestingly, according to both NHK World and the New York Times, the virus spreads most effectively indoors, and places with poor ventilation. The density of the amount of virus inhaled seems to increase the likeliness of sickness, the bigger the dosage, the more likely you will fall ill. And your most likely to take in a large amount, when indoors in close contact with someone who is infected. This is why hospital staff around the world seem so hard hit, and why a rural truck stop, a suburban coffeeshop, are just as dangerous a place to catching contagion as a hip club in the city.

    NHK World is now running public announcements calling for people to ventilate their indoor spaces well, to get air circulating, fresh replacing stale. Ironically, this makes older building that relied heavily on natural ventilation and lightening from before WWII better indoor environments then modern buildings that lean heavily, if not exclusively on mechanical ventilation (where it exists) and lighting. An older office building with tall ceilings and large windows that open wide lets in sunlight and fresh air, while modern building gives you the same stale air from the AC and often office spaces that never sees the light of day.

    When you look at old hospital/sanatorium design, often with big operable windows, high ceilings, and porches to roll out the patients onto for a breath of fresh air, they might do better than more modern hermetically sealed facilities. I know I have often felt ill in modern buildings like my 1970s elementary school, yet have had little issue in a old building like my college architectural studio, which where in a 1920s Beaux-Arts building with lofty ceilings, and big windows that open wide, letting in air and light.

    Pandemics spread in hospitals. Changes in design and protocols can save lives.

    How the Tuberculosis Epidemic Influenced Modernist Architecture

    Sick building syndrome

    • michaelrjames

      Interestingly, according to both NHK World and the New York Times, the virus spreads most effectively indoors, and places with poor ventilation. The density of the amount of virus inhaled seems to increase the likeliness of sickness, the bigger the dosage, the more likely you will fall ill.

      Yes. Also they think, or speculate, that the route of initial infection may account for the wildly different outcomes. If via nasal passage then the person may hardly notice, but if in lungs, and beyond some threshold dose, then serious disease.

      I continue to think that the extraordinary hot and dry conditions in Australia over Dec & Jan (remember, half the country was in flames) played a significant role in slowing its spread in the early days. Even in cities which may not have been burning but were very hot & dry (and getting regularly inundated with smoke from those bushfires, even in sub-tropical Brisbane where bushfires are rare due to humidity and mostly lush growth).

      Note that some of these features may actually make control of this particular virus a bit easier than thought.

      • adirondacker12800

        It’s is extraordinarily dry and comfortably warm indoors in cold climates all winter long.

        • michaelrjames

          Yet influenza is very seasonal, so American attempts to turn winter into summer, at least indoors, hasn’t changed this. It is also seasonal in the subtropical zone where it is more frequent in the rainy season –which I attribute to higher humidity = better survival of airborne virus. There are other factors and theories such as flocks of adirondackers sequestered in their warm indoors happily incubating and transmitting disease freely.

          It is yet to be determined if SARS-CoV-2 is seasonal but it does resemble influenza and even more the other coronavirus diseases like the common cold.

        • adirondacker12800

          Let me see if I have this

          Places that are warm and dry year round it’s the warm dry air.

          Places that are warm all year round but have dry season and wet season, the warm humid air of the wet season makes it worse.

          Places where it is cold in their winter and hot in their summer, the warm dry indoor winter air makes it worse but the warm humid air in summer makes it better.


          • michaelrjames

            Sure, in many ways you have it right.
            Which is to say, after all these years and billions spent on researching these diseases, they still don’t understand what causes seasonality. It correlates with various seasonal factors but it is hard to prove any one factor is the major causative thing. In the tropical zone there is less seasonality but it still peaks in the rainy season, which could be for all kinds of reasons including the simple one of more proximity between people. IMO humidity has to always be a factor because these viruses don’t like to be dry and survival just a bit longer in the air makes all the difference. However in all cases proximity is key (not so much all the crap we keep being warned about, like that steel pole on the metro etc. and least of all soft furnishings.

      • Matthew Hutton

        If you go to Australia from Europe you are almost guaranteed to have to go via Asia (traditionally Hong Kong and Singapore) and Asia has had better disease controls.

