The U.S. Big Deal

The advance estimate (rough draft) of real GDP growth for 2020 I came out this morning: the annualized growth rate was -4.8%.

*************************************

OK. Time for some time series econometrics, following the videos on my YouTube playlist for ECON 3020.

I had estimated with logged quarterly real GDP data an AR(1) in differences, which is equivalent to an ARI(1,1) in levels: basically, Case 5 from the Handbook. Here's the rerun with the new piece of data:

(We should not expect a time series model to change too much when we add one new observation. If it does, this is usually indicative of a structural break, and more time fleshing that out in the software. But this one is fairly consistent with the earlier ones).

The important thing to flesh out of this is the size of the residual for 2020:1. This is our estimate of the shock to the economy (remember that the announced growth rate is a composite of the fitted value (capturing the persistent presence of past shocks) plus the residual (the new shock). Also, don't forget that the actual data is not annualized. The estimate of the residual is -1.9%.

Note that this annualizes to -9.3%. In conjunction with the announced growth of -4.8%, this is telling us that we benefited from the persistence of previous positive shocks that were swamped by the new one.

That shock is big, but not enormous: the caveat here is that most of the effect of COVID-19 on the U.S. economy occurred in the last 3 of the 13 weeks in the quarter.

When are the other big shocks? The biggest one is 1958:1 at -2.7%. That was the during a severe recession (the shortness of the 1957-8 recession is what keeps it from being ranked up there with the big 3 — 1973-5, 1981-2, and 2007-9 — since World War II). The other shocks that were bigger than the recent one are (in order from second worst to not quite as horrible) are 1980:2, 2008:4, 1981:4, 1960:4, 1981:2, and 1949:1. Every one of those is the worst part of a particular recession (1981-2 was a double dip), and most recessions don't have a shock as bad as any of these.

Some of those who did the time series portion of the final exam were surprised at the low AR coefficient, indicating that quarterly shocks are not that persistent. Even so, size matters: any persistence of a big shock is going to matter for a while: since the 2009 trough, the current shock is about 6 times the typical size. So, the hangover/aftershock from the persistence of the shock in quarter II will still be -0.7% (that's the -1.9% times the 0.36 coefficient). We haven't had a shock that big in 6 years: so the aftershock will be bigger than all shocks in the last several years ... and we should expect another big negative shock this quarter to be added on top of that. In 2020 III the aftershock from the current one will still be comparable to the most shocks of the last 5 years, and it will be mixed up with the aftershock from 2020 II, and any new shock in III (and who knows how COVID-19 will be doing in late summer???).

China's Big Deal

China has rescheduled (the national) Lianghui in Beijing for May 22. This is a very good sign that top leaders feel comfortable bringing a bunch of people from the provinces into close contact.

I can't find the page, but I quoted someone about 2 months ago noting 3 things that would need to happen to conclude that China is getting back to normal. This was one of them.

Let's Learn a Tiny Bit More About Logs, To Make a Common Chart a Whole Lot More Functional

Charts like this are common:

Certainly we should be glad that they have a log scale on the vertical axis.†

But charts like this are also rather stupid. The reason is lack of basic understanding of logs (something beyond ... well ... everyone else is using OurWorldInData and they chose the scale).

Consider this equation:

2^a = 10 

Is there an exponent x for which this is true? Of course.

Now, take logs of both sides to get:

alog(2) = log(10)

What this tells us is that any log in base 10 is a times the log of that number in base 2.

Looking above, the vertical axis is not labeled with the logs, but the values 1,000 then 10,000, then 100,000, then 1,000,000 are evenly spaced because in base 10 they are 3, 4, 5, and 6. But in base 2, they'd be multiples of those: a is about 3.32, so they'd be 6.64, 9.97, 13.29, 16.61, and 19.93.

In and of itself, this is not a big deal. What stupid about the chart is the faint lines showing doubling. These have to be added because they can't be read directly off the screen. But if you put everything in base 2 logs, that problem solves itself.

What's worse, those faint lines tell you zero about the situation right now. Take the U.S. for example. We are very close to the line for doubling every 5 days. The thing is, we used to double faster than that, and now we double slower than that, and neither one can be read off the chart.

Here's what I came up with instead:

Now, with the vertical axis in base 2 logs, each tick mark represents doubling. And to figure out how long to double, you just count to the right (in fact, if you're observant, you'll note that the days to double is now the inverse of the slope of each curve). So, for the U.S.:

• We most recently doubled in about 18 days (we're almost at 20 on day 56, and we were almost at 19 on day 38). That's a big difference from that doubling 5 days line we're close to in the top chart. 
• But, before that it was doubling 8 days. We really are flattening the curve.
• The doubling before that was in 5 days.
• The time before that it was 4 days.

It's hard for me to imagine a reason other than innumeracy as the reason no one does their graphs this way.

† Do note that I chose this chart because it happened to be the first one like it that I found when I went looking. I note this because it is rather biased: it compares the U.S. to countries that have "flattened their curves" to show that we're not doing very well. Point taken. But, it does not compare the U.S. to some random selection of countries, so it is missing that there are many countries, that like the U.S., have not had a good outcome with the outbreak. Also, it does not scale by population, which is a form of bias, since the U.S. is much larger than all the countries shown: in particular, the Faroe Islands at the very bottom are actually doing worse on a per capita basis than the U.S. is.

COVID-19 # 60

Kevin Hassett, who is a serious macroeconomist working in the Trump White House remarked this morning that he expected 2020 II to be the worst quarter for real GDP since the Great Depression. There's a reason in ECON 3020 that I make you work with the annual GDP data going back to 1929, rather than the quarterly going back to 1947: it has always been possible to get a repeat of the growth rates seen from 1929-1946.

New Zealand is set to re-open at midnight. They took advantage of their geographic isolation to aggressively combat their outbreak, and have been successul: cases are now in the single digits.

Boris Johnson returned to his office yesterday (10 Downing Street is the official, office-like place, the PM lives at # 11). That's 31 days since testing positive.

King Un still has not been sighted since April 11th. A letter from him was sent to the president of South Africa (although to me that sounds a little like those urban myths of absent schoolkids producing notes signed "Billy's parents").

The Economist had a cool graphic showing that reductions in smartphone movement are income dependent, with the poor reducing less.

The chart starts above zero because the measure is relative to average travel in January, so people were out a bit more as winter faded. The article (free with your email address) asserts that poorer Americans are less able to afford to isolate themselves. No doubt this is true for some. But there are some serious flaws with that. First off, the rich are about 5 times as likely to hold jobs, so one wonders where all the poor were and are going. Second, whether we like to admit it or not, the poor are less likely to follow beneficial directions. But probably most importantly, the measure on income used is by county, so what it's really tracking here is higher income urban areas vs lower income rural areas. Last I checked, most of us live in one of those, and one of the disadvantages is having to drive long distances for some things (and, of course, some of us like it that way).

Georgia began re-opening this week. Caitlyn Rivers noted that their own publicly available data shows that their epidemic is far from under control (click through to see the animated GIF which didn't want to play nice with Blogger). BTW: the large outbreak in the southwestern part of Georgia is associated with a single funeral on February 29th.

Some of you have studied some ethics with Dave Christenson. If so, you'll know what this metaphor is all about:

He is the “Tank Man” of the US. He tries to block the arrogance, anti-scientific temperament of some Americans, which is, however, supported and promoted by the US President. pic.twitter.com/Yyxzm1gso6

— Hu Xijin 胡锡进 (@HuXijin_GT) April 21, 2020

Mahan Ghafari posted a tweet storm using solid data about the Iranian outbreak.

(1/n) We gathered and analysed genetic and epidemiological data related to the #COVID19 outbreak in #Iran to gauge its size in late Feb - early Mar and estimate the likely start of transmissions. I will summarise some of our key findings in this thread. Will be ~8 tweets. https://t.co/0qiEd4Ns5B

— Mahan Ghafari (@Mahan_Ghafari) April 23, 2020

He dates the beginning of the outbreak to January 15th, with a serial time of 3 days. This is consistent with normal flights from China importing the infection fairly early. He also predicts a second wave of outbreaks around June 1st.
I low-level love this:

First graders back to school in Hangzhou, with social distancing headgear

The long horizontal plumes on Song Dynasty toppers were supposedly to prevent officials from conspiring sotto voce with one another while at court—so social distancing was in fact their original function! pic.twitter.com/0AOKsWE1xH

— eileen chengyin chow (@chowleen) April 27, 2020

Kids back in school, and they get to learn something interesting about history. Plus funny hats!

New research on an outbreak at a call center in South Korea:

Another crucial study showing how dangerous indoor situations can be. An alarming 94 (out of 216) people on the same floor in a call center in Korea got infected—an attack rate of 43.5%! But look at the schema. Most infections are on one side of the room. https://t.co/UBX98XfgQs pic.twitter.com/JQsVnDGtA2

— zeynep tufekci (@zeynep) April 24, 2020

Definitely shows how important social distancing is.

Lots of charts in this thread of excess deaths by region:

NEW: a lot of data on reported Covid deaths is highly suspect, so we’ve been looking into excess mortality — how many more people than usual have been dying around the world in recent weeks?

Story by me, @ChrisGiles_ & @valentinaromei (free to read): https://t.co/EiE5Q3OSmR pic.twitter.com/AiTdBnBma9

— John Burn-Murdoch (@jburnmurdoch) April 26, 2020

grib

That Pre-Existing Condition News

The other piece of news we got today is that data from New York City, through April 1, showed that most people who died of COVID-19 had some sort of pre-existing condition.†

Note that this is pretty much exactly what I clued you in to about a month ago. Simpson's Paradox is about the naive conclusions drawn from univariate statistics — like COVID-19 is a disease of the old, or African-Americans are more susceptible — that people make on the basis of their mental models (or statistically limited models).

COVID-19 is not a disease of the old. It's a disease of the sick, who are commonly old.

COVID-19 is not a disease of African-Americans. It's a disease of the sick, who are commonly from disadvantaged circumstances.

COVID-19 is a disease of the sick. Which actually, is pretty much like all the other infectious diseases out there that don't kill you outright and quickly. Ebola is probably not a disease of the sick because it's so effective at killing the healthy. That's why it's so scary. COVID-19 is tamer.

† BTW, the country should have a collective "Duh" moment about this. If pre-existing conditions are so important for COVID-19 mortality, then they should be really important for health insurers. But that's exactly what health insurers have been telling politicians and healthcare advocates for the last 10 years: insuring people with pre-existing conditions is really expensive. Do note that I am not saying that a national healthcare policy should cut those people loose. But I am absolutely saying that politicians and advocates who brush off pre-existing conditions as some meanness or profitability issue on the part of insurance companies ... are probably lying to our faces.

The Good News (for NY) and the Bad News (for Utah)

The big news today is that the state of New York did a carefully designed, randomized, serological  study.

Serology measures the number of people who have been infected with something and have developed antibodies to it.

The data was startling: the point estimates are 14% of people in the state, and 21% within New York City itself have had COVID-19. I have yet to see confidence intervals for that, but with a sample size of 3,000, I'd expect them to be about ± 2%.

The good news, which didn't require much deep thinking, was immediately picked up on in the legacy media: if a lot more people have had the disease than we thought, then its fatality rate is a lot lower too.

The bad news for Utah (and places like it) comes in two parts: one that is bad, and one that's neutral but sounds bad.

The bad one is that if the fatality rate is lower than we thought, then we can reverse the logic and say that places that have not done the serology work probably have a lot more undiagnosed cases than we thought. So COVID-19 is probably more widespread in Utah than it seems.

The more neutral one begins with the idea that we all knew (even though it is hard to admit) that with no immunity, a large fraction of the population of the planet is eventually going to get COVID-19. We now know New York state and New York City are much further along that path than we thought they were. And it makes sense that Utah is too: that's the neutral part. But it doesn't change the fact that we may still have to experience what New York City has gone through: they didn't get to 21% of the population having the disease without mass trench burials.

The important upshot from these statistics is that they will probably be used by policymakers to argue that COVID-19 is not as dangerous as we thought (which is true), but it does not change the fact that any particular location may have to run the gauntlet of a full-blown outbreak.

Let's crunch some numbers.

• New York City currently has 1,727 positive cases per 100K people. 
• But the serology indicate 21,200 people have had COVID-19. 
• So testing has found less than 10% of them, or alternatively that infections are 12 times cases. 
• For Utah that means that while we have 3.5K positive test results, we probably have 42K people who have had the disease. (That's 500 in Iron County, and 2K in Washington County).
• And those 3.5K positives work out to about 11 per 10K. 
• Worst case, our outbreak needs to double 6-7 times to get to the scale of the one in New York City. We had to double 11-12 times already to get from that first initial case to the 3.5K we have now. 
• Our current fatality rate is about 0.1 per 10K people. That's great. But doubling 6 times gets that to 12, and 7 times gets it to 24. That's a much better rate than New York City's 129 per 10K (probably reflecting that we live healthier, and we're more spread out here). But it still gets us to 400-800 deaths in the state (and 5-10 in Iron County, and 20-40 in Washington County).

So here's some food for thought: if our outbreak doubled 11 times already, do we really think it's not going to double 17 times in total? I put it to you that it's magical thinking that it can't.

But it might not. Maybe, just maybe, we are social distancing enough to get R0 down near 1, and we're persistent enough to keep at it. Maybe. May is not going to be fun. Stay safe.