  10. smooth indian

    It is obvious that those using the COVID-19 crisis as a propaganda against dense cities are inclined to discredit the importance of big cities for cultural and ideological reasons. It is conveniently forgotten by such people that whether you live in big cities, small towns or villages, we need to interact with other people. Humans are a social species and while there may be preferences to the amount of interacting any human is comfortable with, it will still be an essential and will involve coming within arms length of other humans. People will still need to visit stores/markets (so much for online shopping🙄), eat in restaurants with friends/family, travel in common modes such as ride-share or flights. Of course good planning/design will help us in reducing the risk of spreading infections. Instead of spreading half-baked theories it would be worthwhile in potting resources and ideas towards minimizing such pandemics.
    Ironically, I am stuck in a small town in India during this crisis. The advantage of being in a small town is that basic supplies such as milk, vegetables, woodgrains and some other necessities are available easily since a lot of it is produced locally. Indians outside of the major cities still prefer to consume locally. I am hearing of people in big cities having a tough time obtaining supplies. These residents of my gated housing complex in Mumbai are prevented from stepping outside their apartments even when they just wanted to have a stroll on the lawns of the complex. All the while I can still step outside my home into the small yard we have to enjoy the different flowers my mother has nurtured.Since the house plot sizes are compact (13 meters X 16 meters) and not much fuss about privacy, we can still interact with our neighbors from our front yard (‘Aangan’ in India) and catch up on the latest gossip.

  11. Eric

    Whatever the rhetoric against density now, I think it will probably be forgotten shortly after coronavirus is eradicated. As precedent I would cite the rhetoric against skyscrapers after 9/11. As soon as 2004, the Burj Dubai was already under construction in the Middle East, and many more supertall skyscrapers have been built after 9/11 than before 9/11.

    • Benjamin Turon

      Its not population density but the density of “social interaction”, including long-distance travel, this pandemic spreading globally at a rapid rate via air travel. New York Times had a map showing how infected college students partying in Florida spread the virus across the nation during Spring Break. Despite this, air travel and long-distance vacations are unlikely to disappear after this pandemic is finally put to rest.

      • Eric

        There’s a lot of rhetoric against population density too – using NYC as the example.

        • Benjamin Turon

          Only link any one has provided of some one broadly attacking the population density of New York City as a long-term issue was of someone NOT attacking the population density of New York City as a long-term issue. Only a governor urging his citizens to “social distance” by reducing density on playgrounds. He said nothing about tearing down skyscrapers, ripping up subways, and force moving the Downstate population to Upstate NY to reduce population density. He said the opposite a few days ago, said the density of NYC was a good thing, and would speed up the economic and social recovery.

          I’m sure some idiot on Fox News or the website of the Reason Foundation has attacked New York for being “too dense”, but those folks attack urban centers and public transit all the time, they’re just temporarily directing fire from San Francisco to Gotham City. I myself have not seen anything in the media I read and watch, from the New York Times to NPR to CBS News, on population density, praising American suburbia. If you have a media piece from people beyond the rightwing nuts-o-sphere (the folks suing to keep churches open for Easter Sunday) advocating for reducing urban density, please provide a link. A “meme” is not “news”, as I keep telling my Trump supporting aunt in Montana.

          • Eric

            This article might be incorrect, but it’s exactly what you’re asking for. Found in about 5 seconds of googling. (I hadn’t seen the article before – I was more thinking of dozens of impromptu comments I had seen across social media in recent weeks.)

          • Benjamin Turon

            OK, that is a good start, but nobody quoted in the NY Times article suggested that the density in New York City, or use of mass transit is a overall long-term problem that should be curtailed after the pandemic is over. I didn’t see Randal O’Toole quoted. There was no long-term advocacy for cities to be like Houston. Also this article seems to be in line with what Mr. Levy stated in his first paragraph:

            “Solly’s paper looks at confirmed infection rates in American metropolitan areas as of late March and finds a significant correlation with density, but no significant correlation between deaths and density.”