† I lived my first 25 years in New York state. A problem with analyzing county level data is how to deal with Long Island. People in New York know how to do these stats properly, but I'm never sure when I see something like this in national media. Anyway, Long Island has 4 counties on it (Kings, Queens, Nassau, and Suffolk). But Kings and Queens are also in New York City. So when someone refers to Long Island, you need to figure out if it's the whole island, or just the suburban parts. I couldn't tell from the sources I used, so I didn't go there with deeper calculations.

COVID-19 # 59

There is nothing known for sure about Kim Jong Un. What is new is that a South Korean legislator noted on April 22nd that Pyongyang was put on a sort'of super-lockdown several days previously. No one knows why. Un skipped public ceremonies for the country's biggest holiday on April 15th. And Un's sister Kim Yo Jung (who has always held a ton of official and unofficial power) was formally elevated to a position as de facto successor on April 11th.

Gas is down to 86 cents a gallon, retail, in eastern Wisconsin.

There's excellent coverage of oil tanker storage weirdness in this piece from Zero Hedge. Over 10% of the world's supertanker ships are now being used for storage, prices to charter those ships are way up, total tanker volume used for storage is believed to be up over 50%, and 20% of the world's daily consumption is already anchored off the California coast with nowhere to go. Also, the U.S. doesn't actually get very much of its oil from Saudi Arabia (we have plenty of our own, and S.A. isn't that close), but there are actually tankers with 7 months full of our normal imports from them on their way here right now: the Saudi's are probably glad to have sold it to someone with a ship and are glad to see it go because they have to figure out what to do with the stuff coming out of the ground today.

In a February 13th post I noted oil rig counts, and that's I'd need to return to that data in a month to see what happened. Here you go:

They've been taking 5-10% of rigs out of production each week for about a month now. Note that taking a rig out of production is a ridiculously huge move. Legally, you have to back fill them with concrete down the pipe. "Capping" them is not just putting a screw top on. They will never be used again, although you can drill a new one right next to the old one when the price goes up again.

California officials have pushed back the first death in the U.S. from community spread of COVID-19 by about 3 weeks to February 6th. So it was "in the wild" here in late January. And The New York Times piece entitled "Solving the Mysteries of Coronavirus with Genetic Fingerprints" reports that the first known case imported into the U.S. — for a while we've known someone flew from Wuhan to Seattle with it on January 15th ... you know, before the government in Beijing knew there was a problem — is responsible on its own for cases in 14 states, including Utah. And, there's an intriguing new possibility: while there are hundreds of strains of the virus now, there were only a handful then, and it seems possible that one was in British Columbia even before it was in Seattle (and if you don't know, Vancouver has a huge expatriate Chinese community, and an international airport almost as big as SeaTac).

In India, families and mobs have attacked doctors for being unable to save patients. Not good folks ... not good.

Oktoberfest (the big one, in Munich) has been cancelled. BTW: it starts in mid-September.

Tyson indefinitely closed the country's largest pork plant  due to absenteeism.

Singapore, widely believed to have the best healthcare in the world, continues to get worse.

Remember that China's CDC also uses a 3-level warning system. In March they got the whole country down to level one. Now Chaoyang, an urban district within the Beijing metropolitan area is back up to level three. Chaoyang is close to the airport, as well as home to many expatriate homes. Well, it's a little worse than that: it's home to almost 4 million people, is one of the districts in the city's core, hosts every embassy but for two, and is the location of the central business district for the city.

Unintended Consequences and the Lucas' Critique

I low-level love this story. Definitely not for any of the people involved specifically, but for the general idea that public policy is a hard thing to do because people adjust their actions and expectations.

Anyway, for the future, when the link may be dead, the city of San Clemente attempted to enforce California's social distancing policies. A problem are cropped up at a local skate park: it was closed, but the kids showed up anyway. So the local policymakers filled it with sand. So the kids switched to dirt bikes.

This relates to one of the most famous ideas in macroeconomics: the Lucas' Critique. In 1976 (and there are bits of this in a 1972† paper too), future Nobel Prize winner Robert E. Lucas Jr. critiqued why policies like those proposed by Keynes often fail. He argued that the design of them often incorporates the idea that people won't react to the policy; that they'll just accept it without changing other aspects of their lives. But Lucas noted this is like a game of strategy: policymakers put in a policy, people figure out the best way to make that constraints in that policy work for them, so then policymakers have to put in new constraints, and so on. The meat of the critique is that the statistical models used to develop policies did not include all that back and forth. Since that time economists have developed much better (and sophisticated) modeling techniques that allow us to envisions better policies. And I'm not saying that government officials don't have access to that stuff. But, what happens is that we all still have our mental models of how policy should work, and they have not advanced on this at all because it's too difficult for us to do the math in our heads.

Back to the skate park: do you think there was a lonely voice in the meeting rooms of the policymakers saying "I think we need to come up with something better than telling kids they're banned from the skate park"?

† This one, "Econometric Testing of the Natural Rate Hypothesis" is hard to find these days. I have a hard copy if anyone needs it. If you want the original, it's in a 1972 book The Econometrics of Price Determination, edited by Otto Eckstein, whose chapters are papers from a conference at the Board of Governors of the Federal Reserve System from 1970.

COVID-19 # 58

Governor Gavin Newsom of California was asked straight up on national TV this morning if schools would open in September in a normal way, and his answer was "No." Do note that's not saying they won't open, but rather that if they do, it won't look like a year ago.

Harvard researchers suggest that fully re-opening the U.S. economy will require 20 million tests per week. That's everyone doing a test, on average, every 2 weeks. They expect that capacity to be available by late July. Note that this doesn't mean that politicians couldn't re-open sooner, but what the doctors say is necessary for it to work well.

Yep. Giving away crude oil.

An economy based on one resource? Uh-oh.
Today oil plummeted 321% amid due to #COVID19 #pandemic.
This means:#Nigeria 2019 exported $509B worth of oil. Now? Losing $88 M/day.#Aramco #SaudiArabia >$850B 2019 oil sales. Now -$500M/dy.
Norway $28B exports in 2019. Now -$48M/day. pic.twitter.com/BAjHhOei0h

— Laurie Garrett (@Laurie_Garrett) April 20, 2020

And it's not the crappy stuff either. Crude oil can be produced at any well, but any price for that oil and its transportation is unofficial and not usually public. When it gets to a location called a price point, oil from different wells is blended and then graded to a benchmark, and the benchmark determines the price it's sold for publicly. WTI and Brent are not the best, but they are very good, and they are sold in the high volumes (it is surprisingly difficult to get any information on volumes by benchmark — no one wants to let anyone else know the true quality and volume each well produces, so they hide info through decentralized sales).

Rioting in Paris today.

Just out from The New York Times is "28,000 Missing Deaths:Tracking the True Toll of the Coronavirus Crisis". They looked at publicly available data on deaths in 7 countries and 3 large cities with major outbreaks. The compared average deaths in past years to currently counted deaths, and subtracted out those associated with COVID-19. This gives them an estimate of how many COVID-19 deaths there are around the world that have not been counted. That's the 28K. And, just to make things worse, their data does not include Italy, China, or Iran at all, and only the largest city in each of the U.S., Turkey, and Indonesia. Need a picture?

Do note that the red is showing total deaths, which is a median of 30% higher in red. This is before subtracting out the ones that have already been identified as COVID-19 deaths.

The University of Texas has an interesting simulation site. You can break it down by state level, and it tells you the probability that daily deaths have peaked.

Interestingly, the Governor of Georgia is going to re-open the state very soon, and this site indicates that it's nearly certain that Georgia is past their peak.

Here's a good quote:

When I asked @mlipsitch recently about the future of the #covid19 pandemic, he told me: „We’ve managed to get to the life raft. But I’m really unclear how we will get to the shore.” Here‘s my story and a short thread about the long way to shore https://t.co/u2n0E6JWU1

— Kai Kupferschmidt (@kakape) April 15, 2020

A good sign from China would be rescheduling Lianghui (the nationwide bureaucrats convention). A still fair to good sign would be having it using Zoom.

Prelude of national #lianghui? Zhejiang’s party leadership decided that cities in the province will convene their “two sessions” - annual meetings of law makers and political advisors - from late April. Measures such as video conferencing would be explored to reduce gathering. pic.twitter.com/MeW0hK4KoD

— Xinqi Su 蘇昕琪 (@XinqiSu) April 20, 2020

Zhejiang (province) is on China's east coast, south of Shanghai. Getting cities to do their Lianghais would be a step in the right direction. As always with China (and good advice more generally), watch what governments do, not what they say.

All exercise facilities in Beijing were closed abruptly today. Rumor is there was a positive test result from someone using one.

Still keeping an eye on factory fires: today in Chengdu and Zhengzhou, yesterday in Wenzhou and Nanchang. Before about a few weeks ago, videos of factory fires were not a thing. Now they are. What's up with that?

COVID-19 # 56

Pretty damn big financial news today. Didja' ever see Crazy Rich Asians? It's about crazy rich asians who happen to be ethnically Chinese families native to Singapore. Anyway, one of those big (#1349 in that linked article), closely-held, businesses went bankrupt  — and in a worse than normal way — over the last few days. Hin Leong Trading (HLT) was in bad enough shape that it pledged collateral to obtain funding ... and then sold the collateral. So its bankers (including HSBC, and # 267 in the linked article) are out way more than they anticipated might be possible. This is the way financial crises get started. HLT has a lot of business lines, but their main business is oil storage and tankering, which I've posted about here and here as being a weak link in international energy infrastructure.

Yeah ... about Singapore ... they used to have it under control. Not so much now:

The gross numbers are not that large in Singapore, it is after all a city-state. Remember not to do comparisons between the curves a graph like this: New York City's rate is still 10 times that of Singapore.

China's Ambassador to Russia released a video discouraging Chinese students in Russia from going home or attempting to do so. Of course, Russia is on the upswing of their outbreak while China is trying to void re-igniting theirs.

FYI:

Philadelphia prematurely ended its quarantine from the Spanish Flu. Some 200K people lined the streets. Within 72 hours, Philadelphia's 31 hospitals filled & the city ended up with 4500 people dying from the flu or its complications within a matter of days. Learn from history

— Arthur Caplan (@ArthurCaplan) April 19, 2020

Philadelphia's and St. Louis's (better) experiences are often contrasted in the literature.

NYC data:

The last time I saw 3 obituary pages with an index in the Sunday @nytimes was after 9/11....the new #COVID19 normal. Many families ask that donations be made to health related institutions in their loved ones' names. pic.twitter.com/S1QLGunazO

— Laurie Garrett (@Laurie_Garrett) April 19, 2020

I posted something similar about Bergamo about a month ago.

Do you remember that ancient history (from all of 10 days ago) when I noted that the U.S. death toll from COVID-19 would soon exceed the Korean War? Yeah ... happened yesterday. We'll pass the entire 8 years of the Vietnam War (a defining part of American culture for 55 years now), probably on Wednesday.

Ummm ... yeah ... have you checked out the numbers in the Central Utah Public Health Department District? There are about twice as many people in those six counties as in Iron County, so it's not like they shouldn't have more cases. And they don't. But they're rising pretty quick, and they have them in 4 of the 6 counties now, and it's not like those communities are all connected through a tight network.

COVID-19 # 57

Hospitals in Tokyo are turning away potential patients. Japan continues to get worse. Russia and Turkey too.

Hospitals in Mexico City are turning away potential patients.

May oil future for West Texas Intermediate (a particularly high quality and high price form of crude oil) hit $2/barrel this morning. Later in the day, some futures prices did drop below zero.

Police in South Africa fired rubber bullets at rioters demanding food.

Some countries feel they are far enough past their peak to begin re-opening: Germany, Denmark, Czechia, Norway, India and Ghana.

Indonesia continues to be in complete denial. Outbreak is believed to be large and growing in Jakarta, but they're still not running many tests.

An outbreak in Harbin in northeastern China has been traced to a college student who returned from the U.S. after her college went online.

There are rumors and counter-rumors all over the internet that North Korean leader Kim Jong Un is dead or near death. It has not been reported that this is related to COVID-19. It was noted on the 15th that King Un did not make a public appearance on a major national holiday.

COVID-19 Working Papers

The European Economics Association is hosting a board to inform others of researchers ongoing efforts in understanding the events of the last 5 months.

The board is here. There were 80 papers when I started counting, but they aren't numbered, and I'm not counting them again.

The invitation to post is here.

Estimating R0

Excellent article by Andy Larsen in the Salt Lake Tribune entitled "This is one key metric Utah should track as the state considers easing coronavirus restrictions"; it's on estimating R0 in real time. It includes estimates (with confidence intervals). Do note that the big data coder is more sophisticated than the epidemiologists. The latter view R0 as a constant, which isn't very applicable to COVID-19, since the whole idea of social distancing is to make R0 go down. I doubt epidemiologists are unaware of either point, but the jargon of the field has yet to catch up. Anyway, here's Rt on a state by state basis:

More importantly for SUU economics students, it includes links to the Python code to run the whole thing.

Of course, you need to get into the safer range (R0 less than 1), and then stay in that range. Utah did well on the first one, but not the second:

 

The article is also the first I've seen to point out that Smithfield executives gave perverse incentives to workers that contributed to their outbreak. That one is going to be in ManEc texts for decades to come, I'm sure (and I'm not sure how I missed it so far).