            More people overall get infected in New York City, but the death rate of the infected doesn’t increase from overall national average, including hard hit rural places, like Albany, Georgia. And while yes, while the article stated that density and use of mass transit likely plays a part in the spread of the virus, so does being a “global city” with a huge amount of international travel. That would seem to be obvious. On the other side of the coin, this evening on the CBS Evening News was a piece on how unprepared rural America is for the pandemic, with less healthcare services per person, including hospital beds. Albany, Georgia might be the prime example, its small hospital was quickly overwhelmed, with no help forth coming for sometime.

            Still, perhaps Mr. Levy should have started this blog post by linking to this article from the NY Times and then name calling; Thomas R. Frieden, the former director of the Centers for Disease Control and Prevention and the New York City health department; Dr. Deborah L. Birx ,Coordinator of the United States Government Activities to Combat HIV/AIDS; Dr. Lee Riley, professor of infectious diseases at the University of California Berkeley School of Public Health; and Dr. Steven Goodman, an epidemiologist at Stanford University… as much of “incompetents” peddling “bullshit” about “density”.

            What I’m really looking for is an article titled: “Why COVID-19 Proves Big Cities are BAD (And why everyone should move to the Suburbs)”.

            Yes, I’m sure there will be one soon on the Cato Insitute’s website: “In the Age of Pandemic: Why we should build more Freeways instead of Trains” by Randal O’Toole. Of course Mr. O’Toole would attack mass transit and praise cars if Martians invaded. “War of the Worlds: Why a SUV will saves you from Mars”… “Why Grover Mills was safer than Manhattan during the 1938 Martian Attack”…

          • RossB

            >>> “Why Grover Mills was safer than Manhattan during the 1938 Martian Attack”…

            Nicely done. I got a chuckle out of that.

  12. Eric

    Here is a very interesting article explaining reasons why Japan has suffered less from coronavirus than other countries without lockdowns. Apparently what is needed for efficient viral spread is not just proximity, but *talking*. And in Japan there is a strong culture of not talking on the subway (and I suspect other public areas too). Masks probably help a bit too.
    View at

    I wonder if Germany also has less public talking than other European countries?

    • Alon Levy

      I haven’t seen any differences in behavior on the subway between New York, Paris, and Berlin – the normal amount of talking plus the occasional panhandler. Few to no buskers here, unlike in New York or Paris, but I don’t think that’s what the difference would be…

      • michaelrjames

        Here’s a tidbit from that Rosenthal article in the NY Times:

        As of Monday, there were more than 13,000 confirmed cases of coronavirus in New York and about 500 in Los Angeles. New York reported 125 deaths; Los Angeles reported seven.
        …. (density … crowding … in NYC … in subway and Time Square etc)…
        Another factor in the differing rates between New York and Los Angeles may be the warmer weather in Southern California, a climate that some early analysis suggests may slow the spread of the virus. Regions with average temperatures above 64.4 degrees Fahrenheit (or 18 degrees Celsius) account for fewer than 6 percent of global cases so far, according to researchers at the Massachusetts Institute of Technology.

        Not just temperature, but .. how does it go? … it never rains in Southern California.
        Sure, correlation is not causation but ….

        • michaelrjames

          Oops, that should have been posted elsewhere … under the thread on seasonality … nevermind.

    • Nicolas Centa

      The Japanese TV is now actually telling people it’s fine to take public transportation as long as you don’t talk (and wash your hands afterwards).

  13. SB

    Even if there isn’t relationship between Coronavirus and density, people’s perception matter and could impact where people live.
    Someone who was previously content with city life might feel different when they feel unsafe every time they have to step outside to do laundry and carry groceries to a small fridge.
    All of sudden detached housing with big yard and car feels appealing.
    Now they may have to move to region where they could afford to live a bigger house but people moving to regions with lower cost of living is already happening and could accelerate because of Covid-19.
    Readers of this blog who are interested in urban design and transit may feel differently but many people living in the city just to live most comfortable life that they can afford and Coronavirus could change how they feel about city life.

    • Alon Levy

      step outside to do laundry

      Not counting staying at hotels (or Airbnbs), the last time I stepped outside to do laundry was in 2011. Most cities are not New York. They’re not run by failures like de Blasio and Cuomo, they have trash-free sidewalks, and they have apartment furnishings commensurate with local GDP per capita.