You could read a lot into the rankings of states, noting that some that want to "re-open" are still at the bad end. But, there's also some that dealt with their outbreaks swiftly and decisively, like Washington, that are still at the bad end. But here's the thing: no one wants to turn out like New York, but New York

End for Spring 2020

Anything newer than this post is not required for ECON 3020 for Spring 2020.

Anything older than this post is required for ECON 3020 for Spring 2020 (typo fixed, thanks TH).

That Stanford Study

You might want to be leery of that Stanford study that's been in the news.

It's in the news because it finds a much higher rate of past infection than would be suspected by positive tests. That would make the fatality rate immediately much lower, but would also suggest that the outbreak is further along and that the population is closer to reaching herd immunity.

Except ... they used  a convenience sample drawn from a Facebook ad. Do you think it's possible that people might be drawn to that ad because they'd had symptoms???

COVID-19 # 55 (Required Parts Highlighted)

China updated the death tolls in Wuhan, adding over a thousand to the total. Better late than never.

Here's an updated version of those common charts:

Nice that it has a logged axis. Not nice that it allows comparison of one country to another (which is usually good, but which with a logged axis can lead to trouble). Not great that it doesn't have deaths as a percentage of population: that would be more informative, but would obviate the need for logging the axis too. The variety of countries shown is good, as well as the extended time frame on the horizontal axis ... but it also tends to show that the story China is telling is somehow different.

Deaths were way up in the U.S. yesterday. People have been trying to rationalize the reasons, and I haven't seen a good one yet. Oh boy.

Do you know anyone that still thinks this?

1/2 New York City, with population of 8.4 million, has 11,477 confirmed and probable #COVID19 deaths. Even if everyone in the city had covid (8.4M people) which is improbable, then 11,477 deaths would put its case fatality rate higher than what's seen in seasonal flu. #NotTheFlu

— Scott Gottlieb, MD (@ScottGottliebMD) April 17, 2020

If you do know someone like this, they'd better that a virus that somehow has spread all over the globe is going to stop before it gets to their town. BTW: deaths were up 10% in Utah ... in the last day, and cases were up 4.5%. We're up to 42 in Washington County, 17 in Iron County, and 15 in the rest of the Southwest and Central Utah Public Health Districts.

The IMHE model is the primary one everyone is using to forecast COVID-19:

New analysis by @meyerslab finds that the popular IHME model underestimated deaths when applied to Italy and Spain. h/t @nataliexdean pic.twitter.com/ynpHDOOFyc

— Caitlin Rivers, PhD (@cmyeaton) April 17, 2020

Not good. BTW: the model  is doing a pretty good job of forecasting deaths in Utah though. It's still predicting a Utah peak late next week.

Best estimates are the U.S. needs to do about 500K tests per day to be able to model the outbreak effectively. But over the last week we've plateaued at about a third of that. As far as states go, Utah is doing OK:

You have to crunch the numbers a bit, but a 500K/day rate would shift a state from one color to the next every 4 days. Currently we're shifting every ten.

One of the things that is worthwhile hoping for is that a lot of people have already been infected and don't know it. There are serology tests for this. Here's what they found in Wuhan:

Wuhan hot spot Zhongnan Hospital did antibody tests on 3,600 security guards, cleaners, doctors and nurses and 5,000 visitors.

about 2.5% positive for exposure. which surprises on the low side, while still being 5X the rate of confirmed cases in city population. https://t.co/OdAljsf0nE

— Antonio Regalado (@antonioregalado) April 17, 2020

While showing a lot of undiagnosed cases, it's not nearly a high enough proportion to limit the spread of future outbreaks.

There's now video of mass burials coming out of Iran.

Gas Prices

As I mentioned in yesterday's post, current rack (wholesale) prices are hard to come by. But I have a source who helped me put together a back-of-the-envelope estimate that the prices for regular unleaded gas at the racks in the area is in the $0.90/gallon range.

The article from Bloomberg about rack prices being as low as $0.12/gallon in the upper Midwest was paywalled yesterday, but I was able to find this free version this morning.

Keep in mind that gasoline is retailed with essentially zero profit. Gas stations make most of their money off of car washes, candy bars, and coffee ... and sometimes even ... repairs. Once you start with the rack price, and add in shipping, plus state tax, plus federal tax ... then you've got to add in the maintenance costs of running the pumps, insurance, labor in the c-store, and there's not much left over. So that $0.90 wholesale price translates into a retail price that's over a dollar higher.

I showed you the chloropleth maintained at GasBuddy earlier in the semester, and here's the current one:

Those darkest purple areas in the upper midwest seem to be mostly in Wisconsin, which matches up with a new twist that I learned from the linked article:

The price decline is especially pronounced for cities at the end of pipeline systems, such as Fargo and Milwaukee.

Cedar City is along a pipeline, not at the end of one. And this is just crazy:

What we are seeing is that a lot of the big pipelines are being used as storage, and the product will just get pushed and pushed until it has no place else to go.

Our pipeline has a storage terminal at the end that can hold 330K barrels, and the one west of town that can hold 200K barrels more. The maximum capacity of the pipe itself is 120K barrels, but typical flow is more like 60K barrels/day. A typical gas station might take 300-600 barrels to fill its underground storage tanks. I do not know how many gas stations are served by this pipeline, but I'd guess about a thousand. So figure 500K barrels is the maximum storage in our neighborhoods. That puts storage in our area at 330+200+500+60=1,090K barrels ... and the industry is so hard up for storage that they're willing to use the other 60K of pipeline capacity to find some more room! Again, I'm guessing at local numbers, but I figure the region served by our pipeline uses 100K barrels per day. So, if all the storage started out empty, and we're down on our consumption by say half, the whole system would fill up to the top with new product again in about 22 days, and they could maybe stretch that by a day by filling the pipeline to the very top. No wonder prices are dropping like crazy.

COVID-19 # 54 (Required Parts Highlighted)

New research shows social distancing is reducing R0.

Those are 95% confidence intervals. When R0 drops below 1, as the estimates indicate it has in southern California, the infectious agent will die out. Of course, we also have to worry about what the estimates say about northern California possibly getting worse.

This chart suggests that Utah is still in pretty good shape (we are not on here, so use your imagination). The curves are estimates of equivalent-mortality lines (level sets, sort of like indifference curves). For example, if you're at the top left, you've tested a lot, but the deaths haven't really hit you yet, and as you go towards the "southeast" you're getting less cases identified, but more deaths from as yet unidentified cases. So the curves capture regions that have comparable outbreaks, even if they do not have comparable testing and enumerating policies. Obviously, your region is worse if you are towards the "northeast". The "tails" indicate the movement over the last week: no state is improving.