      • Herbert

        I never quite understood why some apartments don’t have coin laundromats in the basement… It’s a cheap extra profit and the water and electricity is there anyway…

        Heck, it’s a much more profitable user of space than parking…

      • adirondacker12800

        Not everybody has your tastes or wants to make the same compromises. I know what a “pound laundry” is and if you couldn’t figure out how to have someone else do your laundry, in New York, someone has to lend you their Fodor’s. You can tell how rich a certain age of neighborhoods are, in the Midwest and Northeast by guesstimating the ratio of air conditioners hanging out windows to apartments. Or look at that age of the cars parked on the street. That doesn’t work in neighborhoods that are really rich because those people don’t park on the street. If they own cars at all. There are neighborhoods with rich people who don’t keep cars in the city. That’s likely only a consideration in Boston, New York, Philadelphia, Washington D.C. and Chicago. And the only place I’ve seen large numbers of black sedans loitering around is in Manhattan. Not everybody has your tastes or wants to make the same compromises.

    • Benjamin Turon

      Once again, the problem doesn’t seem to be population density — but very wealthy people who hopscotch around the world on jet aircraft to attend various social and work events, involving large numbers of people. These people are the “super-spreaders” who eventually infect entire regions. The density issue is not apartment buildings verse ranch houses, subways verse freeways, but birthday parties, business conferences, religious gatherings, and well-to-do college students grinding at trendy clubs while on spring break. These jet-setting super spreaders — who like a wild fire leaping large swaths of forests with burning embers– seed the virus in new regions, separated by hundreds and thousands of miles. What good is a big yard and a big card in Houston if your exposed at work from your boss who went to that conference in Boston?

      NEW YORK TIMES — How a Premier U.S. Drug Company Became a Virus ‘Super Spreader’: Biogen employees unwittingly spread the coronavirus from Massachusetts to Indiana, Tennessee and North Carolina.

      “The Biogen meeting was one of the earliest examples in the U.S. of what epidemiologists call “superspreading events” of Covid-19, where a small gathering of people leads to a huge number of infections. Unlike the most infamous clusters of cases stemming from a nursing home outside Seattle or a 40th birthday party in Connecticut, the Biogen cluster happened at a meeting of top health care professionals whose job it was to fight disease, not spread it.

      “Although some other companies canceled international meetings around that time, Biogen never discussed doing so. The outbreak was raging in China but had not yet been declared a worldwide pandemic. As of Feb. 21, the Friday before the meeting, the United States had only 30 confirmed cases, according to data compiled by The Times. Biogen executives in Germany, Switzerland and Italy — where there were just 20 known cases — packed their bags.

      “On the first night, about 175 executives gathered for a buffet dinner and cocktails at the Marriott Long Wharf overlooking Boston Harbor. Colleagues who hadn’t seen each other in a year shook hands and vied for face time with bosses. Europeans gave customary kisses on both cheeks.”

  14. RossB

    By the way, I can counter the “cities are bad” argument quite easily in the United States. If the country was only a teeny bit more densely populated a few years ago, then Hillary Clinton would be President of the United States, instead of Donald Trump. As a result, tens of thousands of people wouldn’t have died in the U. S., and perhaps a lot more lives would have been saved across the globe.

    I still go back to Europe though. Why have so many people in Europe died, given what had widely considered to be good health care systems in various countries?

    • Eric

      Trump is uniquely bad in refusing to listen to science at all when it says something he doesn’t already believe. But nearly all governments outside East Asia, while they respect science in general, did not understand it well enough to really get what exponential growth means. Meanwhile, the US could have ended up much worse than it did, but most state governors were responsible (on the level that Europe was responsible, not on the level of East Asia) which made up for the vacancy in the Oval Office.

        • michaelrjames

          Forget Republican versus Dem, here’s the real reason: add London Breed to this list:

          [NPR] A Forbes story has grabbed a lot of attention over the last 24 hours, and the headline (“What Do Countries With The Best Coronavirus Responses Have In Common? Women Leaders”) is clickbait gold. The piece has been read nearly 3 million times, and gives a nice roundup of how the female leaders of Iceland, Finland, Norway, Denmark, Taiwan, Germany and New Zealand gave a master class in leadership during this pandemic.