Given today's numbers (21 deaths out of 3,206,000 population puts us at) 0.65 on the axis along the top, while (2,683 cases out that population puts us) 8.4 on the scale along the vertical axis. So ... a small circle above and to the left of the T in Tennessee. This comes from a fantastic weekly report from the CDC of China. Data like this, since it is scaled by population does not (and usually should not) be logged.

They also produced one for countries. Comparing these is somewhat difficult. If overlayed on the one above, the one below would look more like landscape format.

China's CDC has been producing these for 3 weeks now. The charts in the two older ones (here and here) are exceptionally sobering.

On a good note, the lengths of the tails suggest that European countries are getting worse faster than U.S. states.

On a bad note, the April 1 edition shows estimates of R0 versus weather. It shows no signs of R0 falling as weather warms up, or of being lower in tropical regions. There is also little difference between northern and southern hemispheres.

Also of interest in the above is that Sweden is currently the world leader in following a policy of not social distancing. They're not doing too great.

The same paper also included this chart:

 

This one requires a little more care to interpret. Going to the right isn't good, but it isn't scaled by population: the U.S. has the 3rd biggest population, so we shouldn't be surprised if it is at the top for number of cases. But it's definitely bad going up; "flattening the curve" is equivalent to moving downwards. I mentioned several days ago that I was really concerned about Russia and Turkey, and this shows why. I was also concerned about Japan, which does not look too bad here (although I suspect if Japan tested more they'd move higher).

Here's the bad news. While that study was released yesterday, it shows data only through April 8th. We've had a week to get worse.

I've been reporting figures for Iran from dissident sites. Now there's this:

Iran Parliament’s research center reporting government was vastly undercounting coronavirus cases and deaths.

Estimating 760,000 cases and 8,500 deaths.

Would make Iran’s caseload highest in the world.

— ian bremmer (@ianbremmer) April 16, 2020

That is a semi-official number indicating 10 times as many cases and twice as many deaths as the official numbers. Meanwhile, the official numbers coming out of Iran today showed the lowest number of deaths in almost 5 weeks. I hate to say I told you so ...

Brazil's minister of health was fired. President Bolsonaro has been downplaying the risks of SARS-CoV-2 after having a mild case after a visit to Mar-a-lago.

There is a video from Douala, the capital of Cameroon showing the aftermath of the top doctor at a hospital dying of SARS-CoV-2. The nurses then ... ran away.

About a third of those surveyed believe the virus was created in a lab. I know it's fun to play around with that idea, but the best evidence is that this is still an urban myth.

There's a new dash board on county level preparedness. On a 1 to 100 scale (with 100 the best), Iron County is a 32. Washington County, with 5 times as many people, is a 12. Washington and neighboring Garfield Counties are the 2 worst in the state. Most frighteningly, they don't predict Utah will peak until August 3rd. I can probably get into a lot of trouble with President Wyatt for noting what that means for Fall semester. There's not much way to check that peak date (no one is forecasting this stuff ... what they are doing is simulating, and that's much sketchier) but they seem fairly accurate for the area around NYC.

Is this sobering enough for you?

Here is the story of today's @ONS mortality statistics in a single chart. The worst week for UK mortality (all causes) since records began.
The blue area shows the minimum and maximum mortality each week since 2010.
The red line shows you this year so far. pic.twitter.com/j71W00tCVy

— Ed Conway (@EdConwaySky) April 14, 2020

Worst week ever is fairly convincing. And the UK is not as bad off as, say, Italy or New York City.

Singapore has been doing a great job. Singapore's cases have been rising ominously for a while, and they were up 20% again today. Uh-oh.

Gasoline is sold wholesale at racks. Our rack is off Highway 56 about 20 miles west of town. Retail prices are typically marked up $0.20 to $0.60 over rack prices. Up to the minute rack price information is sort'of a trade secret; for data you don't have to pay for, you're probably looking at 30+ days out of date. Anyway, Bloomberg reports that storage facilities in the midwest are so full that rack prices have dropped to $0.12 per gallon, implying retail prices under $1.

I am just suspicious at this point, but I am getting leery of the hospitalization numbers (I'm talking mostly about NYC, because what else does anyone look at these days). I think what we may be seeing is the message getting out that unless you are having difficulty breathing you should stay away from the hospitals. This is lightening the load in emergency rooms. What I'm seeing is the load not lightening in ICU's yet.

There's now a third regional group of governors trying to plan out a restart of the economy. I actually think these "beneficial cartels" are sort'of interesting, and a positive thing.

A Very Guarded Quote (Updated a Little)

This happened today: protesters clogged the capital of Michigan (social distancing be damned) to protest ... social distancing.

I am not supporting or condoning this.

But the group's leader made a money quote:

Quarantine is when you restrict movement of sick people. Tyranny is when you restrict the movement of healthy people...

It is worth emphasizing that the world was surprised back in January when China locked down healthy people. This was surprising because [and I am certain that I mentioned this in class as well as in this post] quarantining is what you do with the infected people. What China did was not a quarantine.† There simply was no epidemiological evidence prior to that Chinese policy of any government ever trying that, much less having it be effective.

So, it is worth pointing out and remembering that we still don't know if this is a good idea or not, on net.

† They are called SEIQR models because the population progresses from left to right, and quarantine comes after infection.What comes before infection, susceptibility and exposure, includes nothing about social distancing or quarantining healthy people.

*************************************

My personal opinion is that social distancing and shutting down much of the economy is working, and I support it. So I'm angry at protesters like those in the linked article.

But I will also admit that I do not know or have evidence that social distancing will work, and my feelings on this could be wrong. Lots of policymakers around the world would benefit from realizing this.

Also, paying close attention to the countries (say Sweden) and states (say South Dakota) that are doing things differently is warranted. Recall from basic regression analysis that it is the observations that are not like all the others that have the most explanatory power.

I added the two highlights in green about 12 hours after initially posting this.

More Innumeracy

This is from Reuters, so it's available by clicking a link the world over. Here's a quote:

The United States recorded its first coronavirus fatality on Feb. 29. It took 38 days to reach 10,000 deaths and just nine more days to go from 10,000 fatalities to 30,000. The previous high single-day death toll was 2,364 on Tuesday.

It's true, but the presentation of the math is misleading.

First off, note that going 10K to 30K deaths is tripling in 9 days.

Secondly, how many times would you have to triple to get to 10K from one? Hmmm: 1, 3, 9, 27, 81, 243, 729, 2187, 6561 means that tripling 9 times would get you well past 10K.

Third, to triple 9 times in 38 days is tripling a little more than every 4 days.

Would you rather triple in 4 days or 9 days?

While all of these deaths are tragedies, and the ongoing pandemic is a horrible thing, what we have here is evidence of "flattening the curve" reported as things getting worse because " It took 38 days to reach 10,000 deaths and just nine more days to go from 10,000 fatalities to 30,000."[emphasis added]

COVID-19 # 53 (Required Parts Highlighted)

Interesting twitter storm from an emergency room doctor working in New York City. Not much urban mythology about the U.S. healthcare system in this one.

Here's some arithmetic to put things in perspective. New York City has a population of 8,000,000. Currently they have about over 100,000 positives, and over 10K deaths. So that's 1 out of 80 people has it, and out of those, 1 out of 10 will die. If you were on campus at SUU, 1 out of 80 is roughly 2 positives in the typical student's classes. And, between you and 4 friends, you'd know 10 positives, one of whom would have died already.And with an R nought of 3 and a serial period of a week, it would go from 1 out of 80, to 3, then 9, then 27, then 81 out of 80 within a month. Glad you're at home? And the amazing thing is — in New York City they're feeling a little better about their situation because they felt it could be worse.

New research suggests the intermediate animal that transmitted the virus from bats to humans in Wuhan may have been dogs. The presumption is that feral street dogs at bats or bat droppings, and the virus thrived in them.

Of course, you've probably heard by now that China is requiring scholars at universities to get their COVID-19 and SARS-CoV-2 research approved by the government before submission to academic journals. Last time I checked, peer review meant review by your peers.

Trump is going to defund the WHO. Not good (at this time). But lost in all of the news about this is the strong suspicion that another one of the WHO's large donors told them the same thing about 3 months ago.

Yesterday was Day 1 for Austria starting to allow small businesses to re-open. Stay tuned.

This is personal opinion, but did it seem like the story du jour in the legacy media changed on a dime this week? From New York City is getting worse, to plans to "re-open the economy".

For those of you not getting this, governors have become so important because of the 10th Amendment. It states that any powers not given specifically to the federal government belong to the state governments. And pandemic response was not exactly something that the federal government was planning for. Already, there are two groups of governors planning on working in tandem to "re-open".

Interesting news in China: there have been quite a few factory fires. No one is sure why (China's insurance system does not work the same way as ours, so owner-arson seems unlikely). Not much news coverage of it inside China either. This bears watching.

Fake news: at a press conference Iran's health minister showed an image of an Iranian nurses PPE related bruises. Internet sleuths determined it is a photo of a nurse from Brazil. Meanwhile, dissidents report Iran's death toll to be roughly the same as the U.S., and about 5 times higher than the official toll. That's consistent with the WHO report from a month ago that the number of cases was 5 times larger than the official numbers.

COVID-19 # 52 (Required Parts Highlighted)

The next problem for Americans is going to be rolling closures of "essential services". This week it's a Smithfield Farms production facility in South Dakota, that produces about 5% of the packaged pork products in the country. I seen this as potentially a big problem, but not one so far — what is essential is that places like this don't start to shut down simultaneously. BTW: the "pig farm" in Milford (about 45 minutes north and then northwest of Cedar City) is part of the supply chain for Smithfield Farms.

In China, there were provinces, especially to the north of Wuhan, where the first wave of COVID-19 had little impact. Suifenhe, in Heilongjiang, is not under a lockdown similar to that in Wuhan from 11 weeks ago. This includes building new hospitals. China is blaming this outbreak on travelers from Russia. There may be some truth in that, but it only takes a few seconds on the internet to confirm that most of the population in that area lives on the Chinese side of the border (and epidemics tend to go urban to rural).

The Epoch Times has obtained health department documents from Wuhan during the worst parts of the outbreak. I have stressed two big points about Chinese data this semester. The first is that it tends toward falsehood; not outright lies, but consistent shifts in one direction. Students in past semester seem to get a good handle on that, but they're less good on the second point. This is, that China's data problems start at lower levels of government, are compounded as they go up, and are not typically a national policy coming from Beijing. These new numbers fit that pattern. 

• On January 12th, during the 2 week period when authorities in Wuhan were claiming that cases has plateaued at 41, data showed that 200 people went to the hospital for a fever in just of one of Wuhan's thirteen districts (of course, fever is not something one usually goes to the hospital for). 
• It is not known exactly how many hospitals there are in Wuhan, but 200 is a rough guess. On January 29th, 26 of them were designated for COVID-19 patients only. However, at the same time, hospitals were told they could not report COVID-19 cases unless they were one of those designated hospitals. The city reported 356 admissions to those hospitals that day, while one of the 13 districts reported 150 hospitalizations for COVID-19 outside of those hospitals. This is suggestive of an outbreak 5 times the reported size. That's consistent with the crematoria data which showed an outbreak 12-13 times the reported size. Also, growth rates of 64% were in the official data as recently as 2 days before this move; growth rates dropped steadily after that.
• On February 11th, the columns on the data reporting forms for COVID-19 used by local hospitals were deleted, and did not appear again. That date corresponds to a discrete drop in official daily growth rates of cases by over 20%, followed by continued smooth declines.
• A listing of names of confirmed COVID-19 deaths from March 14th shows that 71% were outside the 26 designated COVID-19 hospitals, and an additional 26% did not have a hospital designated. Only 3% of the reported COVID-19 deaths were from a COVID-19 designated hospital.
• Lastly, a report from a residential compound for hospital employees in Hubei indicated that a third of them were infected.

Again, I'll emphasize that The Epoch Times is not friendly to the government in Beijing, so they are definitely looking for dirt. Having said that, they usually find it, and it tends to be quietly confirmed in our legacy media within a few days.

COVID-19 # 51 (Required Parts Highlighted)

We made the national news, or at least Washington (near St. George) did. And not in a good way (although I was ticked that the made zero mention of the fact that Utahns have a headstart on being prepared for this sort of thing). Anyway, I was down there on Thursday evening actually. I did see quite a few businesses closed, but it was busier than I expected. Also, those big digital signs they post warnings on over the interstates were turned off; to me that seems like an obvious place to note that travel is discouraged, or say how the number of COVID-19 fatalities compares to the highway fatalities they're always advertisign. BTW: cases and hospitalizations in the southwest Utah region were up again today.

Caixin, a publication that does work openly in China, but which is also openly critical of the government, reported today that the government in Beijing was not informed of human-to-human transmission until January 19th (2 days before the decision to quarantine Wuhan was made).

I have seen several versions of this information in the legacy media, and every single one of them has been wrong (they show just one graph, so it's basically just noting that pollution tracks urban areas on a random day). So I went to the source. You need to compare the two images. This is before:

This is after:

What's being shown here is nitrogen dioxide concentrations. These are primarily generated by diesel vehicles and coal burning power plants (not usually located in cities). It is somewhat weather dependent as well (this explains the absence of Columbus, Rochester, and Syracuse in the before picture, and huge fadeout of Boston in the after one). We got chloropleths of this from China about a month ago. Anyway, viewed properly, pollution is way down, especially in NYC. FWIW: I've noticed basic thinking ability seems to go down anytime pollution measures are involved.