          NYC, NY state and of course the presidency could all do with their first female leader. And the MTA?

          • Eric

            Not really. Singapore and South Korea are up there with Taiwan for best response, but have male leaders. Iceland has one of the highest virus rates in the world, about 30% more per capita than Spain. Germany is doing very badly in number of cases, though well (so far) in deaths. Norway and Denmark are mediocre in the statistics. New Zealand is doing worse than Australia (right-wing male leader). I don’t know how such a nonsense article manages to get published.

        • Eric

          I agree that Cuomo and de Blasio were horrible on about the same level as Trump was horrible. That’s why I said MOST state governments, not all. For a long time California and Washington each had more cases than New York. Imagine if those states hadn’t taken their protective measures early, they would now have more cases than New York does. But they did take those protective measures. Most “blue” states (except New York) took those measures early, and most “red” states took those measures later but not disastrously later.

          • adirondacker12800

            South Dakota’s rate, as of the time the graph was drawn and I looked at the graph, sometime in the past 36 hours, was steepest of the states on the graph. It included New York, California and Washington State. And still rising fast. Go ahead whine that Bill DeBlasio and Andrew Cuomo were less than two days behind two other places.

          • Benjamin Turon

            Biggest hotpot in America is a meet packing plant in South Dakota, about 60 infected. The USS Theodore Roosevelt has had about 500 sailors test postived, with one death. A veterans home in Holyhoke Mass has had almost 50 deaths to date, with the National Guard called into take over the facility. Yes density is an issue, but you can find that across America from the NYC Subway to Walmart to churches to tire factories. The county with the highest rate of infection in Wyoming is also the richest county in America (more than Manhattan), Teton County. Moving to the suburbs won’t save you from getting infected from this or a future pandemic. Better governance at the global, federal, state, and local level will. As for New York State, yes they made some mistakes — including bickering — but compare to other states they have done fairly well. Surprisingly they may have done as well as Tokyo and Japan! Most governors have seen big increases in public approval, in stark contrast has been the approval ratings of President Trump and the governor of Florida…

            South Dakota Meat Plant Is Now Country’s Biggest Coronavirus Hot Spot

    • Matthew Hutton

      European leaders assumed the US president (or perhaps the UK prime minister) would take the lead and didn’t have a coherent response when that didn’t happen.

      • Lee Ratner

        After nearly four years of watching Trump and the entire Brexit debacle, they thought that our President or the UK prime minister would take the lead? Are they really that naive? There aren’t some mystical magical forces that will make Trump a serious leader when necessary. He will always be an utterly horrible person.

    • Onux

      RossB, why do you say without Trump perhaps more lives would have saved across the globe? What effect does/has Trump had on virus response in Italy, Spain or elsewhere?

      I agree with you on Europe. Before this conventional wisdom would have said that Western Europe had all of the traits to provide a better response (universal health care, lower inequality, stronger civil service, etc.) In fact there were articles early on that predicted a rough time for the US because it doesn’t have those things, predictions that look silly now that W Europe is so much worse than the US.

      Another question, why does Eastern Europe have infection and death rates so much lower, including EU and Schengen nations such as Finland, Poland, Greece, etc.?

      • RossB

        Trump’s response was terrible from the very beginning. To quote the New York Times: The virus at first took a back seat to a desire not to upset Beijing during trade talks, but later the impulse to score points against Beijing left the world’s two leading powers further divided as they confronted one of the first truly global threats of the 21st century. (

        In contrast, Obama and Bush both worked with other countries to reduce the spread of the disease. Remember how Ebola spread through Europe in 2014, and killed so many people? Neither do I, because (and this time I’m quoting The Atlantic):

        America’s prior investments in global health preparedness—the largest of any nation’s—have already made a tangible difference. In 2010, the CDC helped Uganda set up a new surveillance system for viral hemorrhagic fevers like Ebola and Marburg. Health workers there are now trained to recognize these diseases, and have tools for collecting samples safely. Labs have diagnostic equipment. Response teams are ready to go. “It’s been incredible to watch,” says Inger Damon, who oversaw the CDC’s 2014 Ebola response. “It used to take two weeks to respond to an outbreak. By the time you understood what was going on, you’d have 20 to 30 cases, and eventually hundreds. Now they can respond in two days.” Sixteen outbreaks have been detected since 2010, but they were typically much smaller and shorter than before. Half of them involved just one case.