Excellent piece in The Wall Street Journal entitled "Why Doesn't the Flu Tank the Economy Like COVID-19?"

Instead of gentle waves of cases cascading across the country over a span of six months, like the flu, a tidal wave of Covid-19 cases has swept over a handful of cities in half the time. The concentration of quickly amassing serious infections overwhelmed the affected areas ...

We had a flu pandemic about 10 years ago, and here's the comparison:

In 22 fewer days, Covid-19 infected 11 times as many people and killed 60 times as many. 

That pandemic was far worse than a typical flu season. Even so, while a lot more people get the flu, amongst those that do, you're 8 times more likely to be hospitalized if you get SARS-CoV-2, and your stay will be almost 4 times as long. It is mathematically valid to multiply those and imply that hospital usage will be 30 times higher during COVID-19.

It took me over a day to find a day-by-day total for Utah (yeah Bing, boo Google). Go to the top item on this link, and then click Trending Stats, and mouseover the graph. Anyway, yesterday, my top chart was about doubling rates after a location hits 10 deaths. Utah doesn't qualify for fleshing out that graph just yet, but the daily death total just doubled in 7 days (from 8 to 17), which make us comparable to Washington (state). FWIW, I also did an exhaustive examination of our doubling times for all new deaths (e.g, when we hit 2 how long did it take to get to 4, and so on), and the day were 1, 3, 5, 6, 10, and 7. Together, those suggest that our experience in this state is not much better than what other states have experienced; but we clearly started later, so we'll peak later.

Prime Minister Boris Johnson was discharged today after 8 days and 7 nights in the hospital. He will not be returning to work or even his primary residence right away. As is probably normal with things involving leaders, it turns out that the truth was worse than we were led to believe. Apparently they had been notified he might need hospitalization 3 days prior to ultimate check-in, and his condition  in the hospital deteriorated more deeply and rapidly than people were led to believe.

There are confirmed reports this morning of hospital overload hitting Moscow.

COVID-19 # 50 (Required Parts Highlighted)

As long as you don't use this sort of chart the wrong way (to make a forecast by presuming that each location behaves the same way), it can still be used in interesting ways:

This gives a sense that small improvements in "flattening the curve" can lead to big declines in deaths. And, I just checked the data: Utah just recently had its 10th death, so if we were to go on this graph it would be way down at the left.

Here's a time series graph to help you digest what is going on in New York City:

All of those peaks every winter are the flu. Any thoughts on whether this is just like the flu? Oh ... and sharper photos of those trench graves in New York City show the coffins are stacked three high.

That cool chart that TL brought to me from Flouish has been updated. I could not get it to repost properly, so you'll just have to click through, but COVID-19 has now moved up to be the #1 cause of death in the U.S. The other day it was # 3.

Contact tracing is how South Korea did so well. In it, you figure out everyone a person came in contact with, and you find them and get them quarantined. Here's what happens when you don't:

This is from Chicago. In South Korea they got cellphone data showing where positives had been, and reached out to every other person with a phone who'd been to that place.

New research from the CDC estimates that the R0 of SARS-CoV-2 is much higher than previously thought. Their 95% confidence interval for R0 is 3.8 to 8.9.

Remember federalism? Ian Bremmer makes an excellent point:

Trump isn’t going to re-open the country.

He’ll say the country should open. And most governors—who actually wield the authority—will keep it locked down.

Similar to what’s already been playing out in Brazil.

— ian bremmer (@ianbremmer) April 10, 2020

This is the flip side of some governors (like Herbert in Utah) not shutting down their states. Do note that Utah is now requiring travelers from outside the state to divulge their travel plans once arriving in our state.

This is a personal opinion.

But it's okay to take a moment and just fully realize, this is one of the worst things to have ever happened. And there's no sugar to coat it.https://t.co/LTelcTbOEM

— Brian Resnick (@B_resnick) April 10, 2020

Numbers continue to go up in our region: 52 positives in southwestern Utah and 4 people in the hospital. The much smaller Central Utah Public Health Department (which is likely to send some cases to the hospital in Cedar City) now has a COVID-19 page up and running. They are up to 6 cases and 1 hospitalization. The worst hit is Millard County, which is in the area that probably will send people to Cedar City.

And here's an example of how not to use a log scale.

The math here is correct but the perception it created is incorrect. This invites comparison of the black, blue, and red curves, and the eye is drawn to the distance between them. But, because of the log scale, that distance doesn't mean the same amount on the right and left of the chart.

Some New Visualizations

Here's a couple of new visualizations from Statnews. The yellow circle is current numbers, the larger browner circle is projected numbers:

I dislike these seasonal flu comparisons. Flu is something we do nothing about, and this is what we get. COVID-19 is something we're doing unprecedented things to combat, and that only gets us down in the range of flu. In a month.

Not too fond of this one either. None of these are infectious, communicable, diseases. But, I think it's useful because as a nation we're pretty freaked out about suicides, overdoses, and Alzheimer's.

What Tyler Cowen Thinks

On the internet, Tyler Cowen is the most famous/important economist. Here's what he said about social distancing, shutdowns, re-opening, and the pandemic. It sounds realistic to me.

Putting COVID-19 Deaths In Context: A New Milestone

Part of the problem with COVID-19 is getting your mind around the scale of the numbers involved.

Yes, it's true, car accidents kill more Americans every year. So far. We're probably going to pass that in a couple of weeks (I haven't forecast that one yet).

Yes, it's also true that influenza kills more Americans every year. We're forecast to break that one as well. The thing is with the flu, it's not a single germ — it's a whole family of them. It also produces that death toll over months. The real comparison here will be the number of deaths next winter that are flu-related rather than SARS-CoV-2 related.

I remarked a while back that COVID-19 would overtake 9/11 in deaths. That seems like a long time ago, but it was less than 2 weeks ago that I made that prediction.

As I write this on the evening of April 9th, the total number of U.S. deaths is 16,679. A month ago, on the 9th of March, we'd had 26.

That's more in this one month that ... the entire Afghanistan war, Iraq wars, and fight against ISIS combined. It's actually already more than twice that amount. BTW: that's all conflict related deaths, not just the combat ones.

We're actually closing in the Korean War. We're forecast to pass the Vietnam War as well.

Anyway, I'm not trying to be morbid. I'm trying to come up with a number that puts this month in perspective with some other month. And I think I've finally found it. We've had more deaths in the past month than the U.S. Army (and Air Corps)† had in combat deaths in June 1944 — the month of the D-Day invasion. In fact, we've now passed every single month of World War II, including July 1944 (German counterattacks in Normandy), December 1944 (the Battle of the Bulge), February 1945 (Iwo Jima), and April 1945 (Okinawa).

† I was unable to find month by month death totals for Naval (and Marine) in World War II. They are kept track of by campaign, but every single one of those is already dwarfed by one month of COVID-19.