        And in July 2014, in the midst of the West African Ebola outbreak, those investments very likely prevented a horrific catastrophe that might otherwise still be unfolding today. A Liberian American man brought the virus into Lagos, Nigeria, home to 21 million people and one of Africa’s busiest airports. “If it had gone out of control in Lagos, it would have gone all over Africa for years,” Tom Frieden, the former CDC director, says. “We were right on the edge of the abyss.”

        But Nigeria responded quickly. For years, it had used investments from the U.S. and other countries to build infrastructure for eradicating polio. It had a command center and a crack team of CDC-trained epidemiologists. When Ebola hit Lagos, the team dropped its polio work. It found every person who’d contracted Ebola, and every person with whom those infected had had contact. In only three months, after just 19 cases and eight deaths, it brought Ebola to heel and stopped it from spreading to any other country.

        That is my point. A smarter, better prepared president (i. e. Hillary Clinton) would have done the same thing that Obama did — worked with the country that had the epidemic, and reduced the spread of the outbreak.

        • Onux

          If I understand you correctly, you are saying that the virus got outside of China and infected the world because the US did not work with China to contain it the way the US helped Liberia/Nigeria contain Ebola. I see several problems with this reasoning:

          1. I’m not sure China needs the technical help. Liberia and Nigeria are very poor with medical capacity well below the US. China has one of the most advanced HSR networks in the world, produces patents in quantum computing, and there is a Biosafety Level 4 lab in Wuhan for goodness sakes.
          2. It’s not clear China has a poor response. If Chinese statistics can be believed (unlikely, see below) they have one of the best responses in the world, as good as Australia and New Zealand. They certainly reacted quickly with one of the strictest and longest shutdowns in Wuhan, and managed to sequence and release the virus genome. Even if they are underreporting by 90% they have a death rate better than basically all of the US and Europe, and an infection rate equal or better as well.
          3. China has been very unforthcoming in some ways with the situation, to the point of stories about doctors being disappeared if they don’t toe the party line. The New York Times reported that Chinese officials are spreading a conspiracy theory of the virus being introduced by the US Army. The idea that the Chinese would have welcomed the US with open arms to assist with virus containment on Chinese soil is dubious at best, farcical at worst.
          4. I don’t remember a major Ebola outbreak in 2014, but I do remember SARS in 2003, swine flu in 2009 and MERS in 2012. A hemmoragic fever like Ebola has a very different transmission profile than respiratory diseases like SARS, covid, H1N1, etc. There are no asymptomatic carriers of Ebola, it is very clear who is infected because they are laying on the ground bleeding out of their orifices. I’m not sure response to Ebola can be directly compared to Covid. SARS and MERS however are both corona viruses (Covid-19 is technically SARS-CoV2, in 2003 it was SARS-CoV1) and SARS and swine flu both originated in China. In these cases I see no evidence that Bush/Obama helped contain the outbreak by fostering cooperating with the country with the epidemic (mainly China). On the contrary, all of the diseases got out to other countries, and swine flu spanned the globe killing hundreds of thousands. If you do not blame Obama for the magnitude of global swine flu, you cannot blame Trump for the global impact of covid. His responsibility for the US reaction and performance is of course another thing entirely.

      • RossB

        By the way, it cuts both ways. Italy’s incompetence in handling the crisis has caused American deaths, just as American incompetence has caused European deaths. But American incompetence is due largely to one man: Donald Trump.

  15. Pingback: The Subway is Probably not Why New York is a Disaster Zone | Pedestrian Observations
  16. Pingback: That MIT Study About the Subway Causing COVID Spread is Crap – Streetsblog New York City
  17. Pingback: Automobiles Seeded the Massive Coronavirus Epidemic in New York City - Market Urbanism
  18. Pingback: Subway-Coronavirus Connection Suffers From Lack of Evidence - UCPcouncil
  19. Pingback: Governor Wolf Should Improve his Reopening Plan – Connect-PGH

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.